Emergence in Science and Philosophy (Routledge Studies in the Philosophy of Science)

Emergence in Science and Philosophy

Routledge Studies in the Philosophy of Science

1. Evolution, Rationality and Cognition A Cognitive Science for the Twenty-First Century Edited by António Zilhão 2. Conceptual Systems Harold I. Brown 3. Nancy Cartwright’s Philosophy of Science Edited by Stephan Hartmann, Carl Hoefer, and Luc Bovens 4. Fictions in Science Philosophical Essays on Modeling and Idealization Edited by Mauricio Suárez 5. Karl Popper’s Philosophy of Science Rationality without Foundations Stefano Gattei 6. Emergence in Science and Philosophy Edited by Antonella Corradini and Timothy O’Connor

Emergence in Science and Philosophy

Edited by Antonella Corradini and Timothy O’Connor

New York

London

First published 2010 by Routledge 270 Madison Avenue, New York, NY 10016 Simultaneously published in the UK by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN

Routledge is an imprint of the Taylor & Francis Group, an informa business This edition published in the Taylor & Francis e-Library, 2010. To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to www.eBookstore.tandf.co.uk. © 2010 Taylor & Francis All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Emergence in science and philosophy / edited by Antonella Corradini and Timothy O’Connor. p. cm. — (Routledge studies in the philosophy of science ; v. 6) Includes bibliographical references and index. 1. Emergence (Philosophy) 2. Science—Philosophy. I. Corradini, Antonella. II. O’Connor, Timothy, 1965– Q175.32.E44E445 2010 116—dc22 2009051120 ISBN 0-203-84940-X Master e-book ISBN

ISBN13: 978-0-415-80216-1 (hbk) ISBN13: 978-0-203-84940-8 (ebk)

Contents

List of Figures Introduction

ix xi

ANTONELLA CORRADINI AND TIMOTHY O’CONNOR

PART I Emergence: General Perspectives Part I Introduction

3

ANTONELLA CORRADINI AND TIMOTHY O’CONNOR

1

The Secret Lives of Emergents

7

HONG YU WONG

2

On the Implications of Scientific Composition and Completeness: Or, The Troubles, and Troubles, of NonReductive Physicalism

25

CARL GILLETT

3

Weak Emergence and Context-Sensitive Reduction

46

MARK A. BEDAU

4

Two Varieties of Causal Emergentism

64

MICHELE DI FRANCESCO

5

The Emergence of Group Cognition GEORG THEINER AND TIMOTHY O’CONNOR

78

vi

Contents

PART II Self, Agency, and Free Will Part II Introduction

121

ANTONELLA CORRADINI AND TIMOTHY O’CONNOR

6

Why My Body is Not Me: The Unity Argument for Emergentist Self-Body Dualism

127

E. JONATHAN LOWE

7

What About the Emergence of Consciousness Deserves Puzzlement?

149

MARTINE NIDA-RÜMELIN

8

The Emergence of Rational Souls

163

UWE MEIXNER

9

Are Deliberations and Decisions Emergent, if Free?

180

ACHIM STEPHAN

10 Is Emergentism Refuted by the Neurosciences? The Case of Free Will 190 MARIO DE CARO

PART III Physics, Mathematics, and the Special Sciences Part III Introduction

207

ANTONELLA CORRADINI AND TIMOTHY O’CONNOR

11 Emergence in Physics

213

PATRICK MCGIVERN AND ALEXANDER RUEGER

12 The Emergence of the Intuition of Truth in Mathematical Thought SERGIO GALVAN

233

Contents vii 13 The Emergence of Mind at the Co-Evolutive Level

251

ARTURO CARSETTI

14 Emerging Mental Phenomena: Implications for Psychological Explanation

266

ALESSANDRO ANTONIETTI

15 How Special Are Special Sciences?

289

ANTONELLA CORRADINI

Contributors Index

305 309

Figures

2.1

An ion channel and its protein sub-units.

28

2.2

Blobs and globules in simple aggregations.

34

2.3

Blobs and globules in complex aggregations.

35

2.4

Diagram of the options as understood by the philosophical critics and illustrating the perceived implications of the Argument from Composition.

41

Diagram of the main options illuminated by appreciating the implications of the Conditioned view of aggregation and the possibility of Strong emergence.

42

5.1

Opening the “black box” of group cognition.

94

5.2

Phases of the choreographic development of dance material for Red Rain.

99

2.5

9.1

Keil’s libertarianism.

184

11.1

Representative phase space trajectories for damped oscillator and undamped oscillator.

217

11.2

Diagram 1.

219

11.3

Diagram 2.

219

11.4

Diagram 3.

227

14.1

A non-emergent phenomenon: Take three separate angles.

271

14.2

A non-emergent phenomenon: Combine three angles.

271

14.3

A non-emergent phenomenon: Obtain a triangle.

271

x

Figures

14.4

Another non-emergent phenomenon: Take three round shapes.

271

14.5

Another non-\emergent phenomenon: Obtain a triangle.

272

14.6

An emergent phenomenon: Obtain Kanizsa’s triangle.

272

14.7

Target-source similarity under the conditions of the fi rst experiment concerning on-line rating.

279

Source-target similarity under the conditions of the third experiment concerning on-line rating.

279

Percentages of analogical solutions under different conditions in the series of experiments concerning the awareness of source-target correspondence.

280

14.8 14.9

Introduction Antonella Corradini and Timothy O’Connor

The concept of emergence has seen a significant resurgence in philosophy and a number of sciences in the past couple decades. Yet debates between emergentist and reductionist accounts of specific phenomena, and of visions of the natural world generally, continue to be hampered by imprecision or outright ambiguity in the use of terms. The term ‘emergence’ is clearly evocative for thinkers across the spectrum of those who theorize about the relationship between ‘high-level’ theories, and the real-world properties and dynamics they seek to describe, and theories and phenomena that pertain to more basic physical systems. Evocative, but extremely vague. Emergent phenomena are said to arise out of and be sustained by more basic phenomena, while at the same time exerting a ‘top-down’ control, constraint or some other sort of influence upon those very sustaining processes. To some critics, this has the air of magic, as it seems to suggest a kind of circular causality. (See Kim, 1999, for an argument to this conclusion.) Other critics deem the concept of emergence to be objectionably anti-naturalistic, requiring the onset at particular historical junctures of novel properties and behavior that are discontinuous with the world’s fundamental dynamics. Objections such as these have led many thinkers to construe emergent phenomena as complementary to yet harmonious with the behavior of fundamental physical entities supposed to be uniform in every context, including those involving emergent phenomena. On this view, emergent properties and the patterns to which they give rise are explanatorily selfcontained. They are embedded in nature at a relatively coarse-grained level of structure while not ‘disrupting’ or ‘violating’ the ordinary dynamics of the fi ner-grained (more fundamental) levels. Nature, on this understanding, has a hierarchical structure, with each level of the hierarchy (corresponding to basic physics, chemistry, various levels of biology, and psychology and other information-based sciences) requiring its own concepts and laws to capture the distinctive behavior it exhibits. However, the preceding attempt at reconciling emergence with a (presumed) pervasive causal continuity at the fundamental level can seem to deflate emergence of its initially profound significance. It locates the

xii

Introduction

‘autonomy’ of higher-level sciences in their capacity to describe coarsegrained patterns in the world’s mosaic that, however interesting and useful, do not contribute to driving the world’s evolution. The true causal work, on this objection, is all done at the level of basic physics. On reflection, higher-level sciences appear as mere shorthand in the business of describing the world’s behavior. As it is often put, such an outlook threatens to turn emergence into an epistemological, rather than metaphysical, concept. (See O’Connor & Wong, 2006.) Proposals and criticisms such as those just gestured at constitute, in skeletal form, the basic problematic informing modern discussion of the concept of emergence. It is mirrored by similar controversy over how best to characterize the opposite systematizing impulse, usually given the equally evocative but vague term, “reductionism.” We have collected the chapters in this volume in the belief that much progress has been made in recent years in clarifying the alternatives and the proper terms in which competing claims of evidential support should be advanced. While it is scarcely credible for a partisan to claim that his or her favored view has been more or less established, inadequacies in some older formulations and arguments have been exposed, narrowing the field a bit. The new essays collected here reflect that improved perspective and attempt to advance the debate along one or another front. The volume has three parts. We provide a detailed introduction to each part immediately prior to the chapters in that part. Here, we make but short and general remarks. Part I lays a general ontological foundation. In it, six authors consider different accounts of how we might develop an emergentist picture of nature. Most target avowedly metaphysical (and not merely epistemological) construals of emergence. Collectively, they advance a number of fresh proposals, while being informed by philosophical and scientific discussion to date. Through these chapters, the reader will get a pretty thorough understanding of the range of highly general alternatives that have been floated in recent discussion. In Part II, the authors focus specifically on views concerning the status of mind in the physical world. Reflecting a general trend in contemporary metaphysics to try to refurbish and defend views that were thought dead during the long, dark night of much of the twentieth century when austere empiricism reigned, three of our authors argue for an emergent dualist account of human persons (as does Corradini in Part III of the volume). This is followed by two discussions of how notions of emergence might bear upon the plausibility of the belief that human beings have free will, or metaphysical freedom. Part III of the volume turns to concrete examinations of particular sciences, asking whether and in what sense they indicate the reality of emergent phenomena. Physics, individual psychology, and the special case of mathematical intuition are all considered under the emergentist rubric.

Introduction xiii (Chapter 3 by Bedau and Chapter 5 by Theiner and O’Connor also discuss specific instances of recent scientific theorizing where emergentist themes plausibly have application.) Attempts to show in convincing detail how different general accounts of emergence may or may not have application to the complex mess of going empirical theories are perhaps the most likely sources of fruitful philosophical discussion of emergence in the years to come.

REFERENCES Kim, J. (1999). Making sense of emergence. Philosophical Studies, 95, 3–36. O’Connor, T. & Wong, H.Y. (2006). Emergent properties. In E.N. Zalta (ed.) The Stanford Encyclopedia of Philosophy. Retrieved from http://plato.stanford. edu/entries/properties-emergent/. Accessed: September 15, 2009.

The publication of this book has been supported by the Catholic University of the Sacred Heart of Milan by means of a grant set for publication of scientific reports (research found D3.1, year 2009 and research found D3.2, year 2008).

Part I

Emergence General Perspectives

Part I Introduction Antonella Corradini and Timothy O’Connor

Philosophical and scientific writing about the concept of emergence is animated by the widespread though not universal conviction that wherever the natural world manifests patterns of organized behavior whose description seems to require distinctive descriptive and causal/explanatory concepts, there may be found distinctive natural properties and processes—ones that operate independently of more fundamental processes. But the nature of their distinctive character and causal/explanatory autonomy from more basic phenomena is sharply debated. One of the most fundamental divides arises from the following question: do emergent phenomena entail that theories of basic physics must necessarily be ‘incomplete’, insofar as these theories are developed by studying small-scale microscopic systems exclusively when they are not embedded within organized emergent systems, and so not being subjected to the distinctive forces or constraints of such systems? The chapters in this fi rst part address general questions of this sort, and each proposes a different account of how emergence is best understood for reasons of either conceptual coherence or explanatory fruitfulness. Hong Yu Wong’s contribution explores the coherence of the classical supervenience model of emergence that was fi rst articulated by C. D. Broad and recently resurrected by Kim and others in discussions of non-reductive physicalism. Using Kim’s arguments against emergence in his “Making Sense of Emergence” as a foil, Wong defends the internal consistency of the classical emergentist position. He challenges Kim’s downward causation and causal exclusion arguments as applied to emergent properties. In the second half of his chapter, having set aside causal exclusion worries deriving from Kim’s work, he examines whether there might still be a tension at the heart of the classical emergentist position that comes out of not entirely unrelated considerations. He approaches the issue not through the usual route of what the prospects for a claim about the causal closure of the physical are, but through considering the sense in which emergent properties supervene on basal properties on the classical ontology and whether this is consistent with emergent properties having novel downward causal influence. He attempts to describe a (putative) minimal classical emergentist world where emergents have novel causal powers, but appear to fail

4

Antonella Corradini and Timothy O’Connor

to supervene on basal properties. Consideration of why emergentists must reject this world as an emergentist world brings out why emergent properties supervene on the classical picture. He argues that supervenience in the case of emergent properties must be sui generis—grounded solely on fundamental, non-derivative emergent laws. This sets certain constraints on the explanatory potential of the classical emergentist position, since any causal influence by emergent properties must be consistent with patterns of co-variation between emergents and bases permitted by the fundamental emergent laws. The overarching aim of his chapter is to provide a more accurate picture of the classical emergentist position and its strengths and weaknesses. Carl Gillett’s point of departure is debates not in general metaphysics but among contemporary scientific theorists. He suggests that the battle between reductionism and emergentism is a crucial case where a ‘metaphysics of science’ promises greater illumination than traditional metaphysics carried out in complete abstraction from and disregard of ongoing scientific theorizing. Both reductionism and emergentism, Gillett suggests, are well represented among prominent scientists, while neither label is embraced by an equally large share of philosophers. Perhaps this difference, then, reflects a fundamental difference in understanding how these terms are to be understood. Using the example of the mechanisms involved in potassium ion channels, Gillett tries to show how ‘compositional reductionism’ is a viable approach to special science entities, able to withstand traditional philosophical arguments against reductionism. He further holds that ‘strong emergentism’, as he characterizes it, is a similarly viable option, one that does not suffer from the charges of conceptual or metaphysical ‘spookiness’ often lodged against emergentist hypotheses. Strong emergentism is a form of non-reductive physicalism, accepting that all emergent properties are wholly realized by physical properties and relations, while rejecting usual formulations of the causal completeness of physics. Mark Bedau advocates that philosophers cease arguing about which account of emergence is correct and instead embrace pluralism about emergence. He identifies three distinct kinds of emergence—nominal, weak, and strong—and contends that it is a simple empirical matter which of these may have application. Nominal emergence occurs where there is a macrolevel property that logically or conceptually cannot have application to the micro-level. It is plain that there are such properties (e.g., liquidity). Weak emergence is a more ambitious, dynamical concept, on which the macro-level phenomena are ontologically and causally reducible to micro-phenomena, but only through an especially complex explanation. Thus, the hallmark of weak emergence is what he terms “explanatory incompressibility”—something which admits of degrees (and hence, weak emergence does as well). Offering several illustrative examples, Bedau argues that weak emergence is widespread in nature and scientifically very important. Finally, strong emergence requires irreducible “downward’ causal powers. Contrary to

Part I Introduction

5

some, he contends that this concept is fully intelligible, but “all the evidence today suggests that strong emergence is scientifically irrelevant” (51). Like Bedau, Michele Di Francesco holds that we should recognize a variety of emergence relations, since, he argues, there is no single notion of emergence that plausibly can do all the explanatory work that is foisted upon it by various thinkers motivated by distinct philosophical concerns. His point of departure in thinking about emergence are the platitudes that emergence is not reductionism and emergence is not dualism. He seeks to elucidate two varieties of emergence: moderate emergentism, which shades into non-reductive physicalism of standard sorts, and radical emergentism, which comes close to being an outright form of dualism and is the weakest form of naturalism we may conceive. Finally, Georg Theiner and Timothy O’Connor take as their point of departure the working assumption of a moderate form of non-reductive physicalism. They show how this broad framework may admit three distinct varieties of emergence: emergence as organization dependence, as rooted in the absence of intentional design, and as multiple realizability. They then turn to the central focus of their chapter, which is the controversial thesis of group cognition. They propose that one who endorses a non-reductive physicalist view ought to endorse a “big tent” approach to cognition, one admitting of at least seven component capacities: adaptability, information processing, heed (of one’s environment), intentionality, extension, self-reflexivity, and consciousness. They then argue at length that contemporary work in several scientific domains provides strong evidence for forms of group cognition that exhibit several combinations of these characteristics (except for phenomenal consciousness). Different phenomena score differently on each of the three varieties of emergence identified previously. As with Bedau and Di Francesco, a moral of their discussion is that we may think of emergence in a plurality of ways, each admitting of degrees.

1

The Secret Lives of Emergents* Hong Yu Wong

1. INTRODUCTION In “Making Sense of Emergence”, Jaegwon Kim argues that ontological emergence does more wrong than just ‘betting against physics’; it is incoherent.1 Ontological emergence is the thesis that when aggregates of microphysical properties attain a requisite level of complexity, they generate and (perhaps) sustain emergent natural properties. What is constitutive of ontological emergence is the novel causal influence of emergent natural properties. One significant question is how to understand the notion of novel causal influence. An intuitive gloss is that an emergent property provides a causal contribution that goes beyond causal contributions made by any of the lower level properties had by the system and its parts taken either in isolation or in combination. One way to capture this is to claim that the behavior of an emergent property over time is characterized by a fundamental law. A further, but natural commitment of the emergentist is that once these distinctive properties emerge from basal properties, they can exercise causal influence on properties at the basal levels. The emergence of these new properties is taken to affect the dynamics of properties at the basal level. (This is commonly known as ‘downward causation’.) Kim contends that such a picture cannot be sustained. Kim argues that unless emergence is given a deflationary, epistemological interpretation, it is unworkable, because we cannot make sense of emergent properties having downward causal influence. He considers two varieties of emergent downward causation: synchronic reflexive downward causation and diachronic reflexive downward causation. (Reflexive because emergent properties have causal influence on events involving their own micro-constituents.) Kim defi nes the two varieties thus: Synchronic reflexive downward causation. At a certain time t, a whole, W, has emergent property M, where M emerges from the following configuration of conditions: W has a complete decomposition into parts a1, . . . , an; each ai has property Pi; and relation R holds for the sequence a1, . . . , an. For some aj, W’s having M at t causes aj to have Pj at t. (28)

8

Hong Yu Wong Diachronic refl exive downward causation. As before, W has emergent property M at t, and aj has Pj at t. We now consider the causal effect of W’s having M at t on aj at a later time t + Δt. Suppose, then, that W’s having M at t causes aj to have Q at t + Δt. (29)

Kim thinks the synchronic reflexive variety absurd, for it seemingly involves causal circularity (modulo worries about simultaneous causation). The diachronic variety, however, escapes the circularity worries because of the time delay between the putative cause and effect. Kim surmises that diachronic reflexive downward causation is all that emergentists need. But he argues that diachronic reflexive downward causation is open to his causal exclusion argument: . . . I earlier argued that any upward causation or same-level causation of effect M 2 by cause M 1 presupposes M 1’s causation of M 2’s lower level base, P2 (it is supposed that M 2 is a higher-level property with a lower-level base; M 2 may or may not be an emergent property). But if this is a case of downward emergent causation, M 1 is a higherlevel property and as such it must have an emergent base, P1. Now we are faced with P1’s threat to preempt M 1’s status as a cause of P2 (and hence of M 2). For if causation is understood as nomological (law-based) sufficiency, P1, as M 1’s emergence base, is nomologically sufficient for it, and M 1, as P2’s cause, is nomologically sufficient for P2 . Hence P1 is nomologically sufficient for P2 and hence qualifies as its cause. The same conclusion follows if causation is understood in terms of counterfactuals—roughly, as a condition without which the effect would not have occurred. Moreover, it is not possible to view the situation as involving a causal chain from P1 to P2 with M 1 as an intermediate causal link. The reason is that the emergence relation from P1 to M 1 cannot properly be viewed as causal. This appears to make the emergent property M 1 otiose and dispensable as a cause of P2; it seems that we can explain the occurrence of P2 simply in terms of P1, without invoking M 1 at all. If M 1 is to be retained as a cause of P2 , or of M 2 , a positive argument has to be provided, and we have yet to see one. In my opinion, this simple argument has not so far been overcome by an effective counter-argument. (32)2 In this chapter I will provide the counterargument that Kim asks for. The purpose of this chapter is twofold: (1) to defend the internal consistency of the classical emergentist position, more or less in the form as it was articulated by C. D. Broad (1925), by resisting Kim’s argument and variations on it, but also (2) to make clear the costs of embracing the classical emergentist position. These two aims correspond to the two halves of this chapter. In the fi rst half, our central concern is the possibility of diachronic downward causation by emergent properties. I begin by considering Kim’s

The Secret Lives of Emergents 9 charge of causal circularity against synchronic reflexive downward causation. I then turn to the central issue of diachronic downward causation. I present Kim’s argument, which proceeds in two steps. The fi rst step is commonly known as the ‘downward causation argument’ and the second step the ‘causal exclusion argument’. I challenge each step in turn. This concludes the fi rst half of this chapter. In the second half of this chapter, I consider the price of embracing such a position. I approach the issue not through the usual route of what the prospects for a claim about the causal closure of the physical are, but through considering the sense in which emergent properties supervene on basal properties on the classical ontology and whether this is consistent with emergent properties having novel downward causal influence. I attempt to bring out a tension between having novel downward causal influence and supervenience by considering a putative situation where emergent properties fail to supervene. On the basis of this, I argue that supervenience in the case of emergent properties must be sui generis—grounded solely on fundamental, non-derivative emergent laws. This sets certain constraints on the explanatory potential of the classical emergentist position, since any causal influence by emergent properties must be consistent with patterns of co-variation between emergents and bases permitted by the fundamental emergent laws. My overarching aim is to provide a more accurate picture of the classical emergentist position and its strengths and weaknesses.

2. SYNCHRONIC DOWNWARD CAUSATION Before I discuss diachronic reflexive downward causation, I want to remark on Kim’s discussion of synchronic reflexive downward causation (28–29). Kim argues that synchronic reflexive downward causation is causally circular and hence violates “the causal-power actuality principle”.3 While I fi nd synchronic reflexive downward causation implausible because of worries about simultaneous causation, there are possible ways to understand synchronic reflexive downward causation that do not appear to involve causal circularity. For it seems plausible that the emergent is not dependent on the total microstructural base, a1, . . . , an; where each ai has property Pi; and relation R n holds for the sequence a1, . . . , an; but only on a core microstructural nuclei, a1, . . . , am, where each ai has property Pi; relation R m holds for the sequence a1, . . . , am; m < n; and m, n ∈ ℕ. The emergentist can then suppose that the downward causal effects of the emergent property fall outside the microstructural nuclei; thus there is no causal circularity. For there is reason to suppose that if there truly are robust emergents which ontologically transcend their microstructural bases, these emergents will be invariant to slight perturbations at the microstructural level. The situation presented previously helps make sense of this invariance (cf. the notion of a “protected” property sometimes invoked in scientific discussions of

10

Hong Yu Wong

emergent properties4). Another way to think of this: the core microstructural state is less than maximally specified (i.e., it is a state D which partly constitutes total microstructural states D1, D 2 , . . . , where Dn = D + ρn; and ρn is the remainder microstructural state that with D constitutes total microstructural state Dn) and serves as an emergence base for emergent property E which in turn causes a deterministic evolution of the remainder microstructural state from ρ1 to ρ2 to ρ3 and so on (and thus the total microstructural state from D1 → D 2 → D3 → . . . ). Further variants on this scenario can be imagined.

3. DIACHRONIC DOWNWARD CAUSATION Turn now to consider diachronic reflexive downward causation, which is our central concern. Kim’s argument boils down to this: diachronic activity of an emergent property will inevitably be redundant, since its effects are directly attributable to the conditions which sustain it; thus, emergent properties could not possess the “genuinely novel causal powers” that emergentism demands—powers that “must be capable of making novel causal contributions that go beyond the causal powers of the lower-level basal conditions from which they emerge” (25).5 Kim’s argument proceeds in two steps. The first step attempts to establish that any causal influence of emergent properties, even on other emergent properties, must proceed by downward causal influence on basal properties that the emergent properties supervene on. This step is what is often known as the ‘downward causation argument’. The second step attempts to establish that emergent properties and their base properties compete as causes of subsequent basal level effects, and that, given some auxiliary assumptions, emergent properties are superfluous in explaining why these later basal level effects occurred. This step is commonly known as the ‘causal exclusion argument’. Let us consider each step in turn. For Kim, both upward and same-level causation entail downward causation (23–24). Kim presents a familiar argument to this conclusion (Kim, 1992). Consider a property M1, at non-fundamental level L and time t1, that causes another property M 2 , at non-fundamental level L and time t2. Since M 2 is a property at a non-fundamental level, by hypothesis, it has emergence base, P2 , at t2 and level L–1. Kim sees a tension in this situation because there appear to be two answers to why M 2 is instantiated at t2. First, M 2 is instantiated at t2 because M1 at t1 caused it (ex hypothesi); second, M 2 must of (at least) nomological necessity be instantiated at t2 because its emergence base, P2 , is present. Kim claims that there appear to be two competing causes for the instantiation of M 2 at t2 , and this apparently jeopardizes M1’s causal responsibility for M 2. Kim suggests the only coherent description of the situation where M1’s causal responsibility for M 2 is preserved is this: M1 causes M 2 by causing its emergence base P2. The moral is that emergents can only be brought about by bringing about their supervenience base.

The Secret Lives of Emergents 11 The basis of Kim’s argument is his “causal/explanatory exclusion principle”,6 which says that there cannot be more than one complete and independent explanation for any given event unless we have a case of causal overdetermination, something thought to be rare. However, it is unclear that the causal/explanatory exclusion principle applies here, for it is neither obvious that M1 and P2 are competing causes nor that they are independent. To begin with, one explanation is diachronic (M1 at t1 causes M 2 at t 2), while the other is synchronic (M 2 is necessarily instantiated at t 2 because P2 is instantiated at t 2). Notice that while the relation between M1 and M 2 is causal, the emergence relation between P2 and M 2 —Kim tells us—is non-causal. A fortiori: if P2 isn’t a cause of M 2 , how can it compete as a cause with M1? (The non-causal nature of the relation between an emergent property and its basal conditions is an important premise in the second step of Kim’s argument.) Furthermore, since M1 is a higher-level property, it necessarily emerges from basal conditions P1. It is not implausible to suppose that P1 also causally suffices for P2 (especially given the completeness of physics), in which case M1 and P2 will not be independent.7 Thus we have found no reason that any emergent causation must proceed through downward causation. This is not to say that we do not want downward causal powers for emergent properties on a plausible emergentist position, only that we have found no good reason to restrict emergent causation to downward causation. In fact, we might want downward causation for somewhat independent reasons. Just to take the case of the mental and the physical as an example: 1. our actions (mental) have physical consequences—action chains stretch out into the physical world; 2. some mental-mental chains have physical intermediaries—when we communicate, there are sound waves that bridge the causal gap between us; and 3. there are regularities across levels. There are grounds for thinking that Kim’s worries at the causal exclusion step have some force even if we reject that all emergent causation is downward causation since we have independent reasons to want downward causation and we may want downward causal influence in some of the central cases where we posit emergent properties. If downward causation is something we want in these central cases, then even if Kim’s reasoning at the fi rst step is problematic, we may grant Kim that M1 causes M 2 via causing its emergence base P2 for the sake of discussion and proceed to Kim’s second step. The second step of the argument is a version of the causal exclusion argument. Kim argues that emergent properties are epiphenomenal (and hence emergentism is incoherent). Here is his argument:8

12

Hong Yu Wong

(1) By hypothesis, M1 is nomologically sufficient for P2 (downward causation). (2) By hypothesis, emergent property M1 must have emergent base, P1. (3) Since P1 is M1’s emergence base, P1 is nomologically sufficient for M 1. (4) By transitivity, P1 is nomologically sufficient for P2 . (5) Hence, on an appropriate analysis of causation, P1 qualifies as a cause of P2 . (6) Since the emergence relation is non-causal,9 P1 cannot cause P2 via M 1. (7) We can explain the occurrence of P2 simply in terms of P1, without invoking M1. (8) Thus, barring further argument, M1 is dispensable as a cause of P2 . A few remarks are in order. (1) follows from the earlier argument for downward causation. Kim’s intention is that the argument should be run independent of specifics about causation; assume your favorite theory of causation and run the argument. I think this implausible; as we shall see, the specifics about causation influence the specifics of the argument, but Kim has left the specifics out. Note that Kim’s argument is intended to be metaphysical, not epistemological; Kim means to argue that causal effects of emergent properties are directly (and not just indirectly) attributable to its basal conditions. Note also that Kim’s argument proceeds without invoking the completeness of physics. Hence, epiphenomenality is not forced upon emergent properties by the completeness of physics, where physics is roughly understood as excluding mental and emergent entities.10 Why doesn’t Kim invoke the completeness of physics (so understood)? Because emergentists need not be committed to it.11 Finally, observe that the argument does not contain a premise against systematic overdetermination (so it is an unusual variant); instead, it is a somewhat obscure argument from the transitivity of nomological sufficiency, securing the status of P1 as a cause and then asserting that M 1 is dispensable because M 1 doesn’t do anything that P1 can’t do. I shall now attempt to develop and evaluate the argument. Any plausible nomological sufficiency analysis of causation will say that causation is nomological sufficiency plus some extra factor, most likely spatiotemporal contiguity (Kim, 1973); and similarly for counterfactual analyses. On this count, the argument will rest on the fact that if the extra factor, say spatiotemporal contiguity, holds between M1 and P2 , it will likewise hold between P1 and P2 , since P1 is located where M1 is and is instantiated at the same time when M1 is. By the transitivity of nomological sufficiency, P1 is nomologically sufficient for P2 . But since whatever the extra factor that allows M1 to qualify as a cause of P2 will also apply to P1, P1 now qualifies as a cause of P2 . Kim then makes the observation that M1 has a non-causal relationship with P1. So there is no causal chain going from P1 through M1

The Secret Lives of Emergents 13 to P2; it is not the case that we have a causal chain P1 → M1 → P2 . On the basis of this, he concludes that M1 cannot be a cause of P2 . There’s a gap in the deduction: what excludes M1 as a cause of P2? It is important to note that in the argument the status of P1 as a cause of P2 is indirectly secured. It is unclear that P1 can qualify as a direct cause of P2 independent of assuming the completeness of physics which is not in play in this argument, especially given that the downward causal powers of M1 are emergent ones, while P1 has powers which are always manifest. Thus, it seems open to the emergentist to describe the scenario as one on which M1 emerges from P1 (synchronically) and causes P2 .12 Even if P1 is a direct cause of P2 , why can’t M1 and P1 overdetermine P2? Kim nowhere invokes the overdetermination bugbear in his argument; so why is M1 epiphenomenal? To complete the argument, Kim has to condemn the option of pervasive overdetermination (“it is difficult to believe that all cases of mental causation are cases of overdetermination”), invoking Occam’s razor, so that the emergent properties are banished. Thus construed, Kim’s argument fares better; it doesn’t, however, establish that physical events are entirely determined by prior physical events. The argument doesn’t establish that emergent properties are epiphenomenal, and the emergentist may opt for overdetermination without remorse.13 If Kim wants to condemn emergent properties to an epiphenomenal existence, he has to argue against overdetermination. Elsewhere, Kim (1998) gives the following counterfactual argument against overdetermination: For consider a world in which the physical cause does not occur and which in other respects is as much like our world as possible. The overdetermination approach says that in such a world, the mental cause causes a physical event—namely the principle of causal closure of the physical no longer holds. I do not think we can accept this consequence: that a minimal counterfactual supposition like that can lead to a major change in the world. (Kim, 1998: 45)14 In essence, Kim’s argument is that the class of worlds w 1, where M1 floats free (a “nomological dangler”) and causes a physical event (hence violating closure, a cardinal sin for physicalists), is closer to the actual world than the class of worlds w 2 , where M1 has alternative physical base P1’ and closure is satisfied. This argument cuts no ice against emergentism, for—as Kim notes—emergentists have no love for physical causal closure. But it is also unclear that the argument succeeds even on its own terms, for on the standard similarity metric for evaluating closeness of worlds (from Lewis), there are grounds for claiming w 2 will be closer to the actual world than w 1, for in w 1 a fundamental emergent law would be violated as M1 has no emergent base. Another strategy that Kim might employ is to argue from theoretical economy. Kim might argue against overdetermination on the basis of

14

Hong Yu Wong

Occam’s razor—”don’t multiply entities beyond necessity”. The problem with this is that we have no clear idea of which theory is most economical independent of assuming the completeness of physics. There are three possible cases: (1) overdetermination, where M1 and P1 overdetermine P2; (2) epiphenomenalism, where M1 emerges from P1 but is epiphenomenal, and P1 causes P2; and (3) identity, where M1 is identical to P1. I see no sense in which epiphenomenalism is a more economical theory than overdetermination. Furthermore, if we withhold judgment on the completeness of physics, the epiphenomenalism scenario seems less plausible than the overdetermination scenario. Why postulate a property that does nothing? There is a fi nal argument that might be mustered (not for epiphenomenalism, but for identity): If properties are their causal powers, then, in this context, M1 and P1 have the same power (i.e., the disposition to cause P2); thus M1 and P1 are identical.15 Even if properties had certain causal powers in common, they might not share all their causal powers; if this were the case, then M1 and P1 would be non-identical. It is surely implausible to think that properties only have one causal power each; and even if each property possessed only a single causal power in each causal context, M1 and P1’s causal powers might diverge in causal contexts other than the mind-body context. In any case, Kim doesn’t argue that emergent and basal properties make the same contribution to the causal powers of things that instantiate them in all possible circumstances.16 Kim’s arguments give us no reason to think that emergent properties are epiphenomenal. This concludes the fi rst half of this chapter.

4. EMERGENCE AND SUPERVENIENCE So far we have seen no reason to think that the classical emergentist position is incoherent. Though no internal inconsistency has been demonstrated, one might wonder whether there is a tension at the heart of the view: can we square classical emergentism’s commitments to novel causal powers with supervenience? If emergent properties have novel causal influence and alter the dynamics of complex systems, then there would be effects that would not have come about (or at least would have been less likely) had the system not instantiated these emergent properties. But if that is so, then it becomes unclear why emergent properties supervene—or have to supervene—on basal properties on the classical emergentist picture. After all, if emergent properties can alter the dynamics of complex systems, why do they have to have effects in a way that ensures that there cannot be differences in the emergent properties of a system without differences in that system’s basal properties—which is what supervenience requires? So this raises the question of why emergent properties supervene on basal properties on the classical emergentist picture. I approach this question in two stages. First, I attempt to articulate the orthodox position that emergents must supervene. This is supported by

The Secret Lives of Emergents 15 arguments from Crane and Kim that are examined and rejected. At this stage, we are then left with no understanding of why emergent properties supervene on basal properties, since the reasoning behind the orthodox position is flawed. But what about the tension we were wondering about? Given the failure of the orthodox position, seeing whether and how the tension can be resolved becomes pressing for theorists who want to understand classical emergentism. Second, I describe a toy world z where it appears that emergent properties fail to supervene because of their novel causal powers. Consideration of why emergent properties fail to supervene in z helps us to see just why emergents supervene on the classical picture. The point of this entire exercise is to bring out certain constraints on the explanatory potential of the classical emergentist position. The standard ontology formulates emergence as a synchronic modal-operator strong supervenience relation that holds with nomological necessity between emergent and basal properties; call this supervenience emergentism.17 Several distinct questions in this vicinity should be distinguished. Must emergent properties supervene on basal properties (in some sense)? Can emergents supervene on basal properties? Why do emergent properties supervene, if they do? To these questions orthodoxy answers: “if they do not . . . supervene, then it seems somewhat perverse to describe the properties as ‘emergent’” (Crane, 2001: 212). But the reasons behind the orthodox declaration are opaque. The difficulty of our inquiry is exacerbated by the need to clarify what is meant by emergent in the fi rst place; emergence is a notorious philosophical term of art. We begin by considering arguments from Tim Crane and Jaegwon Kim that emergents must supervene. Tim Crane argues that novelty fails to distinguish emergent from reducible properties.18 The emergentist, he suggests, may respond with a stronger notion of ‘emergent property’: “the notion of a property of a whole whose powers are unrelated to whatever the powers of its parts are” (Crane, 2001: 212). According to Crane, this response denies mereological supervenience. But Crane’s thought is that there surely is some metaphysical relation between emergent and basal properties, since the idea of emergence is that of novel holistic qualities arising from complex basal properties; so the denial is imprudent. It is surely plausible for the emergentist who buys Dependence (the claim that mental properties are properties of physical objects)19 to also buy mereological supervenience, at least for the non-relational properties of the wholes, Crane counsels. His reasons for this are as follows: “For if they do not thus supervene, then it seems somewhat perverse to describe the properties as ‘emergent’. Presumably part of the point of this label is to pick out the sense in which putting a thing’s parts together gives you something new—but not because you have ‘added’ something ‘from the outside’. If emergentism is to be distinguished from dualism and vitalism (which do add something ‘from the outside’), then it must reject this strong notion of emergence. The upshot is that a reasonable emergentist thesis is committed to [mereological supervenience]” (Crane, 2001: 212–213).

16

Hong Yu Wong

Crane’s argument is that emergent properties have to supervene on basal properties because only that would distinguish emergentism from more radical views, such as dualism and vitalism; and since we want to set emergentism apart from these radical positions, ermergent properties must supervene on basal properties. Crane’s premises are suspect. First, supervenience fails to distinguish substance monisms from dualisms; there is no bar to Cartesian dualists, occasionalists, and Leibnizians holding supervenience. (The best way to save appearances in deterministic worlds is to supervene.) Second, there is some sense in which the supervenience of emergents is trivial; if emergent and basal properties exhibit patterns of modal dependent-variation, then, given determinism, some supervenience thesis will characterize their co-variation. We may conclude that supervenience is not forced on the emergentist so as to separate it from more radical positions; supervenience cannot do so, because it comes too cheap. In particular, this argument will not get emergent properties to supervene with nomological necessity on basal properties, as the standard ontology requires. Thus, insofar as Crane is arguing for the supervenience of emergents on the standard ontology, his argument is ineffective. It is clear that his concern is with supervenience emergentism since he argues that non-reductive physicalism and emergentism share the exact same ontological commitments and troubles (but only emergentism possesses the virtue of epistemic humility). Furthermore, concerning Crane’s thesis Dependence, substance monism doesn’t make property supervenience any more or less plausible (calling it ‘Dependence’ doesn’t help); token identity of objects doesn’t imply supervenience of emergent properties on basal properties or its failure. Crane is also too quick in thinking that the stronger notion of emergents entails the denial of mereological supervenience. On the stronger notion, emergent properties are properties of a whole which has powers unrelated to whatever the powers of its parts are. The sense of ‘unrelated’ is tricky. By ‘unrelated’ Crane must mean something stronger than independence, for he writes that “emergence is best understood in terms of the idea of emergent properties having causal powers that are independent of the causal powers of the objects from which they emerge” (Crane, 2001: 207, my italics). Since Crane thinks that emergent properties supervene on basal properties, by ‘independent’ he must mean something like ‘unrelated, except by supervenience’; but note that ‘independent’ can also mean ‘independent variation’, which amounts to no supervenience. Kim’s argument that emergents must supervene rests on similar confusion. While clarifying the notion of ‘emergence’, Kim rejects what Van Gulick calls ‘radical kind emergence’ on which emergent holistic properties are: “1. different in kind from those had by its parts, and 2. [are] of a kind whose nature and existence is not necessitated by the features of its parts, their mode of combination and the law-like regularities governing the features of its parts” (Van Gulick, 2001: 17). Kim argues that the second clause of Van Gulick’s defi nition “amounts to the requirement that an emergent

The Secret Lives of Emergents 17 property of a whole not be determined by the properties and relations characterizing its parts; that is, an emergent property of this . . . kind does not ‘supervene’ on the microstructure of an object . . . The real problem with this kind of emergence is . . . whether it is a form of emergence at all” (Kim, 2006: 192). Kim’s worry is exactly similar to Crane’s; he thinks that Van Gulick’s second clause precludes a systematic connection between basal conditions and emergent phenomena: “If the connection between pain and a neural condition is irregular, haphazard, or coincidental, and not to be relied upon, what reason could there be for saying that pain ‘emerges’ from that neural condition than another? . . . It is clear then that we must consider supervenience as a component of emergence” (Kim, 2006: 193). One way of developing the idea that emergent properties are in some sense independent is O’Connor’s notion of non-structurality. An emergent property is then a property of a composite system that is non-structural; where a property, P, is structural if and only if “proper parts of particulars having P have some property or properties not identical with P, and this state of affairs is constitutive of the state of affairs of the particular’s having P” (O’Connor, 1994: 93; modified from original defi nition in Armstrong, 1978). Non-structural emergents satisfy Crane’s stronger notion of emergence and also qualify as radical kind emergents, yet they appear consistent with supervenience. In fact, O’Connor explicitly incorporates a supervenience component in his early account of emergent non-structural properties. 20 Crane, himself, notes that strong supervenience is consistent with emergent properties having independent causal powers; and hence it is consistent with emergentism. 21 Horgan (1993: 559), too, agrees that emergents could supervene and yet possess basic causal properties. At this point, what we have is mere modal intuition. We have not yet put to rest the tension between their having novel causal powers and emergents supervening that we began this section with. In order to resolve this issue, I will develop an argument that attempts to show that emergents fail to supervene because they have novel causal powers by considering a toy world z (which we might think of as a candidate minimal classical emergentist world) that I will describe. I will use this argument as a stalking horse for unraveling the sense in which emergents supervene; in considering why supervenience fails in z we will be able to tease out why emergents supervene on the classical picture. Within a causal realist framework at least, there is a tension between holding that emergent properties supervene synchronically on their physical base properties and that emergent properties have novel causal powers. It seems that if synchronically emergent properties have novel causal powers (in a robust causal realist sense, as opposed to merely falling under different counterfactuals), then it is possible that the emergent properties will not synchronically supervene on their basal properties. Consider the following toy world z with complex object c and four possible properties of c, structural physical properties, P1 and P2 , and emergent properties M1 and M 2. P1 is M1’s emergence base and P2 is M 2’s. Two

18

Hong Yu Wong

probabilistic laws govern all causation in the toy world Prob(M1 → M 2) = ½ and Prob(P1 → P2) = ½. This situation should be plausible if M1 has novel causal powers. (I take it that something along the lines of these causal laws is about as weak an interpretation as possible of emergent properties having novel causal powers. Emergent properties should at least influence the probabilities of subsequent events if they possess novel causal powers. In z, emergent properties only have intra-level causal influence.) Assume also that when M1 doesn’t cause M 2 , then c continues to instantiate M1, and ditto for P1 (immanent causation). At time t1, complex object c instantiates structural property P1 and emergent property M1. Four scenarios are possible at t2: (1) P1 at t1 causes P2 at t 2 , and M1 is inactive; thus c instantiates M1 and P2 . (2) Both M1 and P1 are inactive. (3) M1 at t1 causes M 2 at t 2 , and P1 at t1 causes P2 at t 2 . (4) M1 at t1 causes M 2 , and P1 is inactive; thus c instantiates M 2 and P1. Scenarios (1) through (3) are unproblematic; but since scenario (4) involves an emergent difference without a physical difference, we may conclude that, in z, emergent properties don’t in general supervene on their basal properties. Supervenience emergentists might object that emergent and basal causal propensities are not wholly independent. But the indeterministic case is interesting because it is consistent with linked propensities, and yet different actual outcomes. Perhaps non-epiphenomenal supervenience emergentists have to say that the probabilistic intra-level laws governing them must be consistent with the impossibility of M 2 without P2 , and reject certain permutations of the outcomes as impossible. Also, the ruling out of M 2 without P2 follows naturally from the (alternative) downward causation picture, in which P1 always causes M 2 by causing P2 (so the laws have to be inter-level). However, it seems that if we add robust realism about causal relations, there remains the question of which of P1 or M1 caused a given subsequent P2 , which will complicate the assignments of probabilities. 22 That emergents fail to supervene on basal properties in z is perhaps unsurprising. The setup in z with the two probabilistic laws governing the evolution of states is consistent with emergent and basal properties being completely independent, as the supervenience emergentist feared. Let the P properties be shapes and M properties colors. A square at t1 has a 50 percent chance of causing a circle at t 2 , and a red thing at t1 has a 50 percent chance of causing a blue thing at t 2 . This won’t force colors to supervene on shapes or the other way round. So we might ask: what does emergence add here? Notice also that contrary to classical emergentism, emergence bases in z fail to even suffice for the respective emergent properties. Rather than z raising questions about the status of emergence bases, we should reject the claim that z constitutes a minimal classical emergentist world. What are missing in z are the appropriate emergent laws that specify fundamental

The Secret Lives of Emergents 19 nomological connections between emergent and basal properties which function as metaphysically fundamental supervenience principles. This is exactly what we should expect on the classical ontology. Broad tells us that if emergentism is correct, then: We should have to recognize aggregates of various orders. And there would be two fundamentally different types of law, which might be called ‘intra-ordinal’ and ‘trans-ordinal’ respectively. A trans-ordinal law would be one which connects the properties of aggregates of adjacent orders. A and B would be adjacent, and in ascending order, if every aggregate of order B is composed of aggregates of order A, and if it has certain properties which no aggregate of order A possesses and which cannot be deduced from the A-properties and the structure of the B-complex by any law of composition which has manifested itself at lower-levels. An intra-ordinal law would be one which connects the properties of aggregates of the same order. A trans-ordinal law would be a statement of the irreducible fact that an aggregate composed of aggregates of the next lower order in such and such proportions and arrangements has such and such characteristic and non-deducible properties. (1925: 77–78) What Kim has missed in arguing that radical kind emergents cannot supervene is that basal properties and relations of parts of the complex object and any intra-ordinal laws (“law-like regularities governing the features of its parts” in Van Gulick’s defi nition) do not necessitate the trans-ordinal laws, which vouchsafe supervenience. 23 That the supervenience of emergents on the standard ontology is secured by emergent laws is clearly brought out by the second clause of McLaughlin’s defi nition of property emergence: If P is a property of w, then P is emergent if and only if (1) P supervenes with nomological necessity, but not with logical necessity, on properties the parts of w have taken separately or in other combinations; and (2) some of the supervenience principles linking properties of the parts of w with w’s having P are fundamental laws. [A law L is a fundamental law if and only if it is not metaphysically necessitated by any other laws, even together with initial conditions.] (1997: 39) This point can be illustrated even in non-standard emergentist ontologies. For Paul Humphreys (1997), emergent properties are produced by singular causal interactions (‘fusions’), so it is not necessary that there are emergent laws akin to C. D. Broad’s trans-ordinal laws. Given that, one can see why it is not necessary that the class of emergents characterized by fusion must supervene, for the supervenience of emergents on the standard ontology is secured by fundamental trans-ordinal laws, and these are not present generally in the fusion cases.24 I have belabored this point about why emergents supervene on the classical ontology only because recent commentators seem to have forgotten this.

20

Hong Yu Wong

5. CONCLUSION We began this chapter by defending the classical emergentist position against Kim’s arguments. There we saw no reason, even when we extended Kim’s arguments, for thinking that the classical emergentist position is incoherent. However, we had lingering suspicions that there might be a tension at the heart of the view between emergent properties possessing novel causal powers and their having to co-vary with basal properties in such a way that there could be no emergent differences without basal differences. The moral of our discussion in the second half of this chapter is that emergent properties don’t supervene on basal properties simpliciter; they supervene given the fundamental emergent laws. Fixing the basal level doesn’t fi x the higher level; fi xing the basal level and the emergent laws fi xes the emergent level. There is, however, a caveat. Emergent properties are non-structural properties, in contrast to resultant complex properties which are structural properties; but both supervene on basal properties. Since emergent properties are non-structural, supervenience in the case of emergent properties must be sui generis; it is not a matter of constitution, identity, realization, causation, or any of the usual relations that ground supervenience; it is a matter of fundamental, non-derivative emergent laws that must be “simply swallowed whole with that philosophic jam which Professor Alexander calls ‘natural piety’” (Broad, 1925: 55). Given this, the explanatory potential of supervenience emergentism is limited to situations involving deeply unexplainable co-variation. They key thing to note is that more flexible interactionist scenarios are precluded since fundamental emergent laws restrict the patterns of co-variation of emergents and bases, and any emergent causation must be consistent with these. (This is the tension that we identified in discussing z.) I now conclude my defense of supervenience emergentism. It is surely consistent and possible, but at the price of the eternal mystery of a sui generis metaphysical determination relation and deep explanatory conservatism—caveat emptor.

NOTES * I am indebted to Tim Crane, Jaegwon Kim, Barry Loewer, Brian McLaughlin, Jonathan Schaffer, and especially Timothy O’Connor and Dean Zimmerman for their feedback on earlier drafts, dating back to 2002. Thanks also to audiences at Brown, Tucson, Paris, and Budapest for their reactions. More recently, in preparing the fi nal version, I have benefited from the counsel of Eleanor Knox and Gerald Vision. 1. Kim (1999). All references are to this paper unless otherwise stated. 2. Note that counterfactuals, unlike nomological sufficiency relations, are not transitive. This may not make a difference in the particular case at issue, however. I have altered Kim’s original notation, where temporal subsequence is expressed by starring the original variable, to a notation with numerical

The Secret Lives of Emergents 21

3.

4.

5.

6.

7.

8.

9.

indexes tied to times of instantiation in order to preserve notational consistency throughout this chapter. The causal-power actuality principle: “For an object, x, to exercise, at time t, the causal/determinative powers it has in virtue of having property P, x must already possess P at t. When x is caused to acquire P at t, it does not already possess P at t and is not capable of exercising the causal/determinative powers inherent in P” (29). “We say that superfluidity, ferromagnetism, metallic conduction, hydrodynamics, and so forth are “protected” properties of matter—generic behavior that is reliably the same one system to the next, regardless of details. There are more sophisticated ways of articulating this idea, such as stable fi xed point of the renormalization group, but these all boil down to descriptions of perturbations of the underlying equations of motion” (Laughlin et al., 1999: 32). It is unclear if Laughlin et al. have an epistemological or an ontological notion of emergence. See also what Wimsatt calls “dynamical autonomy” in complex systems (Wimsatt, 1995). Kim identifies this as the fi fth doctrine of emergentism: “The causal efficacy of emergents: Emergent properties have causal powers of their own—novel causal powers irreducible to the causal powers of their basal constituents” (23). Kim (1989 and elsewhere). More recently, Kim has formulated the Principle of Causal Exclusion thus: “If an event, e, has a sufficient cause, c, at t, no event at t distinct from c can be a cause of e (unless this is a genuine case of causal overdetermination)” (Kim, 2001: 276). Of course there is a question about what counts as ‘independent’. My counterexample is not unlike what Kim calls “supervenient causation”. Similar problems have been noted by a number of commentators; see, e.g., Crisp and Warfield (2001). Kim’s response (2001) to these objections is to extend the exclusion principle to encompass all instances of generation, causal or otherwise. Kim’s driving metaphysical intuition is that “a given thing or event can have at most one generative provenance, only one source from which it derives its being”. Kim attempts to capture this in the following principle: Principle of Determinative/Generative Exclusion: If the existence of an event e, or an instantiation of a property P, is determined/generated by an event c—causally or otherwise—then e’s occurrence is not determined/ generated by any event wholly distinct from or independent of c (unless this is a genuine case of overdetermination). I doubt that Kim’s reasoning here can be saved even if we embrace the stronger “generative exclusion principle”. It is unclear what counts as a “wholly distinct” or “independent” event. Also, though I can just about make sense of causal realist intuitions concerning oomph, non-causal oomph doesn’t make much sense. Kim now prefers to rest the argument on intuition rather than on any exclusion principle (though he does have a new dictum, after Jonathan Edwards, concerning the tension between ‘vertical’ determination and ‘horizontal’ causation); see Kim (2003). For the text of Kim’s argument, see the quotation in the fi rst section of this chapter. Events cause other events; that is, property instantiations at a time cause other properties to be instantiated at a time. I speak of properties causing other properties only for convenience. Kim quotes Lloyd Morgan’s Emergent Evolution. Elsewhere, Kim has construed the relationship between the emergence base and the emergent property as one of either physical realization or supervenience (Kim, 1996: 232). Brian McLaughlin has remarked to me in conversation that the relationship between an emergence base and an emergent property is, in Morgan’s sense,

22

10.

11.

12.

13.

14.

15. 16.

17. 18.

Hong Yu Wong defi nitely not strong supervenience with metaphysical necessity. Nor is it realization as is typically understood by contemporary physicalists. See Spurrett and Papineau (1999). I fi nd Spurrett and Papineau’s non-mental characterization unsatisfactory, for the completeness of physics is in a sense analytic, and serves not to constrain physical causes (or principles such as the conservation of energy), but dictates the expansion of the domain of physics. It is thus preferable to level an inductive argument against emergent entities (cf. Papineau, 2001). Invocations of completeness of physics in this chapter, however, are to be understood in Spurrett and Papineau’s (and Kim’s) nonanalytic sense. “Most emergentists will have no problem with the failure of the physical causal closure; although they may have to tinker with their doctrines somewhere to ensure the overall consistency of their position, they are not likely to shed any tears over the fate of the closure principle. For many emergentists that precisely was the intended consequence of their position” (Kim, 1993: 356); see also Kim (1996: 232). Notice that the argument does not employ the premise of physical causal closure, but instead relies on arguing that P1 (the emergence base of M1) has a good claim to be a cause of P2 (the emergence base of M 2) to generate overdetermination. The main problem is that though P1 has a good claim to be a cause of P2 , it is not obvious that P1, unlike M1, qualifies as a direct cause of P2 . Presumably only (fairly) direct causes can qualify as symmetrical overdeterminers of an effect; otherwise, we would have massive overdetermination in the physical realm anyways from causes of direct causes, and causes of those causes, and so on, undermining the ‘no systematic overdetermination in the actual world’ clause. Metaphysicians might wonder why this counts as overdetermination since M1 and P1 are not independent, unlike the case of two assassins simultaneously shooting at target X. Assuming causal realism, it is not implausible to view M1 and P1 as distinct causes even if they are not independent, since they are distinct properties (presumably, emergence is not identity, thus M1 ≠ P1) and hence constitute different events (where events are property instantiations at a time). Kim’s argument against overdetermination is part of his Supervenience Argument against non-reductive physicalism (NRP) (Kim, 1998: 38–46). I am here exploring different ways Kim might argue and don’t mean to imply that Kim would use this argument against overdetermining emergent properties. But I want to stress that Kim needs to argue against overdetermination— even in the case of emergence. Elsewhere Kim has expressed his sympathies for a causal theory of properties à la Shoemaker (1980). See Kim (1998: 56 and 129, fn. 46). Here is another route to non-identity: M1 and P1 differ in their modal properties, each figuring in different counterfactuals; thus, M1 ≠ P1 (by Leibniz’s law). Theorists who claim constitution is weaker than identity frequent this route. This raises questions about property identity that will take us too far afield. See, e.g., Kim (1999), McLaughlin (1997), O’Connor (1994), Van Cleve (1990). Note that the standard ontology is typically supplemented with a clause about novel causal powers. On his defi nition of novelty, modified from Spencer-Smith (1995), which is: a property P is novel in x if x has P, and there are no determinates P’ of the same determinable as P, such that all constituents of x have P’ (Crane, 2001: 211). As the defi nition is meant to be weak enough not to exclude properties like wetness and liquidity, its failure to uniquely characterize emergents is unsurprising.

The Secret Lives of Emergents 23 19. The use of ‘Dependence’ to denote a substance monist thesis is somewhat odd. Crane’s minimal characterization of emergentism consists of Dependence and Distinctness theses; the latter asserts that mental properties are distinct from physical properties. 20. For O’Connor, a property P is an emergent property of a (mereologically complex) object o iff: (1) P supervenes on properties of the parts of o; (2) P is not had by any of the object’s parts; (3) P is distinct from any structural property of o; and (4) P has direct (downward) determinative influence on the pattern of behaviour involving o’s parts (1994: 98). 21. The remark is odd given that he earlier argued that emergents must supervene. 22. Note that a variant of the argument may not be applied to physicalism; while probabilistic physicalism allows for indeterministic evolution of states, the entailment relations are not at all chancy. By physicalism, I mean the thesis that any minimal physical duplicate of our world is a duplicate simpliciter (Loewer 2001). 23. That is not to say that there could not be radical kind emergents which do not supervene. But again there are different senses of supervenience. There might be radical emergents which supervene in a nonclassical sense (e.g., diachronically), and it is not difficult to conjure radical emergents that can run modally amok. The latter are caused by their ‘bases’ and then lead separate lives. 24. See the Appendix to Wong (2006) for further discussion of this point.

REFERENCES Armstrong, D. (1978). Universals and Scientific Realism, vols. 1–2. London: Macmillan. Broad, C. D. (1925). The Mind and Its Place in Nature. London: Routledge and Kegan Paul. Crisp, T., & Warfield, T. (2001). Kim’s master argument. Noûs, 35(2), 304–316. Crane, T. (2001). The significance of emergence. In C. Gillett & B. Loewer (Eds.), Physicalism and Its Discontents (pp. 207–224). Cambridge: Cambridge University Press. Hogan, T. (1993). From supervenience to superdupervenience: meeting the demands of a material world. Mind, 102: 555–586. Humphreys, P. (1997). How properties emerge. Philosophy of Science, 64, 1–17. Kim, J. (1973). Causation, nomic subsumption, and the concept of event. In Supervenience and Mind (pp. 3–21). Cambridge: Cambridge University Press. . (1989). Mechanism, purpose, and explanatory exclusion. In J. Kim, Supervenience and Mind (pp. 237–264). Cambridge: Cambridge University Press. . (1992). “Downward causation” in emergentism and nonreductive physicalism. In A. Beckermann, H. Flohr, & J. Kim (Eds.), Emergence or Reduction? (pp. 119–138). Berlin: de Gruyter. . (1993). Supervenience and Mind. Cambridge: Cambridge University Press. . (1996). Philosophy of Mind. Boulder, CO: Westview Press. . (1998). Mind in a Physical World. Cambridge, MA: MIT Press. . (1999). Making sense of emergence. Philosophical Studies, 95, 3–36. . (2001). Mental causation and consciousness: the two mind-body problems for the physicalist. In C. Gillett & B. Loewer (Eds.), Physicalism and Its Discontents (pp. 271–283). Cambridge: Cambridge University Press.

24

Hong Yu Wong

. (2003). Blocking causal drainage, and other chores with mental causation. Philosophy and Phenomenological Research, 67(1), 151–176. . (2006). Being realistic about emergence. In P. Clayton & P. Davies (Eds.), The Re-Emergence of Emergence (pp. 189–202). Oxford: Oxford University Press. Laughlin, R. B., Pines, D., Schmalian, J., Stojkovic, B. P., & Wolynes, P. (1999). The middle way. Proceedings of the National Academy of Sciences, 97(1), 32–37. Loewer, B. (2001). From physics to physicalism. In C. Gillett & B. Loewer (Eds.), Physicalism and Its Discontents (pp. 37–56). Cambridge: Cambridge University Press. McLaughlin, B. (1997). Emergence and supervenience. Intellectica, 1997(2), 25, 25–43. O’Connor, T. (1994). Emergent properties. American Philosophical Quarterly, 31, 91–104. Papineau, D. (2001). The rise of physicalism. In C. Gillett & B. Loewer (Eds.), Physicalism and Its Discontents (pp. 3–36). Cambridge: Cambridge University Press. Shoemaker, S. (1980). Causality and properties. Reprinted in H. Mellor and A. Oliver (Eds.), Properties (pp. 228–254). Oxford: Oxford University Press. Spencer-Smith, R. (1995). Reductionism and emergent properties. Proceedings of the Aristotelian Society, 95, 113–129. Spurrett, D., & Papineau, D. (1999). A note on the completeness of physics. Analysis, 59(1), 25–29. Van Cleve, J. (1990). Mind-dust or magic? Panpsychism versus emergence. Philosophical Perspectives, 4, 216–226. Van Gulick, R. (2001). Reduction, emergence and other recent options on the mind/body problem: a philosophic overview. Journal of Consciousness Studies, 8(9–10), 1–34. Wimsatt, W. (1995). The ontology of complex systems. Canadian Journal of Philosophy, 20, 564–590. Wong, H. Y. (2006). Emergents from fusion. Philosophy of Science, 73, 345–367.

2

On the Implications of Scientific Composition and Completeness Or, The Troubles, and Troubles, of Non-Reductive Physicalism Carl Gillett

1. INTRODUCTION One of grandest contemporary battles in the sciences, continuing a bitterly fought debate that has raged over the last half century, is between what I will roughly term ‘scientific reductionists’ and ‘scientific emergentists’. At the core of scientific reductionism is the idea that higher-level entities are all composite entities which are ‘nothing but’ their components—framed in a slogan: ‘Wholes are nothing but their parts’. Scientific reductionists, like Steven Weinberg (1994), thus press a special distinction for fundamental physics and “dream”, to use Weinberg’s term, of a ‘Final Theory’: a simple set of laws governing the behaviors of the fundamental entities in simple aggregations which are claimed, in principle, to provide the complete account of everything. In contrast, scientific emergentists have pressed conflicting views of scientific composition. Researchers like Philip Anderson (1972), Stuart Kauffman (1995), Ilya Prigogine (Prigogine & Stengers 1996), or Robert Laughlin (2005), most famously accept positions that are sloganized in the claim that ‘Wholes are more than the sum of their parts’, but less widely noticed is their equally distinctive claim that ‘Parts behave differently in wholes’. Given these commitments, scientific emergentists claim that some composed entities are what I term ‘Strongly’ emergent in being composed entities which are still efficacious in the world through their determinative influence upon their components.1 Scientific emergentists consequently argue that a range of higher-level sciences studying Strongly emergent entities, and not just particle physics, are needed to explore the “frontiers of science”. Oddly, although a large number of researchers in the sciences are ‘reductionists’, in philosophy ‘reduction’ is now notoriously a dirty word. Similarly, despite ever-increasing numbers of scientists professing to be ‘emergentists’, few philosophers espouse ‘emergentism’ and most dismiss such positions as incoherent. One might not think this disconnect between scientists and philosophers is so problematic, but I intend to argue that it has damaged the scientific debates, leaving them without crucial theoretical aid, and has left philosophers missing the deeper

26

Carl Gillett

issues, and the strongest positions, about the overall structure of nature and the sciences. My strategy in illuminating these claims is to vary our approach in two ways. First, I address the issues through a pursuit of the ‘metaphysics of science’, that is the abstract examination of ontological issues as they arise within the sciences and real scientific cases. And, second, I use this stance to follow the lead of the scientists to explore the arguments and positions these writers have recently articulated about ‘reduction’ and ‘emergence’. To this end, in Section 2, I begin by looking at a concrete scientific case of a kind which provides common ground to both sides. Each group endorses the importance of so-called ‘inter-level mechanistic explanations’ that explain higher-level entities in terms of the lower-level entities scientists take to compose them. Using our scientific case, I draw out the scientific concepts of composition, and their features, that underpin these explanations. (I should note that throughout the paper by ‘composition’ I mean only the concepts posited in such inter-level mechanistic explanations, and I thus use the term in the narrower scientific usage that stretches back at least to Sir Isaac Newton.) Using the resulting framework for scientific composition, in Section 3, I highlight the obvious kind of argument that uses scientific composition to drive the reductionist claim that ‘Wholes are nothing but their parts’. I frame the simplest form such reasoning takes, in the ‘Argument from Composition’, and the resulting reductionism, in what I term ‘Compositional reduction’. This scientific reductionism is ontological in nature and I note how it is very different from the forms philosophers of science have focused upon, as well as being resistant to recent philosophical critiques of ‘reduction’. Philosophical responses fail against Compositional reductionism, but I show in Section 4 that scientific emergentists have had more promising ideas. I highlight the centrality of the scientific emergentist’s distinctive view of aggregation, in what I term the ‘Conditioned view’, that I show underlies their claim that ‘Parts behave differently in wholes’. Using a scientific thought experiment, I then show that Strong emergence is possible when we have Conditioned aggregation and that the Argument from Composition is actually invalid. Though these are important results, I outline another argument for Compositional reductionism, in Section 5, built around a previously implicit thesis that is a precondition for the existence of the reductionist’s Final Theory and related claims such as the so-called ‘Completeness of Physics’. This thesis is the common view of aggregation that I term the ‘Simple’ view and I show how it underpins a valid defense of Compositional reductionism. However, I note that scientific emergentists will challenge the soundness of this argument since they reject the Simple view of aggregation. Our work thus illuminates that recent scientific debates center around differing views of the kind of aggregation, varieties of determination, and

Scientific Composition and Completeness Implications 27 components that we fi nd in nature. And, in concluding the chapter, I connect to recent philosophical discussions by using our fi ndings to show how philosophers have fallen into a false dichotomy of options where we really have a trichotomy of positions. For the Strong emergentism of working scientists has been overlooked and offers a viable version of non-reductive physicalism. In contrast, our work establishes that what I term the ‘Standard’ non-reductive physicalism endorsed by philosophers is such that we cannot even imagine how it could be true. Reflection on recent scientific views thus leads us to the deeper issues over ‘reduction’ and ‘emergence’ in both the sciences and philosophy.

2. SCIENTIFIC COMPOSITION AND ITS FEATURES Common ground for both sides in the scientific debates is the importance of mechanistic explanation and I am going to use a concrete example to illustrate the concepts of composition used in such explanations. I therefore need a metaphysical framework to guide my work. Since our primary focus will be on properties, I follow recent work in the metaphysics of science and assume a weakened version of Shoemaker’s ‘causal theory’ of properties under which a property is individuated by the causal powers it potentially contributes in this world, under certain conditions, to the individuals in which it is instantiated. 2 For reasons that will become apparent, I am primarily concerned with properties I term ‘causally efficacious’—properties whose instantiation actually determines the contribution of causal powers to an individual.3 With this machinery in place, consider a prominent example of mechanistic explanation in the neurosciences. We know that, under appropriate background conditions, a voltage-sensitive potassium ion channel plays a key role in a neuron due to its property of being a voltage-sensitive gate contributing the backward-looking power of opening in response to a change in the charge of surrounding cells. Roderick MacKinnon won the Nobel prize for developing a compelling mechanistic explanation of how such ion channels do this by illuminating the relevant chemical and spatial properties/relations of the complex protein molecules that are ‘sub-units’ (i.e., parts) of these channels. Figure 2.1 provides a simplified illustration of the higher- and lower-level mechanisms whose implementation relations such fi ndings allow us to understand. Basically, when the charge in surrounding cells changes, then the backward-looking powers, to change relative spatial position, of each of the sub-units are manifested together, and the sub-units all swivel to adopt new spatial relations to each other. These numerous lower-level mechanisms together implement the higher-level process of the ion channel opening. And, in addition, we have a range of other compositional relations: we have part-whole, and constitution, relations between the ion channel and

28

Carl Gillett

Figure 2.1 An ion channel and its protein sub-units.

protein sub-units, so-called ‘realization’ relations between the properties of these individuals, and what I term ‘comprising’ between their powers. To better understand these various compositional notions, let me note a few of their common features. First, even a cursory inspection shows that the compositional relations posited between powers, properties, individuals, and processes are all synchronous, do not occur between wholly distinct entities, and do not involve the mediation of force and/or the transfer of energy. I thus use the term ‘non-causal determination’ to refer to the compositional relations involved in mechanistic explanations. Second, the entities bearing compositional relations in this case, and the many like it in the sciences, are usually qualitatively distinct. Thus, for instance, the properties and relations of the protein sub-units contribute no common powers with the property of the ion channel—the protein subunits have powers such as changing their relative spatial positions, under certain conditions, but not the powers of opening or allowing speedy passage to potassium ions. It is very common for higher-level properties to be ascribed such ‘novel’ powers not had by any of their realizers and this commonplace feature of compositional relations underpins what we may term ‘qualitative’ emergence—properties of higher-level individuals not had by the lower-level individuals which are their parts. It may initially be puzzling how entities of one kind could compose entities of very different kinds; consequently, it is important to note a third

Scientific Composition and Completeness Implications 29 general feature of such compositional relations. Given their interconnections, the various ‘packages’ of powers, properties, individuals, and mechanisms studied by lower-level sciences together compose the qualitatively different powers, property instances, individuals and mechanisms studied by higher-level sciences. Compositional relations in the sciences are thus often ‘many-one’, and although each is qualitatively distinct from the composed entity, the deeper point is that numerous lower-level component entities can together non-causally result in the qualitatively different entity they are taken to compose.4 To further sharpen our understanding of scientific notions of composition we could present more precise accounts of each compositional relation we fi nd in such cases. But let us limit ourselves to a brief account of realization relations, since this will aid our later discussion. This is what I have elsewhere termed ‘Dimensioned’ realization (Gillett, 2002a) which we may frame in this simple ‘thumbnail’ account: (Realization) Property/relation instance(s) F1–Fn realize an instance of a property G, in an individual s under condition $, if and only if s has powers that are individuative of an instance of G in virtue of the powers contributed, under $, by F1–Fn to s or s’s constituent(s), but not vice versa. Amongst other features, the Dimensioned view allows for both the qualitative distinctness of realizer and realized properties and their being instantiated in distinct individuals, and also for the many-one character of scientific realization. For the account earns its name because we have ‘dimensions’ in the form of distinct lower- and higher-level powers, properties and individuals, as well as the distinct processes they ground. Though this account of realization will be useful, I should note that any successful account of scientific composition must accommodate the three general features just highlighted. In the next section I argue that such features suffice to drive the type of scientific reductionism I focus upon, so the resulting reductionism does not depend upon the details of my view of realization. (Similar points also hold about my later discussion of Strong emergence.) Let us therefore turn to the reductionism we fi nd in the sciences.

3. ONTOLOGICAL REDUCTIONISM AND THE ARGUMENT FROM COMPOSITION Using our observations about scientific composition, we can now return to the intuition of reductionist scientists that by illuminating how certain higher-level entities are composed we thereby show that such entities are really ‘nothing but’ their lower-level components. Central to this intuition is the feature of a mechanistic explanation that given the nature of the components we find in an aggregation we can account for all the powers and properties of

30

Carl Gillett

individuals at the lower and higher levels. But scientific reductionists suggest, after reflecting upon the nature of these explanations, and the compositional relations they posit, we can therefore see that we should not be committed to both component and composed entities. For example, we should no longer accept both ion channels and also protein sub-units, or accept the property of being a voltage-sensitive gate and also the chemical properties and relations of the proteins. Our mechanistic explanations show we can account for the truth of the relevant special science statements commonly understood to refer to such composed entities by simply taking the components to be their sole truthmakers. Thus, given our mechanistic explanations, one can, for instance, account for the powers of the ion channel’s properties using the powers contributed by the properties and relations of the protein sub-units, but not vice versa. As a result, the reductionists conclude that we should only take the component entities to exist, since reflection upon our mechanistic explanations shows these are the only entities we need to account for all the powers of individuals. Here we have a intuitive form of ontological parsimony reasoning driven by the nature of mechanistic explanations, and the compositional relations they posit, to the conclusion that there really are no composed entities—that ‘Wholes are nothing but their parts’. And we should note that such reasoning works with qualitatively emergent entities. Such qualitative emergence goes along with the composition which is the engine of such reductionist reasoning—thus scientific reductionists actually embrace qualitative emergence to run their arguments. As we shall see, there are a number of ways one can more precisely articulate such reasoning, but let us start with a formulation that simply attempts to use the nature of compositional relations alone. Since we have an account of realization between properties, I focus on a version of such an argument based around realization. But we can produce similar reasoning focused on the comprising of powers, constitution of individuals, or implementation of processes. Given the nature of realization in the sciences, a scientific reductionist may argue that it is ontologically profligate to take any realized property instance to determine the contribution of causal powers, and hence to be causally efficacious, in addition to its realizer property instances. For the reductionist claims that, given the nature of the realization relation, we can account for all the causal powers of individuals simply using the contributions of powers by the realizer property instances of these individuals, or their constituents, rather than also as contributions from realized property instances. But we cannot account for all causal powers of individuals simply as contributions by realized property instances. If we assume that the causal powers of individuals are not overdetermined, then appealing to Occam’s razor the reductionist argues that we should accept the existence of no more causally efficacious property instances than we need to account for the causal powers of individuals. The proponent of this simple argument thus concludes that we should only accept that realizer property

Scientific Composition and Completeness Implications 31 instances contribute powers to individuals and hence only accept that realizer property instances are causally efficacious. But it is also plausibly true that the only property instances that exist are those that make a difference to the causal powers contributed to individuals. We may thus further conclude that there are only realizer property instances. Let us call this the ‘Argument from Composition’. And from this point on I will use the terms ‘ontological’ or ‘Compositional’ reductionism to refer only to the brand of reduction that utilizes the Argument from Composition or the analogous parsimony argument I detail in Section 5, whether focused on powers, individuals, processes, or properties. It should be obvious that Compositional reductionism is rather different from earlier understandings of scientific reductionism, since the Argument from Composition is not focused on theories or law statements, and their relations, like the semantic reductionism of the Positivists. Instead, Compositional reductionism argues to an ontological conclusion about the putative implications of the compositional relations between entities posited by mechanistic explanations. The Argument from Composition, and the Compositional reductionism it supports, are thus, as Steven Weinberg puts it nicely, “not concerned . . . so much with what scientists do, . . . as with the logical order built into nature itself” (Weinberg, 1994: 45–46). Ultimately, I contend that Compositional reductionism is best understood as a combination of ontological reductionism and semantic, or ‘theory’, antireductionism in what Weinberg terms a “compromising” reductionism.5 This brand of scientific reductionism therefore poses a far stiffer challenge than the ‘reduction’ broached by philosophers, since it can embrace both the indispensability of higher sciences and their predicates, as well as the phenomenon of multiple realization, to turn aside the main philosophical objections to ‘reduction’.6 A great deal more can be said about the commitments of such a position on ontology, laws, theories, and methodology, but let us instead turn to scientific emergentism which, unlike mainstream philosophy, offers more promise of an effective response to Compositional reduction.

4. CONDITIONED AGGREGATION AND ITS IMPLICATIONS: THE EMERGENTIST’S NOVEL VIEW OF COMPOSITION AND STRONG EMERGENCE Our work reconstructing scientific reductionism allows us to understand the overwhelming opposition to the very possibility of Strong emergence. The Argument from Composition, and related reasoning, is taken to show that, like square-circles, Strongly emergent properties offer a combination of properties we will never fi nd. However, just like scientific reductionism, scant philosophical attention has been paid to the ideas that scientific emergentists offer about Strong emergence. Focusing on empirical research

32

Carl Gillett

in a range of areas, writers like Anderson, Kauffman, Prigogine, Laughlin, and a number of others, have given us interesting reasons of just the kind needed to speak to the challenge of scientific reductionism.7 To illustrate these ideas, I focus on the work of Laughlin, the Nobel prize-winning condensed matter physicist, since his discussion is the most recent and detailed in nature. The concrete cases that Laughlin examines are unsurprisingly drawn from condensed matter physics, but he claims similar points apply to phenomena in chemistry, biology, and other areas. Like Anderson, Laughlin focuses primarily on examples of symmetry breaking, and he explains what he sees in such cases when he tells us: The idea of symmetry breaking is simple: matter collectively and spontaneously acquires a property or preference not present in the underlying rules themselves. For example, when atoms order into a crystal, they acquire preferred positions, even though there was nothing preferred about these positions before the crystal formed. When a piece of iron becomes magnetic, the magnetism spontaneously selects a direction in which to point. These effects are important because they prove that organizational principles can give primitive matter a mind of its own and empower it to make decisions . . . Symmetry breaking provides a simple convincing example of how nature can become richly complex all on its own despite having underlying rules that are simple. (Laughlin, 2005: 44) There is a lot to unpack in this passage. First, Laughlin implies that concrete cases of ‘emergence’ are to be found where we now know that the laws that hold of components in simpler aggregations cannot alone exhaustively determine the behavior, and powers, of the components in more complex aggregations. Thus ‘Parts behave differently in wholes’ is true of cases of symmetry breaking and so, second, we need to posit further fundamental laws, or “organizational principles”, to explain the novel behaviors and powers of the components. Third, we should note Laughlin’s claim that the composed entity “spontaneously”, or synchronously, influences the features of its components. Thus we have a composed entity that is still efficacious and hence Strongly emergent. Notice that, given his claims about components, Laughlin is espousing a novel picture of aggregation and, as a result, an original view of the nature of components as they aggregate in certain complex, composed entities. Crucially, rather than components obeying the same laws in all aggregations, and hence contributing the same powers and behaving in the same ways across all aggregations, Laughlin suggests that our empirical fi ndings establish aggregation is sometimes rather different. In the aggregations we fi nd in complex composed entities, Laughlin claims, components sometimes contribute powers, and behave differently, than the laws governing simpler aggregations would alone determine. Thus, for example, atoms have

Scientific Composition and Completeness Implications 33 preferred positions in crystals, says Laughlin, which the laws holding of simpler aggregations do not determine. These different powers do not violate the range of powers allowed by the lower-level laws applying in simpler aggregations, but Laughlin claims our empirical fi ndings show that such laws also do not determine the further ‘conditioned’ powers and behaviors we fi nd in the complex aggregations where new laws are operative.8 Let us call this the ‘Conditioned’ view of aggregation. Lastly, we must mark how this view of aggregation allows for the existence of the determination that Laughlin claims between the composed entity and its components, for it is the novel powers of the components that Laughlin contends that the composed entity determines. But how do these strikingly different claims, about aggregation, determination, and components, help to establish the possibility of Strong emergence? The glimmer of a possibility with Conditioned aggregation is that it allows that component entities, like lower-level realizer properties, only contribute certain powers when aggregating into and composing a certain ‘whole’ such as a realized property. As a result, where we have Conditioned aggregation a realized property instance can be efficacious not by contributing powers itself, but by determining the contributions of powers by other property instances in its own realizers (i.e., its own components). Though composed, such a realized property would still determine the powers of individuals and be efficacious. Scientific emergentists like Laughlin, through their Conditioned view of aggregation, may thus have illuminated a missed option, and the suspicion is obviously that the reductionist’s Argument from Composition is invalid. To better evaluate whether these ideas establish the possibility of strong emergence, we now need to more carefully examine whether we can reconstruct these insights in a coherent fashion. To assess such theoretical claims, scientists like Einstein or Galileo often use a kind of thought experiment where well-confi rmed scientific concepts apply and then one considers what is, and is not, possible. In effect, we have seen that scientific reductionists deny that concepts of scientific composition, when they comprehensively hold of some situation, allow Strong emergence. I am therefore going to explore a scientific thought experiment which applies our framework for scientific concepts of composition to a scenario of the type envisioned by scientists like Laughlin in order to see whether we can have Strong emergence when concepts of scientific composition comprehensively apply. I should emphasize that I am considering a scientific thought experiment and not considering a real case. To reinforce the latter point, let us therefore name the individuals we look at in our thought experiment ‘blobs’ and ‘globules’. I shall assume that blobs are individuals that can exist in isolation or in simple aggregations, as well as in complex aggregations. In simple aggregations, blobs bond with each other in a certain way and also have the property, call it P1, in a certain kind of weak force exerted on other blobs that contributes a range of powers on individuals. In simple aggregations,

34

Carl Gillett

let us assume, the property P1 contributes to blobs, amongst other powers, that of weakly attracting other blobs in some random direction. In addition, let us assume that when aggregated in the right way in simple aggregations, then blobs constitute the individuals I am terming ‘globules’. The simple globules that result have a range of realized properties, including the realized property of exerting a weak attractive force on globules in all directions—for this is simply the attractive force that results from the powers together of all the instances of property P1 found in the blobs that constitute such a simple globule. The situation in simple aggregations of blobs and globules is outlined in Figure 2.2. In addition, of course, we are also stipulating that the scenario involves Conditioned aggregation. So let us further stipulate that at a certain point in the aggregation of blobs we get complex globules with the property of exerting a very strong attractive force on globules in a particular direction. Let us call this property of complex globules ‘H’. Furthermore, let us also stipulate that when a globule has the property H, then the instances of property P1 in its constituent blobs continue to contribute the power of attracting blobs with a certain magnitude, but also contribute the further power of attracting blobs only in the same preferred direction as the other blobs in the relevant globule. That is, the property P1 of the blobs contributes the powers it previously contributed, but also one power that is different from that which it contributes in simpler aggregations—a power which instances of P1 would not contribute if the laws applying to simpler

Figure 2.2 Blobs and globules in simple aggregations.

Scientific Composition and Completeness Implications

35

aggregations solely determined their contributions of powers in complex globules. In the complex globules, blobs thus continue to bond with each other and continue to exert a weak force with the same magnitude, but now the force of each blob is exerted only in the same preferred direction as that of other blobs constituting the globule. Finally, following Laughlin, let us also stipulate that property H of a complex globule non-causally determines that instances of property P1 contribute this new power. Our full scenario is outlined in Figure 2.3, with the simple aggregations of blobs and resulting globules on the left, and the complex aggregations of blobs and associated globules on the right. So let us now briefly examine what holds true in our scenario for Conditioned aggregation. First, we should note that all of the properties of the complex globule are realized. The globule’s properties like the attractive force H, or shape or size, have powers which are completely comprised by the powers of the constituent blobs. Consider property H and its power to strongly attract globules in a certain direction. This power is comprised by the many powers of attracting blobs in a certain direction had by the constituent blobs that constitute the globule. Thus it appears that property H is fully realized, for all its powers are composed by powers of its constituents. Similar points obviously hold for the other properties of the globule, including more mundane properties like shape or size. It thus appears that concepts of scientific composition comprehensively apply to the higher level properties in our scenario. What about the causal efficacy of H in the complex globules? The properties of the blobs contribute to them the power to attract blobs in a preferred

Figure 2.3 Blobs and globules in complex aggregations.

36

Carl Gillett

direction. And this power would not be contributed if H were absent from the globule in which the blobs having these realizer properties are to be found. Furthermore, we have stipulated that H non-causally determines the novel powers contributed by its realizers. Thus the property H does make a difference to the causal powers of individuals—in this case the powers of component individuals. H is thus a joint cause, with the properties of the constituent blobs, of any lower-level effects that result from the relevant conditional powers. Consequently, the realized property instance H in our scenario is causally efficacious. One may establish that a set of statements is logically coherent by trying to conceive of a situation, using a conceptual representation, in which these statements are all true. And with our scenario for Conditioned aggregation we have successfully outlined just such a representation in which it is both true that a property instance is realized, but nonetheless also true that this realized instance is causally efficacious. The thought experiment thus provides good reasons to think that Strong emergent property instances can exist when scientific concepts of composition comprehensively apply.9 Given these points, the thought experiment establishes that the Argument from Composition is invalid, since the premises of the Argument are all apparently true in our scenario and its conclusion is false. We can therefore see that the lack of the attention from philosophers has hurt scientific debates. By letting scientific composition and the scientific reductionist’s arguments about it go unexamined, many writers have been left endorsing an invalid argument. Another, less obvious, oversight by philosophers is also worth emphasizing. For in our thought experiment we have the kind of synchronous ‘downward’ determination relations emphasized by scientists like Laughlin. Once again, in understanding this type of determination our framework for composition is useful. First, we should note that the ‘downward’ determination between the Strongly emergent instance and its realizers is non-causal, since it is synchronous, holds between entities that are not wholly distinct, and does not involve the transfer of energy or mediation of force. Second, although non-causal, we should note that the ‘downward’ determination is also not compositional—for comprising between powers is plausibly a necessary condition for a scientific composition relation. Although the realizers, including P1, have powers that comprise the powers of H, the reverse is not true—the properties of the globule contribute no powers that comprise powers of the blobs. The ‘downward’ determinative relation between the Strongly emergent property instance and its realizers is non-causal but not compositional in nature. Scientists have thus apparently been concerned with a kind of non-compositional, non-causal determination overlooked by philosophers. And we therefore need a new term to mark this neglected class of determination relations. Combining the Greek words ‘macro’ and ‘chresis’, where the latter is roughly the Greek for ‘use’, we get the term ‘machresis’ for the general

Scientific Composition and Completeness Implications 37 phenomenon of composed entities non-causally, and non-compositionally, determining the nature of their components. Obviously, it is an intriguing, and very important, question whether we fi nd this type of Strong emergence in the ‘emergent phenomena’ scientists have recently highlighted, whether superconductivity, the phases of matter, symmetry breaking of all kinds, Benard cells in chemistry, the shoaling of fish and flocking of birds, and many, many more examples. Similarly, it is interesting how machresis relates to the cases of so-called ‘self-organization’ found with ‘emergent’ phenomena in chemical systems, biological systems like ant colonies, or social systems such as traffic patterns, amongst others. Only a careful examination of each of these cases can resolve such questions and potentially fi ll out our understanding of Strong emergence and machretic determination. However, we cannot pursue such case studies here. Equally clear is that a bevy of objections can be leveled against my defense of the possibility of Strong emergence. But let us also put those to one side here as well.10 Instead, I want to consider another argument apparently deployed by scientific reductionists, since this reasoning is valid and also gets us to the deeper issues.

5. THE REDUCTIONIST’S WORLDVIEW AND ANOTHER ARGUMENT FOR COMPOSITIONAL REDUCTION The scientific reductionists’ “dream” of a Final Theory is the hope that the laws illuminated by studying the ontologically fundamental components either in isolation or simple aggregations reveals all the fundamental laws of nature. Having considered the scientific emergentists’ position, an assumption about aggregation that underlies this aspiration is apparent. If one assumes that the laws governing the fundamental entities in isolation, or simple aggregations, are a complete set of fundamental laws, then one assumes two things. First, that when they aggregate, component entities have their powers determined by other entities in the same ways they are determined in isolation or simple aggregations. For one is assuming the laws governing these situations are all the fundamental laws. And, second, one is assuming that other component entities are the only entities that ever determine the powers of component entities. Since the laws the reductionist points to about fundamental entities as candidates for fundamental laws only refer to such entities. The resulting picture of aggregation is a natural one. It takes aggregation to be regular in its nature across all aggregations. And it takes aggregation to be tidy in only involving component entities as the sole determinants. We thus have a simple picture, of a regular phenomenon only involving one type of entity, so I term this the ‘Simple’ view of aggregation. And although the Simple view is rarely made explicit by scientific reductionists, it is arguably one of their foundational ontological assumptions. For we have just

38

Carl Gillett

seen that the truth of the Simple view is a precondition of the existence of a Final Theory, and, perhaps unsurprisingly, it is also a precondition of the connected thesis philosophers term the ‘Completeness of Physics’ and frame thus: (Completeness of Physics) The entities of physics are determined, in so far as they are determined, solely by other microphysical entities according to the laws governing their behavior in isolation or simple aggregations. Usually this thesis is stated by reference simply to the ‘laws of physics’, but we have just seen that the relevant laws are taken to describe the behaviors of the fundamental entities in isolation or simple aggregations. Note why the Simple view of aggregation must be true in order for the Completeness of Physics to be true. For the Completeness of Physics is true only if other component entities (i.e., other entities of physics) solely determine the powers contributed by other fundamental components—hence presupposing the Simple view. As well as illuminating a key ontological assumption of the reductionist, appreciating the Simple view of aggregation is also important because it can be used by scientific reductionists to buttress their earlier reasoning to turn aside even the response based around Strong emergence. Recall that the scientific emergentist’s lesson was that there are two ways to be efficacious and make a difference to the causal powers of individuals: a property can either contribute powers itself that bring about certain effects, or it can determine the powers contributed by other properties, thus becoming a joint cause of the resulting effects of such powers. The Argument from Composition failed to exclude the latter role for realized properties as our scenario for Conditioned aggregation exposed. But assuming the truth of the Simple view of aggregation does exclude this kind of possibility. For the Simple view of aggregation implies that only other component entities, like realizers, determine the powers contributed by components such as realizer properties—thus excluding a realized property from playing a machretic determinative role and blocking Conditioned aggregation. If we build the Simple view onto the premises of the Argument from Composition, then we plausibly avoid the emergentists’ challenge. Let us more precisely frame such an argument by again focusing on realization. Our reductionist argues that it is ontologically profl igate to take any realized property instance to determine the causal powers of individuals in addition to realizer property instances. For the reductionist claims that there are two ways for a property instance to determine the powers of individuals: either the property instance contributes powers to individuals itself or the property instance determines the powers contributed to individuals by other properties. But given the nature of the realization relation, we can account for all the property instances that

Scientific Composition and Completeness Implications 39 contribute powers through the realizer property instances of individuals, or their constituents, rather than also taking realized property instances to contribute powers. We must thus conclude that realized properties do not contribute powers and that only realizer properties do so. Furthermore, assuming the Simple view of aggregation, we also know that the powers contributed by the realizer properties are only determined, insofar as they are determined, solely by other realizer properties. So realized properties do not determine the powers of the realizer properties that are the only properties that contribute powers. If we assume that the causal powers of individuals are not overdetermined, then appealing to Occam’s razor the reductionists argue that we should accept the existence of no more property instances than we need to account for the causal powers of individuals. The proponent of this simple argument thus concludes that we should only accept that realizer property instances determine the powers of individuals, either by contributing such powers or determining the powers contributed by other properties. But it is also plausibly true that the only property instances that exist are those that make a difference to the causal powers contributed to individuals. We may thus further conclude that there are only realizer property instances. Let us call the latter the ‘Argument from Composition and Completeness’, since it adds to the Argument from Composition the Simple view of aggregation which stands behind the complete set of laws the scientific reductionist looks for in a Final Theory. And it appears that the Argument from Composition and Completeness is plausibly valid, since this reasoning shuts off the theoretical opening exploited by the scientific emergentists against the Argument from Composition. However, we should immediately note that scientific emergentists will challenge this new reasoning in another respect, for our work shows why they will attack the soundness of the Argument from Composition and Completeness. As we have seen, scientific emergentists claim to have empirical evidence that with many phenomena the Conditioned view of aggregation applies, and machretic determination is operative, and so the Simple view is simply mistaken. Our work thus shows that the battle between scientific reductionists and scientific emergentists is not over whether we have mechanistic explanations for higher-level phenomena—both sides agree that we do. Nor is the debate over whether all higher-level phenomena are composed— again, both sides agree that such entities are all composed. Rather, the deeper issues of the scientific debate concern the character of the component and composed entities, the ‘parts’ and ‘wholes’, that our empirical fi ndings reveal in nature. Specifically, the scientific reductionist presses the Simple view of aggregation and claims components are only ever determined by other components, thus being led to deny there are any composed entities. In contrast, scientific emergentists claim that Conditioned aggregation exists and that in some cases Strongly emergent composed entities machretically determine their components.

40

Carl Gillett

6. CONCLUSION: THE SCIENTIFIC POSITIONS, PHILOSOPHICAL OPTIONS, AND DEEPER ISSUES Rather than beginning with the work of philosophers, my approach has been to start with the sciences, both in the nature of their fi ndings and concepts, and in the kinds of positions defended by working scientists themselves. And I want to conclude by showing that the significance of our work is not simply confi ned to the sciences but also has implications for philosophical debates. I therefore propose to finish the chapter by connecting our work to recent debates over the viability of ontologically non-reductive physicalism. In particular, I want to examine the prospects for what I term the ‘Standard’ non-reductive physicalism of philosophers which blithely endorses both universal composition and the Completeness of Physics, as well as the existence of efficacious composed entities. In response to Standard non-reductive physicalism, in one strand of their work, critics such as Jaegwon Kim have argued at length that non-reductive physicalists face grave problems because the nature of composition (articulated through Kim’s favored framework) means that universal composition is incompatible with the existence of efficacious composed entities.11 Though he uses different arguments, based around supervenience or causal exclusion, Kim thus defends a conclusion similar to that of the Argument from Composition. The resulting picture about the tenable combinations of theses, and hence viable positions, is outlined in Figure 2.4. Perhaps unsurprisingly, given the similar conclusions his reasoning leads to, Kim agrees with the scientific proponents of the Argument from Composition in taking the top-left box, which combines composition and efficacious higher-level entities, to be untenable. Furthermore, Kim (1992) has argued at length that the philosophical Standard non-reductive physicalism is akin to the emergentism so often defended by scientists—Kim thus puts both Strong emergentism and the philosophers’ ‘Standard’ non-reductive physicalism together. Putting the bottom-right box to one side because it is hard to evaluate what the position comes to, Kim thus concludes, like scientific reductionists, that our only options are either an empirically implausible anti-physicalism or reductive physicalism like Compositional reductionism. As a consequence, Kim suggests that the orthodox position in much of philosophy must actually be abandoned and the only option for physicalists is an ontological reductionism. Such is one highly influential view of our options. However, our work shows that this account of the viable positions is actually mistaken. Although Kim’s reasoning is distinct from the Argument from Composition, its conclusion is that one should not accept both universal composition and the existence of efficacious higher-level entities. But our scenario for Conditioned aggregation establishes that compositional relations can hold universally and composed entities can still be efficacious. Thus our

Scientific Composition and Completeness Implications 41

Figure 2.4 Diagram of the options as understood by the philosophical critics and illustrating the perceived implications of the Argument from Composition. (Boxes with dots indicate untenable combinations of theses; COMP is acceptance of evidence for universal composition in nature; HCE is endorsement of the existence of efficacious higher-level entities).

fi ndings about Strong emergence plausibly establish that Kim’s reasoning is also likely to be invalid or otherwise flawed. What we have seen is that rather than a pair of incompatible theses, in composition and the existence of efficacious higher-level entities, there is actually a trio of incompatible theses in universal composition, the existence of efficacious higher-level entities and the truth of the Simple view of aggregation. Consequently,

42

Carl Gillett

there is trichotomy of positions we may adopt in response to this incompatibility, including the overlooked option presented by scientists in their Strong emergentism. The combinations of compatible theses, and viable positions, are therefore more accurately represented by Figure 2.5. As the table in Figure 2.5 shows, reflecting the Argument from Composition and Completeness, Standard non-reductive physicalism is untenable. And this confi rms another strand of the argumentation from critics like

Figure 2.5 Diagram of the main options illuminated by appreciating the implications of the Conditioned view of aggregation and the possibility of Strong emergence. (Boxes with dots indicate untenable combinations of theses; COMP is acceptance of evidence for universal composition in nature; HCE is endorsement of the existence of efficacious higher-level entities; SA is the Simple view, and CA the Conditioned view, of aggregation ).

Scientific Composition and Completeness Implications 43 Kim, and others such as Barry Loewer (Forthcoming), who have pressed the point that the Completeness of Physics is incompatible with composition and the existence of efficacious higher-level entities. However, such critics have again concluded that we must consequently choose either reductive physicalism or anti-physicalism—thus implying that the only choice for physicalists is ontological reduction. But as our work establishes, and as Figure 2.5 makes plain, this is actually a false dichotomy—for the third option offered by Strong emergentism has been overlooked. Strong emergentism, reflecting the Argument from Composition and Completeness, diverges from Standard non-reductive physicalism in its commitments, but the positions of working scientists like Laughlin, Anderson, and Prigogine among others are still a form of ontologically non-reductive physicalism. These researchers endorse universal composition of all higher scientific phenomena whilst also offering an account of how these composed entities can still be efficacious. Obviously, Strong scientific emergentism has a price—no substantive ontological conclusion comes for free, after all—in a commitment to the Conditioned View of aggregation and the existence of machretic determination. A corollary is also obviously that one must abandon both the Simple view and the Completeness of Physics. However, as we saw previously, we can understand how the Strong emergentist version of ontologically non-reductive physicalism could be true in a world where scientific concepts of composition apply comprehensively. And scientific emergentists argue that a range of scientific cases establish that aggregation is Conditioned, and that machresis and Strong emergence exist. In contrast, we cannot even imagine how Standard non-reductive physicalism could be true. To conclude, we have found that recent scientific debates do indeed have implications for philosophical discussions. We have found that we face a very different set of issues and positions than philosophers have thought. On one side, I have suggested that scientific reductionism, in Compositional reduction’s blend of ontic reductionism and semantic antireductionism, resists extant philosophical responses and arguably presents one of the two viable ‘grand pictures’ of nature and the sciences. On the other side, I have shown that the Strong emergence endorsed by scientific emergentism can exist, contrary to almost universal philosophical skepticism, and offers the other viable ‘grand picture’ of nature and the sciences. In fact, we have found that Conditioned aggregation, machretic determination, and the Strong emergence they enable underpin the only defensible form of ontologically non-reductive physicalism. Defending such a Strong scientific emergentism involves the careful examination of concrete cases of ‘emergence’ using the better theoretical frameworks we have begun to provide. Such a project is yet to be shown to be successful. In contrast, of course, the Standard non-reductive physicalism so popular amongst philosophers is such that we cannot even imagine how it could ever be true.12

44

Carl Gillett

NOTES 1. There are obviously other concepts of ‘emergence’ other than that of Strong emergence. See Gillett (2002b; 2006; Unpublished-b: Chapter 5), for a detailed survey and comparison of these concepts. Furthermore, many working scientists do indeed endorse concepts of ‘emergence’ other than Strong emergence. My contention here is only that a number of prominent scientific emergentist are best understood as endorsing Strong emergence and that their ideas have important implications which I examine later in this chapter. 2. Shoemaker (1980). 3. The reader should also note that I use ‘entity’ as a catch-all referring to powers, properties, individuals, processes, etc. 4. For a fuller survey of all the features of scientific composition, see my (2003c; Unpublished-a; Unpublished-b: Chapter 2). 5. Weinberg (1994: 53). 6. The famous arguments against ‘reduction’ are found in papers such as Fodor (1974) or Kitcher (1984). See Gillett (2007) for a sketch of how Compositional reductionism avoids these critiques and Chapter 4 of my (Unpublished-b) for a more detailed overview. 7. See Philip Anderson (1972), Stuart Kauffman (1995), Ilya Prigogine & Isabel Stengers (1996), or Robert Laughlin (2005), for accessible presentations of these researchers core ontological claims. 8. See Gillett (2003b; Unpublished-b: Chapter 8) for a discussion of the forms that laws take when we have Conditioned aggregation. 9. For more precise, though more abstract, characterizations of a scenario involving Conditioned aggregation see Gillett (2003a; 2003b; 2006). 10. For examinations of a variety of objections to my arguments for the possibility of Strong emergence see Gillett (2003a; 2003b; 2006; Unpublished-b: Chapter 6). 11. This is the conclusion of the explanatory or causal exclusion arguments found in the papers in Kim (1993) and also the Supervenience Argument presented in Kim (1998). 12. Thanks to the editors and the conference audience at the University of Milan for helpful comments. The paper is based on material and figures drawn from Chapters 2, 3, and 6 of my (Unpublished-b), as well as the articles cited in this chapter.

REFERENCES Anderson, P. (1972). More is different. Science, 177, 393–396. Fodor, J. (1974). Special sciences: or, the disunity of science as a working hypothesis. Synthese, 28, 97–115. Gillett, C. (2002a). The dimensions of realization: A critique of the standard view. Analysis, 62, 316–323. . (2002b). The varieties of emergence. Their purposes, obligations and importance. Grazer Philosophische Studien, 65, 89–115. . (2003a). Non-reductive realization and non-reductive identity: What physicalism does not entail. In S. Walter & H.-D. Heckmann (Eds.), Physicalism and Mental Causation (pp. 31–58). Exeter: Imprint Academic. . (2003b). Strong emergence as a defense of non-reductive physicalism: A physicalist metaphysics for ‘downward’ determination. Special issue on emergence. Principia, 6(1), 89–120.

Scientific Composition and Completeness Implications 45 . (2003c). The metaphysics of realization, multiple realizability and the special sciences. Journal of Philosophy, 100(11), 591–603. . (2006). The hidden battles over emergence. In P. Clayton (Ed.), The Oxford Handbook of Religion and Science (pp. 801–819). Oxford: Oxford University Press. . (2007). Understanding the new reductionism: The metaphysics of science and compositional reduction. Journal of Philosophy, vol. CIV, 193–216. . (Unpublished-a). Making sense of levels in the sciences: Composing powers, properties, parts and processes. . (Unpublished-b). The Roots of Reduction and Fruits of Emergence. Kauffman, S. (1995). At Home in the Universe. New York: Oxford University Press. Kim, J. (1992). ‘Downward causation’ in emergentism and nonreductive physicalism. In A. Beckermann, H. Flohr, and J. Kim (Eds.), Emergence or Reduction? Essays on the Prospects of Nonreductive Physicalism (pp. 119–138). New York: De Gruyter. . (1993). Supervenience and Mind: Selected Philosophical Essays. New York: Cambridge University Press. . (1998). Mind in a Physical World. Cambridge, MA: MIT Press. Kitcher, P. (1984). 1953 and all that: A tale of two sciences. Philosophical Review, 93, 335–373. Loewer, B. Forthcoming: “Why is there anything other than physics?”. Synthese. Laughlin, R. (2005). A Different Universe: Remaking Physics from the Bottom Down. New York: Basic Books. Prigogine, I., & Stengers, I. (1996). The End of Certainty. New York: Free Press. Shoemaker, S. (1980). Causality and properties. Reprinted in H. Mellor & A. Oliver (Eds.), Properties (pp. 228–254). Oxford: Oxford University Press. Weinberg, S, (1994). Dreams of a Final Theory. New York: Pantheon.

3

Weak Emergence and Context-Sensitive Reduction Mark A. Bedau

Emergence and reduction are unquestionably connected, and many people assume that they are inconsistent. I disagree. After distinguishing nominal, strong, and weak emergence, I explain how one kind of emergence—what I call “weak” emergence—is not merely consistent with reduction; it actually depends on a particular context-sensitive form of micro-causal reduction. Weak emergence and context-sensitive reduction go hand in hand. All this illustrates a form of pluralism about emergence, which holds that nature can contain many distinct kinds of emergence. Rather than looking for the one single true view of emergence, we should fi nd them all.

1. REDUCTIONISM Reductionism generally is the view that the macro is the way it is in virtue of how things are at the micro. Reduction involves the connection between macro and micro domains of phenomena: Macro reduces to micro. The macro domain involves various macro entities with various macro properties; the micro domain involves a large population of micro entities with various micro properties. Typically the micro entities constitute the macro entities, but that is not required; maybe physical states or processes in the brain and nervous system do not “constitute” mental states or processes. The attraction of reductionism is that it unifies and connects different domains of science. Cell biology is unified with molecular biology; thermodynamics is unified with statistical mechanics; and so on. Scientific unification itself is a kind of scientific progress. Reduction also shrinks the number of fundamental facts or principles on which all other explanations must rest, a point emphasized especially by Weinberg (1987). Decreasing the number of explanatory starting points makes explanations a little simpler, and therefore a little better. That is another kind of scientific progress. Reductionism comes in many different forms. To untangle the interconnections between emergence and reduction, one must distinguish certain dimensions of reductionism. One dimension concerns the macro domain and specifies what kind of macro thing is being reduced. The options

Weak Emergence and Context-Sensitive Reduction 47 include macro entities, macro causes, and macro explanations, among others. A second dimension concerns the kind of micro information used to articulate the macro reductions. There are a number of options. The micro information might be static and concern only the exact same moment as the macro information, or it might be dynamic and concern how macro states derive from earlier micro states. In addition, the micro explanations might be context-free and depend only on the previous state of the micro element itself, or they might be context-sensitive because the state of a micro element depends on the previous states of other local micro elements (e.g., its neighbors in a cellular automaton). A third dimension concerns the explanatory modality of the reduction of macro to micro. Is micro reduction of the macro merely logically possible, merely feasible in principle, or actually feasible in practice? These are the three most important dimensions of reductionism from the point of view of emergence.

2. THE TWIN HALLMARKS OF EMERGENCE The reductionistic picture of science is challenged by certain examples of apparent emergent phenomena. The typical examples include a striking variety of cases, and many of them have roots that go back at least to C. D. Broad (1925). One set of examples concerns certain properties of physical systems. For example, the liquidity and transparency of water are sometimes said to emerge from the properties of oxygen and hydrogen in structured collections of water molecules. Life is another common theme in examples of apparent emergence, as in the relationship between a living organism and the molecules that constitute it at a given moment. The organism is composed of nothing except those molecules, but those very same molecules would not constitute an organism if they were totally rearranged, so the living organism seems to emerge from the molecules. The mind is a third rich source of potential examples of emergence. Our mental lives consist of an autonomous, coherent flow of mental states (beliefs, desires, memories, fears, hopes, etc.), and these all presumably somehow emerge out of the swarm of biochemical and electrical activity involving our neurons and central nervous system. These examples illustrate the twin hallmarks of emergence (Bedau, 1997, 2003). All emergence involves macro-level1 phenomena that (1) arise from and depend on some more basic, micro-level phenomena, and yet (2) are simultaneously autonomous from that micro-level base on which they depend. The topic of emergence is fascinating and controversial in part because emergence seems to be widespread, and yet the twin hallmarks of emergence seem opaque and perhaps even incoherent. The two hallmarks are admittedly vague, and this allows for a kind of pluralism about emergence (see Section 3). Phenomena might be dependent on underlying processes in many ways, and they might be autonomous

48

Mark A. Bedau

from underlying processes in many ways. Any way of simultaneously meeting both hallmarks seems to be a good candidate notion of emergence. The hallmarks structure and unify the different notions of emergence and provide a framework for comparing them. The hallmarks also help explain the controversy over emergence. Viewing macro phenomena as both dependent on and autonomous from their micro bases seems metaphysically problematic: either inconsistent, or illegitimate, or with mysterious and unexplainable powers. The problem of emergence is to defi ne a notion of emergence that is metaphysically reasonable and scientifically useful.

3. PLURALISM ABOUT EMERGENCE Pluralism about emergence enables one to defi ne a new notion of emergence merely by specifying an interpretation of “dependent” and “autonomous” in the twin hallmarks of emergence. Different notions of dependence and autonomy generate different notions of emergence, and many of these notions are logically and metaphysically coherent. But not all are philosophically interesting or useful in empirical science. A compelling defense of some kind of emergence should make it more than a philosophical curiosity. Only if it plays a central and constructive role in our explanation of nature can a notion of emergence fully solve the problem of emergence. One of my preferred solutions to the problem of emergence treats many apparent emergent phenomena as genuine examples of a certain kind of “weak” emergence (explained in Section 4). This form of weak emergence applies to real examples of apparent emergent phenomena, and it is free from problematic metaphysical implications. More than that, it plays a central role in our best current scientific explanations of apparent emergent phenomena. So, a strong case can be made for adopting this kind of weak emergence. There are many conceptions of emergence (see, e.g., Bedau and Humphreys, 2008), and this is not the occasion for a careful categorization.2 I will briefly describe three different kinds of emergence (Bedau, 1997, 2003, 2008): nominal emergence, weak emergence, and strong emergence. Each of these conceptions is broad and includes many different instances. I introduce nominal, weak, and strong emergence here in order to illustrate what I mean by pluralism about emergence. I am a pluralist in the sense that (i) there are a number of prima facie coherent conceptions of emergence (i.e., interpretations of “dependent” and “autonomous”), and (ii) each conception of emergence must make its own case for itself, on its own merits. I am also an optimistic pluralist, since I believe that some conceptions of emergence will prove very useful in empirical science. Convincing credentials for one kind of emergence do not rub off on other kinds of emergence; each kind is independent. My defense of weak emergence (Bedau, 1997, 2003, 2008) is an example of how to defend a conception of

Weak Emergence and Context-Sensitive Reduction 49 emergence. Pluralism about emergence involves mounting the same kind of argument that I have used to defend weak emergence: (i) to defi ne it precisely by specifying the kind of dependence and autonomy between macro and micro domains, (ii) to show that those kinds of dependence and autonomy have no negative metaphysical implications, and (iii) to show that the resulting combination helps science understand some interesting aspect of reality. Macro entities and micro entities each have various kinds of properties. Some of the properties that characterize a macro entity can also apply to its micro constituents; others cannot. For example, consider micelles. These are clusters of amphiphilic polymers arranged in such a way that their hydrophilic ends are on the outside and their hydrophobic tails are on the inside. Those polymers are themselves composed of hydrophilic and hydrophobic monomers. In this context, the micelles are macro objects, while the individual monomeric molecules are micro objects. The micelles and the monomers both have certain kinds of physical properties in common (having a location, mass, etc.). By contrast, some of the properties of micelles (such as their permeability) are the kind of properties that monomers simply cannot possess. Here is another example: The constituent molecules in a cup of water, considered individually, cannot have properties like fluidity or transparency, though these properties do apply to the whole cup of water. Nominal emergence involves a macro property that logically or conceptually cannot be a property of the micro entities in question. This is an especially simple and bare notion of an emergent property; a macro property is simply a property that can apply to macro objects (of the kind in question) but not to the micro objects (of the kind in question). We saw some examples of nominal emergence earlier: A macro-level micelle has the property of permeability, but permeability makes no sense for the individual polymers that make up the micelle. A cup of water is transparent, but transparency makes no sense when applied to the individual water molecules that fill the cup. Harré (1985) and Baas (1994), among others, have called attention to nominal emergence. Nominal emergence easily explains the two hallmarks of emergence. Macro-level emergent phenomena are dependent on micro-level phenomena in the straightforward sense that wholes are dependent on their constituents, and emergent phenomena are autonomous from underlying phenomena in the straightforward sense that emergent properties do not apply to the underlying entities. When dependence and autonomy are understood in these ways, there is no problem in seeing how emergent phenomena could simultaneously be both dependent on, and autonomous from, their underlying bases. The notion of nominal emergence is very broad. It applies to a large number of intuitive examples of emergent phenomena and corresponds to the rather compelling picture of reality consisting of a hierarchy of levels.

50

Mark A. Bedau

Its breadth can be viewed as a weakness, though, for nominal emergence applies to all macro-level properties that are not possessed by micro-level entities. Macro properties are traditionally classified into two kinds: genuine emergent properties and mere “resultant” properties, where resultant properties could be viewed as those that can be predicted and explained from the properties of the components. For example, a circle consists of a collection of points, and the individual points have no shape. So being a circle is a property of a “whole” but not its constituent “parts”—that is, it is a nominal emergent property. However, if you know that all the points in a geometrical figure are equidistant from a given point, then you can derive that the figure is a circle. So, according to the preceding conception of resultant properties, being a circle is a property that is both nominally emergent and resultant. Unless one adopts a different conception of resultant properties, one can distinguish emergent from resultant properties only with more narrowly defi ned kinds of emergence. The most stringent conception of emergence, which I call strong emergence, adds the requirement that emergent properties are supervenient properties with irreducible causal powers. These macro-causal powers have effects at both macro and micro levels, and macro-to-micro effects are termed “downward” causation. We saw earlier that micro-determination of the macro is one of the hallmarks of emergence, and supervenience is a popular contemporary interpretation of this determination. Supervenience explains the sense in which emergent properties depend on their underlying bases, and irreducible macro-causal power explains the sense in which they are autonomous from their underlying bases. These irreducible causal powers give emergent properties the dramatic form of ontological novelty that many people associate with the most puzzling kinds of emergent phenomena, such as qualia and consciousness. In fact, most of the contemporary interest in strong emergence (e.g., Chalmers, 1996; Kim, 1992, 1997, 1999; O’Connor, 1994) arises out of concerns to account for those aspects of mental life like the qualitative aspects of consciousness that most resist reductionistic analysis. The supervenient causal powers that characterize strong emergence are the source of its most pressing problems. One problem is the so-called “exclusion” argument emphasized by Kim (1992, 1997, 1999). This is the worry that emergent macro-causal powers would compete with micro-causal powers for causal influence over micro events, and that the more fundamental microcausal powers would always win this competition. I will examine downward emergent causation at length later in this chapter. The exclusion argument aside, the very notion of strongly emergent causal powers is problematic to some people. By definition, such causal powers cannot be explained in terms of the aggregation of the micro-level potentialities; they are primitive or “brute” natural powers that arise inexplicably with the existence of certain macro-level entities. This contravenes “causal fundamentalism”—the idea that macro-causal powers supervene on and are determined by micro-causal

Weak Emergence and Context-Sensitive Reduction 51 powers, that is, the doctrine that “the macro is the way it is in virtue of how things are at the micro” (Jackson & Pettit, 1992: 5). Many naturalistically inclined philosophers (e.g., Jackson & Pettit) find causal fundamentalism compelling, so they would accordingly be skeptical about any form of emergence that contravenes causal fundamentalism. Still, causal fundamentalism is not a necessary truth, and strong emergence should be embraced if it has compelling enough supporting evidence. But this is where the fi nal problem with strong emergence arises. All the evidence today suggests that strong emergence is scientifically irrelevant. Virtually all attempts to provide scientific evidence for strong emergence focus on one isolated moribund example: Sperry’s explanation of consciousness from forty years ago (e.g., Sperry, 1969). The scientific irrelevance of strong emergence is easy to understand, given that strongly emergent causal powers would be brute (i.e., unexplainable) natural phenomena that could at best play a primitive role in science. We should accept brute natural phenomena only if we are convinced that they cannot be explained. The problem is that strong emergence seems like it always can be explained away. 3

4. WEAK EMERGENCE AS EXPLANATORY INCOMPRESSIBILITY Poised between nominal and strong emergence is an intermediate notion, which I call weak emergence. The term ‘weak’ is intended to highlight the contrast with the “strong”, irreducible macro-causal powers characteristic of strong emergence (Bedau, 1997, 2003, 2008). Weak emergence (as I mean it) can be explained in various ways.4 My explanation here uses the concept of explanatory incompressibility. Previously I have defi ned weak emergence using the concept of un-derivability except by simulation (Bedau, 1997, 2003). These two defi nitions are similarly indirect, and they are essentially equivalent. I support them both. Here I focus on explanatory incompressibility in order to highlight that weak emergence applies not just to computer simulations, but also to a great many natural systems, especially those that motivated the original discussions of emergence by the British emergentists (McLaughlin, 1992). I defend a version of weak emergence that is distinctively dynamical, because this kind of weak emergence concerns the complex dynamic processes by which certain global phenomena are generated. The dynamical nature of weak emergence helps explain the distinctive role that computer simulations play in both explaining natural emergent phenomena and artificially generating interesting new examples of emergent phenomena. Emergence always involves a certain kind of relationship between global or macro phenomena and local or micro phenomena. Specifically, emergent macro phenomena somehow both depend on, and are autonomous from, micro phenomena.

52

Mark A. Bedau

The characteristic feature of weak emergence, in general, is that the macro is ontologically and causally reducible to the micro in principle, but the reductive micro explanation is especially complex (e.g., Simon, 1996). Different kinds of explanatory complexity create different versions of weak emergence. 5 Here I defi ne weak emergence in terms of macro states that are un-derivable except by simulation with macro states that have only incompressible explanations. Throughout I tacitly assume that we seek only true, exact, and complete explanations of how macro properties are generated from prior micro properties over time. My defi nition of weak emergence in a macro property is this: If P is a macro property of some system S, then P is weakly emergent if and only if P has a generative explanation from all of S’s prior micro facts, but only in an incompressible way (see next paragraph). This defi nition defi nes weakly emergent macro properties by the distinctive way in which we explain how they are generated from underlying micro facts. Analogous accounts can be given for weakly emergent entities, events, phenomena, or patterns (which involve weakly emergent properties). The basic idea of an incompressible generative explanation is simple. An explanation is generative just in case it exactly and correctly explains how macro events unfold over time (i.e., how they are generated dynamically). The generative explanation assumes complete information about both the micro-causal dynamics that drive the system and the system’s earlier micro states and boundary conditions. The explanation works simply by tracing through the temporal details in the complex web of micro-level causal interactions that ultimately generate the macro events. This kind of explanation is appropriate for any system with (global) macro features that depend on (local) micro features in certain complex ways. In particular, it is appropriate if we can describe the system’s macro features at a given time by appropriately conjoining or aggregating or summing the (local) micro features that exist at the same time. This is a synchronic reduction of macro to micro. Now, by starting with a completely specified initial condition, and by tracing forward in time through the network of local micro-level causal interactions, the system’s macro features (which are an aggregation of micro features at a given time) can be explained from immediately preceding aggregations of micro features. I shall describe the process of explaining the generation of a system’s macro behavior by aggregating and iterating the earlier local micro interactions over time, for short, as “crawling the microcausal web”. Incompressible explanations cannot be replaced without explanatory loss by shorter explanations. If an explanation of some macro property of some system is incompressible, then there is no shortcut generative explanation of that macro property that is true, complete, accurate, and able to avoid crawling the micro-causal web.6 Hence the temporal signature of incompressible explanations: Explaining later behavior requires a

Weak Emergence and Context-Sensitive Reduction 53 proportional amount of additional explanatory effort (Crutchfield et al., 1986). On the other hand, if the explanation is compressible, then explaining the macro property arbitrarily far into the future takes some fi xed and fi nite amount of explanatory effort, no matter how far into the future the explanation reaches. The required explanatory effort is capped; explaining later behavior takes no more explanatory effort than explaining earlier behavior. A compressible explanation can achieve this economy because it avoids the incrementally growing cost of crawling the micro-causal web. The simplest and clearest examples of compressible explanations involve cellular automata. A cellular automaton (or CA) is a regular spatial lattice of “cells”, each of which can be in one of a fi nite number of states. The lattice typically has one, two, or three spatial dimensions. The state of each cell in the lattice is updated simultaneously in discrete time steps. The behavior of each cell is governed by a fi nite-state machine that outputs its own next state, given as input the states of the cells within some fi nite, local neighborhood of the lattice. Typically all cells in the lattice are governed by the same fi nite-state machine, and typically the fi nite-state machine is deterministic. Now, consider the exceedingly trivial cellular automaton that I call All Life, which is completely governed by the following very simple local causal rule: A cell is alive at a given time whether or not it or any of its neighbors were alive or dead at the previous moment. It is trivial to give a shortcut explanation of any macro property of All Life arbitrarily far into the future, because one can see that all cells will be alive at all future times, no matter what the initial aggregate local configuration of the cellular automaton. All Life is atypical compared with the cellular automata usually discussed, such as the Game of Life (GOL). The most interesting cellular automata are those like GOL that are complex and known to have incompressible explanations.7 The behavior of complex cellular automata typically cannot be explained except by crawling the micro-causal web. The same holds throughout many other kinds of computational systems, such as soft artificial life systems like Ray’s Tierra, Holland’s Echo, and Packard’s Bugs.8 The more simulations of natural complex adaptive systems we study, the more weak emergence we will fi nd. This is a contingent empirical claim, but it accords with all the evidence available today. What would make a system weakly emergent? That is, what makes a system’s explanation incompressible? The answer, presumably, is the complexity of the system’s micro-causal network. If we could describe a certain structure of micro-causal networks that exactly captured those whose explanations are incompressible, then we could defi ne weakly emergent systems as those with that micro-causal network structure. But today we know no micro-causal structure that can directly defi ne weak emergence. Nevertheless, our empirical experience with weak emergence

54

Mark A. Bedau

suggests that incompressible explanations are due to two structural features of complex systems. First, weakly emergent behavior typically comes from massively parallel micro-level populations of autonomous agents that interact independently with their neighbors and their local environment. Second, the interactions among such agents and their environments are typically non-linear and synergistic, so that the behavior of an agent is highly sensitive to its local context, including the states of neighboring agents. These two factors make the system’s behavior impossible to predict, even given complete prior micro information, without crawling the micro-causal web. On the positive side, crawling the micro-causal web is guaranteed to derive all the system’s macro behavior, at least in principle. So, weakly emergent phenomena need an incompressible explanation when a system’s macro properties are generated by context-sensitive micro-level rules for micro-level agents. For this reason, weak emergence depends on context-sensitive reduction. One might ask exactly how much micro-causal complexity is sufficient for weak emergence, and exactly how much is necessary. This question makes sense if there is a bright line separating weakly emergent properties from merely resultant properties, but the truth about weak emergence is more complicated: Emergence comes in degrees. Assad and Packard (1992: 232) describe a scale for degrees of emergence, ranging over behavior that is “immediately deducible upon inspection of the specification or rules generating it”, to behavior that is “deducible in hindsight from the specification after observing the behavior”, and continuing to behavior that is “deducible in theory, but its elucidation is prohibitively difficult”, and fi nally reaching behavior that is “impossible to deduce from the specification”. Explanatory incompressibility can be arrayed into similar stages. So, since weak emergence depends on explanatory incompressibility, weak emergence also comes in stages or degrees.9 The paradigm case of weak emergence involves properties with incompressible explanations. A lower degree of emergence involves properties with compressible explanations that are so complicated that in practice no one can use the explanation except with a computer simulation. A higher degree of emergence involves properties that can be simulated in theory, but not in any fi nite simulation. These examples illustrate some of the sorts of ways in which weak emergence comes in degrees.10 The implication is that there is no sharp distinction between emergent and non-emergent phenomena, at least not for weak emergence defi ned via explanatory incompressibility. Instead, different phenomena, including natural phenomena, are more or less emergent, depending on their degree of explanatory incompressibility. Viewing emergence as a matter of degree contrasts sharply with the dichotomous defi nitions typically favored by philosophers such as McLaughlin (1992) and Kim (1999) but conforms to the approach suggested by Wimsatt (1986, 1997, 2000) and Bedau (1997, 2003, 2009).

Weak Emergence and Context-Sensitive Reduction 55 5. WEAK EMERGENCE DEPENDS ON REDUCTIONISM Embracing ontological and causal reduction permits weak emergence to avoid one of the traditional complaints against emergence. J. J. C. Smart (1963), for example, rejected emergentism because he favored the reductionistic view of the natural world as a very complicated mechanism. However, weak emergentism depends on exactly this view. Rather than rejecting reduction, weak emergentism requires (ontological and causal) reduction, for these are what make the complicated mechanisms work. The question whether any apparently emergent macro-level phenomena still resist micro-level explanation, even in principle, is controversial. What is uncontroversial is that weak emergence applies only to those phenomena that do have micro-level explanations (of a certain complex sort). So weak emergence would not cover any phenomenon the explanation of which is irreducible in principle. The possibility of completely explaining weakly emergent phenomena by crawling the micro-causal web entails that weak emergence is consistent with reductionism. Many philosophers and scientists assume that emergence and reduction are incompatible. One typical form of reductionism is mereological supervenience (Kim, 1978), the view that wholes are completely determined, ontologically and causally, by their parts. And it is certainly true that some kinds of emergence are incompatible with reduction; for example, strong forms of emergence are often defi ned in terms of reductive failure (e.g., Kim, 1999). However, weak emergence differs from strong emergence because it is consistent with many forms of reduction.11 To see this, consider reducing ontologies, causal relations, and explanations. Each concrete physical embodiment of weak emergence is ontologically nothing more than some kind of aggregation of smaller embodied objects. For example, the ontological substance of a traffic jam is nothing more than a certain aggregation of cars on a road. Similarly, the ontological substance of a vesicle is nothing more than a certain closed-bilayer aggregation of amphiphilic molecules in water. Furthermore, the causes and effects of each concrete instance of any kind of weakly emergent macro phenomenon can be fully explained by referring to only the aggregation of the causes and effects operating at the micro level, at least in principle (Hanczyc et al., 2003). So, each example of macro-level weak emergence is ontologically and causally reducible to micro-level phenomena. However, in practice, typically nobody can understand or follow such a complex micro-causal reduction unless they simulate the micro-causal web on a computer. In a wealth of interesting cases, studied in fields like soft artificial life, computer simulations make it possible to crawl simulated micro-causal webs and discover the resulting emergent macro patterns.12 The distinction between explanations or reductions that hold only in principle, versus those that also hold in practice, deserves further discussion. A reductive generative explanation of macro from micro might exist,

56

Mark A. Bedau

in principle, but be unhelpful for explaining weakly emergent phenomena, in practice, for a variety of reasons. One is that some relevant micro-level details required for the explanation might be unknown and inaccessible. So, weakly emergent phenomena always have complete and accurate explanations solely from micro-phenomena, at least in principle. It is easy to see why it is typically impossible for anyone to grasp or understand how the emergent macro phenomena unfold from the micro, in practice, without resorting to computer simulations. This leads to the special connection between weak emergence and computer simulations, discussed elsewhere (Bedau, 2008). The distinction between explanation and reduction in principle and in practice helps explain how weak emergence fits the two hallmarks of emergence mentioned earlier: the dependence of the macro on the micro, and the autonomy of the macro from the micro. In cases of weak emergence, the macro depends on the micro because, in principle, each instance of the macro ontologically and causally is nothing more than the aggregation of micro-causal elements. For example, the ontological and causal state of a cellular automaton macro structure is nothing more than the aggregation of the ontological and causal states of its micro constituents. At the same time, weak emergence exhibits a kind of macro autonomy because of the incompressibility of the micro-causal generative explanation of the macro structure. Being incompressible, the explanation is useless in practice, except insofar as it serves as the basis for a good simulation of the system. This gives weakly emergent properties at least a form of epistemological autonomy; they require macro-level explanation. I argue elsewhere that this epistemological autonomy is not merely epistemological (2003) and does not imply that weak emergence is merely in the mind of the beholder (2008). The subtle way in which weak emergence balances principles and practices is summarized with the awkward but apt notion of in-principle irreducibility in practice. Although weakly emergent phenomena have true, complete, and exact micro-level generative explanations, at least in principle, incompressibility makes the explanations useless, in practice. In practice, we have no alternative but to simulate the system’s micro-level behavior, if we want to observe what macro behavior will emerge. This is a practical limitation, a limitation on irreducibility in practice. Furthermore, this practical limitation holds in principle for any naturalistic epistemic agent that is trying to explain the behavior of complex systems. My slogan that weakly emergent phenomena are “in principle irreducible in practice” summarizes these two points. As an aside, we should note that in many contexts an especially important subspecies of weak emergence is robust weak emergence. Weak emergence is robust when it involves causally salient law-like patterns involving weakly emergent macro properties (Bedau, 2003). These robust emergent patterns recur as statistical regularities. Being typical or generic,

Weak Emergence and Context-Sensitive Reduction 57 they describe regularities and hence can be used to provide explanations. Many properties of the emergent patterns are insensitive to the details of the local micro interactions that produce the patterns, so the emergent patterns have multiple realizations. One interesting special case of robust weak emergence is the set of physical systems that exhibit what physicists call “universal” behavior, especially around phase transitions, such as when a solid melts into a liquid.13 Physicists in some instances have mathematically proved that the critical behavior of some large class of physical systems is insensitive to almost all details about the system, but in most cases one has merely empirical evidence that a physical system exhibits universal behavior. Nevertheless, this empirical evidence can be very strong (see, e.g., Stanley, 1971). Since the evidence is empirical, sometimes we are wrong when we think we have strong evidence that a system’s behavior is weakly emergent. But this is not a weakness in the notion of weak emergence. With all empirically justified claims we run the risk of being wrong. Stephen Weinberg’s (1987) vigorously defended reductionism illustrates how reasonable forms of reduction are closely related to both nominal and weak emergence. Weinberg wants to establish that “particle physics ... is in some sense more fundamental than other areas of physics” (347), in a way that is uncontroversial when properly understood. Weinberg carefully rejects certain other forms of reductionism, such as the claim that other sciences have no autonomy and all will be absorbed into elementary particle physics, saying: Even within physics itself, leaving aside biology, we certainly don’t look forward to the extinction of thermodynamics and hydrodynamics as separate sciences; we don’t even imagine that they are going to be reduced to molecular physics, much less to elementary particle physics. After all, even if you knew everything about water molecules and you had a computer good enough to follow how every molecule in a glass of water moved in space, all you would have would be a mountain of computer tape. How in that mountain of computer tape would you ever recognize the properties that interest you about the water, properties like vorticity, turbulence, entropy and temperature? (1987: 349) Now, Weinberg might never have entertained the concept of nominal emergence, as it is defi ned earlier in this chapter. Nevertheless, when he says that describing macro phenomena requires new emergent concepts that are undefi ned for micro-level objects, he is talking about nominal emergence. Consider a sample of water that is transparent and has some specific temperature. The water is composed of (virtually) nothing more than individual water molecules, and an individual water molecule cannot be transparent or have a temperature. Those properties are defi ned only for large aggregates of water molecules. Transparency and temperature of water are

58

Mark A. Bedau

two classic examples of emergence, and they fit the notion of nominal emergence perfectly. Furthermore, it is obvious in this example that nominal emergence involves no spooky metaphysics. The only entities that exist are the molecules of water in the cup; that is, the water in the cup. The reductionism that Weinberg defends entails that the laws (“generalizations”) of fundamental science explain non-fundamental laws. Weinberg’s example is that Newton’s laws, which are more fundamental than Kepler’s laws in some obvious and intuitive sense, explain Kepler’s laws. It is an empirical fact about nature that if we keep asking why some empirical fact obtains, our ultimate explanations refer to the behavior of the very small. Things need not have turned out this way, but the evidence continually and increasingly suggests that it is true. This attitude is clearly expressed in the following quote: In the same way, even though new concepts ‘emerge’ when we deal with fluids or many-body systems, we understand perfectly well that hydrodynamics and thermodynamics are what they are because of the principles of microscopic physics. No one thinks that the phenomena of phase transitions or chaos . . . could have been understood on the basis of atomic physics without creative new ideas, but does anyone doubt that real materials exhibit these phenomena because of the properties of the particles of which the materials are composed? (Weinberg, 1987: 350–351) Aside from another appearance of nominal emergence, the attitude in this passage is that the macro is the way it is solely because of what happens at the micro. Weinberg emphasizes that this reduction typically holds only in principle. By contrast, in practice we cannot work out the chemical behavior of DNA by solving Schrödinger’s equation, but we can in principle. So, Weinberg’s reductionism is not only consistent with nominal and weak emergence; it invokes and depends on them.

6. CONCLUSION This chapter has defended a form of weak emergence that is based on the notion of explanatory incompressibility. Weak emergence is the macrolevel mark of incompressible complexity in the network of micro-causal interactions. When the objective micro-causal web is sufficiently complex, explanations of the network’s macro behavior inevitably become incompressible. This explanation is consistent with many forms of naturalism, for it depends on only the dynamical micro-causal processes underlying complex phenomena. Complex causal processes occur not just in computer simulations or systems; they also occur throughout the natural world. Weak emergence is widespread in nature.

Weak Emergence and Context-Sensitive Reduction 59 A world with weak emergence contains important roles for reduction. First, as noted previously, the scientific unification provided by successful reduction is an important kind of scientific advance, because it shrinks the set of basic, brute facts about nature that we must assume. Also, tracing through all the explanatory arrows (crawling the micro-causal web) is a positive aspiration, not an evil to exterminate, for the contingent details are always illuminating. Of course, the relevant explanatory connections are often context-dependent and highly synergistic; in these cases of weak emergence, crawling the micro-causal web is the only way to explain the unfolding events. In a straightforward sense, illustrated by the denial of vitalism in biology, contemporary science agrees that the macro is the way it is because of how things are at the micro. Ernst Mayr, the archenemy of reductionism in biology, accepted this form of reductionism: “Every biologist is fully aware of the fact that molecular biology has demonstrated decisively that all processes in living organisms can be explained in terms of physics and chemistry” (1982, cited in Weinberg, 1987: 352). The huge advances in molecular biology in the past sixty years are among the great success stories in the history of science, and they support micro-causal reductionism. But it is important to recognize that those micro-causal reductions are typically context-sensitive, and involve many synergies among the micro elements. It is all too easy to forget that accepting these kinds of reasonable reductionism is completely consistent with also accepting weak emergence. I join Anderson (1972), Fodor (1974), Laughlin and Pines (2000), Mayr (1982), Putnam (1967), and others in accepting the scientific legitimacy, autonomy, and worth of today’s explanations of emergent phenomena in non-fundamental sciences. Fundamental sciences are not more important than non-fundamental sciences, and there is no reason to think that they will ever absorb and do all the work of the non-fundamental sciences. The universe contains epistemologically autonomous domains of phenomena. How epistemological autonomy is connected with ontological autonomy is a larger question discussed elsewhere (Bedau, 2008). People often mistake strong emergence for all emergence, so they often mistake reduction for the contrary of emergence. Even if we reject strong emergence, being pluralists, we still accept nominal and weak emergence. Weak emergence has an especially close connection with reduction, since it involves phenomena that are reducible in principle but also “in principle irreducible in practice”. Furthermore, in practice systems that are in principle irreducible in practice can be simulated on a computer, and the simulation works by crawling a simulated micro-causal web. This empirical process fi nally reveals what weakly emergent properties were latent in the micro-causal contingencies and the context-sensitive micro-causal rules.

60

Mark A. Bedau

NOTES 1. Although my point in the text is unaffected by this, note that this example really involves multiple levels of emergence, for we could split these levels more fi nely into macro (micelles), meso (polymers), and micro (monomers). See Rasmussen et al. (2001) for an analysis and a model of this situation with multiple levels of emergence. 2. Space considerations have made me omit some kinds of emergence, such as emergence relative to subjective observer surprise (Ronald et al., 1999). 3. The leading contender for genuine strong emergence is conscious mental states. Our inability to have found any plausible micro-level explanation for conscious mental states might reflect just our ignorance, but another possibility is that these phenomena really are strongly emergent. 4. The different variants of weak emergence in the literature include those by Wimsatt (1986, 1997), Rueger (2000a, 2000b), Boogerd et al. (2005), and Bedau (1997, 2003, 2008). Different conceptions of weak emergence focus on different kinds of explanatory complexity, but all agree that weak emergence involves some distinctive kind of explanatory complexity. The material in this and the next section is adapted from Bedau (2008). 5. Unfortunately, the phrase ‘weak emergence’ is used in different ways by different philosophers. For example, what I call “weak emergence” differs from what is called “weak emergence” by A. Stephan (2006) or by K. Balog (personal communication), and it does not apply to typical Braitenberg’s vehicles (an example of weak emergence from A. Beckermann, personal communication). 6. The explanation can apply to indeterministic systems by including complete information about the system’s indeterministic micro-state changes in the information from which the explanation is sought. System boundary conditions are handled in a similar fashion (see Bedau, 1997). My notion of incompressible explanation is closely connected with Chaitin’s notion of random sequence (Chaitin, 1975, 1988) and Wolfram’s subsequent notion of incompressible computation (Wolfram, 1985, 2002), as well as the notion of a dynamical system that must be simulated to discover its generic properties (Crutchfield et al., 1986). 7. The classic reference on the Game of Life is Berlekamp et al. (2004), and Wolfram (1985, 2002) are important references on cellular automata in general. For more on emergence and cellular automata, see Bedau (1997, 2003) and the references therein. 8. E.g., Bedau (1997) reviews the supple adaptation to the edge of disorder that emerges from Packard’s Bugs model. 9. See, also, Bedau (2003: 163). 10. In this chapter I will not take a stand on how to measure amounts of weak emergence. Paul Hovda (2008) developed one precise and explicit formal defi nition of degrees of weak emergence. Hovda defi nes the amount of simulation effort needed to derive something, which could also be interpreted as the amount of effort required for something’s generative explanation. 11. William Wimsatt (1986, 1997) often emphasizes this point. 12. It should be noted that in many cases we still do not know how to explain some natural regularities or patterns we seem to see in nature. One good example is the arrow of complexity in the evolution of life on Earth (Bedau, 2009). 13. See, e.g., Laughlin and Pines (2000), Laughlin (2006). Batterman (2002) is the fi rst philosopher to emphasize the connection between emergence and universality in physics.

Weak Emergence and Context-Sensitive Reduction 61 REFERENCES Anderson, P. W. (1972). More is different. Science, 177, 393–396. Assad, A., & Packard, N. H. (1992). Emergence. In M. A. Bedau & P. Humphreys (Eds.) (2008), Emergence: Contemporary Readings in Philosophy and Science (pp. 231–234). Cambridge, MA: MIT Press (citations refer to this edition). Originally appeared as Section 2 of Emergent colonization in an artificial ecology, in F. Varela & P. Bourgine (Eds.), Towards a Practice of Autonomous Systems: Proceedings of the First European Conference on Artificial Life (pp.143–152). Cambridge, MA: MIT Press. Baas, N. A. (1994). Emergence, hierarchies, and hyperstructures. In C. G. Langton (Ed.), Artificial Life III (pp. 515 –537). Redwood City, CA: Addison-Wesley. Batterman, R. (2002). The Devil in the Details. New York: Oxford University Press. Bedau, M. A. (1997). Weak emergence. Philosophical Perspectives, 11, 375–399. . (2003). Downward causation and autonomy in weak emergence. Principia Revista Internacional de Epistemologica, 6, 5–50. Reprinted in M. A. Bedau & P. Humphreys (Eds.) (2008), Emergence: Contemporary Readings in Philosophy and Science (pp. 155–188). Cambridge, MA: MIT Press (citations refer to this edition). . (2008). Is weak emergence just in the mind? Minds and Machines, 18, 443–459. . (2009). The evolution of complexity. In A. Barberousse, M. Morange, & T. Pradeu (Eds.), Mapping the Future of Biology: Evolving Concepts and Theories (pp. 111–130). Berlin-Heidelberg-New-York: Springer. Bedau, M. A., & Humphreys, P. (Eds.). (2008). Emergence: Contemporary Readings in Philosophy and Science. Cambridge, MA: MIT Press. Berlekamp, E. R., Conway, J. H., & Guy, R. K. (2004). What is life? In E. R Berlekamp, J. H. Conway, & R. K. Guy (Eds.), Winning Ways for Your Mathematical Plays (2nd ed., Vol. 4, pp. 927–961). Wellseley, MA: A K Peters. Boogerd, F. C., Bruggeman, F. J., Richardson, R. C., Stephan, A. & Westerhoff, H. V. (2005). Emergence and its place in nature: A case study of biochemical networks. Synthese, 145, 131–164. Broad, C. D. (1925). The Mind and Its Place in Nature. London: Routledge and Kegan Paul. Chaitin, G. J. (1975). Randomness and mathematical proof. Scientific American, 232 (May), 47–53. . (1988). Randomness in arithmetic. Scientific American, 259 (July), 80–85. Chalmers, D. J. (1996). The Conscious Mind: In Search of a Fundamental Theory. New York: Oxford University Press. Crutchfield, J. P., Farmer, J. D., Packard, N. H. & Shaw, R. S. (1986). Chaos. Scientific American, 255, 46–57. Fodor, J. (1974). Special sciences: Or the disunity of science as a working hypothesis. Synthese, 28, 97–115. Harré, R. (1985). The Philosophies of Science. Oxford: Oxford University Press. Hanczyc, M. M., Fujikawa, S. M., & Szostak, J. W. (2003). Experimental models of primitive cellular components: Encapsulation, growth, and division. Science, 302, 618–622. Hovda, P. (2008). Quantifying weak emergence. Minds and Machines, 18, 461–473. Jackson, F., & Pettit, P. (1992). In defense of explanatory ecumenism. Economics and Philosophy, 8, 1–21. Kim, J. (1978). Supervenience and nomological incommensurables. American Philosophical Quarterly, 15, 149–156.

62

Mark A. Bedau

. (1992). Downward-causation in emergentism and nonreductive physicalism. In A. Beckermann, H. Flohr, & J. Kim (Eds.), Emergence or Reduction? Essays on the Prospects of Nonreductive Physicalism (pp. 119–138). Berlin: de Gruyter. . (1997). The mind-body problem: Taking stock after forty years. Philosophical Perspectives, 11, 185–207. . (1999). Making sense of emergence. Philosophical Studies, 95, 3–36. Laughlin, R. (2006). A Different Universe: Reinventing Physics from the Bottom Down. New York: Basic Books. Laughlin, R., & Pines, D. (2000). The theory of everything. Proceedings of the National Academy of Science U.S.A., 97, 28–31. Mayr, E. (1982). The Growth of Biological Thought. Cambridge, MA: Harvard University Press. McLaughlin, B. P. (1992). The rise and fall of British emergentism. In A. Beckerman, H. Flohr, & J. Kim (Eds.), Emergence or Reduction? Essays on the Prospects of Nonreductive Physicalism (pp. 49–93). Berlin: de Gruyter. Reprinted in M. A. Bedau & P. Humphreys (Eds.) (2008), Emergence: Contemporary Readings in Philosophy and Science (pp. 19–59). Cambridge, MA: MIT Press. O’Connor, T. (1994). Emergent properties. American Philosophical Quarterly, 31, 91–104. Putnam, H. (1967). Psychological predicates. In W. H. Capitan & D. D. Merrill (Eds.), Art, Mind, and Religion (pp. 37–48). Pittsburgh: University of Pittsburgh Press. Rasmussen, S., Baas, N. A., Mayer, B., Nillson, M., & Olesen, M. W. (2001). Ansatz for dynamical hierarchies. Artificial Life, 7, 329–353. Reprinted in M. A. Bedau & P. Humphreys (Eds.) (2008), Emergence: Contemporary Readings in Philosophy and Science (pp. 305–334). Cambridge, MA: MIT Press (citations refer to this edition). Ronald, E. M. A., Sipper, M., & Capcarrère, M. S. (1999). Design, observation, surprise! A test of emergence. Artificial Life, 5, 225–239. Rueger, A. (2000a). Robust supervenience and emergence. Philosophy of Science, 67, 466–489. . (2000b). Physical emergence, diachronic and synchronic. Synthese, 124, 297–322. Simon, H. (1996). The Sciences of the Artificial (3rd ed.). Cambridge, MA: MIT Press. Smart, J. J. C. (1963). Philosophy and Scientific Realism. London: Routledge and Keagan Paul. Sperry, R. W. (1969). A modified concept of consciousness. Psychological Review, 76, 532–536. Stanley, H. E. (1971). Introduction to Phase Transitions and Critical Phenomena. New York: Oxford University Press. Stephan, A. (2006). The dual role of ‘emergence’ in the philosophy of mind and in cognitive science. Synthese, 151, 485–498. Weinberg, S. (1987). Newtonianism, reductionism, and the art of congressional testimony. Science, 330, 433–437. Reprinted in M. A. Bedau & P. Humphreys (Eds.) (2008), Emergence: Contemporary Readings in Philosophy and Science (pp. 345–357). Cambridge, MA: MIT Press (citations refer to this edition). Wimsatt, W. C. (1986). Forms of aggregativity. In A. Donagan, A. N. Perovich, Jr., & M. V. Wedin (Eds.), Human Nature and Natural Knowledge (pp. 259–291). Dordrecht: Reidel. . (1997). Aggregativity: Reductive heuristics for fi nding emergence. Philosophy of Science, 64, S3720–S384. . (2000). Emergence as nonaggregativity and the biases of reductionisms. Foundations of Science, 5, 269–297.

Weak Emergence and Context-Sensitive Reduction 63 Wolfram, S. (1985). Undecidability and intractability in theoretical physics. Physical Review Letters, 54, 735–738. . (2002). A New Kind of Science. Champaign, IL: Wolfram Media.

4

Two Varieties of Causal Emergentism Michele Di Francesco

1. INTRODUCTION: DEPENDENCE AND AUTONOMY OF MENTAL PHENOMENA Contemporary philosophy of mind is characterized by the difficult attempt to conciliate in a unified theory of the psychophysical nexus two apparently contradictory aspects of mental phenomena: dependence and autonomy. On the one hand—avoiding the well-known difficulties of Cartesian substance dualism—widespread contemporary respect for naturalism requires mental phenomena to be grounded in physical phenomena. We take the very existence of mentality in our universe to depend on the physical organisation of the world. On the other hand, many reasons have been advanced to claim that mental phenomena cannot be identified or reduced to physical phenomena, and a certain degree of autonomy should therefore be accorded to mental reality. In this sense we may take reductive physicalism and Cartesian substance dualism as the two extremes of a continuum of philosophical views—where the position of each view in the continuum is determined by the degrees of dependence/autonomy it allows. Dependence and autonomy can be seen as the fundamental ingredients of our philosophical recipes: according to the degree of ‘materialistic’ dependence or ‘Cartesian’ autonomy we allow, we may cook the philosophical dishes we find more palatable. Among the various recipes recently proposed, emergentism has been the object of increasing interest. We shall describe it more accurately later, but for the moment we may say that emergentism rejects the idea of the division of our universe into two independent ontological regions, and in particular rejects the idea that mental phenomena should be considered as something external to the material world, something added to it “from outside”,1 so to speak. At the same time, emergentism assumes that our natural world contains irreducible mental properties, that is, even though everything that exists is physical, there are physical entities that have irreducible mental properties. In particular, when our physical reality arrives at a certain degree of complexity, it reveals new features whose nature and existence is not explicable by mere knowledge of the lower (physical) level, and whose action exhibits new genuine causal powers.

Two Varieties of Causal Emergentism 65 There are many reasons for contemporary interest in emergentism. Some come more directly from the scientific field, for example from the application of the complex system theory to mental phenomena, typical of the externalist and enactive models of cognition.2 Others are motivated by more philosophically oriented reflections about the prospects of physicalism (and in particular non-reductive physicalism).3 In this chapter I am mainly interested in the latter. In particular, reference to emergentism may be helpful in dealing with what I take to be two contradictory trends that may be found in the present discussion of the mental. The fi rst is the physicalistic reading of the extraordinary progress of the brain-sciences (in particular in the fields of consciousness research and social neuroscience), where an eventual ‘materialistic’ explanation of the mind in terms of brainmechanism is envisaged by many top scientists. The second comes from the persisting difficulties posed to the possibility to reduce the mind to the physical by what we may call the hard problems of reductionism: the difficulty of offering a reductive analysis of crucial mentalistic notions such as phenomenal consciousness, intentionality, subjectivity and selfhood, and rational agency. In fact, I take this kind of tension as the best example of the difficulty in conciliating dependence and autonomy of the mental. My main question, then, will regard whether emergentism can be taken as a coherent position from which to solve this difficulty. (I am thus mainly interested in the coherence of emergentism, an issue that precedes that of its truth.) My answer will be negative: it is disputable that we can associate to the label “emergentism” a single view, capable of reconciling dependence and autonomy in an equilibrated and coherent way. Rather there are many forms of emergentism, which tend to stress the former or the latter aspect of the psychophysical nexus: in particular, a moderate reading of emergentism tends to assimilate it to (non-reductive) physicalism, while a radical one falls short of full-blooded dualism. I shall speak in this connection of moderate and radical emergentism. Even though it is important to distinguish these two forms of emergentism (which differ in substantial aspects), I do not consider the presence of a plurality of emergence relations as a bad thing, since I claim that we should be prepared to adopt different models of the foundational links among levels of reality, according to the particular way we experience them. (To give just one example, we should be prepared for the possibility of explaining the relation between the sub-personal and personal level of mental phenomena in a different way from the relation between a brain state and the representational content associated to it.) I shall proceed in the following way: fi rst, I shall introduce two platitudes that should characterize emergentism: (a) emergentism is not dualism; (b) emergentism is not reductionism. I shall then endeavor to show that the attempt to conciliate them—by means of notions like the “causal inheritance principle”—leads to interesting consequences for our view of the relation between emergence, reduction, and causality, the most interesting being the possibility of conceiving a continuum of emergentistic views,

66

Michele Di Francesco

in place of a single notion of emergence. My fi nal suggestion will be that we should accept the existence of a plurality of emergence relations, each of them sensitive to a different way of conceiving the relations among the different levels of reality.

2. EMERGENTISM IS NOT DUALISM. FROM SUPERVENIENCE TO NOVELTY The fundamental idea that lies behind emergentism is easy to state; as Jaegwon Kim puts it, “the intuitive idea of an emergent property stems from the thought that a purely physical system composed exclusively of bits of matter, when it reaches a certain degree of complexity in its structural organisation, can begin to exhibit genuinely novel properties not possessed by its simpler constituents” (2006: 548). Emergent properties, then, are ‘irreducible’ (to ‘basal’ properties), ‘novel’ and ‘unpredictable’ (in terms of previous physical knowledge), and “they are not ‘explainable’ or ‘mechanistically reducible’ in terms of their underlying properties” (Kim, 1999: 5). This characterization works mainly in terms of irreducibility and novelty. Another crucial aspect of emergence concerns causality: emergent properties have genuine causal powers: “emergents bring into the world new causal powers of their own, and, in particular, [ . . . ] they have powers to influence and control the direction of the lower processes” (Kim, 1999: 5–6). In this perspective, emergentism involves “downward causation”. In fact, reference to causality should be considered as crucial to characterizing emergentism. In his seminal paper on British emergentism Brian McLaughlin (1992: 50) takes emergentism as “a view about the causal structure of reality”. Emergentism claims that reality is structured in different levels, and that when a given level reaches a certain degree of complexity, new causal powers emerge. As we shall see, one reason to concentrate on causality is that reference to the causal efficacy of mental properties works as a sort of litmus test to distinguish radical emergentism from weaker forms of non-reductive physicalism. Irreducibility, novelty, and causal efficacy easily differentiate emergentism from reductive physicalism. According to emergentism, mental properties cannot be reduced to physical properties; a purely physical description of the world is not complete. However, it is important to remember that emergentism is not dualism. Emergent properties are “not added from outside”; as Tim Crane put it, “the rough idea is that these features of objects are genuinely ‘novel’ in the sense that they are not purely ‘consequences’ of microscopic parts, and yet they are not ‘added from outside’ in the way that is claimed by (for example) a Cartesian conception of mental properties, or a vitalist conception of biological properties” (2007: 197). The next question, then, is how to explain the idea that, even though they are not purely consequences of lower-level properties, emergent properties are grounded

Two Varieties of Causal Emergentism 67 in them. A standard answer to this question appeals to the notion of supervenience. Emergent properties nomically supervene on physical properties. They are not logically supervenient from physical properties, since their existence is not deducible from physics alone,4 but it is a fact about our universe that certain configurations of matter bring out certain (emergent) mental phenomena. Crane (2007) claims that no form of physicalism (even non-reductive physicalism) should take supervenience as a brute fact. This is an important point. But what about emergentism? In a certain sense it may be stated that, contrary to physicalism, it is natural for emergentism to take the existence of supervenience as a brute fact—to be accepted with “natural piety”, to quote Alexander’s famous dictum (1920, II: 47). However, the thesis that supervenience cannot be explained within physics, so to speak, should be made compatible with the anti-dualistic constraints that require that emergent properties are not taken from outside. This seems to require a link between ‘basal’ and emergent properties. Such a link is hard to fi nd if we concentrate on the notions of novelty, irreducibility, and unpredictability—all characteristics of emergent properties that render them inexplicable in physical terms. But if we approach the issue from the point of view of causality, we find that something more can be said about the “brute fact” of supervenience.

3. EMERGENTISM AND CAUSAL INHERITANCE Assuming then that higher-level features of the world are not “added from outside”, but supervene on a physical basis, we may ask whether we should try to explain supervenience, or take it as a brute fact. The claim that supervenience needs explanation seems reasonable within a physicalistic perspective. What about emergentism? My answer will be (a) that even emergentism can address the question of the explanation of supervenience, and (b) that non-reductive physicalism and emergentism differ to the extent of the required explanation. In fact, we may posit a continuum of ‘explanations’, ranging from the physical reductive explanation up to the mere reference to supervenience as a brute fact. As we noted before, we may take reductive physicalism as committed to the idea of “cosmic hermeneutics”, the possibility of the deduction of all the truths about the world from microphysics (Crane, 2007: 194). Since they deny cosmic hermeneutics, however, non-reductive physicalism and emergentism should offer a different solution to the problem of supervenience.5 The main difficulty is that all the ingredients for the existence of emergent mental phenomena should be present at the (lower) physical level, but their existence is detectable only at the (higher) mental level, where they exhibit new genuine causal powers. A way to explain this puzzling state of affairs may be to adopt what Jaegwon Kim (1999) calls the causal inheritance principle (CIP), according to which the causal powers of the emerging

68

Michele Di Francesco

properties are the product of the causal powers of the “basal” properties. Kim (1999) characterizes his version of CIP as a thesis about the identity of causal powers between an emergent functional property and its physical realizers: “If a functional property E is instantiated on a given occasion in virtue of one of its realizers Q being instantiated, then the causal powers of this instance of E are identical with the causal powers of this instance of Q” (Kim, 1999: 16). The causal powers of emergent properties, then, are the same as their basal properties. This seems to imply that the only activities endowed with causal efficacy are those of the lower level—even if they can be detected and described only at the higher level. In this case there is no need for extra ingredients to explain the causal efficacy of emergent properties (and any reference to downward causation amounts to a mere façon de parler). This model of causal inheritance seems to be workable when there is the possibility to give a functional analysis of a determinate emergent property, that is when (a) we know how to describe in functional terms the essential traits of the emergent phenomena we are dealing with, and (b) we know some of its physical realizers, at least in principle. This is a weak sense of emergentism, however, not only because the entire causal job is done at the level of physical realizers, but also because the functional analysis of emergent properties is not always available. In the case of functional analysis, what makes the higher-order properties emergent is their irreducibility and novelty, the necessity to collect under new higher-order concepts many classes of phenomena which are not homogeneous from the physical point of view, but play the same functional role. It may be disputable whether this position (we may label it minimal emergentism) deserves the name of emergentism, however. The reason is that to understand the supervenience relation in terms of functional realization is in some sense at odds with a bold reading of the emergence relation: one of the main motivations of emergentism is in fact the idea that emergent properties cannot be located (i.e. are not realized) at the lower level from which they supervene;6 they are detectable only at the higher level where their causal powers manifest themselves. If, on the contrary, we maintain that emergent properties are functionally realized, their causal efficacy is at risk. As Kim himself puts it, in a different but related context, higher-order mental connections that have the appearance of causal connections should be explained as not really causal: These regularities are by no means accidental, in a clear sense they are law-based and may even support appropriate counterfactuals. However, if we understand the difference between genuine, productive and generative causal processes, on the one hand, and the non-causal regularities that are observed because they are parasitic on real causal processes, we are in a position to understand the picture [ . . . ]. (Kim, 1998: 45)

Two Varieties of Causal Emergentism 69 In this perspective, the only “genuine, productive and generative causal processes” are those operating at the base level, while higher-order mental properties are “parasitic on real causal processes”. This way of assuring that emergent properties are not “added from outside”, then, seems at odds with the second platitude mentioned earlier: that emergentism is not reductionism. It is true that minimal emergentism may reject conceptual reduction because the mental idiom is necessary to collect under a single concept an “open” plurality of physical realizers of certain functional roles. But from the point of view of causality, the causal efficacy of mental properties is explained away (and in this sense reduced) by reference to physical properties. This kind of view can be assimilated then to a form of non-reductive physicalism, but it is really difficult to claim that it is able to satisfy all the philosophical reasons that motivate recent interest in emergentism. A major problem is that when dealing with what we have labeled the “hard problems” of reductionism, such as phenomenal consciousness, intentionality, subjectivity and selfhood, and rational agency, functional analysis is not available (according to the emergence-oriented philosophers, at least). This is a well-known point, often made with reference to the problem of phenomenal consciousness. But it can be generalized to all those cases in which there are deep differences of explicative style between base and emergent levels. We encounter differences of this kind, for example, when we move from the neurobiological to the intentional description of mentality (connecting ‘syntactical’ brain functioning with ‘semantical’ mental content), or when we look for the conceptual link that connects the cognitive to the phenomenal aspects of consciousness; or, more generally, when we switch from the sub-personal to the personal way to account for selfhood and agency.7 In all these cases, we may fi nd some correlation between levels (even neural correlates in certain circumstances), but we are quite often far from any functional analysis. The best we can do is to signal the presence of a correlation, and to express it in terms of an “emergence law” (Kim, 2006: 551). If certain brain structures are activated, then a certain phenomenal state emerges; when I consciously undertake a certain decision, certain brain structures are more active than usual; a fi rst-personal perspective emerges from a dynamical interaction between bodily self-awareness and ‘ecological’ cues to action, and so on.8 In all these cases, we have to take emergence as a brute fact. We may even predict emergence, but the prediction cannot be made ‘within physics’; we have to employ bridge laws (“emergence laws”) that incorporate our previous knowledge of the emergent structure of the world.9

4. MODERATE AND RADICAL EMERGENTISM We seem then to be confronted with two alternative views of emergentism, a moderate and a radical one. The former accepts the causal inheritance

70

Michele Di Francesco

principle (in its functional reading): the causal powers of the emerging properties are the product of the causal powers of the “basic” properties. The basic (physical) level is the only level that exhibits “genuine, productive and generative causal processes”. The latter, radical emergentism, denies causal inheritance and takes the emergence of new causally efficacious properties as a brute fact. Higher-level systems exhibit new kinds of causal organization, and such organization derives from the action of new kinds of properties, whose existence is not only undetectable at the basic level but should be considered as representing new emergent features of the world. Moderate emergentism may be viewed as an attempt to bring together a physicalistic ontology with the plurality of explicative styles, conceptual tools, and languages that we fi nd in our current investigations of reality.10 Radical emergentism takes this epistemological pluralism as a cue for the ontological reality of emergent properties. The causal inheritance principle, then, is rejected: the new causal powers are truly novel, and the involved downward causation is independent of causal intercourse at the basic level. Radical emergence is a kind of emergence that undermines the (causal) unity of the world that we associate with physicalism.11 The difference between moderate and radical emergentism should not be underestimated. They reflect very different conceptions of the psychophysical nexus. Moderate emergentism reflects our willingness to recognize the explicative merits of higher-order conceptual descriptions, typical of a given level of analysis and investigation of the world, and at the same time to be loyal to a physicalistic ontology. In the case of radical emergentism our desire to be loyal to a physicalistic ontology leaves room for the necessity to interpret the indispensability of the higher-order conceptual descriptions as the signal of the presence of ‘ontological novelties’ that characterize the different ‘levels’ or ‘domains’ of reality. The distinction between moderate and radical emergentism admits intermediate positions, however, since we may offer ‘reductive’ and ‘non-reductive’ readings of the causal inheritance principle. To consider the possibility of these intermediate positions, we may refer to Sidney Shoemaker’s notion of latent properties. Shoemaker (2002) distinguishes between manifest and latent properties of a whole. Manifest properties are both present and detectable at the ‘basic’ level; latent properties are present but not detectable. Latent properties are grounded on the base level, but remain unrealized until the system in which they occur reaches the required complexity (Shoemaker, 2002: 54). In other words, manifest properties have causal powers that can be detected at a lower degree of complexity, while latent properties emerge and can be detected only when the system acquires the required degree of complexity. According to Shoemaker, then, when we say that the emergent properties of a whole supervene on the properties of its parts, and that the new causal powers of the emergent whole are inherited from the lower level—we should count both manifest and latent properties among these properties:

Two Varieties of Causal Emergentism 71 The component entities have powers that, collectively, determine the instantiation of the emergent property when they are combined in the emergence engendering way. But, these being cases of emergence, these cannot all be powers that manifest themselves when the components are not combined in emergence engendering ways. (Shoemaker, 2002: 55) In this connection, a new sense can be attributed to the causal inheritance thesis: emergent properties are causally efficacious because they derive their causal powers from “basic” properties. However, we have two kinds of basic properties: manifest and latent. And the novelty involved in the emergent properties is justified by the peculiar nature of latent properties, which express their causal powers only at the emergent level. Both physicalism and emergentism, then, claim that micro-facts determine macro-facts and that micro-causal powers determine macro-causal powers, with the crucial difference that, on the emergentist view, micro-facts include the instantiation of micro-latent properties. A critical aspect of this proposal is the relation envisaged between emergent and physical micro-structural properties. If we wish to exploit the idea of latent properties in order to bridge the explanatory gap, we will need something like a determination relation: “as a matter of nomological necessity [ . . . ] any micro-entity having that micro-manifest power, or that set of micro-manifest powers, also has that micro-latent power” (Shoemaker, 2002: 57). One question that should be answered, however, concerns the nature of the nomological nexus here postulated. Are the laws of nature that express it to be understood as basic physical laws? Or should we rather conceive them as ‘emergence laws’ of some sort? This issue is connected to the question of the ontological status of latent properties. If we take them as physical properties, we obtain a new variety of moderate emergentism which accepts an epistemological but not an ontological reading of causal pluralism (latent properties cannot be detected, but are still present at the physical level). The claim that latent properties are physical, however, is disputable, as they are individuated only with reference to the emerging system in which they are detectable.12 If we consider them as physical, they work as “hidden variables”: on the basis of our physicalistic assumptions, we claim that there must be unknown physical determinants of every mental phenomenon. Latent properties may, then, be suspected of having been introduced as an ad hoc hypothesis just to defend our physicalistic bias against emergentistic evidence. This is particularly true in all cases in which there is a strong heterogeneity of explanatory style between higher-level and lowerlevel causal intercourse. A further obstacle to conceiving latent properties as physical is that, since they are individuated with reference to the emergent level, in a certain sense they depend ontologically, in their metaphysical essence, on the existence of emergent levels of reality at which the new causal organisation of the world manifests itself. Taking a very different

72

Michele Di Francesco

perspective from Shoemaker’s, we may even think of a “double aspect” reading of the notion of latent property, according to which the possession of certain physical properties is ipso facto the possession of correlative mental properties (although this kind of property manifests itself only at certain degrees on complexity). From the point of view of the physical description of the world, this double aspect would be taken as a brute fact. Of course such a radical alternative lends the notion of latent properties an unquestionable (and perhaps unpalatable) dualistic flavor. It may be useful to sum up the different relations between the causal inheritance principle and emergentism we have presented thus far. We may roughly depict four alternatives: 1. Causal inheritance depends on physical causal interactions, and is in principle explainable within physics. Causal pluralism reflects only our present state of incomplete knowledge of how to provide a functional analysis of mental properties. 2. Causal inheritance depends on physical causal interactions, but to be explained, requires reference to latent properties, whose action can be detected only at the higher levels. An epistemic reading of causal pluralism is tolerated. 3. Latent properties are ‘non-physical’ properties that are already present at the basic level; higher-order causal powers are inherited—but they do not originate from physical properties (alone). 4. Higher-order causal powers are grounded on physical phenomena, but such grounding is a brute fact. Causal inheritance does not occur.

5. EMERGENCE AND CAUSALITY IN A PLURALISTIC UNIVERSE All this talk about the causal inheritance principle, latent properties and the like shows the tension between the contrasting requirements of dependence and autonomy described at the beginning of this paper. The adoption of the causal inheritance principle avoids the risk of taking emergent properties as “added from outside” the natural world. But it is in contrast—in the original Kimian version, at least—with the requirement of autonomy. In the ‘realizer’ reading of it, CIP guarantees that we may appeal to causal power at the physical level to explain higher-level causality. Even if we are unable to unify (and describe) in a single macroconnection all the micro-facts that act as the supervenience basis of higher-level causal connections, we are granted that they exist. We have, then, ontological causal monism together with epistemological causal pluralism. Things are not substantially different in case we adopt a reading of the idea of “latent” properties which considers them as nomologically necessitated by physical laws. In both cases we may speak of two members

Two Varieties of Causal Emergentism 73 of the same family of emergentism: moderate emergentism. According to this view, anti-reductionism and causal pluralism receive an epistemological reading, while the substantial (causal) unity of the world is preserved at the physical level. For this reason this moderate emergentism may be considered as a form of non-reductive physicalism. A different kind of emergentism, radical emergentism, collects together a family of positions that either reject the causal inheritance principle, or give it a non-physicalistic reading. Here any commitment to physicalism seems to be lost—and it would be imprudent to define such a position as a form of non-reductive physicalism, its acceptance of the supervenience thesis notwithstanding. The main reason is that radical emergentism does take supervenience as a “brute fact”,13 and it is committed to ontological causal pluralism. One lasting connection with “physicalism” lies in its commitment to the generality of physics thesis (“the claim that all objects have physical properties”, that is “that the laws of physics apply unrestrictedly across the universe; there are no regions where these laws fail or break down” (Crane, 2007: 200)). Together with its acceptance of the supervenience of the mental (as a brute fact), adherence to the generality of physics leads radical emergentism to the rejection of Cartesian dualism, and in this sense may be associated to one of the weakest forms of naturalism we can think of.14 Radical emergentism is connected to an ontological reading of causal pluralism. That is, it is ready to give a strict ontological reading of statements such as the following: The world, it seems, runs in parallel, at many levels of descriptions. You may fi nd that perplexing; you certainly aren’t obliged to like it. But I do think we had all better learn to live with it. (Fodor, 1997: 162) The idea here is to read Jerry Fodor’s claim as stating that at each level of description of the natural order there are laws of nature that are expressions of causally efficacious and irreducible emergent properties.15 We should note that in this perspective a limited unity of the world is preserved by the required “parallelism” among levels. We may take a more extreme direction, however, taking advantage of such radical appeals to the disunity of the world as the following passage from Nancy Cartwright: We live in a dappled world, a world rich in different things, with different natures, behaving in different ways. The laws that describe this world are a patchwork, not a pyramid. They do not take after the simple, elegant and abstract structure of a system of axioms and theorems. Rather they look like—and steadfastly stick to looking like—science as we know it: apportioned into disciplines, apparently arbitrarily grown up; governing different sets of properties and different levels of abstractions. (1999: 1)

74

Michele Di Francesco

This ontological reading of the disunity of science is quite extreme. Shifting from the level-model to the patchwork-model of emergence may be seen as a very hazardous and dubious move: if we interpret in a literal and factual sense the idea of Nature as “governed in different domains by different systems of law not necessarily related to each other in any systematic or uniform way” (1999: 31), we may lose sight of the lasting sense of the unity of the world that seems to characterize (classical) emergentism when it describes the development from physical simplicity to higher-phenomena complexity as a natural ‘evolution’ of our universe. This change of attitude may be motivated by a conception of the relation between ontology and epistemology different from the usual one: a conception that favors a concrete analysis of our explicative practices against metaphysical bias.16 If we believe that our metaphysical assumptions should follow our explicative practices, and not the other way around, we should be prepared to transfer pluralism from our analysis of what real science looks like to our overall world view. A denial of the causal unity of the world, however, may be considered an unwise and imprudent hypothesis. But there is an alternative, even for radical emergentism: to look for the unity of the world not at the bottom, but at the top of it, so to speak; it is at the emergent level that all causal powers merge together. In this perspective, to consider the physical world is to look at it in abstraction from all the higher-level causal interactions. Such a view is still radical, since it requires an interpretation of the emerging systems as ‘more real’, more effective, than their physical components, as far as causality is concerned. From the top-level perspective, when physical phenomena contribute to the rise of new causal efficacious properties, they have causal powers they would not have if they were not part of an emergent system. This may require the association of the idea of “levels of reality” to the idea of “degrees of reality”, taking higher levels as fundamental. The thesis that the unity of our world is gained at the emergent level and not the physical is quite unusual, and seems to reverse the traditional explicative route from everyday experience and its (physical) basis. But anyone prepared to give the (purported) impossibility of offering a reductive analysis of causally efficacious emergent properties an ontological reading should be prepared to take this inversion into consideration. An accurate discussion of this issue would lead us far beyond the scope of this paper. So I shall limit myself to the less ambitious conclusion that (as far as causality is concerned) we should carefully distinguish between moderate and radical emergentism. More generally the notion of “emergence” reveals a fundamental ambiguity, which is connected to the problematic ambition to preserve both the dependence and autonomy of mental phenomena. Moderate emergentism can be seen as an attempt to safeguard—within an anti-reductionist framework—a view of the world as determined by its physical constituents. Radical emergentism does not fit suitably with the very idea of determination, and in this sense undermines the (causal) unity of the world. In a certain sense, the presence

Two Varieties of Causal Emergentism 75 of this ambiguity of the notion of emergence is unfortunate, since it may engender confusion. In another sense, however, the presence of a plurality of conceptual tools for analyzing the psychophysical connection may be considered as an opportunity. In this (optimistic) reading, we may take emergentism (or better, the plurality of ways of conceiving emergence) as a flexible conceptual tool in all those cases in which we look for the ontological background that best fits with the actual state of knowledge in a certain area. In this sense, once we avoid ambiguities and recognize that a unified reading is not available, there seem not to be principled reasons to choose a priori among the different readings of emergentism. We should rather take a more empirical and open-minded attitude, and be prepared to adopt different models of the foundational links among levels of reality, according to the particular way we experience them. NOTES 1. Cf. Crane (2007: 197). See also Crane (2001), Kim (1999, 2006), McLaughlin (1992), O’Connor and Wong (2006). 2. Cf. Clark (1997), O’Connor and Wong (2006), Thompson and Varela (2001). 3. Non-reductive physicalism claims that everything that exists is physical, but there are physical entities that have irreducible mental properties. Non-reductive physicalism adopts the supervenience thesis, too. It claims that there cannot be two possible worlds that are physically identical but differ in their mental properties. Non-reductive physicalism is still quite popular, but it has recently come under attack by many philosophers who questioned its internal coherence. Cf. Crane (2001, 2007), Kim (1998). 4. Cf. McLaughlin (1992), O’Connor and Wong (2006). Contrary to emergentism, reductive physicalism postulates the possibility of “cosmic hermeneutics” (Horgan, 1984; Jackson, 1998), the idea that “there can be an a priori deduction of all the truths about the world from the microphysical truths about it plus the conceptual truths about non-physical concepts” (Crane, 2007: 194). 5. Even if I start from Crane’s analysis of the problem, what follows is not Crane’s solution. Crane argues that the non-reductive physicalist has to explain supervenience, but the only option open to her is to bridge the “explanatory gap”. This option is not available, however, since a non-reductive physicalist does not believe that the gap can be closed by a reductive strategy. For a general comment on Crane’s analysis, cf. Di Francesco (Forthcoming). 6. Cf. Shoemaker (2002: 57). 7. All these examples of inexplicability may be disputed by the reductionist, of course. But they are taken seriously by emergentism. They even motivate the interest in emergentism; so if minimal emergentism is not able to take them into consideration, it can hardly be considered apt to satisfy all the emergentistic objectives. 8. For a discussion of the notion of emergence in the last two cases, cf., respectively, Di Francesco, Motterlini, and Colombo (2007) and Di Francesco (2008). 9. Cf. Kim (1999, 2006) and Crane (2001, 2007) for important further reflections on these issues.

76

Michele Di Francesco

10. Cf. Di Francesco (2005) for an analysis of the relation between emergentism and causal pluralism. 11. In this sense radical emergentism falls dangerously close to dualism. We shall try to offer a less metaphysically dangerous reading of radical emergentism later on. 12. Shoemaker himself examines this question (2002: 60) but offers a different solution; for a critical discussion cf. Di Francesco (2005: 112–114). 13. Or at least a fact that cannot be explained ‘within physics’. 14. On the possibility of moderate forms of naturalism, cf. De Caro and Macarthur (2004). 15. I am not attributing this reading to Fodor himself. 16. Cf. for example, Baker (1993), Burge (1993), Dupré (1993). On the same issue, cf. Di Francesco (2003, 2005). Reference to science ‘as it is’ may of course be contrasted with reference to science as it should (or will) be. But here emergentism would accuse its physicalistic opponent of unjustified metaphysical bias.

REFERENCES Alexander, S. (1920). Space, Time, and Deity. 2 vols. London: Macmillan. Baker, L. R. (1993). Metaphysics and mental causation. In J. Heil & A. Mele (Eds.), Mental Causation (ch. 6). Oxford: Clarendon Press. Burge, T. (1993). Mind-body causation and explanatory practice. In J. Heil & A. Mele (Eds.), Mental Causation (ch. 7). Oxford: Clarendon Press. Cartwright, N. (1999). The Dappled World. A Study of the Boundaries of Science. Cambridge: Cambridge University Press. Clark, A. (1997). Being There. Putting Brain, Body and World Together Again. Cambridge, MA: MIT Press. Crane, T. (2001). The significance of emergence. In C. Gillett & B. Lower (Eds.), Physicalism and Its Discontents (pp. 207–224). Cambridge: Cambridge University Press. . (2007). Cosmic hermeneutics vs. emergence: The challenge of the explanatory gap. In M. Beaney, C. Penco, & M. Vignolo (Eds.), Explaining the Mental. Naturalist and Non-Naturalist Approaches to Mental Acts and Processes (pp. 192–206). Newcastle: Cambridge Scholar Publishing. De Caro, M., & Macarthur, D. (Eds.). (2004). Naturalism in Question. Cambridge, MA: Harvard University Press. Di Francesco, M. (2003). What is consciousness for? From epiphenomenalism to causal efficacy. Proceedings of the International Workshop, Exploring Consciousness. Humanities, Natural Science, Religion (pp. 85–106). Milan: Fondazione Carlo Erba. . (2005). Filling the gap, or jumping over it? Emergentism and naturalism. Epistemologia, XXVIII, 93–120. . (2008). Consciousness and the self. Functional Neurology, 23(4), 179– 187. . (Forthcoming). Explanation, emergence and causality. A comment on Crane. In C. Macdonald & G. Macdonald (Eds.), Emergence in Mind. Oxford: Oxford University Press. Di Francesco, M., Motterlini, M., & Colombo, M. (2007). In search of the neurobiological basis of decision-making. Explanation, reduction and emergence. Functional Neurology, 22(4), 197–294.

Two Varieties of Causal Emergentism 77 Dupré, J. (1993). The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Cambridge, MA: Harvard University Press. Fodor, J. (1997). Special sciences: Still autonomous after all these years. Philosophical Perspectives, 11, 149–163. Horgan, T. (1984). Supervenience and cosmic hermeneutics. Southern Journal of Philosophy, 22 (Spindel Conference Supplement on Supervenience), 19–38. Jackson, F. (1998). From Metaphysics to Ethics. Oxford: Oxford University Press. Kim, J. (1998). Mind in a Physical World. Cambridge, MA: MIT Press. . (1999). Making sense of emergence. Philosophical Studies. 95, 3–36. . (2006). Emergence: Core ideas and issues. Synthese, 151(3), 347–354. McLaughin, B. (1992). The rise and fall of British emergentism. In A. Beckermann, H. Flohr, & J. Kim (Eds.), Emergence or Reduction? Essays on the Prospects of Nonreductive Physicalism (pp. 49–93). Berlin, New York: de Gruyter. O’Connor, T., & Wong, H. Y. (2006). Emergent properties. In: E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. Retrieved from http://plato.stanford. edu/entries/properties-emergent/; accessed on June 19, 2009. Shoemaker, S. (2002). Kim on emergence. Philosophical Studies, 108, 53–63. Thompson, E., &Varela, F. J. (2001). Radical embodiment, neural dynamics and consciousness. Trends in Cognitive Sciences 5(10), 418–425.

5

The Emergence of Group Cognition Georg Theiner1 and Timothy O’Connor

1. INTRODUCTION The Group Mind Thesis—understood as the claim that groups as a whole can be the subjects of mental states—was a popular idea in the intellectual landscape of the late nineteenth and early twentieth centuries. 2 For many scientists and philosophers of that period, it provided a succinct expression of what they perceived to be two characteristic features of groups: on the one hand, their ability to function as collective agents who can have intentions, make decisions, and pursue their own goals; on the other hand, the idea that groups are emergent wholes which are more than the sum of its members. Combine the two features, and the functional analogies between individual and group behavior strongly suggest adopting an intentional stance towards both. But the group mind thesis fell out of grace with the rise of behaviorism and operationalism—no doubt expedited by the fact that some of its traditional expressions trafficked freely in unexplained mentalistic and vitalistic idioms that were rightly considered to be at odds with a scientifically informed worldview. As Wegner et al. (1985: 254–256) point out in their brief history of the group mind concept, the main problem was that the group mind seemed to lack its own body. Hence it remained unclear where to look for its properties, and how to measure them. One way to summarize the precarious ontological status of group minds is in the form of a theoretical dilemma. If the group mind is nothing over and above the collection of individual minds and the group processes by which they interact, an appeal to group minds appears to be superfluous. However, if the group mind is something over and above all these things, it appears to imply a collective version of mind-body dualism. This raises the familiar question of how the group mind exercises its causal influence on individual group members. Suggested answers included the mediation of a genetic “ectoplasm” (Jung, 1922) or telepathic communication (McDougall, 1920). Neither horn of the dilemma makes the idea of group minds seem very attractive. Despite its historical ballast, the idea that groups can have cognitive properties of their own has recently gained new ascendancy in a wide

The Emergence of Group Cognition 79 range of disciplines concerned with group behavior. Economists and political scientists continue to explore the relationships between individual and group rationality (List, 2008; Pettit, 2003; Satz & Ferejohn, 1994). Sociologists, anthropologists, and historians fi nd it useful to express generalizations about social groups in terms of their collective memory (Burke, 1989; Le Goff, 1992). Social psychologists studying problem solving and decision making in small groups increasingly embrace the view of groups as information processors (Larson & Christensen, 1993; Hinsz, Tindale, & Vollrath, 1997). Organizational scientists study the memory and learning processes of fi rms and organizations (Argote, 1999; Sandelands & Stablein, 1987; Walsh & Ungson, 1991). Evolutionary biologists have revived the idea that groups can evolve into adaptive units of cognition as a result of group-selection (D. S. Wilson, 1997, 2002; D. S. Wilson, van Vugt, & O’Gorman, 2008). Recent studies of animal behavior have revealed a number of collective decision-making mechanisms that are shared across a wide range of group types such as swarming ants, schooling fish, flocking birds, and even humans (Bonabeau, Dorigo, & Theraulaz, 1999; Hölldobler & E. O. Wilson, 1990; Seeley, 1995). The framework of distributed cognition has been used to study the dynamics of collaborative work practices which are socially, technologically, and temporally distributed, and whose coordination is mediated by rich situational, material, and organizational constraints. (Hollan, Hutchins, & Kirsh, 2000; Hutchins, 1991, 1995a, 1995b). It has recently been embraced by some philosophers of science as a unifying framework to overcome the present hiatus between “rationalist” and “social-constructionist” approaches to scientific cognition (Giere, 2002; Giere & Moffat, 2003; Nersessian, 2006). Finally, philosophers seeking a conceptual analysis of collective intentionality have tied their accounts to the recognition of groups as intentional subjects in their own right (Gilbert, 1989; Schmitt, 2003a; Tollefsen, 2004). If we take these proliferating appeals to group cognition at face value, do they run into the same embarrassments that plagued traditional versions of the group mind thesis? In a series of papers and monographs examining the social dimension of cognition, Rob Wilson (2001a, 2004, 2005) has argued the contemporary appeals to group cognition remain fraught with unnecessary ontological commitments that have no real explanatory value. First, it is not sufficient to show that groups can collectively perform actions which are explained by psychological processes if all these processes are reducible to forms of “socially manifested” individual cognition. Wilson’s way of putting his point with respect to Hutchins’s (1995a) analysis of ship navigation as a form of socially distributed cognition is that “[t]he statement ‘the crew saw the oncoming ship and decided to change direction’ might be made true simply by individual-level psychological facts, together with other, nonpsychological facts about social organization” (2004: 291). Considering David Sloan Wilson’s (1997, 2002) analysis of groups as adaptive decision-making units in their own right, Rob Wilson goes on to argue that he “would

80

Georg Theiner and Timothy O’Connor

seemingly need to show that this functions at the group-level by individuals relinquishing their own decision-making activities. For it is only by doing so that he could point to a group-level psychological characteristic that is, in the relevant sense, emergent from individual-level activity” (2004: 297). The point of Wilson’s argument is to cast the contemporary proponent of group cognition in a similar mold as the older collective psychology tradition, in which the emergence of group-level activity was taken to degrade and corrupt individual cognitive abilities (e.g., Le Bon’s 1895/1960 suspected transformation of autonomous individuals into “maddening crowds”). Clearly, what drives much of the current philosophical interest in the idea of group cognition is its appeal to the manifestation of psychological properties—understood broadly to include states, processes, and dispositions—that are in some important yet elusive sense emergent with respect to the minds of individual group members. Like the group mind thesis, the term “emergence” had long fallen into disrepute, only to be rehabilitated for its seeming applicability to a wide variety of empirical phenomena as well as its usefulness in articulating certain claims in metaphysics, particularly claims concerning the ontology of the mind. However, care needs to be taken in understanding the usage of particular authors. It is apparent that there are at least three families of concepts that are at play in different contexts. (See O’Connor & Wong, 2000, for further discussion.) First, some intend a purely epistemological concept: emergence in this broad sense implies unpredictability (in some sense) from a certain vantage point. Second, there are modest metaphysical concepts: emergent properties of certain complex systems are taken to be real and nonidentical to structures of underlying properties and to make a distinctive causal contribution to the world, yet this is explicated in such a way as to be consistent with a broadly (albeit non-reductive) physicalism. Finally, there are strong metaphysical concepts: emergent properties whose manifestation is explicitly avowed as being inconsistent with one or other defining feature of physicalism—either the causal completeness (or “closure”) of physics (Gillett’s 2006 “strong emergence”) or both completeness and the realization of all macro-level features (O’Connor & Wong’s 2005 notion of “ontological emergence”; see also O’Connor & Churchill, forthcoming). As these last citations indicate, one of us is friendly to the possible application of a strong, physicalism-negating concept of emergence. But such a view is very much a minority view in contemporary philosophy of mind and metaphysics. (Some doubt its coherence, thinking it must collapse into an outright mind-body dualism.) For the present chapter, we ask the reader to assume that, when it comes to human cognition and consciousness, strong metaphysical varieties of emergence, on the one hand, and austerely reductive or eliminativist views, on the other, are all off the menu of serious options on present evidence. That is to say, we will suppose for the sake of argument the correctness of the majority view that human mentality is a wholly physical phenomenon yet emergent in some modest metaphysical sense.

The Emergence of Group Cognition 81 Our goal will then be to address a set of related, conditional questions: If human mentality is real yet emergent in a modest metaphysical sense only, then - what would it mean for a group to have emergent cognitive states? - is this even a metaphysically coherent view? - relative to which notion of emergence do we have reason to believe that certain groups in fact have emergent cognitive states? We will argue that, given our central assumption, evidence from a wide variety of social science domains makes it plausible that there are group cognitive states and processes no less metaphysically emergent than human cognitive (and other special science) states and processes. We leave it to the reader to draw your own conclusion: some will follow one of us in supposing that the consequent is to be embraced as a surprising and enlightening empirical discovery. Others may follow the second of us in taking the truth of the conditional to provide significant reason for doubting its antecedent.

2. TOWARD A CONCEPTUAL FRAMEWORK FOR ANALYZING GROUP COGNITION The undeniable sex appeal of the idea that groups can have emergent psychological properties is frequently purchased at the expense of conceptual clarity and rigor. In this section, we propose a conceptual framework for analyzing group cognition. For philosophers of mind, there is an inherent danger to remain narrowly focused on the question, relative to which notion of the mental can groups be considered the subjects of mental states (e.g., can groups have consciousness?). Adopting a ‘big tent’ approach, we treat the notion of cognition as a theoretical term, and break it down into several distinct capacities that we take to be indicative of cognitive systems. But even if we confi ne ourselves to a predominantly functional understanding of cognition, it remains an open question whether groups can have emergent cognitive properties in their own right. To address that question, we flesh out three different features that emergence can plausibly be taken to signify in the context of group cognition: dependence on the social organization and interactions among individuals; the manifestation of unintended cognitive effects at a group level; and the multiple realizability of cognitive properties by different types of group structures.

2.1. A ‘Big Tent’ Approach to Cognition A promising strategy for making sense of the idea of group cognition is one that has been quite successfully employed in cognitive science, where

82

Georg Theiner and Timothy O’Connor

the concept of cognition is treated as a theoretical term that is largely defi ned by its role in our explanatory practices. In cognitive science and related fields, the relevant meaning of cognition is partly inspired by but nevertheless to be distinguished from what we would ordinarily consider as instances of mental states or activities. 3 In this section, we likewise propose to employ a flexible notion of cognition that we can plausibly consider as common ground in the present debate over group cognition. Bearing in mind the conditional nature of our enterprise that we have outlined earlier, there are several desiderata for a suitably ecumenical, ‘big tent‘ approach to cognition. First, we need a notion of cognition that is not bio-centric (i.e., a notion which is not essentially tied to the physical substrate of cognitive processes occurring inside the sheath of biological organisms). In philosophy, the putative substrate-neutrality of mental properties is commonly associated with functionalist theories of mind (Block 1980, 1996; Fodor, 1968; Lewis, 1972; Lycan, 1987). Broadly speaking, functionalist theories of mind claim that what makes something a mental state of a particular type does not depend on its intrinsic material constitution, but rather on the way it functions in the system to which it belongs. A second desideratum is to analyze the notion of cognition as a cluster concept which subsumes a more or less loosely knit family of capacities that we can distinguish for taxonomic purposes (for similar approaches, see Chadderdon, 2008; Dennett, 1996; Poirier & Chicoisne, 2006). The possession of each capacity enables its bearer to engage in a distinctive range of behaviors that we associate with signs of intelligence. This divideand-conquer strategy is necessary to keep all parties of the debate from drawing cheap but unreliable inferences about the occurrence of group cognition. For instance, when we consider the rationality of ant colonies as unitary decision makers in their own right (Edwards & Pratt, 2009), we do not wish to imply that they possess a collective form of consciousness. Consequently, the absence of consciousness from the catalogue of grouplevel cognitive properties does not by itself refute the broader idea of group cognition. In defense of this approach, we should emphasize that our goal here is not to provide a reductive analysis of cognition but to provide a set of diagnostic criteria that allow us to classify and compare various systems in terms of their cognitive prowess. The criteria that we shall propose are not meant to be mutually exclusive or jointly exhaustive. To exemplify the ‘big tent’ approach we have in mind, consider the following list of capacities that have all been discussed in the literature as characteristic features of cognition. We shall say that system S is cognitive to the extent that

1. AD (adaptability): S can adapt its behavior to changing environments. 2. IP (information-processing): S can process information from its environment.

The Emergence of Group Cognition 83 3. H (heed): S can selectively and purposefully attend to its environment. 4. IT (intentionality): S can create internal representations of its environment. 5. E (extension): S can modify its environment through the creation of artifacts. 6. R (self-refl exivity): S can become aware of itself as a cognitive agent. 7. C (consciousness): S can have conscious experiences of itself and the world. Moving from the possession of one or more of these capacities to the subject of cognitive properties, the associated notion of a “group mind” should thus not be considered as an all-or-nothing phenomenon, but one that admits of degrees. Cognitive systems which exhibit more of the relevant cognitive capacities are ranked higher in terms of their “mindfulness”. It is a desirable consequence of our list that individual human beings are very “mindful” creatures, although we do not assume that the cognitive profile of human beings always provides the gold standard of what it means to be cognitive. In this paper, we are only concerned with the fi rst five conditions. But for the sake of completeness, let us say a few words about the remaining two conditions on our list. Condition R is borrowed from recent philosophical discussions of epistemic agency (Burge, 2000). It refers to the essentially indexical capacity of deliberate epistemic agents to think of themselves in fi rst-person terms. In his discussion of group minds, Pettit (2003) has argued that certain groups can not only be intentional systems in their own right, but would also qualify as institutional persons. As opposed to the former, Pettit takes it as a mark of persons that they can be held responsible for failures to unify their intentional states and actions in a way that complies with rational norms. For persons to assume this kind of responsibility, they must actively avow and acknowledge their intentional mental states and actions as their own. Burge (2000) has suggested that the characteristic immediacy by which one is moved to think and act in accordance with one’s own reasons (but not anybody else’s) is based on understanding the fi rst-person concept. Against epistemic agent individualism, Tollefsen (2004) has argued quite convincingly that groups can be subject to the same rational assessment when they self-consciously act from a fi rst-person plural point of view. If Pettit and Tollefsen are right, their examples show how certain collectivities can satisfy condition R. Finally, let us explain how consciousness fits into this picture. Clearly, the pre-theoretical notion of consciousness is ambiguous and admits of different interpretations (Chalmers, 1996; Tye, 1995). For instance, it is sometimes used to denote phenomena which are already captured under conditions IT, H, or R. Condition C is meant instead to cover the phenomenal aspects of consciousness. Consider the distinction between what

84

Georg Theiner and Timothy O’Connor

Block (1995) has called A-consciousness and P-consciousness. Phenomenal (P-) consciousness describes the character of experience; the phenomenally conscious aspect of a state is what it is like to be in that state. The mark of access (A-) consciousness, by contrast, is being directly available for global executive control, such as being poised for use in reasoning and rationally guiding speech and action. As we understand condition C, it truly applies only to cognitive systems which have the capacity to entertain P-conscious mental states. Our implicit assumption here is that A-consciousness can be reduced to a composite of other conditions—especially IP, IT, H, R, and perhaps E—in which case it would be redundant to assign it a level of its own. We currently see no compelling evidence that there are any groups which satisfy condition C. Views of human mentality that reject physicalism may suppose that this acknowledgment is no small concession, but one that gives the whole game away: it is open to the dualist to suppose that true mentality is constitutively tied to the capacity for conscious awareness. But this sort of view, it seems to us, would be very implausible for a physicalist, for whom conscious mental states are just one special variety of physically realized state playing certain causal roles in a person’s mental economy. We will, in any case, return to the issue of consciousness and the status of group minds at the end of the chapter.

2.2. What’s Emergent About Group Cognition? 2.2.1. Emergence1 as Organization-dependence The intuition that a distinction must be made between genuine systems (e.g., a biological organism) and mere aggregates (e.g., a heap of stone) is epitomized in the popular slogan that systems are emergent wholes which are “more than the sum of their parts”. Taking this slogan in a very uncompromising sense forces a choice between “holism” and “atomism”, with the holist supposing unity-conferring emergent properties beyond the reach of mechanistic explanation. Thus explicated, holism is committed to emergent properties in the strong metaphysical sense that is abjured here. How instead should a physicalist who takes seriously the reality and importance of the system/aggregate distinction seek to elucidate it? In a series of papers, Bill Wimsatt (1974, 1986, 1994, 1997; collected in Wimsatt, 2007) proposes that the emergence1 of complex system properties is defi ned as a failure of “aggregativity”, considered as a strong form of organization-dependence. (See especially Wimsatt, 1986.) Let s1 to sm stand for the m components of a system S (relative to some decomposition D); p1 to pn for the n properties of S’s components; and F for the organization or mode of interaction between pi(sj), such that a system property P(S) is determined by the composition function: P(S) = F[pi(sj) for i = 1 to n, and j = 1 to m]. For P(S) to be purely aggregative, it must satisfy the following conditions 1–4 (for a given decomposition D of S); otherwise, it exhibits degrees of emergence1.

The Emergence of Group Cognition 85 1. IS: P(S) is invariant under the inter-substitution of parts of S, or any other parts taken from a relevantly similar domain. 2. QS: P(S) remains qualitatively similar (differing only in value) under the addition or subtraction of parts. 3. DR: P(S) is invariant under the decomposition and re-aggregation of parts. 4. CI: There are no cooperative or inhibitory interactions among parts. Consequently, we can say that a group S instantiates a cognitive property P(S) just in case P(S) is emergent1 relative to a decomposition of S into its members, their behavioral and psychological properties, and their modes of social interaction.4 Let us briefly elaborate on some features of this analysis. First, since Wimsatt’s defi nition of emergence1 presupposes the existence of a composition function, emergent1 properties of a system can in principle be mechanistically explained in terms of the system’s components, their properties, and the totality of their interactions. We note that Wimsatt refers to the suggested type of componential analysis as a form of “reductive” (1997: S373) explanation, because he considers the composition function as an “equation” which yields “an inter-level synthetic identity, with the lower level specification a realization or instantiation of the system property” (1997: S376). 5 (Whether the term ‘reductive’ is an apt one in this context will be hotly disputed, but we need not consider it here.) Second, complex systems with emergent1 properties fail to be “near-decomposable” in the sense of Simon (1969). Since the properties of such systems are largely dependent upon the interactions between parts that do not perform any tasks which can recognizably be associated with a functional decomposition of the whole, the classic twin strategies of decomposition and localization fall short (Bechtel & Richardson, 1993). This means that their behavior cannot be properly understood by fi rst dividing up the entire system into a number of independently working component units, characterizing the contributions of these units as if they were isolated from each other, and then adding up their contributions by associating them with specific aspects of what the system does as a whole. Wimsatt has argued that expressions of nothing-but-tery, such as ‘the mind is nothing but neural activity’ or ‘social behavior is nothing but the actions of individuals’, are a result of such functional localization fallacies (1997: S382–S383). They reflect our disposition to use the assumption of near-decomposability as a powerful kind of meta-heuristic when we seek out mechanistic explanations, because aggregative decompositions afford regularities which are less context-dependent, and support simpler theories and models. However, this does not put emergent1 properties beyond the wider scope of more sophisticated (and surely empirically more adequate) mechanistic models that are better equipped to deal with interactional complexity (Bechtel, 2006; Bechtel & Abrahamsen, 2002).

86

Georg Theiner and Timothy O’Connor

2.2.2. Emergence2 as the Absence of Intentional Design A recurring thread in the fabric of social life is that the behavior of individuals is often not a good predictor of its collective consequences. Public benefits can flow from selfish intentions. Adam Smith’s (1776) conception of the “invisible hand” famously refers to the idea that a community of traders, acting purely from self-interest, is driven by the competitive forces of the marketplace to furnish goods for all, at an affordable price. Conversely, public vices can spring from private virtues. The “tragedy of the commons” (Hardin, 1968) illustrates a scenario in which a community of independently acting rational individuals suffers detrimental long-term consequences that are not in anybody’s self-interest. Collective actions that generate effects other than what individuals intended or expected stir our natural curiosity, because they violate untutored intuitions about how local behavioral rules scale to the global properties of inter-connected wholes (Resnick, 1994). We shall say that emergent 2 cognitive properties arise from the local interactions between many individuals, but without being planned or purposefully designed by any of these individuals (or some central planning agency), and which those individuals may even fail to notice.6 Let us briefly characterize a few of the roles which the notion of emergence2 has played in the social sciences. As our two examples indicate, some of the historically most influential statements of emergence2 come from economics, where it has been used to advocate a reductionist doctrine known as methodological individualism (Popper, 1957; Watkins, 1957). This might come as a surprise at fi rst, since methodological individualists contend that all social phenomena can and should be explained in terms of the actions of individuals and how they are interrelated. However, in order to carve out a niche for economics as an autonomous discipline vis-à-vis psychology, individualist economists pointed to the occurrence of undirected collective effects that can spring from the actions of the many. For instance, Hayek wrote that “the conscious action of many men produce undesigned results [ . . . ] regularities which are not the result of anybody’s design. If social phenomena showed no order except in so far as they were consciously designed, there would indeed be no room for theoretical sciences of society and there would be, as is often argued, only problems of psychology” (1942: 288, cited after Sawyer 2005: 43). Far from showing an essential incompleteness of individuallevel explanations, Hayek thus took the absence of intentional design as an impetus to discover the laws by which emergent 2 social phenomena arise (e.g., Hayek’s (1944) “compositive method”). A different intellectual tradition pre-occupied with emergent2 social phenomena is known as (the sociology of) collective behavior (Blumer, 1939; Lang & Lang, 1961; Park & Burgess, 1921). In this tradition, the term ‘collective behavior’ refers to a class of social phenomena that are not shaped by pre-established social structures (e.g., laws, institutions, conventions),

The Emergence of Group Cognition 87 but arise “spontaneously” from people behaving en masse, guided only by simple and purely local concerns. Paradigmatic examples of collective behavior (in this sense) are mob actions, riots, rumors, mass delusions, and fads, although Park also advanced the stronger claim that “institutions and social structures of every sort may be regarded as products of collective action” (1927: 733). Third, the remarkable coherence and synchronization of collective animal behavior such as swarming ants, flocking birds, and schooling fish has long stirred the imaginations of scientists and philosophers. “They must think collectively, all at the same, or at least in streaks or patches—a square yard or so of an idea, a flash out of so many brains”, wrote the field naturalist Edmund Selous (1931) upon observing the splendor of tens of thousands of starlings coming to roost (cited after Couzin, 2007: 715). In similar vein, the entomologist William Morton Wheeler (1920, reprinted in Wheeler 1939) coined the term ‘super-organism’ to denote the high degree of co-dependence and functional integration among eusocial insects (e.g., ants, bees). Arguably, what makes the attribution of emergent 2 psychological capacities—including perception, planning, and decision making—to “super-organismic” groups so intuitively compelling is the stark discontinuity between the complex collective behavior that we observe and the rudimentary cognitive resources of individual members. Finally, the concept of emergence2 has also fueled the new wave of “mechanists” in contemporary social science who reject the traditional deductive-nomological covering law model of explanation (e.g., Hempel, 1965) in favor of generative, process-oriented approaches (Alexander & Giesen, 1987; Macy & Willer, 2002; Sawyer, 2004). Unlike some traditional versions of individualism, the new mechanists do not deny the reality of higher-level social structures, but re-emphasize the need to provide “bottom-up” explanations of the link between micro-social interactions and macro-social patterns. Since the 1990s, a powerful computational methodology by which researchers have studied the mechanisms of social emergence2 is known as agent-based modeling (Epstein & Axtell, 1996; Goldstone & Janssen, 2005; Miller & Page, 2007). The concepts and theoretical tools on which agent-based models are based stem largely from complexity theory (Bak, 1996; Ball, 1998; Holland, 1975, 1995; Kauffman, 1993). A fundamental insight of complexity theory has been that a set of relatively simple rules governing the behavior of decentralized components can give rise to qualitatively novel, global patterns of organization, without being regulated by an outside source or managed by a centralized controller. Emergent2 ordering processes such as phase transitions, non-equilibrium bifurcations, and power-law distributions are generic phenomena that occur in the same way over a wide range of superficially diverse systems. Agent-based models demonstrate how the inexorable dynamics of collective phenomena such as traffic jams, crowd movements, market fluctuations, the growth of fi rms and cities, the formation of alliances, and the evolution

88

Georg Theiner and Timothy O’Connor

of cooperation can emerge2 from a few underlying regularities as long as they are followed by a great many people (Ball, 2004). 2.2.3. Emergence3 as Multiple Realizability In metaphysics and philosophy of science, there has been considerable debate over the question of whether physical realization is an ontological dependence relation which supports ontological reductions (Fodor, 1974, 1997; Horgan, 1993; Kim, 1989, 1992, 1998; Lewis, 1972; O’Connor & Churchill, forthcoming; Shoemaker, 2007; Van Gulick, 2001). A cluster of influential arguments against the reduction of higher-level properties to their lower-level realizers is based on the premise—commonly referred to as multiple realizability—that functional system properties can in principle be instantiated by indefi nitely many distinct physical structures (Block, 1997; Fodor, 1974, 1997; Gillet, 2003; Shapiro, 2000; Sober, 1999).7 Arguments for the multiply realizable nature of mental states are commonly associated with functionalist theories of mind (Block, 1996). Broadly speaking, functionalists claim that what makes something a mental state of a particular type does not essentially depend on its intrinsic material constitution, but rather on the way it functions in the system to which it belongs. Functionalism is thus consistent with the idea that individuals and groups can be sufficiently alike in their functional organization that they share the same mental states, even though the mechanisms by which these mental states are realized are obviously quite different in each case. The implication that groups could think if they are properly organized has often been used by critics to chastise functionalism for its unabashed “liberalism” (Block, 1978; Searle, 1980, 1992). However, what some see as a fundamental “bug” of functionalism that needs to be fi xed, others embrace as a theoretical virtue that we ought to preserve. Our ecumenical ‘big tent’ approach to cognition departs from traditional versions of functionalism in at least two crucial respects. First, since we do not claim that phenomenally conscious mental states fall within the purview of functionalism, the most pressing philosophical objections against functionalism do not carry much weight against the understanding of group cognition that is presented here. Second, functional characterizations of group cognition that derive from a conceptual analysis of folk-psychological concepts often remain on a very coarse-grained level, and thus appear to be explanatorily “shallow”. However, this deficit simply reflects the fact that they are conceived out of the armchair, without paying any attention to empirical facts about the abilities of groups to solve cognitive tasks. Let us briefly revisit three fields of group research in which the functional equivalence between individual and group cognition has served as an organizing framework. First, in the social psychology of small group performance, there has been a growing trend to consider groups as the seats of cognition (e.g., problem solving, judgment, inference, and decision making) and knowledge in their

The Emergence of Group Cognition 89 own right.8 Their functional approach to understanding group problemsolving is predicated on a view of groups as information processors. For instance, Larson and Christensen (1993: 6) emphasize that they use the term “social cognition” “at the group level of analysis to refer to the social processes (e.g. introducing information into a group discussion) that relate to the acquisition, storage, transmission, manipulation and use of information for the purpose of creating a group-level intellective product. [ . . . ] At the group level of analysis, cognition is a social phenomenon”. In their focused review of research on small-group performance, Hinsz et al. (1997) employ a generic information-processing model as an organizing framework that is directly borrowed from cognitive psychology. In this model, a group obtains information through interaction with its environment. The context in which certain information is acquired, during the attention phase, influences the choice of processing objectives. The encoding process involves the selective transformation of information into representations, which can be stored in and retrieved from memory components. In the processing work space, information integration and schematic processing occur on the basis of a variety of rules, strategies, heuristics, and procedures. After enough information has been processed to meet the relevant objective, the cognitive agent makes a response which changes the situation and may lead to feedback that informs the agent about these changes. All of these processes are potentially subject to modification through learning (Hinsz et al., 1997: 43).9 Second, another research framework in which the realm of cognition is extended from an individual to a collective unit of analysis is the theory of distributed cognition (Hollan et al., 2000; Hutchins, 1995a, 1995b; Norman, 1991). It can be characterized by two main theoretical commitments (Hollan et al., 2000: 3). First, the boundaries of cognitive systems are delimited by the functional relationships among its constitutive elements, rather than the intuitive biological boundaries of the individual. Second, there are no intrinsic constraints on the range of mechanisms that may be assumed to participate in cognitive processes. For instance, Hutchins (1995a) provided a detailed study of ship navigation crews as socially and technologically distributed cognitive systems. By extending David Marr’s (1982) tripartite computational analysis of cognition from the individual to the collective unit of analysis, Hutchins was able to gain novel insights into the ways in which individuals, artifacts, representational media, and the environment are coordinated in the context of navigation tasks.10 An important function of social organization—together with the structure added by the concrete context of activity—is to determine the flow of information within the crew. Summarizing his analysis, Hutchins concludes that “organized groups may have cognitive properties that differ from those of the individuals who constitute the group. These differences arise from both the effects of interactions with technology and the effects of a social distribution of cognitive labor. The system formed by the navigation team

90

Georg Theiner and Timothy O’Connor

can be thought of as a computational machine in which social organization is computational architecture” (1995a: 228). Third, in the field of animal cognition, it has been shown that collective decision making of grouping animals has some important features in common with neural mechanisms of decision making in the brain (Couzin, 2009). This is particularly striking in the case of ant colonies, which function as unitary decision makers in the context of foraging for resources, choosing a place to live, or constructing a nest. Passino, Seeley, and Visscher (2008) highlight a number of interesting functional parallels between the synchronized patterns of rhythmic activity displayed by ant swarms and neural networks. Ants exhibit a “neuron-like” behavior insofar as inactive ants have a low propensity to become spontaneously active, but can become excited by other ants with whom they come into contact. Similar to the beyond-threshold depolarization of a neuron, ants seem to temporally integrate the inputs they receive, and start moving if their activation exceeds a certain level. Conversely, ants are prone to lapse back into inactivity if their activation is not sufficiently reinforced, and even exhibit a short refractory period (similar to neurons) before they can be reactivated—a mechanism which keeps the swarm from getting permanently “locked” into an excitatory state. Another functional parallel concerns the role of rhythmic oscillation for input selection. It has been shown that rhythmic neuronal network activity is an energy-efficient way to elevate the brain into discrete windows of high responsiveness to external stimuli. Similarly, the periodical synchronization of ant activity might provide privileged windows of opportunity for the swarm to respond to external foraging opportunities, or efficiently allocate workers for maintenance tasks within the nest.

3. RATING GROUP COGNITION By now, it should be clear that the multi-faceted notion of emergent group cognition that we propose is neither trivial nor shrouded in metaphysical mystery. To sharpen the focus of our analysis, we now look at three welldocumented cases of group-level activity through the lens of our theoretical framework, and rate each of them in terms of their respective degrees of cognition and emergence1–3. We have chosen these toy examples not because we believe that they are equally spectacular instances of its kind, but because they enable us to test the discriminatory capacity of our diagnostic tools— especially where the suggested criteria pull us in different directions.

3.1. Distributed Problem Solving in Groups In order to investigate the impact of social network structure on the distributed problem-solving abilities of groups, Kearns, Suri, and Montfort (2006) studied groups who were attempting to solve the graph-coloring

The Emergence of Group Cognition 91 problem. A familiar instance of this problem arises if one has to fi nd a way to color a map of the United States with the smallest possible number of colors such that no two adjacent states may share the same color. They chose the problem as an abstract model of social settings in which it is desirable to distinguish one’s behavior from that of one’s neighbors. Examples of these settings include the scheduling of events in a limited number of rooms, selecting a ringtone that differs from one’s friends, or the differentiation of expertise within a social organization. In the experiment, subjects had to collectively solve a number of coloring problems as part of a network that had one of six possible topologies. Each of the chosen network topologies, which corresponded to recently proposed models of network formation, belonged to one of two families. Members of the cycle-based family, all of which required a minimum of two colors, included a simple cycle, two “small-world” networks (Watts & Strogatz, 1998), and a more centralized leader cycle containing two privileged nodes. The other two topologies were generated according to the “preferential attachment” model (Barabási & Albert, 1999), with two (minimum of three colors) or three (minimum of four colors) links initially added to each node. The subjects were connected through a computer platform which provided them with either local or global information about the structure and current coloring state of their network, but without receiving any strategic hints of how to play. The ability of groups to solve the coloring problem was strongly affected by the topology of their network. Within the cycle-based family, for which coloring was generally easier, a smaller average shortest-path length led to reduced solution times. This means that while the addition of links complicates the coordination problem faced by individuals (because they must take into account a larger number of neighbors), it evidently has the opposite effect for the group as a whole by reducing the number of links coloring conflicts must travel through the network in order to be resolved. Similarly, the effects of varying the locality of information provided to the subjects again depended on network structure. Whenever individuals seem to have a strong intuitive grasp of the collective effort that is required to converge on one of the optimal solutions, a high-information view was beneficial. However, in more complex situations where this is not the case (e.g., for preferential attachment networks), giving individuals more information about the collective state of their network significantly hampered their performance as a group. As a possible explanation, Kearns et al. (2006: 826–827) point out that the time and effort subjects spend on attending to, and perhaps even trying to influence (e.g., by signaling behavior) more distant network activity may effectively distract them from doing their own local subtask. Let us briefly evaluate the example as a putative instance of emergent group cognition. To begin with, we take it as uncontroversial that graph coloring (and all formally equivalent instances of the same problem) is a computationally hard task that requires a substantial amount of intelligence if it is performed by a human being or a machine (Jensen & Toft,

92

Georg Theiner and Timothy O’Connor

1995). Based on parity considerations, we further claim that one ought to ascribe the corresponding level of intelligence to a group if its members are collectively responsible for solving the same optimization problem. At the cognitive minimum, our diagnostic criteria suggest that groups which can solve the coloring problem are adaptive information-processing units in their own right (see also Gureckis & Goldstone, 2006). This achievement ranks relatively high in terms of emergence1. The coloring pattern means that condition IS is restricted to subsets of individuals with the same color. Furthermore, the selective influence of different network topologies implies a violation of condition DR, and condition CI fails because every individual’s choices are constrained by the choices of her neighbors. The resulting pattern of differentiation is an emergent 2 effect of repeated cycles of social interactions. In particular, we have seen two unexpected ways in which the regularities governing individual- and group-level behaviors pull into opposite directions. Finally, in virtue of the formal equivalence between social coordination tasks that conform to the coloring problem, any generalizations about the problem-solving abilities of groups that are couched purely in relational terms of their network structure are multiply realizable and thus count as emergent3.

3.2. Transactive Memory Systems in Groups When people regularly have to remember things together as a group—as intimate couples, families, or work teams do—they tend to develop a division of cognitive labor, assuming that each member can reliably access the desired information from others on a need-to-know basis. To study the functional organization of memory as a group-level phenomenon, Daniel Wegner (Wegner, 1986, 1995; Wegner et al., 1985) introduced the notion of a transactive memory system (TMS). A TMS generally consists of two components: a representational component which is the sum total of individual memories, including transactive meta-memories about who knows what, and a procedural component which includes all direct and indirect communication processes (“transactions”) by which members cooperatively allocate, encode, retrieve, elaborate, and share information. For instance, allocating memory items and encoding responsibilities, the semantic elaboration of memories in group discussion, and interactive cueing are physically constitutive vehicles of transactive memory procedures which partly occur outside people’s heads. Wegner claimed that a TMS is “a knowledgeacquiring, knowledge-holding, and knowledge-using system that is greater than the sum of its individual member systems” (1986: 256). If Wegner is right, how can we track a group-level cognitive construct such as TMS, other than measuring group performance (e.g., collective recall)? To answer this question, consider a study by Liang, Moreland, and Argote (1995), conducted within the assembly-task paradigm.11 The goal of their study was to show how the experience of working together as a

The Emergence of Group Cognition 93 group can induce a TMS that improves group performance. As predicted, they found that groups whose members were trained together to assemble radios recalled more steps of the procedure and produced radios with fewer errors than when trained alone. But to infer that TMS acted as a mediator of group behavior, they had to open the “black box” of group cognition (Figure 5.1). In their analysis, TMS is treated as a (second-order) latent variable hypothesized to underlie three (fi rst-order) cognitive manifestations that were found to be positively correlated with each other and group performance: memory differentiation (M1), the tendency of group members to specialize in recalling distinct aspects of the assembly process; task credibility (M2), how much members trusted one another’s expertise (associated with behavioral measures such as less need to claim expertise, better acceptance of procedural suggestions, less criticism); and task coordination (M3), the ability of group members to work together more smoothly (measured, e.g., by less need for explicit planning, fewer misunderstandings, greater cooperation). The combined scores on each fi rstorder factor were used to create a TMS-index. For three other social, but non-cognitive variables (task motivation, group cohesion, and social identity), no correlation with group performance was found except for social identity. Finally, a multiple regression analysis confi rmed that TMS, but not social identity, mediated the influence of group training on work performance. So far, we have looked at TMS as collective repositories of task-specific knowledge that make groups better at doing the kinds of things which they were initially trained to do. What evidence do we have that TMS are not just collective memory systems, but also collective learning systems? Lewis, Lange, and Gillis (2005) studied the effect of TMS on group learning, learning transfer, and adaptation to changing task demands. Underlying their “learning-by-doing” framework is the assumption that every task performance that is mediated by a group’s previously induced TMS at the same time provides a learning environment which changes formerly established TMS structures and processes, and thereby enables the group to acquire new problem-solving skills. Whenever a group utilizes its existing TMS to carry out the task for which it was originally trained (e.g., assembling a radio), it undergoes a re-organization as a result of practice which prepares the group to transfer its knowledge to similar tasks in the same domain. First, group members get a chance to revise and recalibrate their beliefs about who knows what, based on immediate feedback for their performance as individuals and as a group. Second, group members have an opportunity to elaborate and contextualize their transactive memories, based on how their own task-related knowledge relates to other members’ jobs, roles, and expertise. Shared conceptualizations of interrole knowledge have been shown to enhance group coordination and performance (Marks et al., 2002). Third, greater

94 Georg Theiner and Timothy O’Connor experience also creates more specific expectations about how transactive memory procedures are likely to unfold, which leads to a regularization of habitual practices. Moreover, groups with a history of applying their TMS to a variety of related tasks should also be primed to develop a deeper understanding of the task domain. First, the interactive cueing processes which are characteristic of efficient TMSs prompt group members to draw explicit comparisons across tasks, which heightens their ability to recognize structural commonalities. The analogical encoding of two different but structurally similar problems promotes the occurrence of knowledge transfer and abstract understanding (Gentner, Loewenstein, & Thompson, 2003). Second, the growing refi nement of shared higher-order knowledge supports a process of collective induction (Laughlin, 1999) by which groups can collectively infer general principles underlying the task domain. In sum, the “learning-by-doing” framework implies that groups with active TMS should outperform those with no prior TMS on similar follow-up tasks, and are more likely to demonstrate abstract knowledge about the task domain. Consistent with earlier research on TMS in intimate couples (Hollingshead, 1998; Wegner, Erber, & Raymond, 1991), it also entails that groups which experience a disruption of a prior TMS (e.g., as a result of membership change) perform worse on subsequent learning tasks than those which have never developed a TMS. Even though Lewis et al. (2005) did not fi nd in their experiment a significant effect of TMS on learning transfer when it was previously utilized in a single task only, their fi ndings generally confi rmed the suggested predictions.12 A group whose members collectively enact a TMS manifests several of the cognitive capacities we have mentioned. First, the ability of groups with a TMS to learn new tasks provides strong evidence that TMS can adapt to changing environments. Second, it is easy to see how we can apply a generic information-processing model as the one discussed earlier (Section 2.2.3) to describe the functional organization of memory as a group-level

Figure 5.1 Opening the “black box” of Group Cognition.

The Emergence of Group Cognition 95 phenomenon. The task-specific TMS that we have described is most plausibly construed as an instance of procedural memory, because it concerns an implicit knowledge of how to assemble a radio as a group. However, after the occurrence of TMS learning, it also encompasses a body of declarative knowledge about principles of the underlying domain. Third, the allocation of members’ time and effort to acquire and retain specific knowledge about the assembly process reveals how much attention the group is paying to different categories of information. Do TMS exhibit a collective form of intentionality? Understanding attributions of collective intentionality in the context of explaining collective actions has been a topic of considerable controversy in analytic philosophy of action.13 Individualists hold that collective intentionality can be analyzed in terms of an interlocking complex of appropriately shared individual intentions together with a mutual awareness of each other’s intentions (e.g., Bratman, 1993, 1997); anti-individualists deny this claim. There are two main camps among anti-individualists who differ in what they take to be the subject of collective intentionality. According to the “singularist” option, collective intentional states form a special class of intentional states that are directed toward the performance of a group action, but possessed by the individuals who intend to act their part in the pursuit of this goal (e.g., Searle 1990, 1995). The other, “pluralist” option is to hold that collective intentional states are literally properties of collective agents (e.g., Gilbert, 1989; Schmitt, 2003a; Tollefsen, 2004). While we agree with the pluralist claim that groups can be genuine subjects of collective intentionality (if our guiding assumption is correct), our reasoning is importantly different from standard accounts of anti-individualism. Standard arguments for anti-individualism essentially hinge on the propositional content of mental states that are directed toward doing something together as a group. In contrast, our claim is concerned not with the intentional content, but the physically constitutive vehicles of intentional mental states (Hurley, 1998). What underpins the collective intentionality of TMS-states is not (just) that they are directed toward group actions, but that the memory resources which are jointly sufficient to realize any of these states are distributed across the members of the group. The relevant sense of distribution here, we shall now argue, is that of emergence1, or organization-dependence. As a group-level psychological property, consider a three-man team with an established TMS for assembling a radio, a mixed body of partly declarative, partly procedural memory about a complex task that no team member knows how to perform individually. For instance, imagine that member A knows how to insert all the mechanical components into the circuit board, B knows how to handle the electronic components, and C knows how to connect each component to all the others in the proper manner. Because of their differentiated expertise, condition IS fails. It might be argued that each expert could in principle be replaced by hiring and training an individual

96

Georg Theiner and Timothy O’Connor

taken from “a domain of relevantly similar parts”. However, as Wimsatt (1986: 262) has pointed out in a related context, the relevant equivalence class of parts must be characterized in terms of the intrinsic properties which the substituted parts have regardless of their organizational arrangement. By defi nition, this excludes the differentiation of expertise which is imposed socially by a division of labor. Condition QS fails, because taking out enough members who possess critical but unshared knowledge would effectively lesion the TMS and drastically reduce group performance. Since the TMS of groups which have been trained together markedly differs from the TMS of groups whose members are trained individually (and then put together), condition DR is violated. Finally, conditions CI fails, since members’ awareness of how expertise is distributed affects their individual likelihood of acquiring, recalling, and communicating memory items pertaining to specific categories of information. The collective cognitive activities of teams with a TMS are not emergent2 to the same degree as emergent1. To see that, let us recall that the collections of simple agents in physics-based models of collective behavior merely perform pre-specified jobs from their limited local perspective, but without being aware of any collective goals and purposes. Our team members, on the other hand, actively collaborate in the sense that their behavior is actively structured by their awareness that there is a collective task to be accomplished. This awareness introduces an extra layer of higher-order cognitive requirements on the part of individuals which is not directly related to the instrumental knowledge about their subtask, but concerns the need to coordinate their individual contributions with each other. For instance, individual team members must remember each other’s expertise and trustworthiness to offer suggestions and criticisms of each others’ work, and represent complex hierarchical plans so they can communicate about how to integrate their individual contributions into the collective work flow. In short, agents must be capable of observing and representing complex global aspects of social structure—a cognitive process which Castelfranchi (1998) has aptly termed immergence. Relative to the rich immergent knowledge held by individual members, the properties of a TMS are not very emergent 2 . What now remains to be shown is that this realization relation satisfies the criteria for emergence3. Fodor (1974, 1989, 1997) has argued that functional higher-level properties cannot be reduced to the disjunction of their actual and possible lower-level realizers, because the two are not nomologically coextensive. The gist of his argument is that a metaphysically “gerrymandered” disjunction of natural kinds is not itself a natural kind, and thus cannot equally figure in the expression of genuine causal laws.14 Fodor’s argument, if sound, generalizes to the expression of lawful regularities about groups that are couched in terms of their development of a TMS. First, there do not seem to be any constraints specifiable within individual psychology that would make any particular memory states nomologically

The Emergence of Group Cognition 97 necessary for the realization of a TMS at a group level. What makes a person’s transactive memories constitutive for the operation of a TMS is determined relationally by the social role she has been assigned to play in the group context of a memory task. But for any given task, a group can always in principle adopt one of indefi nitely many divisions of cognitive labor that are sufficient to get the job done (even though not all of them will be equally effective in practice). Transactive memories are thus good candidates for socially manifested cognitive processes (i.e., cognitive processes of individuals that can be realized only insofar as those individuals participate in groups of a certain kind) (Barnier et al., 2008; R. Wilson, 2004: ch. 8, 2005).15 Second, there are also no intrinsic constraints on the kinds of social interactions which can serve as the realization of transactive memory procedures. For instance, groups are known to employ to a number of different strategies to accomplish the essential tasks of transactive memory updating, information allocation, and retrieval coordination (Wegner, 1995).16

3.3. Collaborative Creativity in Groups Our third and fi nal example concerns the collaborative group processes that go into the choreography of a contemporary dance performance. By its very nature, the art of contemporary dance poses unique difficulties but also opportunities for insight into the dynamics of creativity (K. Stevens et al., 2000). First, the medium of contemporary dance is the collective movement of bodies through time and space while they continuously interact with one another. As an essentially multi-modal form of artistic behavior, dance contains visual, motor, tactile, aural, kinaesthetic, cognitive, sensual, evocative, affective, spatial, temporal, dynamic, and rhythmic elements that must all be coordinated in real time. Second, because of the ephemeral nature of movement material (compared, e.g., to visual or plastic arts), there are usually few preserved records of the processes which lead to the development of the fi nal work. Third, an increasingly typical feature is the interactive nature of “dance making” in which choreographers and dancers collaborate, often in a highly improvisational manner, in the course of creating a dance. Based on a careful analysis of annotated video recordings and journal entries recorded over a period of roughly six months, C. Stevens et al. (2003) have offered us a fascinating window into the complex psychological processes that underlie the inception, development, and refinement of dance material— processes to which they collectively refer to as choreographic cognition.17 A significant part of their analysis is framed in terms of the Geneplore model of creative cognition (Finke, Ward, & Smith, 1996), which assumes that creative cognition involves two distinct phases: generation and exploration. During an initial phase, generative processes such as memory retrieval, association, synthesis, analogical transfer, and categorical reduction yield “pre-inventive”

98

Georg Theiner and Timothy O’Connor

structures. In the second phase, the properties of these structures are then explored, combined, discarded, modified, and re-interpreted in novel ways. The resulting structures of these exploratory activities continue to be focused or expanded, depending on the desired degree of refinement, and edited until they reach their final form. A creative process typically involves multiple iterations of generation/exploration. The most salient feature of their analysis is that the choreographic processes described by the Geneplore model of creative cognition do not exclusively reside inside the head of the individual choreographer. Instead, they are co-constituted by the interactions between the choreographer and the dance ensemble, the dancers and their bodies, and the artistic props and recording devices on which they rely (Figure 5.2). The creative dynamics of choreographic cognition reflects the transformations of these interactions over time—a process that Sawyer (2003b, 2003c) has dubbed “collaborative emergence”. Manifestations of collaborative emergence receive a high score on all three dimensions of emergence that we have distinguished. First, the classification of the mental representations and operations outlined in the Geneplore model are essentially substrate-neutral (emergent3). Hence it is consistent with their functional specification that they are realized by a socially and technologically extended activity system that stretches beyond the boundaries of the individual (Barnier et al., 2008; Clark, 2003, 2008; Hutchins, 1995a; Sawyer & Greeno, 2009; Theiner, 2008; R. Wilson & Clark, 2009). For instance, the composition of new movement patterns by combining, intersecting, and merging individually developed phrases can be seen as a manifestation of attribute finding. The speculative experimentation of dancers with the book/spine paper sculpture—considered as a pre-inventive structure—corresponds to an instance of functional inference. Choreographing a dynamic DNA-like helix pattern made up of five dancers in constant motion proved to be a particularly complex task that required a great amount of hypothesis testing which consumed many hours of joint experimentation. A preliminary analysis of the task was carried out on paper, with the help of a color braid and its visual representation on paper. Later, a trial-and-error process was executed on the dance floor, using colored tape to mark out the different strands of movement (Stevens et al., 2003: 316–318). Second, moving the relevant unit of choreographic analysis from the individual to the group reveals the intricate ways in which the creation of a dance performance is distributed (i.e., emergent1) along three principal dimensions, each of which provides its own forms of “scaffolding” (Hollan et al., 2000; Sutton, 2006).18 Choreographic cognition is distributed in space, insofar as bodily and environmental resources transform the nature of the cognitive load hoisted upon biological brains; over time, insofar as the outcomes of earlier stages of cognitive processing transform the task demands during the later stages; and socially, insofar as membership in the ensemble (as well as more temporary interactions between dyads and triads)

The Emergence of Group Cognition 99 transforms the cognitive tasks faced by each individual. Consider how misguided it would be to analyze the creative dynamics of the choreographic process by fi rst trying to understand the properties of people and artifacts in isolation, and then stitching them together in a purely aggregative fashion. In the creation of Red Rain, the generation, exploration, selection, and refi nement of dance material was achieved through all sorts of collaborative “entanglements”, including the discussion of texts, images, and cues provided by the choreographer that are mapped into exploratory movement patterns and subsequently adopted, modified, verbally paraphrased and critiqued by others; the interpersonal synchronization of bodily movements; the perceptual spread of “contagious” movement patterns across dancers; the development of shared kinesthetic memories; the playful appropriation and integration of artistic props (e.g., beans, wax, a paper sculpture); or the public viewing of video recordings made of improvised dance phrases displayed at an earlier stage. Third, since novelty is an essential mark of creativity, the inherent creativity of choreographic cognition implies a higher degree of emergence2 which appeared to be lacking in our previous two examples.19 At the beginning of the choreographic process, there is often little more than a few vague images which a choreographer might bring to the table, but no representation of a fi xed plot, no sense of how these images can be combined to construct a narrative which can be transformed into bodily movement, no pre-formatted sequence of moves, and so on. Improvisation is a crucial aspect of creative processes—one that is frequently sparked by collaborative

Figure 5.2 Phases of the choreographic development of dance material for Red Rain, classified in terms of the Geneplore model of creative cognition (shown in uppercase). A number of different phrases of movement were in development at any given time. From C. Stevens et al., (2003: 312).

100 Georg Theiner and Timothy O’Connor interactions. For instance, during week 5, the choreographer introduced an improvisation task to “perturb” habitual movement patterns that were embodied in the dancers’ kinesthetic and muscular memories. This was achieved by externalizing cognitive and motor control processes that would usually run smoothly inside the sheath of the biological body into the social environment—by letting other dancers dictate which and how the parts of one’s body ought to be moved. Each dancer was forced to respond to impromptu verbal cues issued by other dancers (e.g., “right elbow behind back, shoulders tilting, left hand reaching”). The elicited movement patterns were recorded on video, reviewed by the choreographer, discussed with the dancers, and later re-enacted by the dancers from video observations before some of them would eventually be incorporated into the dance (C. Stevens et al., 2003: 304–305). A fi nal point about the diachronic dimension of “collaborative emergence” is worth emphasizing. 20 Although we can retrospectively (at least in principle) identify each participant’s contributions to the incremental process of creating a new dance, many of its effects on the emerging 2 organization are not explicitly intended or planned as such by the time these contributions are made. This is because the full cognitive significance of “dance-making” actions and intentions frequently depends on the subsequent flow of activities, and may thus not be revealed until a much later phase of choreographic cognition. For instance, some of the improvised movement patterns produced in the fi rst couple of weeks (e.g., in response to dripping blood or the pulsing in-and-out motif) acted as recurrent “seeds” that inspired and framed choreographic processes at much later stages before they assumed their fi nal shape (C. Stevens et al., 2003: 306–308).

4. OBJECTIONS AND REPLIES Let us restate the ontological commitments of our analysis of group cognition. As we have emphasized from the outset, our account is contingent on the plausibility of non-reductive physicalism, in particular as it has been applied to the realm of the mental (Block, 1997; Davidson, 1970; Fodor, 1974, 1989; Horgan, 1993) and the social (Kincaid, 1997; Pettit, 1993; Sawyer, 2002, 2003a). We claim that, on that assumption, it is likewise plausible that groups can be the bearers of psychological properties which are distinct from the psychological properties had by its members. Second, we grant that group-level psychological facts are metaphysically determined (“realized”) by the totality of individual-level psychological facts, together with other, non-psychological facts about their social and material organization. At a minimum, this implies that any two groups which are composed of the exact same members participating in all the same social interactions cannot differ in their psychological properties—a relatively weak variety of ontological dependence commonly known as

The Emergence of Group Cognition 101 global supervenience. Third, the supervenience of group cognition on its lower-level realization is consistent with its emergence in each of the senses we have outlined—its organization-dependence, its unpredictability from the standpoint of individual cognition, and its multiple realizability. In virtue of the latter, the relation between group-cognitive properties and their lower-level realization must be one of token-identity (or perhaps token-realization)21, but not type-identity in the stronger sense that is required for inter-theoretic reduction. With these assumptions in place, we now consider two basic metaphysical objections to the very idea of group cognition.

4.1. Is Group Cognition Epiphenomenal? The fi rst objection concerns the causal efficacy of emergent psychological properties that are instantiated by groups as a whole. The supervenience of group cognition on its lower-level realization arguably implies what we may call the causal completeness of the individual level. 22 By this, we mean the claim that for every aspect of an individual’s behavior, there is a sufficient cause which refers only to properties of that individual and her social interactions, but does not take into account any psychological or non-psychological (e.g., behavioral) properties of the entire group to which that individual belongs. Suppose further that we rule out an individual’s behavior being systematically overdetermined by both individual- and group-level causes. Then it appears that there is no real causal work for the psychological properties of groups to do, because any influence it could exercise on individuals is effectively “screened off” by its lower-level realizers. Does this show that the phenomenon of group cognition—even if it is real—is epiphenomenal? In response, let us observe that this objection is very similar to the socalled “causal exclusion” problem in the philosophy of mind, where it has been invoked to demonstrate the causal irrelevance of mental properties vis-à-vis their physical realizers (Kim, 1989, 1992, 1998; Malcolm, 1968). An important similarity is the portrayal of the relationship between higher-level properties and their lower-level realizers as causal competitors in the production of behavior. This assumption requires that both properties figure in causal explanations of one and the same tokens of behavior. However, this is not always the case. For instance, we attribute a TMS to a group in order to explain how the group performs as a whole (e.g., how accurately it assembles a radio), whereas we refer to properties of individuals (including their opportunities for social interaction) to explain how they perform as parts of the group. If we can conceive of individual- and group-level psychological properties as links of two separate causal chains influencing the behavior of entities at different levels of composition, and if, further, these chains do not stand in a zero-sum competition, the present form of the objection does not get off the ground.

102

Georg Theiner and Timothy O’Connor

However, we can further press the objection by reflecting on the fact that group behavior itself is a social (albeit non-psychological) property which is realized by the psychological and behavioral properties of individuals, together with their modes of social interaction. What the causal completeness of the individual level implies, then, is that there is in principle a complete causal account of this collective state of affairs which does not refer to the cognitive properties Mg of groups, but is couched entirely in terms of the lower-level realization of Mg. This follows from the (here presumed) fact that the causal powers associated with Mg are fully determined by the causal powers of its lower-level realization. Therefore, one might still conclude that Mg cannot make a non-redundant causal impact on group behavior (Kim, 1998, 2005; O’Connor & Churchill, Forthcoming). A familiar theme in response to this objection has been to argue that the inter-level relationship between mental properties and their physical realization is not one of rivalry, but of compatibility (see, e.g., Block, 2003; Horgan, 1993; Jackson & Pettit, 1990; Kim, 1984; Shoemaker, 2001, 2007). There are several ways one might seek to cash out this idea. One way is to stress that the metaphysical connection between mental properties and their physical realizers is sufficiently intimate so that the former “inherit” the causal powers of the latter, rather than being excluded by them. Insofar as mental properties exercise their causal influence through their physical realizers, it is claimed, the charge of overdetermination is thereby avoided. Taking this line further, one can distinguish the causal relevance of higher-level properties from the causal productivity of the physical mechanisms by means of which the former are realized.23 Such a strategy can be extended to accommodate the causal relevance of groupcognitive properties. For instance, consider the causal impact of a TMS on group performance in terms of Jackson and Pettit’s (1990) program model of causal relevance. According to this model, the occurrence of a higherlevel property “programs for” a certain same-level effect by ensuring the presence of suitable lower-level realizers which bring about the specified outcome.24 In the case of TMS, we can grant that the mechanism in virtue of which a TMS is causally relevant for explaining the performance of a group involves the appropriate cognitive abilities, motivations, and opportunities for social interaction at the level of individuals. A second, related response to the exclusion worry appeals to counterfactual dependence patterns in order to ground the causal relevance of higher-level event-types (Horgan, 1993; LePore & Loewer, 1989; Loewer, 2002; McLaughlin, 1989). For instance, Horgan (1993) argues that a single event-token can be causally relevant on more than one level, in virtue of falling under multiple event-types which are not nomologically coextensive. Assuming that something like counterfactual dependency suffices for causation or causal relevance, properties at multiple levels could thus each be causal because they all support different counterfactual dependence patterns among events. Now consider the claim that if a group were to develop

The Emergence of Group Cognition 103 a TMS, it would be able to assemble a radio accurately. Is this true, or is the causal relevance of TMS preempted by the occurrence of its lowerlevel realizer? Let us apply the criteria for preemption suggested by Loewer (2002). From the causal completeness of the individual level, it follows that the radio assembly would have been equally accurate if it were somehow performed by a member-by-member duplicate of the group with the same social organization S, but without a TMS.25 But now consider a situation in which a TMS develops, but without its actual lower-level realization. Assuming that the properties of a TMS are multiply realizable, the closest possible world in which this can happen is one in which the functions of the TMS are performed by a group with a slightly different social (and corresponding individual-psychological) configuration S*. Since it remains true that a group whose TMS is realized by S* can assemble the radio accurately, it follows that S does not preempt the causal relevance of TMS. (See Loewer, 2002, for further discussion.) Now, it may well be the case that neither of these proposed strategies for harmonizing the causal completeness of physics with mental-physical realization is entirely neutral with respect to theories of the nature of causation. O’Connor and Churchill (forthcoming) argue that harmonization fails on a non-reductive production account of particular. If that is correct, then our guiding assumption carries with it an implicit rejection of this metaphysical theory. (Again, one who is attracted to the non-reductive physicalist account of human mentality has a choice here between modus ponens or modus tollens, a question on which we are neutral here.)

4.2. How Robust are Group Cognizers? The second objection to the thesis of emergent group cognitive states and processes that we consider here concerns the metaphysical unity of groups. Individual biological subjects possess a robust systemic unity defi ned in terms of the biological persistence of the body and/or its most vital component, the brain. This unity has spatial, temporal, and functional aspects. Owing to these features, its persistence through time (on the relevant time scale) is reasonably well defi ned in biological terms. The objection’s premise is that nothing like this is feasible for an attempted account of the persistence of groups having allegedly emergent cognitive states, and it concludes that there simply are not any persisting groups of the sort that could serve as the subjects of those states. This objection raises a large issue that we cannot fully explore here. However, we can briefly make some points that go a substantial ways, we think, toward challenging the suggested starkness of the contrast between individual and group cognizers. To begin with, we want to draw attention to a similar debate that is currently raging over the “extended mind” thesis (Clark, 2003, 2008; Clark & Chalmers, 1998; R. Wilson, 2004). 26 In their original statement of this thesis, Clark and Chalmers (1998) suggested

104

Georg Theiner and Timothy O’Connor

a number of ‘coupling conditions’ which the causal intercourse between a biological organism and capacity-enhancing bio-external artifacts (e.g., notebook, iPhone, or neural implants) ought to satisfy in order to constitute a unified agent’s extended cognitive apparatus. Their conditions require that the external resource should be reliably available and typically deployed when confronted with the task at hand, that any relevant information contained in the resource should be easily accessible, and that any information retrieved from the external device should be more or less automatically endorsed, and treated as a trustworthy source. In response to the presently explored objection, we maintain that similar principles of systemindividuation are at least equally convincing if they apply to the cognitively significant interactions among people (see also Tollefsen, 2006). We consider two sorts of cases in which groups can be collectively coupled, to be understood as two poles in a continuum from the fleeting to the more permanent. In the fi rst sort, ‘disbanding’ happens regularly, insofar as cognitive activity does not more or less continuously persist, but is instead sharply episodic. Examples would be work assembly lines, dance troupes, and philosophy departments. Individuals assembled in one of these groups function at particular times as a collective cognitive system and then go their separate ways, no longer causally coupled with one another toward a common task. Even so, this regular disbanding and re-forming does not ruin the integrity of the system over time as long as the functional organization remains invariant across such episodes, such that the collective plausibly may continue to persist even though individual members don’t persist. There will of course be borderline cases of more radical change where it is quite unclear whether we are dealing with the very same group, but analogous problems beset accounts of the persistence of thinking organisms. (The strong, physicalism-denying emergentist may well claim an advantage here, since the emergence of ontologically basic properties seems to be an all or nothing affair. But we are proceeding on the premise that this view is off the table.) Granted, the existence of certain kinds of intermittent group cognizers, unlike individual cognizers, is tied down to time-on-(cognitive-) task. But it is not obvious that this difference has the metaphysical import that the objection assumes. On the other side of the spectrum, there is a second sort of group cognizers whose persistence is continuous, because disbanding would seriously limit, if not relinquish, the ability of its parts to function in isolation. Examples of this sort are tightly integrated, long-standing social groups whose individual members (and their sub-assemblies) have come to behave as a kind of collective “super-organism” (D. S. Wilson, 2002; D. S. Wilson, van Vugt, & O’Gorman, 2008; D. S. Wilson & E. O. Wilson, 2007). In biology, the existence of such “major evolutionary transitions” (Maynard Smith & Szathmary, 1995) was originally posited to explain the formation of eukaryotic cells as highly interdependent symbiotic associations of prokaryotic (bacterial) cells. But the idea was later generalized to encompass

The Emergence of Group Cognition 105 to the formation of chromosomes, multi-cellular organisms, social insect colonies, and the evolution of human sociality in our ancestral environments. In each of these rare but momentous transitions, the cooperation of entities at a lower level has strongly beneficial effects for their arrangement as a whole, due in large part to the differentiation of functions. Hence a major evolutionary transition can occur if between-group selection outweighs within-group selection. This requires the existence of lower-level mechanisms that continually suppress competition among selfish individual units, and thus enable the collective to overcome the problem of “free riders” (Sober & D. S. Wilson, 1998). For human groups, especially smallscale human societies that have historically dominated the evolution of sociality in our species, this job is performed by fairly strong mechanisms of social control. These are believed to include a variety of biological and cultural co-adaptations such as religion (D. S. Wilson, 2002), social emotions like anger and guilt (Haidt, 2007), practices of shaming and gossip (Richerson & Boyd, 2005), but also the joy of synchronized movement in song and dance (Haidt, Seder, & Kesebir, 2008). In sum, emergent1 social systems whose members are deeply interdependent for their physical, cognitive, as well as emotional functioning as individuals are also the most likely candidates to act as maximally robust group cognizers.

5. CONCLUSION Can groups think? In this paper, we have probed into a wide range of contemporary frameworks that seem to be committed to giving an affi rmative answer to this question. A common methodological thrust of these research programs is to overcome the strong individualist bias in psychology and cognitive science, and to give a wider berth to environmental (social, material, and cultural) factors in shaping and participating in cognitive processes (see, e.g., the papers in Robbins & Aydede, 2009). As philosophers, our primary goal has been to offer an ontological reconstruction of what it would mean for a group of people interacting with one another to have emergent cognitive properties. Based on the predominantly functional understanding of cognition that is proffered in the theoretical frameworks we examined, we have outlined three different features that emergence can plausibly be taken to signify in the context of group cognition: its dependence on the social organization and interactions among individuals; the manifestation of unintended cognitive effects at a group level; and the multiple realizability of cognitive properties by different types of group structures. From a metaphysician’s bird’s-eye view, all of these features are compatible with various brands of non-reductive physicalism. Hence our advocacy of the idea that groups can think is conditional twice over: fi rst, on the explanatory scope of the research frameworks we have discussed; second, on the ontological coherence of non-reductive physicalism as a general view of the

106 Georg Theiner and Timothy O’Connor mental. Accordingly, our updated portrayal of emergent group cognition departs from metaphysically stronger versions of the Group Mind Thesis in several crucial respects. First, during the heydays of the Group Mind Thesis, the idea of a “collective psychology” was often used to emphasize the essential like-mindedness of people in groups—”the mass movements of a crowd, the majority decisions of the electorate, or the sweetly homogeneous mindlessness of people in love” (Wegner, 1986: 206). While we certainly agree that the social alignment of shared concepts, attitudes, and values can be important prerequisites of group cognition, it is rather the integration of diverse but complementary contributions that directly underpins the emergent cognitive properties of groups. Second, the dilemma that group minds are either “nothing but” collections of individual minds or doomed to inhabit preternatural realms (“telepathy”) rests on an inadequate understanding of interactional complexity. Based on the suggested reading of emergence1 as organization-dependence, there is no reason to think that group-level properties must always have a purely aggregative decomposition relative to which we can neglect the social interactions and communication processes by which individuals coordinate their behavior. Moreover, if group-level cognitive properties are multiply realizable, the non-reductive physicalist can grant that all cognitive facts about a group supervene globally on the totality of individual-level psychological facts about its members, plus other, non-psychological facts about their social organization (R. Wilson, 2004; cf. Section 1 in this chapter). This is because the global supervenience of group cognition on its total physical realization (Shoemaker, 2007) is consistent with its emergence in all of the three senses we have sketched. Third, group cognition is here taken to occur without any dubious sort of collective consciousness. Group cognitive states and processes of the sort suggested by contemporary empirical theories do not entail that there is a conscious, self-aware subject of them. As we emphasized in the beginning, we favor the ‘big tent’ approach to cognition that is reflected in much of recent cognitive science. How we should think about phenomenal consciousness in particular is quite unsettled. But it seems clear that none of the group cognition–friendly theories give reason to posit collective consciousness on any of the most promising philosophical accounts of the nature and function of consciousness (whether physicalist or dualist). Keeping this point in mind, it is worth mentioning a possible asymmetry between the “extended mind” thesis (Clark & Chalmers, 1998) and the kind of “group mind” thesis considered here. One might argue that a move to the group mind thesis involves a riskier generalization of the strategy underlying the case for extended individual cognition, precisely because the former (but not the latter) invokes the existence of a subject who is not capable of having phenomenal consciousness. This suggestion, which is due to David Chalmers (personal communication), is most convincing if

The Emergence of Group Cognition 107 one accepts the following two claims: fi rst, that the presence of or potential for conscious awareness is the single most uncontroversial—if not indubitable—criterion for the existence of a mind; and second, that the conscious mental states of such a subject, contrary to its non-conscious cognitive states and processes, are not likewise extended. For instance, it may turn out that the physical basis of consciousness requires certain forms of highbandwidth signal processing and fi ne temporal coordination that are (at least currently) not supported by organism/environment-interactions, nor by interactions between people (for such hybrid views, see Chalmers, 2008; Clark, Submitted). If both these claims were accepted, the “group mind” thesis, at least where it does not involve implausible posits of group consciousness, would be a nonstarter. It is thus not surprising that Chalmers, a dualist about phenomenal consciousness, should be attracted to the consciousness criterion for genuine mentality. But it is not an equally plausible position for a physicalist to take. From a physicalist perspective, conscious states are not intrinsically distinctive (as on dualism), but are merely a particular variety of complex, physically realized mental states. On many physicalist views, consciousness is ultimately understood as some sort of representational capacity restricted to sophisticated, highly complex cognitive systems—something characteristic of particularly advanced cognitive systems, rather than a necessary condition for any cognition to take place. In our reconstruction, we have already granted that the repertoire of cognitive capacities displayed by groups need not—and typically does not—live up to the full-fledged mentality of individual human beings. Consistent with an incrementalist, ‘big tent’ approach to cognitive complexity, the physicalist should thus be quite comfortable with admitting the existence of collective cognitive systems, while remaining agnostic about group minds as indicated by the consciousness criterion.

NOTES 1. The research of the fi rst author has been supported by an Izaak Walton Killam Post-doctoral Fellowship. 2. For historical overviews, see Wegner (1986), Wegner, Giuliano, and Hertel (1985), R. Wilson (2004: ch. 11). 3. See, e.g., Chomsky (1980) and Thagard (1996) for discussion. 4. An application of Wimsatt’s conception of emergence1 to the idea of group cognition (understood as socially distributed cognition) is also offered by Poirier and Chicoisne (2006). 5. As Wimsatt (1994: 8n.) points out, the relevant notion of identity that his conception of ontological reduction implies falls somewhere between tokenidentity (which he considers as too weak a requirement for reduction) and general type-identity (which he considers as too strong a requirement for reduction) between higher-level properties and their lower-level realizers. We shall get to the putative multiple realizability of higher-level types (kinds,

108 Georg Theiner and Timothy O’Connor

6.

7.

8. 9.

10.

11. 12.

13. 14.

properties) as a potential obstacle for ontological type-reduction in Section 2.2.3. Our notion of emergence2 is related to several other conceptions of emergence that have been discussed in the literature, in particular Chalmers’s (2006) notion of “weak emergence”, Clark’s (2001: 114) notion of emergence as “unprogrammed functionality”, as well as the spectrum of weaker vs. stronger versions of “diachronic emergence” discussed in Rueger (2000) and Stephan (2006). Wimsatt (1981, 1986, 1994) has drawn attention to the fact that the multiple realizability of macro-properties is a rather natural consequence of functional redundancy in the compositional organization of complex systems, and not incompatible with mechanistic explanations of their behavior in terms of its parts and their interactions. As Wimsatt (1981) points out, redundancy is an important design feature of many complex systems, such as built-in hardware redundancy in computers, excess capacities of cells in organisms, and bilaterally symmetric redundant organs in animals. Redundancy ensures a limited form of “sub-aggregativity” between subsets of actual parts (1986). If we further consider that Wimsatt’s condition IS also involves the counterfactual replacement with possible parts taken “from a relevantly similar domain”, the modal implications of multiple realizability become apparent. See, e.g., Cicourel (1990), Hinsz et al. (1997), Larsen and Christensen (1993), Mohammed and Dumville (2001), Propp (1999), Stasser (1999), Wegner et al. (1985). As Hinsz et al. (1997: 44n.) point out, their model does not necessarily depict a blueprint of the cognitive architecture by which information is actually processed (in individuals or in groups), which is likely to be less sequential and more highly interrelated than the model suggests. But this does not diminish its analytic value for identifying various factors and mechanisms which underlie the collective performance of cognitive tasks. Working on human vision, Marr proposed to split the task of explaining how a system (e.g., a cash register or a brain) processes information into three levels. On Marr’s level-1, the goal is to give a functional analysis of the behavior to be performed. This requires that we pin down the precise input-output function of the task at hand and provide a procedural characterization of the subtasks which the system has to carry out. Marr’s level-2 is concerned with describing how the relevant inputs and outputs are represented, and which algorithms are used to compute the required mappings. Finally, level-3 concerns the implementation of these computational processes in a physical substrate. See Moreland (1999) for an overview. These fi ndings are consistent with research on learning at the individual level, which suggests that experience with only one task may not be sufficient for a subject to detect the underlying analogies between superficially dissimilar problems (Gentner et al., 2003). Further analysis by Lewis et al. revealed a significant interactive effect between group type and expertise stability. An intact prior TMS was most beneficial for learning transfer in groups whose members had been reassigned but kept their expertise specialization. On the other hand, an intact prior TMS was detrimental for groups when expertise stability across tasks was low—either because members had been reassigned or the groups abandoned their initial distribution of expertise (see also Lewis et al., 2007). See Schmitt (2003b) for an overview of this debate. Fodor’s argument against reduction has been heavily scrutinized (Kim, 1989, 1992; Polger, 2004). We do not wish to adjudicate the current state of this

The Emergence of Group Cognition 109

15.

16.

17. 18.

19. 20. 21. 22.

23.

debate here, since our reasoning is conditional upon the viability of the nonreductive physicalist’s gambit. That perspective’s chief motivation is precisely the one given by Fodor. As applied to the domain of the social, Ruben (1985: ch. 3) has argued that the argument from multiple realizability is ultimately more compelling for what he calls “variable” social properties (e.g., being a mayor) than it is for mental properties. Since the manifestation of these social properties is partly determined by social conventions, there are even fewer lawful constraints on the class of mental and/or physical properties which are sufficient for their realization. Sawyer (2002, 2003a) has leaned heavily on Fodor’s argument in his defense of the causal-explanatory autonomy of social properties vis-à-vis individuals and their interactions, proposing a theory of social emergence3 that he calls non-reductive individualism. An implication of this “active” kind of externalism is that the individualbound portions of these interactions are not metaphysically sufficient for socially manifested cognitive processes of the requisite sort to occur (R. Wilson, 2001b, 2004). Wegner (1986: 190–191) gives a particularly nice example of transactive retrieval coordination that involves both intrapersonal and interpersonal components and is distributed over internal and external storage spaces: “A client asks the boss for information, for instance, that the boss has no idea about—but thinks the secretary may know. [ . . . ] it may be that the secretary fails to fi nd the item internally, perhaps fi nding instead some other information related to the label. As it turns out, perhaps the secretary recalls that the boss asked for this information at another time and reports to her boss: ‘I gave that to you last Tuesday’. The boss may now be able to use the new lead to retrieve some item internally or externally. He might now recall that the information he asked for Tuesday was in the top desk drawer in a fi le labeled ‘THIS IS IT’.” Surely there is nothing nomologically necessary about transactive retrieval coordination happening in this idiosyncratic way. The study describes the choreographic realization of dance work carried out during 1999–2000 that involved choreographer Anna Smith and eight professional dancers. It led to the production of Red Rain. While the focus of our paper has been the social dimension of cognitive extension, we believe that the gist of our analysis of distribution as a form of emergence1 carries over to all three dimensions of distributed cognitive systems. For present purposes, we rely on a fairly broad conception of creativity defi ned as “a socially recognized achievement in which there are novel products” (Barron & Harrington, 1981: 442). This point is forcefully raised by Sawyer (2003b: 18) as an argument for the irreducibility of “collaborative emergence” to the actions and intentions of individual actors. For further discussion, see Sawyer (2003c). For discussion of the merits of token-realization, see Shoemaker (2007) and Pereboom (2002). For a discussion of closure principles with respect to the physical properties and their relationship to mental properties, see, e.g., Kim (1998, 2005), Lowe (1993), Melnyk (2003), and O’Connor and Churchill (forthcoming). For a discussion of the causal completeness of the individual level with respect to the social, see particularly Kincaid (1997), Pettit (1993), and Sawyer (2003a). This requires that we depart from the “homogeneity assumption” (Crane, 1995) that mental and physical causation are theoretically on a par. Notice that what counts as “higher” and “lower” levels of realization in a given context is itself dependent on the phenomenon that we seek to explain.

110

Georg Theiner and Timothy O’Connor

24. Jackson and Pettit’s program model is very similar in spirit to Kim’s (1984) notion of supervenient causation. See Menzies (2007) for a recent discussion. 25. Whether this situation is even metaphysically possible at all depends on the exact sense in which we take the psychological states of groups to supervene on their lower-level realizers. 26. For a critical evaluation of this thesis, see Adams and Aizawa (2008) and Rupert (2009), as well as the responses offered on behalf of the “extended mind” by Clark (2008, in press), R. Wilson (in press), and R. Wilson and Clark (2009).

REFERENCES Adams, F., & Aizawa, K. (2008). The Bounds of Cognition. Oxford: Blackwell. Alexander, J. C., & Giesen, B. (1987). From reduction to linkage: The long view of the micro-macro link. In J. C. Alexander, B. Giesen, R. Munch, & N. J. Smelser (Eds.), The Micro-Macro Link (pp. 1–42). Berkeley: University of California Press. Argote, L. (1999). Organizational Learning: Creating, Retaining and Transferring Knowledge. Norwell, MA: Kluwer. Bak, P. (1996). How Nature Works: The Science of Self-Organized Criticality. Cambridge, MA: MIT Press. Ball, P. (1998). The Self-Made Tapestry: Pattern Formation in Nature. Oxford: Oxford University Press. . (2004). Critical Mass: How One Thing Leads to Another. New York: Farrar, Straus & Giroux. Barabási, A.-L., & Albert, R. (1999). Emergence of scaling in random networks. Science, 286, 509–512. Barnier, A. J., Sutton, J., Harris, C. B., & Wilson, R. A. (2008). A conceptual and empirical framework for the social distribution of cognition: The case of memory. Cognitive Systems Research 9(1), 33–51. Barron, F., & Harrington, D. M. (1981). Creativity, intelligence, and personality. Annual Review of Psychology, 32, 439–476. Bechtel, W. (2006). Discovering Cell Mechanisms: The Creation of Modern Cell Biology. Cambridge: Cambridge University Press. Bechtel, W., & Abrahamsen, A. (2002). Connectionism and the Mind: Parallel Processing, Dynamics, and Evolution in Networks (2nd ed.). Oxford: Basil Blackwell. Bechtel, W., & Richardson, R. C. (1993). Discovering Complexity: Decomposition and Localization as Strategies in Scientifi c Research. Princeton, NJ: Princeton University Press. Block, N. (1978). Troubles with functionalism. In W. Savage (Ed.), Minnesota Studies in the Philosophy of Science (Vol. IX, pp. 261–325). Minneapolis: University of Minnesota Press. . (1980). Introduction: What is functionalism? In N. Block (Ed.), Readings in Philosophy of Psychology (Vol. 1). Cambridge, MA: Harvard University Press. . (1995). On a confusion about a function of consciousness. Behavioral and Brain Sciences, 18, 227–272. . (1996). What is functionalism? In The Encyclopaedia of Philosophy Supplement. New York: MacMillan Reference Books. . (1997). Anti-reductionism slaps back. In J. Tomberlin (Ed.), Philosophical Perspectives 11: Mind, Causation, and World. Boston: Blackwell.

The Emergence of Group Cognition 111 . (2003). Do causal powers drain away? Philosophy and Phenomenological Research, 67(1), 133–150. Blumer, H. (1939). Collective behavior. In R. E. Park (Ed.), An Outline of the Principles of Sociology (pp. 219–280). New York: Barnes & Noble. Bonabeau, E., Dorigo, M., & Theraulaz, G. (1999). Swarm Intelligence: From Natural to Artificial Systems. New York: Oxford University Press. Bratman, M. (1993). Shared intention. Ethics, 104, 97–113. . (1997). I intend that we J. In R. Tuomela & G. Holmstrom-Hintikka (Eds.), Contemporary Action Theory (Vol. 2, pp. 49–63). Dordrecht: Kluwer. Burge, T. (2000). Reason and the fi rst person. In C. Wright, B. Smith, & C. Macdonald (Eds.), Knowing Our Own Minds. Oxford: Oxford University Press. Burke, P. (1989). History as social memory. In T. Butler (Ed.), Memory: History, Culture, and the Mind. Oxford: Blackwell. Castelfranchi, C. (1998). Simulating with cognitive agents: The importance of cognitive emergence. In J. Sichman, R. Conte, & N. Gilbert (Eds.), Multi-Agent Systems and Agent-Based Simulation (pp. 26–44). Berlin: Springer. Chadderdon, G. L. (2008). Assessing machine volition: An ordinal scale for rating artificial and natural systems. Adaptive Behavior, 16, 246–263. Chalmers, D. (1996). The Conscious Mind: In Search of a Fundamental Theory. New York: Oxford University Press. . (2006). Strong and weak emergence. In P. Clayton & P. Davies (Eds.), The Re-emergence of Emergence: The Emergentist Hypothesis from Science to Religion (pp. 244–256). Oxford: Oxford University Press. . (2008). Foreword to A. Clark, Supersizing the Mind: Embodiment, Action, and Cognitive Extension (pp. vii–xiv). New York: Oxford University Press. Chomsky, N. (1980). Rules and Representations. Oxford: Blackwell. Cicourel, A. V. (1990). The integration of distributed knowledge in collaborative medical diagnosis. In J. R. Galegher, R. E. Kraut, & C. Egido (Eds.), Intellectual Teamwork: Social and Technological Foundations of Cooperative Work. Hillsdale, NJ: L. Erlbaum Associates. Clark, A. (2001). Mindware: An Introduction to the Philosophy of Cognitive Science. New York: Oxford University Press. . (2003). Natural-Born Cyborgs: Minds, Technologies, and the Future of Human Intelligence. New York: Oxford University Press. . (2008). Supersizing the Mind: Embodiment, Action, and Cognitive Extension. New York: Oxford University Press. . (In press). Coupling, constitution, and the cognitive kind: A reply to Adams and Aizawa. In R. Menary (Ed.), The Extended Mind. Cambridge, MA: MIT Press. . (Submitted). Spreading the joy? Why the machinery of consciousness is (probably) still in the head. Clark, A., & Chalmers, D. (1998). The Extended Mind. Analysis, 58(1), 7–19. Couzin, I. D. (2007). Collective minds. Nature, 455, 715. . (2009). Collective cognition in animal groups. Trends in Cognitive Sciences, 13(1), 36–43. Crane, T. (1995). The mental causation debate. Proceedings of the Aristotelian Society, 69, 211–236. Davidson, D. (1970). Mental events. In L. Foster & J. W. Swanson (Eds.), Experience and Theory. London: Duckworth. Dennett, D. (1996). Kinds of Minds: Toward an Understanding of Consciousness. New York: Basic Books. Edwards, S. C., & Pratt, S. C. (2009). Rationality in collective decision-making by ant colonies. Proceedings of the Royal Society of London, B, 276, 3655–3661.

112

Georg Theiner and Timothy O’Connor

Epstein, J. M., & Axtell, R. (1996). Growing Artificial Societies: Social Science from the Bottom Up. Cambridge, MA: MIT Press. Finke, R. A., Ward, T. B., & Smith, S. M. (1996). Creative Cognition: Theory, Research, and Applications. Cambridge, MA: MIT Press. Fodor, J. (1968). Psychological Explanation. New York: Random House. . (1974). Special sciences (or: the disunity of science as a working hypothesis). Synthese: An International Journal for Epistemology, Methodology and Philosophy of Science, 28, 97–115. . (1989). Making mind matter more. Philosophical Topics, 67(1), 59–79. . (1997). Special sciences: Still autonomous after all these years. Philosophical Perspectives, 11, 149–163. Gentner, D., Loewenstein, J., & Thompson, L. (2003). Learning and transfer: A general role for analogical encoding. Journal of Educational Psychology, 95(2), 393–408. Giere, R. (2002). Scientific cognition as distributed cognition. In P. Carruthers, S. P. Stich, & M. Siegal (Eds.), The Cognitive Basis of Science (pp. 285–299). Cambridge: Cambridge University Press. Giere, R., & Moffatt, B. (2003). Distributed cognition: Where the cognitive and the social merge. Social Studies of Science, 33(2), 301–310. Gilbert, M. (1989). On Social Facts. London: Routledge. Gillett, C. (2003). The metaphysics of realization, multiple realizability and the special sciences. Journal of Philosophy, 100, 591–603. . (2006). Samuel Alexander’s emergentism: or, higher causation for physicalists. 153, 261–296. Goldstone, R., & Janssen, M. A. (2005). Computational models of collective behavior. Trends in Cognitive Science, 9, 424–430. Gureckis, T. M., & Goldstone, R. (2006). Thinking in groups. Pragmatics & Cognition, 14(2), 293–311. Haidt, J. (2007). The new synthesis in moral psychology. Science, 316, 998–1002. Haidt, J., Seder, P., & Kesebir, S. (2008). Hive psychology, happiness, and public policy. Journal of Legal Studies, 37, 133–156. Hardin, G. (1968). The tragedy of the commons. Science, 162, 1243–1248. Hayek, F. A. v. (1942). Scientism and the study of society, Part 1. Economica, 9, 267–291. . (1944). Scientism and the study of society, Part 3. Economica, 11, 27–37. Hempel, C. G. (1965). Aspects of Scientifi c Explanation and Other Essays in the Philosophy of Science. New York: The Free Press. Hinsz, V. B., Tindale, R. S., & Vollrath, D. A. (1997). The emerging conceptualization of groups as information processors. Psychological Bulletin, 121(1), 43–64. Hollan, J., Hutchins, E., & Kirsh, D. (2000). Distributed cognition: Toward a new foundation for human-computer interaction research. ACM Transactions on Computer-Human Interaction, Special issue on human-computer interaction in the new millennium, Part 2, 7(2), 174–196. Holland, J. H. (1975). Adaptation in Natural and Artificial Systems. Ann Arbor: University of Michigan Press. . (1995). Hidden Order: How Adaptation Builds Complexity. Reading, MA: Addison-Wesley. Hölldobler, B., & Wilson, E. O. (1990). The Ants. Cambridge, MA: Belknap. Hollingshead, A. B. (1998). Communication, learning, and retrieval in transactive memory systems. Journal of Experimental Social Psychology, 34, 432–442. Horgan, T. (1993). Nonreductive materialism and the explanatory autonomy of psychology. In S. Wagner & R. Warner (Eds.), Naturalism: A Critical Appraisal (pp. 295–320). South Bend, IN: Notre Dame University Press.

The Emergence of Group Cognition 113 Hurley, S. L. (1998). Vehicles, contents, conceptual structure, and externalism. Analysis, 58, 1–6. Hutchins, E. (1995a). Cognition in the Wild. Cambridge, MA: MIT Press. . (1995b). How a cockpit remembers its speed. Cognitive Science, 19, 265– 288. . (1991). The social organization of distributed cognition. In L. B. Resnick, J. M. Levine, & S. D. Teasley (Eds.), Perspectives on Socially Shared Cognition (pp. 283–307). Washington, DC: American Psychological Association. Jackson, F., & Pettit, P. (1990). Program explanation: A general perspective. Analysis, 50, 107–117. Jensen, T. R., & Toft, B. (1995). Graph Coloring Problems. New York: Wiley. Jung, C. G. (1922). Collected Papers on Analytical Psychology. London: Bailliere, Tindall, and Cox. Kauffman, S. (1993). The Origins of Order: Self-Organization and Selection in Evolution. New York: Oxford University Press. Kearns, M., Suri, S., & Montfort, N. (2006). An experimental study of the coloring problem on human subject networks. Science, 313, 824–827. Kennedy, J., & Eberhart, R. C. (2001). Swarm Intelligence. San Francisco: Morgan Kaufmann/Academic Press. Kim, J. (1984). Supervenience and supervenient causation. Southern Journal of Philosophy Supplement, 22, 45–56. . (1989). The myth of nonreductive physicalism. Proceedings and Addresses of the American Philosophical Association, 63, 31–47. . (1992). Multiple realization and the metaphysics of reduction. Philosophy and Phenomenological Research, 52, 1–26. . (1998). Mind in a Physical World. Cambridge, MA: MIT Press. . (2005). Physicalism, Or Something Near Enough. Princeton, NJ: Princeton University Press. Kincaid, H. (1997). Individualism and the Unity of Science. New York: Rowman & Littlefield. Lang, K., & Lang, G. (1961). Collective Dynamics. New York: Thomas Y. Crowell Co. Larson, J. R., & Christensen, C. (1993). Groups as problem-solving units: Toward a new meaning of social cognition. British Journal of Social Psychology, 32(1), 5–30. Laughlin, P. R. (1999). Collective induction: Twelve postulates. Organizational Behavior and Human Decision Processes, 80, 50–69. Le Bon, G. (1960). The Crowd. A Study of the Popular Mind. New York: Viking Press. Le Goff, J. (1992). History and Memory. New York: Columbia University Press. LePore, E., & Loewer, B. M. (1989). More on making mind matter. Philosophical Topics, 17(1), 175–191. Lewis, D. (1972). Psychophysical and theoretical identifications. Australasian Journal of Philosophy, 50, 249–258. Lewis, K., Belliveau, M., Herndon, B., & Keller, J. (2007). Group cognition, membership change, and performance: Investigating the benefits and detriments of collective knowledge. Organizational Behavior and Human Decision Processes, 103, 159–178. Lewis, K., Lange, D., & Gillis, L. (2005). Transactive memory systems, learning, and learning transfer. Organization Science, 16(6), 581–598. Liang, D. W., Moreland, R. L., & Argote, L. (1995). Group versus individual training and group performance: The mediating role of transactive memory. Personality and Social Psychology Bulletin, 21, 384–393.

114

Georg Theiner and Timothy O’Connor

List, C. (2008). Distributed cognition: A perspective from social choice theory. In M. Albert, D. Schmidtchen, & S. Voigt (Eds.), Scientific Competition: Theory and Policy. Tübingen: Mohr Siebeck. Loewer, B. (2002). Comments on Jaegwon Kim’s ‘Mind and the Physical World’. Philosophy and Phenomenological Research, 65(3), 655–662. Lowe, E. J. (1993). The causal autonomy of the mental. Mind, 102, 629–644. Lycan, W. G. (1987). Consciousness. Cambridge, MA: MIT Press. Macy, M. W., & Willer, R. (2002). From factors to actors: Computational sociology and agent-based modeling. Annual Review of Sociology, 28(1), 143–166. Malcolm, N. (1968). The conceivability of mechanism. The Philosophical Review, 77, 45–72. Marks, M. A., Sabella, M. J., Burke, C. S., & Zaccaro, S. J. (2002). The impact of cross-training on team effectiveness. Journal of Applied Psychology, 87(1), 3–13. Marr, D. (1982). Vision. San Francisco: W. H. Freeman. Maynard Smith, J., & Szathmary, E. (1995). The Major Transitions in Evolution. Oxford: Oxford University Press. McDougall, W. (1920). The Group Mind. New York, London: G. P. Putnam’s Sons. McLaughlin, B. (1989). Type epiphenomenalism, type dualism, and the causal priority of the physical. Philosophical Perspectives, 3, 109–135. Melnyk, A. (2003). A Physicalist Manifesto: Thoroughly Modern Materialism. Cambridge: Cambridge University Press. Menzies, P. (2007). Mental causation on the program model. In G. Brennan, R. Goodin, & M. Smith (Eds.), The Common Mind: Essays in Honour of Philip Pettit (pp. 28–54). Oxford: Oxford University Press. Miller, J. H., & Page, S. E. (2007). Complex Adaptive Systems: An Introduction to Computational Models of Social Life. Princeton, NJ: Princeton University Press. Mohammed, S., & Dumville, B. (2001). Team mental models in a team knowledge framework: Expanding theory and measurement across disciplinary boundaries. Journal of Organizational Behavior 22, 89–106. Moreland, R. L. (1999). Transactive memory: Learning who knows what in work groups and organizations. In L. Thompson, J. Levine, & D. Messick (Eds.), Shared Cognition in Organizations: The Management of Knowledge. Mahwah, NJ: Erlbaum. Nersessian, N. J. (2006). The cognitive-cultural systems of the research laboratory. Organization Studies, 27(1), 125–145. Norman, D. (1991). Cognitive artifacts. In J. M. Carroll (Ed.), Designing Interaction: Psychology at the Human-Computer Interface (pp. 17–38). Cambridge: Cambridge University Press. O’Connor, T., & Churchill, J. (Forthcoming). Is nonreductive physicalism viable within a causal powers metaphysic? In G. Macdonald & C. Macdonald (Eds.), Emergence in Mind. Oxford: Oxford University Press. O’Connor, T., & Wong, H. Y. (2006). Emergent properties. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. Retrieved from http://plato.stanford. edu/entries/properties-emergent/ . (2005). The metaphysics of emergence. Noûs, 39, 659–679. Park, R. E. (1927). Human nature and collective behavior. American Journal of Sociology, 32, 733–741. Park, R. E., & Burgess, E. W. (1921). Introduction to the science of sociology. Chicago: University of Chicago Press. Passino, K. M., Seeley, T. D., & Visscher, P. K. (2008). Swarm cognition in honey bees. Behavioral Ecology and Sociobiology, 62, 401–414.

The Emergence of Group Cognition 115 Pereboom, D. (2002). Robust nonreductive materialism. Journal of Philosophy, 99, 499–531. Pettit, P. (1993). The Common Mind: An Essay on Psychology, Society, and Politics. New York: Oxford University Press. . (2003). Groups with minds of their own. In F. Schmitt (Ed.), Socializing Metaphysics (pp. 167–194). Oxford: Rowan and Littlefield. Poirier, P., & Chicoisne, G. (2006). A framework for thinking about distributed cognition. Pragmatics & Cognition, 14(2), 215–234. Polger, T. W. (2004). Natural Minds. Cambridge, MA: MIT Press. Popper, K. (1957). The Poverty of Historicism. London: Routledge & Kegan Paul. Propp, K. (1999). Collective information processing in groups. In L. Frey, D. S. Gouran, & M. S. Poole (Eds.), Handbook of Group Communication Theory and Research (pp. 225–250). Thousand Oaks, CA: Sage. Resnick, M. (1994). Turtles, Termites, and Traffic Jams: Explorations in Massively Parallel Microworlds. Cambridge, MA: MIT Press. Richerson, P. J., & Boyd, R. (2005). Not By Genes Alone: How Culture Transformed Human Evolution. Chicago: University of Chicago Press. Robbins, P., & Aydede, M. (Eds.). (1009). Cambridge Handbook of Situated Cognition. Cambridge, MA: Cambridge University Press. Ruben, D.-H. (1985). The Metaphysics of the Social World. London: Routledge & Kegan Paul. Rueger, A. (2000). Physical emergence, diachronic and synchronic. Synthese, 124(3), 297–322. Rupert, R. (2009). Cognitive Systems and the Extended Mind. New York: Oxford University Press. Sandelands, L. E., & Stablein, R. E. (1987). The concept of organization mind. In S. Bacharach & N. DiTomasco (Eds.), Research in the Sociology of Organizations (pp. 135–161). Greenwich, CT: JAI Press. Satz, D. M., & Ferejohn, J. (1994). Rational choice and social theory. Journal of Philosophy, 91(2), 71–87. Sawyer, R. K. (2002). Nonreductive individualism. Part I—Supervenience and wild disjunction. Philosophy of the Social Sciences, 32, 537–559. . (2003a). Nonreductive individualism. Part II—Social causation. Philosophy of the Social Sciences, 33, 203–224. . (2003b). Group Creativity: Music, Theater, Collaboration. Mahwah, NJ: Erlbaum. . (2003c). Improvised Dialogues: Emergence and Creativity in Conversation. Westport, CT: Greenwood. . (2004). The Mechanisms of Emergence. Philosophy of the Social Sciences, 34, 352–370. . (2005). Social Emergence: Societies as Complex Systems. Cambridge, MA: Cambridge University Press. Sawyer, R. K., & Greeno, J. (2009). Situativity and learning. In P. Robbins & M. Aydede (Eds.), The Cambridge Handbook of Situated Cognition (pp. 347–367). New York: Cambridge University Press. Schmitt, F. (2003a). Joint action: From individualism to supraindividualism. In F. Schmitt (Ed.), Socializing Metaphysics (pp. 129–166). Oxford: Rowan and Littlefield. . (2003b). Socializing metaphysics: An introduction. In F. Schmitt (Ed.), Socializing Metaphysics (pp. 1–37). Oxford: Rowan and Littlefield. Searle, J. (1980). Minds, brains, and programs. Behavioral and Brain Sciences, 3(3), 417–457.

116

Georg Theiner and Timothy O’Connor

. (1990). Collective intentions and actions. In P. R. Cohen, J. L. Morgan, & M. E. Pollack (Eds.), Intentions in Communication. Cambridge, MA: MIT Press. . (1992). The Rediscovery of the Mind. Cambridge, MA: MIT Press. . (1995). The Construction of Social Reality. New York: Free Press. Seeley, T. D. (1995). The Wisdom of the Hive: The Social Physiology of Honey Bee Colonies. Cambridge, MA: Harvard University Press. Selous, E. (1931). Thought-Transference (Or What?) in Birds. London: Constable & Co. Shapiro, L. (2000). Multiple realizations. Journal of Philosophy, 97, 635–654. Shoemaker, S. (2001). Realization and mental causation. In C. Gillett & B. Loewer (Eds.), Physicalism and Its Discontents. Cambridge, MA: Cambridge University Press. . (2007). Physical Realization. Oxford: Oxford University Press. Simon, H. (1969). The Sciences of the Artificial. Cambridge, MA: MIT Press. Smith, A. (1776). An Inquiry into the Nature and Causes of the Wealth of Nations. London: Methuen and Co. Sober, E. (1999). The multiple realizability argument against reductionism. Philosophy of Science, 66, 542–564. Sober, E., & Wilson, D. S. (1998). Unto Others: The Evolution and Psychology of Unselfi sh Behavior. Cambridge, MA: Harvard University Press. Stasser, G. (1999). The uncertain role of unshared information in collective choice. In L. Thompson, J. Levine, & D. Messick (Eds.), Shared Knowledge in Organizations. Hillsdale, NJ: Erlbaum. Stephan, A. (2006). The dual role of ‘emergence’ in the philosophy of mind and in cognitive science. Synthese, 151(3), 485–498. Stevens, C., Malloch, S., McKechnie, S., & Steven, N. (2003). Choreographic cognition: The time-course and phenomenology of creating a dance. Pragmatics & Cognition, 11(2), 297–326. Stevens, K., McKechnie, S., Malloch, S., & Petocz, A. (2000). Choreographic cognition: Composing time and space. In C. Woods, G. Luck, R. Brochard, F. Seddon, J. A. Sloboda, & S. O’Neill (Eds.), Proceedings of the 6th International Conference on Music Perception & Cognition. Keele, UK: Department of Psychology, Keele University, August 2000. Sutton, J. (2006). Distributed cognition: Domains and dimensions. Pragmatics & Cognition, 14(2), 235–247. Thagard, P. (1996). Mind: Introduction to Cognitive Science. Cambridge, MA: MIT Press. Theiner, G. (2008). From Extended Minds to Group Minds: Rethinking the Boundaries of the Mental (Ph.D. Thesis). Indiana University, Bloomington. Tollefsen, D. (2004). Collective epistemic agency. Southwest Philosophy Review, 20(1), 55–66. . (2006). From extended mind to collective mind. Cognitive Systems Research 7(2–3), 140–150. Tye, M. (1995). Ten Problems of Consciousness: A Representational Theory of the Phenomenal Mind. Cambridge, MA: MIT Press. Van Gulick, R. (2001). Reduction, emergence and other recent options on the mind/body problem: A philosophic overview. Journal of Consciousness Studies, 8(9–10), 1–34. Walsh, J. P., & Ungson, G. R. (1991). Organizational memory. Academy of Management Review, 16, 57–91. Watkins, J. W. (1957). Historical explanation in the social sciences. British Journal for the Philosophy of Science, 8, 104–117. Watts, D. J., & Strogatz, S. H. (1998). Collective dynamics of “small-world” networks. Nature, 393, 440–442.

The Emergence of Group Cognition 117 Wegner, D. M. (1986). Transactive memory: A contemporary analysis of the group mind. In B. Mullen & G. R. Goethals (Eds.), Theories of Group Behavior (pp. 185–208). New York: Springer. . (1995). A computer network model of human transactive memory. Social Cognition, 13, 319–339. Wegner, D. M., Erber, R., & Raymond, P. (1991). Transactive memory in close relationships. Journal of Personality and Social Psychology, 61, 923–929. Wegner, D. M., Giuliano, T., & Hertel, P. (1985). Cognitive interdependence in close relationships. In W. J. Ickes (Ed.), Compatible and Incompatible Relationships (pp. 253–276). New York: Springer. Wheeler, G. (1920). The termitodoxa, or biology and society. Scientific Monthly, 10, 113–124. Wheeler, W. M., (1939). Essays in Philosophical Biology. (Selected by G. H. Parker) Cambridge, MA: Harvard University Press. Wilson, D. S. (1997). Incorporating group selection into the adaptationist program: A case study involving human decision making. In J. A. Simpson & D. T. Kenrick (Eds.), Evolutionary Social Psychology. Mahwah, NJ: Erlbaum. . (2002). Darwin’s Cathedral: Evolution, Religion, and the Nature of Society. Chicago: University of Chicago Press. Wilson, D. S., van Vugt, M., & O’Gorman, R. (2008). Multilevel selection theory and major evolutionary transitions: Implications for Psychological Science. Current Directions in Psychological Science, 17, 6–9. Wilson, D. S., & Wilson, E. O. (2007). Rethinking the theoretical foundation of sociobiology. Quarterly Review of Biology, 82, 327–348. Wilson, R. (2001a). Group-level cognition. Philosophy of Science, 68(Supplement), 262–273. . (2001b). Two views of realization. Philosophical Studies, 104(1), 1–31. . (2004). Boundaries of the Mind: The Individual in the Fragile Sciences. Cambridge: Cambridge University Press. . (2005). Collective memory, group minds, and the extended mind thesis. Cognitive Processes, 6(4), 227–236. . (In press). Extended vision. In N. Gangopadhyay, M. Madary & F. Spicer (Eds.), Perception, Action and Consciousness. New York: Oxford University Press. Wilson, R., & Clark, A. (2009). How to situate cognition: Letting nature take its course. In M. Aydede & P. Robbins (Eds.), The Cambridge Handbook of Situated Cognition (pp. 55–77). Cambridge, MA: Cambridge University Press. Wimsatt, W. C. (1974). Complexity and Organization. In K. Schaffner & R. Cohen (Eds.), Proceedings of the 1972 Biennial Meeting of the Philosophy of Science Association, (pp. 67–86). . (1981). Robustness, reliability, and overdetermination. In M. Brewer & B. Collins (Eds.), Scientific Inquiry and the Social Sciences (pp. 124–163). San Francisco: Jossey-Bass. . (1986). Forms of aggregativity. In M. G. Grene, A. Donagan, A. N. Perovich, & M. V. Wedin (Eds.), Human Nature and Natural Knowledge (pp. 259–291). Dordrecht: Reidel. . (1994). The ontology of complex systems: Levels, perspectives and causal thickets. In M. Matthen & R. X. Ware (Eds.), Biology and Society: Refl ections on Methodology (pp. 207–274). . (1997). Aggregativity: Reductive heuristics for fi nding emergence. Proceedings of the 1996 Biennial Meetings of the Philosophy of Science Association, Part 2, S372–S384. . (2007). Re-Engineering Philosophy for Limited Beings: Piecewise Approximations to Reality. Cambridge, MA: Harvard University Press.

Part II

Self, Agency, and Free Will

Part II Introduction Antonella Corradini and Timothy O’Connor

One of the fields in which the concept of emergence plays a particularly important role is the philosophy of mind. It is not difficult to see why this is so: to understand what the mind is requires addressing the problem of the relationship between mind and body and, in particular, between mind and brain. What sort of relationship holds among the neurophysiological processes which take place in the brain and certain mental states occurring in the agent’s mind? What process in the brain corresponds to the performance of a certain higher-lever mental activity? When a decision or choice occurs where should we locate it? Is it to be located in the brain or, instead, in a non-physical mind or field of conscious awareness which the brain sustains? Moreover, what might it mean, exactly, to say that the brain ‘supports’ the choice? As is well known, there are many different approaches to these issues, and they stand in contrast to each other mainly in how the difference between brain states and mental states is conceived. The options range from the most extreme naturalistic positions, according to which a mental state is reducible to a cerebral neurophysiological state, to dualistic positions which view mental states as irreducible to physical states due to the fact that they are states of ontologically independent entities. However, between naturalistic perspectives inclining towards monism and the strongest dualistic positions there is still room for intermediate points of view. Many of these fi nd in emergence a conceptual category which can be usefully employed to avoid both naturalistic reductionism and forms of dualism which posit subjects of experience wholly distinct from physical reality. If mind is understood as an entity or set of capacities which emerges from the brain, then it will be both ontologically and explanatorily irreducible to the cerebral microstructure from which it emerges, and it will be endowed with genuine and autonomous causal powers. On the emergentistic construal, however, the mind has not only emerged from the brain, but it is continuously sustained by it and bound up with it in a single causal system. Therefore, emergence is a natural candidate to represent a point of equilibrium between the implausibly extreme naturalistic or dualistic positions. Present-day discussion in the philosophy of mind, however, shows how complex the emergentist program has become and how many different

122 Antonella Corradini and Timothy O’Connor kinds of emergentism are championed by contemporary philosophers of mind. The chapters presented in Part II aim at clarifying the role that specific variants of emergentism might play in issues such as how to ground the independent ontological status of the human self and how to safeguard some of its essential dimensions such as agency and free will. E. Jonathan Lowe’s contribution addresses the most important question of the metaphysics of mind: “What am I?” and answers this question by affi rming that we are selves or persons. However, this answer will not be satisfactory unless we are able to convincingly argue in favor of the thesis that to be a person is a fundamental ontological category, not identifiable with and not reducible to other categories, such as the category of being an animate or inanimate bodily substance or of being an immaterial substance like the soul. Now, according to Lowe, thoughts cannot exist without their subjects. This is the starting point of the “unity argument” that the author puts forward in order to show the need for an irreducible category of being a self. The central premise of the argument consists in maintaining that all and only my thoughts have just one thing, me, as their unique subject: thoughts can be individuated only by their subject and can exist only if so individuated. But, if this is true, does this individuating function of the self hold for all the other potential candidates for being identical with me? Lowe shows that this is not true in particular for the brain, insofar as the relationship that thoughts have to their subject is different than that they have to the brain. Our thoughts do not depend from the brain in a holistic way. Different thoughts often depend from different parts of the brain. But, while my thoughts depend in a distributive way on different parts of my brain, they depend in a collective way on the subject. This implies that brain B and a proper part of it, E, are equally valid candidates to be identical with me. But the both of them cannot be identical with me, because, then, thoughts would not have a unique subject. The consequence is that I am not a body and not even a brain. The argument, therefore, results in a dualism of self and body. However, Lowe stresses that his position is not equivalent to a Cartesian dualism, but it is close to an emergentistic dualism. The self represents something new and irreducible relative to the body and all its parts, but it is not fully independent from it. Martine Nida-Rümelin begins by noting that, on any view, conscious experience must have occurred in the natural world for the fi rst time long after the world’s beginning, deep into its cosmic evolution. And she takes it that it is very plausible that this is a puzzling—deeply puzzling—phenomenon. The burden of her paper is to clarify the nature of the state of puzzlement that the emergence of consciousness properly engenders and just what it is about conscious experience that is most deeply puzzling. She suggests that we should distinguish the emotion of puzzlement or amazement occasioned by reflection on the emergence of consciousness in a physical world from fascination. This is important because these distinct sorts of reaction will prompt investigations of consciousness under distinct constraints; in

Part II Introduction

123

the former case of amazement, one will not properly insist from the outset that an adequate account of consciousness must be consistent with physicalism. She then presents an argument that the proper object of amazement does not concern the qualities of conscious experience (“qualia”), as the standard view has it. Instead, she proposes, what is most fundamentally amazing is that there should have come to be, at a specific moment in the evolution of the physical world, a subject of conscious experience. In his chapter Uwe Meixner argues in favor of the existence of rational souls, and for the naturalness of their existence, as new ontological kinds. He moves from the phenomenological evidence that the subject of the totality of the consciousness states of a certain animal at a given moment cannot be the animal itself or its brain. The subject of an animal’s consciousness cannot be identified with a physical individual because this subject, in contrast with any physical individual, is at the origin of a point of view from which she sees the world, but which does not belong to the world. To put it in other words, consciousness always involves the indexicality of its subject. In a second step, Meixner aims to show that the subject of a certain animal’s consciousness is also its soul. To this end, the author turns again to the phenomenological self-evidence, in order to maintain that the self, besides being a non-physical individual, is also without inherent temporal dimension. Meixner further argues in favor of the active nature and the nonphysical causal capabilities of the self, which he traces back to the existence and plausibility of a non-physical form of agent causation. Finally, on the basis of these argumentative steps, and of a precise defi nition of the concept of a soul, he comes to the conclusion that the subject of a certain animal’s consciousness is the soul of that animal. In the fi nal part of his chapter, moreover, Meixner tries to show that the attribution of causal capabilities to the soul allows us to account for the conscious interests of many animals: it is a requisite of minimal rationality that the animals who have these interests can pursue them without being “crushed” by the determinism of the physical world. The argument is completed by the consideration that the non-physical causality of the agent yields an evolutionary advantage for the organism which possesses it. Thus, the emergence of rational souls, that is to say, of non-physical active rational substances who are the subjects of the animals’ consciousness, turns out to be a natural expectation. The main purpose of Achim Stephan’s chapter is to analyze the relationship between the most representative positions in the free will debate and the correlated perspectives in the mind-body discussion. To illustrate this issue, the author takes as examples three different points of view on free will held by three contemporary German philosophers, Ansgar Beckermann, Wolf Singer, and Geert Keil. While the fi rst is a compatibilist, both Singer and Keil are incompatibilists, who, however, support two opposed viewpoints. In Singer’s view, incompatibilism combined with the presumed truth of a deterministic view of the brain implies the rebuttal of free will. For Keil, on the contrary, incompatibilism combined with an indeterministic view of the

124

Antonella Corradini and Timothy O’Connor

brain is what we need to account for free will. An interesting aspect of the discussion is that all three philosophers declare themselves to be physicalists. This leads us to the central point of Stephan’s discussion, which is to say whether physicalism should take an emergentist (i.e., a non-reductivist) shape in order to allow for free will. As physicalists, all three authors agree on the thesis that the micro-level synchronically determines the macrolevel. Singer, as a hard determinist who denies the reality of free will and mental causation, sees no difficulty in deriving from this that all seemingly free decisions and actions are reducible to their neurophysiological components. A compatibilist like Beckermann sticks to a reductive explanation, whose main function is to give an account of how the processes underlying our free actions are sensitive to arguments. This is in fact one of the fundamental conditions for a decision or action to be considered as free. It seems therefore that only a libertarian perspective can be matched with a non-reductive explanation of the micro-macro relationship, at least in its diachronic version. Indeed, within the frame of a libertarian concept of free will, the property of being a free decision or action is an emergent property, not reducible to its micro-elements, since it is a holistic characteristic of the whole system. By arguing in this way, have we perhaps reached any result with respect to the free will debate? Stephan argues that we have not. The permanent changes the brain undergoes make it impossible to perform an experiment to decide which of the three positions discussed is the right one. Rather, the decision will depend on how much weight we are ready to give either to the manifest image or to the scientific image of the world. But this does not amount to providing a theoretical solution to the problem of free will, which, therefore, remains unsolved. In the fi nal chapter of Part II, Mario De Caro defends a liberal naturalistic approach to the problem of free will, understood as an emergent capacity of the human agent. According to emergentism the self possesses genuine causal powers which are ontologically irreducible to the less-complex properties of the lower, neurophysiological level, on which, however, these powers depend for their occurrence. This means that the self manifests new properties relative to those of the lower levels, which are, however, rooted in these lower-level properties. Once he has observed that emergentism must fight on many battle fronts (e.g., against orthodox scientific naturalism, supernaturalism, and non-emergentist /non-reductive naturalism), De Caro decides to concentrate on what he deems to be the most serious challenge at the moment, eliminativism as allegedly supported by neuroscientific evidence. The author takes as an example a neuroscientific experiment recently carried out by Soon in order to show that the neurophysiological processes underlying decisions occur much earlier than the moment in which the agent becomes aware of taking her decision. Although this experiment improves upon Libet’s original experiment in a number of ways, it can be criticized from many points of view, both methodological and philosophical. As regards the latter, the authors of the experiment seem

Part II Introduction

125

to presuppose that free will is an illusion. According to De Caro, however, in no way do they succeed in disproving our commonsensical belief that we are free. Among the many critical remarks put forward by De Caro, one in particular should be mentioned here: the actions which are deemed to be free must not necessarily be conscious. There can be non-conscious free actions, as in the cases where we do not attach particular importance to an action or we perform it repeatedly. For this reason Soon’s experiment is relevant not in order to exclude the possibility that certain actions be free, but only to show that decisional processes are subjected to limitations as regards their being conscious processes. The author generalizes to the conclusion that neuroscience at present does not provide any significant reason to prefer reductionism or eliminativism to emergentism.

6

Why My Body is Not Me The Unity Argument for Emergentist Self-Body Dualism E. Jonathan Lowe

Philosophers have long pondered the nature of the self, their judgments concerning its existence and identity ranging from outright nihilism—its dismissal as a sheer illusion—to affi rmations of its absolute centrality to the whole of reality and our knowledge of it. This enormous variety of opinions makes one wonder if all these philosophers can really have meant the same in speaking of ‘selves’—a term which is, in any case, hardly in widespread everyday use. And yet all such talk is centered ultimately upon a linguistic phenomenon that is ubiquitous: the first-person pronoun, ‘I’—a term which always seems to have a perfectly defi nite and indisputably real reference whenever it is employed, referring as it does to whomsoever or whatsoever is uttering it on any occasion of its use. But for all its ubiquity and seeming familiarity, this pronoun occasions some of the most puzzling problems in metaphysics and philosophical psychology that we ever encounter. The most challenging of these problems is the topic of the present chapter.

1. THE QUESTION: ‘WHAT AM I?’—AND WHY IT IS PUZZLING Very arguably the most important question in the metaphysics of mind, which needs to be answered before all others, is the question that Descartes raises early in the Meditations: ‘What am I?’—or, in its plural form, ‘What are we?’1 The plural form of the question presents some difficulties, however, which don’t seem to attach to the singular form: in particular, the difficulty of deciding what or whom to include under the term ‘we’. It seems that I need to decide in advance what kind of thing I fundamentally am in order to decide which is the relevant plurality to regard as the referent of ‘we’ in the plural form of the question—for whoever says ‘we’ intends to include him- or herself within that plurality, together with certain others relevantly like himor herself. It is clear, anyway, that Descartes’s question is just the question of what kind of thing I fundamentally am, where ‘kind’ here, I suggest, should be understood to mean ontological category. This latter notion needs to be spelled out, of course, but it should be clear that we are not talking now about merely superficial classifications—so that ‘philosopher’, for example, would not be a good answer to Descartes’s question, whereas ‘animal’ might be, as might also, perhaps, be ‘material object’.2

128 E. Jonathan Lowe Descartes’ question can seem an odd one, for this reason: he raises it only after having satisfied himself that he certainly exists. Thus, he represents himself as knowing with certainty that he is, without yet knowing what he is. In short, he claims to know of his own existence before knowing, or at least knowing clearly, his essence—for to know a thing’s essence is precisely to know what it is, in Descartes’s sense of this phrase, which goes back ultimately to Aristotle’s. The oddity lies in the fact that one would normally suppose that one cannot know that a certain thing, X, exists unless one has at least some grasp of what it is that one takes X to be. If I were just to assert ‘Gronash exists’, but professed to have no idea whatever what Gronash is, I would rightly be accused of complete vacuity. My assertion would amount to nothing more than the claim that something, I know not what, exists. Perhaps I do indeed know this, because I know that something rather than nothing exists. But then I shouldn’t try to make this claim by saying ‘Gronash exists’, because that misleadingly suggests that I know, regarding some specific thing, that it exists.

2. THE BEGINNINGS OF AN ANSWER IN THE CONCEPT OF THE SELF Part of the solution to this conundrum is that we don’t know absolutely nothing about the referent of ‘I’, prior to asking the question ‘What am I?’ I know, for example, that I am the thing asking this question about myself, and hence that I am the sort of thing that can ask intelligible questions—so I know, as Descartes points out, that I am at least a thing that thinks, and indeed that thinks about itself. In short, I am at least a self-reflecting subject of thought—a ‘self’ or ‘person’, to use those technical philosophical terms. A self or person—at least as I propose to use those terms—is just, by definition, a self-refl ective subject of thought, to which I would add that it is thereby also a subject of perception and an agent of actions, on the grounds that only a perceiving agent can be capable of thought. Demonstrating this latter point is no simple matter, however, so I shall just leave the claim as an undefended assertion for present purposes.3 But then why isn’t it sufficient to answer Descartes’s question—‘What am I?’—with the perfectly simple response, ‘I am a self or person’? Well, that would be an adequate answer, I think, if we could establish that ‘person’ denotes a fundamental ontological category, comparable with ‘animal’ or ‘material object’. But the trouble is that this is not just obviously so. And Descartes himself rightly acknowledges this. It may seem, rather, that ‘self’ or ‘person’ is just a status that can be conferred upon something of some fundamental kind if it satisfies certain conditions—or, to use terminology that is currently popular in this sort of context, if it occupies a certain ‘functional role’—specifically, if it does everything that is necessary to qualify as a self-reflective subject of thought. Then our attention will

Why My Body is Not Me

129

be directed to the question of what kind or kinds of things can satisfy this condition. Can an animal, for instance, or a material object? (Of course, it may be said that animals are themselves just a kind of material object—in which case, consider instead the category inanimate material object, which presumably includes such things as mechanical or electronic robots.) If it should turn out that things of fundamentally different kinds can qualify as selves or persons, then this will imply that ‘self’ does not denote a fundamental kind and that my answer to Descartes’s question need not be the same as that of every other self or person. By the same token, it will then turn out that by ‘we’ I should mean things of the same fundamental kind as me, not necessarily things of other kinds which also qualify as selves or persons. That need not have the alarming implication that I should have no moral concern for such ‘alien’ persons, because it may be that it is still personhood that qualifies something for warranting moral concern, not what kind of thing it is.

3. HOW CAN WE DETERMINE THE ONTOLOGICAL STATUS OF THE SELF? So, how do we decide whether or not ‘self’ or ‘person’ denotes a fundamental ontological category? Well, one obvious way of proceeding is to see whether or not we can rule out all alternative possibilities. So let us begin by examining whether we can at all plausibly regard the term ‘self’ as picking out a certain ‘functional role’ which things of various different kinds could potentially occupy—rather as very differently shaped and composed material objects can occupy the functional role of the bishop in a game of chess. Now, we can see well enough how an arbitrarily shaped piece of wood or metal could occupy the bishop’s role in a chess-game, since its occupying that role is simply a matter of its being moved around the chessboard in accordance with the rules governing the movement of a bishop. So the idea would be that if something, of whatever fundamental kind, could be a subject of thoughts about itself, then it could thereby qualify as being a self or person, in virtue of occupying the ‘self’ role. If an animal—a living organism—could do that, then it could be a person, and if an animal’s brain could do that, then it could be, and so forth. But, it may be asked, what about my own actual case? Our question now seems to reduce to something like this: what is it that does my thinking? Whatever it is, that thing is me, and what I am is the kind of thing that it is. Thus, if it is a living organism that does my thinking, then I am that organism and what I am is a living organism. Similarly, if it is a brain that does my thinking, then I am that brain and what I am is a brain. Similarly again, if it is an immaterial spirit that does my thinking, then I am that spirit and what I am is a spirit. And if any such answer is correct, then it will follow that ‘self’ or ‘person’ certainly does not denote a fundamental ontological category.

130

E. Jonathan Lowe

4. AN UNACCEPTABLE ANSWER TO THE QUESTION: NIHILISM What, however, if we fi nd that no such answer is forthcoming? Suppose I fi nd that, plausibly, there is no entity that belongs to an independently acknowledged ontological category that can plausibly be identified with me. One imaginable response would be the nihilist one of deciding that, after all, I don’t exist—there is no such thing as ‘me’. But any such answer teeters on the edge of at least pragmatic inconsistency and seems at best a desperate last resort. Descartes’s confidence in his own existence does not seem to be misplaced. Part of the reason for this is that it seems incoherent to suppose that there could be thoughts without a subject of those thoughts, so that to deny one’s existence is to deny also the existence of one’s thoughts—and yet the very raising of the question ‘What am I?’ seems to presuppose the reality of my thoughts, since it actually expresses one such thought. But, it may be asked, why must thoughts have a subject? Ultimately, I think they must because thoughts can only be individuated by their subjects—and unless they can be individuated, little sense can be attached to the idea that numerically distinct thoughts really exist at all. By ‘individuated’ here I don’t just mean, of course, singled out in thought, but singled out in reality.4 Something in reality must make one individual thought distinct from another—and the mere content of a thought, even taken together with its time and place of occurrence (if indeed thoughts can be said to be spatially located), does not do this. Distinct thoughts, with the same contents, occurring at the same space-time locations, can very plausibly exist, at least in principle. No one disputes, I suppose, that distinct thoughts with the same contents can occur at the same time, since you and I might now be having thoughts with the same content. Now, if one could uncontentiously identify a thought with, say, a neural event, and argue that no two neural events realizing the same thought-content could occur in the same place and the same time, then one could argue that thoughts are individuated by their contents and space-time locations. But that is a very big ‘if’. Quite apart from the difficulty of establishing that a thought is identifiable with a neural event, it simply doesn’t appear to be the case that events of the same kind can’t occur in the same place and the same time. Events aren’t like material objects in this respect—they don’t exclude others of their kind from their space-time locations. They aren’t ‘impenetrable’. For instance, three armies could be fighting three different battles in the same place at the same time, each participating pairwise in two battles with each of the other two, A with B, B with C, and C with A. This is quite different, for example, from the case in which A and B are fighting together in an alliance against C. Here we see that the battles occurring in a given space-time location can only be individuated in terms of the armies that are fighting them. The same

Why My Body is Not Me

131

applies, mutatis mutandis, to all events and hence to thoughts in particular. They are individuated in terms of their subjects—the things having those thoughts. Two different subjects can’t have numerically the same thought. So we can’t deny the existence of subjects of thought—things occupying the role of ‘self’—without denying the distinctness and thus the very existence of individual thoughts.

5. NARROWING DOWN THE POSSIBLE CANDIDATES So nihilism with regard to the self, and myself in particular, seems not to be a coherent option (though this is an issue that I shall return to later, when I discuss the idea that the self is a ‘fiction’ created by the brain). I am something—and so a thing of some kind. But suppose, again, that I can discover no independently acknowledged kind of thing that is capable of occupying the ‘self’ role in my case. Suppose, for example, I can rule out the possibility that my organism, or my brain, or my putative immaterial spirit, and so on could be the thing that does my thinking: what then should I conclude concerning what I am? I think I would have to conclude that I am just fundamentally a self or person: that these terms must after all denote a fundamental ontological category, to which I belong. Now, obviously, to argue conclusively in this fashion I would have fi rst to identify every possible independently acknowledgeable ontological category and rule it out as one to which I could belong—and this is, to say the least, a very tall order. First of all, I don’t know what the complete inventory of such categories is, since there may be kinds of thing that have never yet been dreamt of in my philosophy, or anyone else’s. So I am going to try to make things easier for myself by assuming that the most likely candidates for identity with me, amongst those things that we independently acknowledge to exist, are all going to be bodily entities of some kind. For—setting aside external world skepticism and idealism—we are pretty confident that bodily entities do exist and indeed that we have bodies. We also take ourselves—or, at least, scientists—to know quite a lot about these bodies and how they function. We know, for instance, that our bodies are living organisms composed of organic matter, organized in various ways into various parts and systems, including organs (such as the liver and brain) and cells (such as blood cells and nerve cells). What I want to rule out, then, is my identity with anything of any of these kinds—organisms, their organs, their cells, or various assemblies of such cells into systems of various kinds, such as central nervous systems. In taking this approach, I admit that I am being somewhat unfair to those who believe that we are, or are at least partly composed of, immaterial souls or spirits. But the fact is that I don’t have a clear enough grasp of what such an entity is supposed to be in order to get to grips adequately with this view. Furthermore, it is certainly not uncontentious that such

132

E. Jonathan Lowe

things exist. If we did know clearly what they are and were reasonably confident that they do exist, then, certainly, it would be incumbent on me to take seriously the possibility that I am an immaterial soul or spirit. But as matters presently stand, I don’t think it is.

6. SOME ARGUMENTS THAT WILL NOT SERVE OUR PURPOSE So I take it that the challenge before me is to show that I am not identifiable with anything that we may reasonably categorize as bodily in nature. Now, prior to trying to rise to this challenge, I want to rule out certain other attempts to establish this conclusion. First of all, there are Descartes’s own arguments to this effect, of which there are basically two: his conceivability argument and his divisibility argument. The conceivability argument proceeds from the premise that my existence without a body of any kind is at least conceivable and so (he thinks) possible, to the conclusion that I am not identifiable with anything bodily. This falls foul of the objection that what is conceivable is not necessarily really possible—unless one simply uses ‘conceivable’ as a synonym for ‘possible’, in which case the argument is question-begging. The divisibility argument proceeds from the premise that all bodily things are divisible but I am not. However, the claim that I am not divisible again seems question-begging in this context and, in any case, needs to be supported by some sort of argument. For my own part, I am in fact strongly inclined to think that I am indivisible, but I don’t think that this claim can be supported without an independent argument that I am not identifiable with anything bodily.5 One other argument that I shall not appeal to here is one that I have elsewhere defended and called the replacement argument.6 This argument proceeds from the premise that I could survive even if all the parts of my existing, biological body were to be replaced—gradually, of course—by nonbiological substitutes, provided that this could be done without destroying my capacity for self-reflective thought. However, the trouble with this argument is, first, that it makes an empirical claim for which, at present, we have no firm corroborative evidence, however much it might seem to be borne out by examples from science fiction literature. Even more fundamentally, though, the problem with this argument is that it can’t rule out the possibility that I am something ‘bodily’ in a more extended sense of the term. It would show at best that I could not be identified with a particular living organism or living organic part of one, such as my current living brain—where by ‘living’ I mean ‘alive’ in the biological sense of the term. But it wouldn’t rule out the possibility that I am some kind of bodily system, such as my central nervous system, since it may be argued that the identity of such a system is not necessarily tied to its having a specifically biological constitution. Indeed, there is even a sense of ‘brain’, perhaps, in which we could talk of the same

Why My Body is Not Me

133

‘brain’ surviving a change from being composed of organic cells to being composed of silicon-based electronic circuits, provided that it functioned in the same way (although I suspect that this is, really, to think of the brain as being a kind of system rather than a kind of object). So, while I think that the replacement argument is not entirely without merits, it does not really serve the purpose I now have of showing that I am not to be identified with anything that can deservedly be said to be ‘bodily’ (excluding, of course, that uncontentious sense of ‘bodily’ in which I am a ‘bodily’ thing in virtue of having a body, or being ‘embodied’).

7. THE UNITY ARGUMENT: PRELIMINARY CONSIDERATIONS The argument that I think will serve this purpose better than any other is what I call the unity argument, so let us now proceed to examine it.7 The key premise of the unity argument appeals to a thesis that I defended a little while ago. This is that each of my thoughts has just one subject, and all of them have the same subject—that subject being, of course, me, whatever I am. Moreover, no thought that is not a thought of mine can have this same subject—it must be a thought of a numerically distinct subject, someone else. In sum all and only my thoughts have just one thing, me, as their unique subject. Furthermore, in having me as their subject, all of these thoughts depend upon me for their very existence—they do so because they depend upon me for their identity, and anything’s identity is obviously essential to it. In saying that each of my thoughts depends upon me for its identity, I am simply appealing once more to the fact that thoughts are individuated, at least in part, by their subjects—which thought a particular thought is being determined, at least in part, by whose thought it is. Of course, a thought’s subject doesn’t completely determine the identity of the thought, since the same subject may have many different thoughts, both over time and at the same time. Plausibly, however, a thought’s identity is completely determined by its subject, time, and content. Its place doesn’t seem to play a significant role in this respect, if indeed thoughts can really be said to have spatial locations at all. I am inclined to say that a thought’s place, if any, is just to be identified with the place of its subject. Thus, my present thoughts are located, if anywhere, wherever it is that I am presently located. No fi ner-grained spatial location can be assigned to a thought than this. Some will perhaps object that my thoughts must at least be located in my brain— but why, unless indeed I am my brain? If I am not my brain, then my brain is not the subject of my thoughts—and, while it may nonetheless be agreed that I need a brain in order to have thoughts, that doesn’t imply that my thoughts are in my brain, any more than the fact that I need legs in order to run implies that my running is in my legs.

134

E. Jonathan Lowe

Now, having agreed, for the foregoing reasons, that all and only my thoughts depend upon me for their very existence—that is, that without my existence they wouldn’t exist—we can proceed to ask whether bodily things of various specifiable kinds are such that all and only my thoughts couldn’t exist without a unique thing of one of those kinds existing. Let X be a candidate thing of one of these kinds. If we can establish that it is not the case that all and only my thoughts depend for their existence on X, then we have ruled out the possibility that I am X. So, what are the candidate things of these putative kinds? What kinds of ‘bodily’ thing might we hope to identify a person or self—and, more particularly, me—with? One obvious possibility is whole living organisms, that is, individual animals. (I assume that plants may be ruled out, and, in any case, I surely couldn’t be identified with any plant, since there seems to be none in my vicinity that is even a possible candidate.) Another possibility is some kind of bodily organ of an animal, such as, most obviously, animal brains—hence, in my case, a certain human brain. Another possibility is some distinguished part of such an organ, such as a human cerebral cortex. Yet another is some bodily system of an animal, such as a human central nervous system. However, we have to be careful about how, exactly, we understand talk of a ‘system’ in this sense. It would seem that a subject of thoughts must be a thing of some kind, by which I mean a concrete object of some kind—for it has to be something that has properties and exists in time (and probably in space too) and that can participate in events. It cannot, therefore, be a purely abstract entity. Often, when we talk of ‘systems’, we do mean to talk about purely abstract entities—for instance, the Dewey decimal system of library cataloguing. An example of a concrete system, by contrast, would be our solar system, consisting of the sun and the various planetary and other bodies orbiting it. A concrete system in this sense must certainly consist of concrete objects or things, even if it is not just the sum of those things, since it is necessary also for the existence of the system that the things in question should be related or interact in certain characteristic and relatively stable ways. If all the planets of the solar system were to be attracted away from the sun by other stars, then our solar system would no longer exist. It is an open question, perhaps, whether we should really call such a concrete system a concrete object in its own right, but I am not averse to saying precisely this and will assume for present purposes that it is legitimate to do so. Here, however, I just want to emphasize that a concrete system must certainly have various concrete objects amongst its parts. Thus, a human central nervous system will include many billions of nerve cells amongst its parts, just as a human brain and cerebral cortex do. If we call brains concrete objects rather than concrete systems, then this might either be a distinction without much of a real difference or else, perhaps, we do so because the parts of a system are restricted to things which play some systemic role in the system, whereas the parts of an object need not be. For instance, it might be held that the only cells involved in a human central

Why My Body is Not Me

135

nervous system are nerve cells which play a systemic role in such a system, whereas a human brain can include as genuine proper parts cells which are not crucial to the functioning of the brain as such.

8. RULING OUT THE ‘ANIMAL’ ANSWER Anyway, we have now narrowed down the kinds of ‘bodily’ things that might, with any degree of plausibility, be identified with me. So let us see whether any one of them passes the crucial test of being such that all and only my thoughts depend upon it for their existence. We can begin with the whole living organism or animal body that I call ‘mine’. Call it A. Do all and only my thoughts depend for their existence upon A? Of course, Descartes would answer emphatically ‘No’, because he believes that I and all my thoughts could exist without my having any animal body at all— that I could survive complete disembodiment—basing this on the alleged conceivability of such an eventuality. But we are not easily persuaded of the truth of this. Let us then concede, at least for the sake of argument, that I could not exist without a body of any kind. And, since A is a body of some kind, let us again ask whether it passes the test for being identical with me. It seems not, for the following reason. If even just one of my thoughts, call it T, could exist without A existing, then A is not identical with me—since T couldn’t exist without me existing—and yet it does seem that at least some of my thoughts meet this condition. Remember that A is supposed to be an entire human animal, which I take to be something that is not only living but which has, of necessity, certain distinctive kinds of bodily organs and a certain distinctive structure or ‘body-plan’. A human heart, for instance, cannot qualify as an entire human animal, even if it is kept alive by some sort of life-support machine. Nor is the concrete system consisting of the living heart attached to and supported by the machine an entire human animal. Now, is it the case that none of my thoughts could exist without a certain entire human animal existing? Surely not. Very plausibly, for instance, a great many of my thoughts could exist even if all that were left of my animal body, following some horrible accident, were my head, kept alive by a life-support machine. A living human head, whether or not attached to a life-support machine— and, indeed, whether or not attached to the rest of an entire human animal— cannot by any reasonable standard be said to be an entire human animal. At this point, philosophers who favor the answer that I am indeed identical with A are sometimes prone to argue as follows.8 They sometimes urge that, while perhaps a living human head which had never been ‘attached’ to the rest of a normal human animal body might not qualify as an ‘entire human animal’, nonetheless, in the dreadful circumstances just envisaged, my human animal, A, would continue to be an ‘entire human animal’, and so continue to exist, but simply be radically reduced in size and radically

136

E. Jonathan Lowe

mutilated in form and structure. If that were so, of course, then we couldn’t conclude, from the possibility of such an accident, that some of my thoughts could exist without A existing—for A would still exist in the envisaged scenario. Now, this looks very much like a purely ad hoc maneuver to save the favored thesis. Such philosophers do, however, have a supplementary line of argument in support of their position. They tend to say that a normal, entire human animal does not literally have a ‘head’, conceived as a real part of that animal, and thus as something which could survive the destruction of the animal. What they deny is that things like animals really have any ‘undetached parts’, with things like heads and arms and various bodily organs, such as the heart and liver, being examples of such supposed parts.9 If that is so, of course, then my animal cannot cease to exist while just its head carries on existing, because there is literally no such thing as my animal’s head. Hence, if anything bodily goes on existing after my dreadful accident, it seems that it must just be my animal, A, albeit in a horribly reduced and mutilated form. Now, I have some sympathy for the idea that it is a mistake to identify one of the ‘parts’ of a living human animal as being its ‘head’, if only because it seems to be rather arbitrary how we delineate the boundaries of such a thing. Where, for instance, does my head stop and my neck begin? On the other hand, this kind of worry is rather less pressing in the case of human organs, such as the heart and liver, whose boundaries seem less arbitrary, because those organs are associated with specific biological functions, unlike the ‘head’ and the ‘neck’. Be that as it may, however, it still seems to me utterly compelling that, whether or not the bodily thing that survives my dreadful accident can properly be called my ‘head’, this thing, whatever it is—and it surely is something—is not by any reasonable standard an entire human animal. The accident surely brings about the destruction of my human animal, A, even though it does not necessarily bring about my destruction—because some of my thoughts can still exist in these circumstances. Hence, I cannot be identical with A. It really doesn’t matter precisely how we categorize the bodily thing that survives the accident, so long as we recognize that we can by no stretch of the imagination categorize it properly as being an entire human animal, which is what A is. Hence, A does not survive the accident and, consequently, cannot be identical with me.

9. RULING OUT THE ‘BRAIN’ ANSWER: A FIRST ATTEMPT We seem, then, to have eliminated one candidate answer to the question ‘What am I?’ I am not my ‘entire human animal’, A, because it is not the case that all and only my thoughts depend for their existence on A. However, many ‘scientifically’ minded people would not profess to be at all surprised by this conclusion, because they think that there is a much better bodily candidate for identity with me, namely, my brain. After all, they may

Why My Body is Not Me

137

point out, it seems clear as a matter of empirical fact that one needs a brain to be able to think, whereas one does not need a heart or a liver or arms or legs—because, they may say, it is the function of the brain to be the ‘organ’ of thought, whereas the heart and liver have quite different functions and the arms and legs, which serve several different purposes, are certainly not essentially involved in thinking. Very well, but is it then the case that all and only my thoughts depend for their existence on my brain—call it B? If they don’t, then I am not identical with B, for now familiar reasons. Now, even if the boundaries of the brain are easier to identify in a principled way than the boundaries of the ‘head’ or the ‘neck’, it is still true that there is some vagueness involved in specifying them. Consequently, there may be, perhaps, in at least some cases, ‘no fact of the matter’ as to whether a certain nerve cell really is a part of my brain. In other cases, there will be no question about this. A nerve cell in my foot is certainly not a part of my brain, whereas a functioning nerve cell in the middle of my cerebral cortex almost certainly is (setting aside barring bizarre cases such as, perhaps, one involving the recent transplant of a nerve from my foot into my brain). However, I don’t want—here, at least—to try to capitalize on such vagueness in order to argue that I am not identical with B: a strategy which, if it worked, could equally be applied in the case of A, of course. There may well be some philosophical mileage in this sort of consideration, since it seems highly questionable that there is any vagueness about the identity of the self.10 But let us set that consideration aside for present purposes. Let us agree that my brain, B, defi nitely exists and pretend, at least, that there is no indeterminacy as regards which bodily thing B is. Let us also agree that if I had no brain at all, none of my thoughts would exist. Does this imply that all and only my thoughts depend for their existence on my brain, B, as they need to if I am to be identified with B? I am going to disregard here any appeal to science fiction scenarios involving ‘teletransportation’ and the like, for one of the reasons that I set aside the replacement argument earlier—they are simply too speculative at present. But can we perhaps construct an argument regarding B’s candidacy for identity with me along the same lines as the argument used earlier to rule out A’s candidacy? In other words, can it be argued that at least some of my thoughts could still exist even if B didn’t exist, on the grounds that it would be enough for these thoughts to exist that some part of B—such as, very plausibly, the cerebral cortex, call it C—still existed (aided by a life-support machine, if need be) and that C could not be regarded as being identical with B itself in a much reduced and mutilated form? I think that this is fairly plausible, in fact. Thought does seem to be associated in particular with the cortex and by no reasonable standard can a human cortex on its own be described as an entire human brain. However, it should be clear that this manner of ruling out bodily candidates for identity with me cannot be reiterated indefi nitely. A single neuron, for example, or even an assembly of several hundred neurons, cannot at all plausibly be claimed to

138

E. Jonathan Lowe

be sufficient to support thought of any kind at all. Moreover, once we get down, say, to the level of individual lobes of the cortex, there comes a point at which further reduction in size does not plausibly leave us with some functionally distinct part of a human brain which cannot be regarded as being just a reduced and mutilated version of some higher-level part. With that fact in mind, we would perhaps be wise not even to pursue the current strategy for eliminating potential bodily candidates even as far as C, but leave it at B, my brain as a whole.

10. A NEW STRATEGY: PRELIMINARY CONSIDERATIONS So we are now left looking for some new strategy to rule out the candidacy of B for identity with me. Here two further pertinent facts may be called upon. The fi rst is that, as far as we know, our thoughts do not depend on our brains in a completely holistic fashion, with every thought depending on the whole brain. Rather, different thoughts very often depend on different parts of the brain. This is particularly clear in the case of commissurotomy patients, whose severed cerebral hemispheres fairly evidently support different thought processes—a fact that is further confi rmed by patients who have undergone hemispherectomy, leaving them with just one hemisphere but still with a capacity for thought. Secondly, we should not lose sight of the kind of dependency that we are really concerned with at present—the ontological dependency of a thought upon its subject, which arises because thoughts are partly individuated by their subjects and thus have their identity determined by those subjects. But is this the kind of dependency that is involved when thoughts are said to be dependent on the brain, or on various parts of the brain? Very plausibly not. The most we can say, it seems, is that without a brain there can be no capacity for thought and hence no individual thoughts. This doesn’t imply that the brain, or any part of it, is something that serves in any way to individuate a thought, in the sense in which the thought’s subject does. Moreover, another thing that we know about human brains is that they are fairly plastic or flexible, being capable of large-scale ‘re-wiring’, particularly in response to damage. Thoughts which might be supported by one area of the brain prior to such damage may later be supported by another. All of these considerations suggest that the kind of ‘dependency’ that my thoughts have upon my brain is quite different in structure and character from the ontological dependency that all my thoughts have upon me, as their unique subject. However, I concede that this line of argument, as it stands, may appear to be less than fully compelling to someone favorable to the hypothesis that I am my brain, B—so let us take another tack, albeit one that draws in part upon the same considerations. This new line of argument will require me to reject the view that things like B have no ‘undetached parts’. However, this rejection is something that I am already strongly committed to and fi nd

Why My Body is Not Me

139

extremely plausible. B surely has, as undetached parts, at least such parts as individual neurons and various assemblies of these. Indeed, B itself, let us suppose, is wholly composed of individual neurons (ignoring, for the purposes of argument, the various other kinds of cell that we actually fi nd within human brains). There may be some indeterminacy regarding which individual neurons belong to B, even though there is no such question concerning the vast majority of them. For instance, it might be debatable whether a certain neuron belonged properly to B or was one of the optic nerves. But let us, once more, ignore such issues of vagueness here, and pretend that B is composed of a perfectly determinate set of neurons—a very large set, of course. Then we can proceed to argue as follows.

11. THE NEW STRATEGY FOR IMPLEMENTING THE UNITY ARGUMENT Allowing as we are that B has undetached parts, we can surely identify some small part of B, call it D, which is wholly composed by some small subset of the neurons composing B. By the same token, we can identify a much larger part of B, call it E, which is D’s ‘complement’ in B—that is to say, E is composed by all and only the neurons in B that do not help to compose D. (If it is worried that a ‘complement’ of one object in another is somehow a less than fully bona fide object, then we could instead take D to be the complement of E in B, thereby forestalling any objection to the ensuing argument that might be made on the grounds that E is somehow inferior to B with regard to its status as an ‘object’. Since we have this move to fall back on if required, however, I shall not in fact make it in the argument that follows.) Furthermore, let us select D on the following basis. First, D should be a part of B which is such that it is not the case that all of my thoughts depend upon it. Since—as was remarked earlier—my thoughts, to the extent that they depend on B, very plausibly do not depend on it in an entirely holistic way, with every thought depending on the whole of B, we have every reason to believe that a small part of B, such as D, can meet this condition. (Later, however, we shall consider the repercussions of denying this assumption.) Second, let us suppose D to be small enough that no thought of mine depends on D but not on D’s complement in B, E. Again, we have every reason to suppose that a small part of B, such as D, which is composed by only a small subset of B’s neurons, can satisfy this condition, since it is very plausible to suppose that a small assembly of neurons—say, a few hundred of them—is not sufficient on its own to support thought of any kind. (Again, however, we shall later consider the consequences of denying this assumption.) With these assumptions in place, we can conclude, first of all, that every thought of mine that depends on B also depends on E. Why? Because the only material difference between B and E consists in B’s inclusion of D—and we

140

E. Jonathan Lowe

have assumed, as we are plausibly entitled to, that no thought of mine that depends on D does not also depend on E. Hence, there can be no thought of mine that depends on B but not on E in virtue of depending on D—and also no thought of mine that depends on B but not on E for any other reason, because B’s inclusion of D is the only material difference between B and E. Thus, as I say, every thought of mine that depends on B also depends on E. Equally, however, every thought of mine that depends on E very plausibly also depends on B, simply because E is a proper part of B. (In any case, since we are arguing against philosophers who hold that all my thoughts depend on B—since they contend that I am B—we are dialectically entitled to make this assumption here.) Hence, in particular, if all my thoughts depend on B, then it is also true that all my thoughts depend on E, and vice versa. That being so, however, it seems clear that dependency considerations do not favor B’s candidacy for identity with me any more than they favor E’s candidacy. In respect of such considerations, B and E are in fact equally good candidates for identity with me, to the extent that either of them is a candidate at all. The mere facts that B is bigger than and includes E cannot be deemed at all relevant in this regard, for these facts have nothing to do with the dependency relations between the thoughts in question and the objects B and E. If those other facts concerning size and parthood relations were relevant, they would equally count against the candidacy of B for identity with me, at least while B is an undetached part of A—for they would count in favor of the candidacy of A, my entire human animal. However, we have already ruled out A’s candidacy, which shows that those facts are indeed irrelevant for present purposes. Might it not be objected here that a relevant difference between B and E is the following? B includes a part, D, on which, we have allowed, some of my thoughts depend, even if all of those thoughts also depend on E—whereas E, of course, does not include that part. But this objection, if sound, would imply that the best candidate for identity with me would in fact be the largest object, O, every part P of which is such that some thoughts of mine depend on P, even if P is a part of O on which no thoughts of mine depend without also depending on P’s complement in O. In that case, however, B itself would be ruled out as the best candidate for identity with me. Consider, for instance, the object O that includes as proper parts B and some blood vessel, V, leading into B. Very plausibly, some of my thoughts depend on V, since if the blood-flows that it carries were cut off, my thought-capacity would be diminished—and yet V is not the sort of thing on which any thought of mine can depend without also depending on B, since V is obviously not sufficient on its own to support thought of any kind. So, in this regard, V stands to B just as D stands to E. Consequently, if B trumps E on account of including D, so too does O trump B on account of including V. Indeed, this method of argument is quite likely to reinstate A, my entire human animal, as the best candidate for identity with me. But, once more, we have already ruled out A, so this method of argument must be flawed and cannot be used to prefer the candidacy of B over that of E.

Why My Body is Not Me

141

Now, clearly, if B and E are equally good candidates for identity with me, then they rule each other out for identity with me. For B and E are numerically distinct objects, being composed of different sets of neurons and having different shapes and sizes, whereas I am just one thing. So B and E cannot both be identical with me. But neither can we with any good reason say that B rather than E, or that E rather than B, is identical with me, for that would be to ignore the fact that they are equally good candidates, so that it would be arbitrary to favor one over the other. The only remaining option available, however, is to say that neither B nor E is identical with me. It is no good contending that B and E are both subjects of my thoughts, for this would deprive those thoughts of a unique subject and thereby deprive them of their identity and hence of their existence. If B and E are subjects of thoughts at all, then they are subjects of numerically distinct thoughts, since B and E are numerically distinct objects.

12. HOW THE NEW STRATEGY MAY BE REITERATED We are now in possession of an argumentative strategy which can be reiterated, if need be, as often as is desired. Suppose, for instance, that it were contended—contrary to our earlier assumptions—that there is in fact some small part of my brain, D, composed perhaps of just a few hundred neurons, which is such that all of my thoughts depend upon it and, indeed, that no thought of mine depends on D’s complement in B, E. Thus, D would apparently be a part of my brain such that, if it were entirely destroyed, I would lose my capacity for thought altogether. (Of course, this is a highly implausible hypothesis, but let us set that fact aside for the purposes of argument.) Could it then be plausibly contended that I am identical with D? Certainly not: for D, even though it contains just a few hundred neurons, is still a very complex object, compositionally speaking. Each neuron in D is composed of many sub-cellular objects, such as mitochondria, and these in turn are composed of complex organic molecules, which in turn are composed of many sub-atomic particles. It is utterly implausible to suppose that there is no small part of D, call it F, whose complement in D—call it G—is not just as good a candidate as D itself for identity with me, in terms of the dependency relations that obtain between my thoughts and D and G respectively. For, even supposing that if D were entirely destroyed, then so too would be my capacity for thought, it cannot plausibly be supposed that the destruction of D’s small part, F, would have this consequence. Furthermore, even if this were supposed, we could then run again the same line of argument with respect to F itself and some small part of it. Sooner or later we shall have to arrive at some small part of my brain, such as a certain individual neuron, N, which is such that it would be completely implausible to suppose that the entire destruction of that object is both necessary and sufficient for the complete destruction of my capacity for thought.

142

E. Jonathan Lowe

What all of this brings out is simply the fact that my capacity for thought, and hence my thoughts themselves, do not depend collectively on any single bodily object, neither my brain as a whole, B, nor any distinguishable part of it, such as D, F or N. Rather, my thoughts depend in a distributive fashion on various different and overlapping parts of my brain. But this simply isn’t the way in which my thoughts all depend upon me, as their subject. Rather, they all depend collectively on me, such that were I to cease to exist, then so would they—each and every one of them. This is why I call the present line of argument the unity argument, because it appeals to the unified way in which all of my thoughts depend upon me as their subject, in contrast with the thoroughly disunified way in which they depend individually on different parts of my brain. This fundamental structural difference in the nature of the dependency relations involved is sufficient to establish the non-identity of me with any of these bodily things, even if we abstract from the distinction between ‘causal’ and ‘ontological’ dependency. This, then, is a strategy which can be used to argue against the identification of any part of my body with me—including, indeed, its sole ‘proper’ part, my entire human animal, A (even though we are already in possession of another argument against my identity with A). Now, I mentioned earlier certain other bodily candidates for identity with me, namely, certain bodily systems, such as my central nervous system (CNS). However, the same strategy can be deployed once again with respect to these, given that, for reasons provided earlier, we are regarding such systems as concrete ones, possessing various bodily objects as parts—in the case of my CNS, obviously, nerve cells. For, just as we can consider large proper ‘undetached’ parts of complex bodily parts—for instance, E above as a large proper part of my brain, B—so, similarly, we can consider large sub-systems of any given bodily system, such as my CNS. And, once again, since it is surely not the case that all of my thoughts depend holistically on the whole of any such system, it will follow that all of my thoughts in fact depend on various large sub-systems of such a system just as much as they do upon the whole system. From this it will follow, once more, that the whole system and such a sub-system are equally good candidates for identity with me and so in fact also equally bad ones, for they rule each other out. Thus, the distinction between bodily parts and bodily systems is not really relevant as far as the applicability of the unity argument is concerned. Given that, between them, bodily parts and bodily systems exhaust all the possible bodily candidates for identity with me, we can conclude that I am nothing bodily at all. This isn’t to say that I am something that could, even in principle, be disembodied, just that I am not myself a body, nor a part of a body, nor a bodily system, nor a part of such a system.

13. AN EMERGENTIST ANSWER TO THE QUESTION What, then, am I? I am just a self or person—a subject of thought and agent of actions. It seems that we must regard this as a basic ontological category,

Why My Body is Not Me

143

not just a ‘functional role’ which can be occupied by things of various more fundamental kinds. We can neither identify any particular self with an entity of some more fundamental kind, nor reduce selves to such entities by showing that they can occupy the ‘functional role’ of the self. Indeed, we may usefully compare these positions concerning the ontological status of selves with analogous positions in the metaphysics of mental properties— identity theories, psychophysical reductionism, and dualism, this last position maintaining that mental properties are distinct from and irreducible to physical ones.11 Analogously, then, I am saying that selves are distinct from and irreducible to bodies and bodily systems. So this is a dualism of self and body. But nothing that I have said in defense of it supports a Cartesian dualism of self and body, which maintains that selves are not only distinct from and irreducible to bodies and bodily systems, but in addition that selves are ontologically separable from bodies and bodily systems, in the sense that selves can exist in a completely disembodied state. The analogous position in the metaphysics of mental properties would be to say that these can be exemplified quite independently of the exemplification of any physical properties—a claim to which the Cartesian substance dualist is in fact obviously committed. But my arguments in favor of self-body dualism leave open the possibility that mental properties can only be exemplified at all if they are co-exemplified with suitable physical properties. This suggests a position regarding the ontological status of the self which deserves to be called emergentist. And, indeed, I am happy to call my view of the self an emergentist one, provided that we understand that such a view can be called ‘physicalist’ at best only in a rather anodyne and misleading sense. Certainly, though, I am happy to concede that it is very probably the case that selves can come into existence only in certain complex physical circumstances, in the absence of which they very probably cannot continue to exist. Selves and their exemplifications of mental properties may, then, very well depend ontologically upon the existence of bodies and their exemplifications of physical properties, but not in a way that implies that the former are identical with or reducible to the latter. Furthermore, this kind of ontological dependency need not exclude the possibility that the causal powers of selves, as agents of actions and subjects of perception, are also neither identical with nor reducible to the causal powers of bodies or bodily systems of any kind. Indeed, I think there are strong arguments in favor of the causal autonomy of selves—by which I mean their possession of distinct and irreducible causal powers—although I shall not go into them here.12

14. FICTIONALISM CONCERNING THE SELF DEFEATED Now, I anticipate that skepticism regarding self-body dualism, even of my non-Cartesian kind, will not have been entirely dissipated by my

144

E. Jonathan Lowe

preceding arguments and elucidations. Some ‘hard-nosed’ physicalists will urge that the ‘self’ is just a fi ction—one that the brain itself somehow creates. The suggestion is that thoughts present themselves as belonging to, or being thought by, a person or self, often because reference to such an ‘owner’ is explicit or implicit in the content of a thought. Certain thoughts present themselves as ‘me’–thoughts, such as the thought that I now feel hungry—which is quite different from the impersonal thought that there exist now feelings of hunger. It may be maintained that it has proved advantageous, in evolutionary terms, for the brains of certain animals to generate such personal or ‘me’–thoughts, perhaps because they prompt more immediate and appropriately focused action—for instance, a brain that creates an I-feel-hungry-now thought is perhaps more likely to receive rapid sustenance than one that merely creates a there-nowexist-feelings-of-hunger thought. However, this sort of theory comes up once more against the diffi culty of explaining how thoughts are individuated, if not at least partly in terms of their subjects. It will not do to say that they are individuated in terms of their causes and effects, because at least some, and in fact very many, of the causes and effects of any given thought will be other thoughts—for the most part, other thoughts of the same subject. Hence, this proposal falls victim to an implicit circularity in its account of the individuation of thoughts.13 In fact, the notion of a thought without a thinker seems, on reflection, to be barely coherent. To regard thinkers as fictions really requires one to regard thoughts as fictions too—and this way lies manifest absurdity for the theory being propounded. For the very notion of a fiction presupposes the genuine applicability of the notion of thought: a fiction just is something that is merely thought of as existing or being the case, when really it does not or is not. Thus, the Land of Oz is a fiction, because it is thought of—by readers of the book or viewers of the film—as existing, when in reality it does not exist. To say that thoughts are fictions is, correspondingly, to say that they are thought of as existing, when really they do not—but then at least the thoughts of their existing must exist, whence it is false to say that thoughts are just fictions. A fi ctionalist theory of thoughts is a contradiction in terms. And, as I say, a fictionalist theory of thinkers is committed, ultimately, to a fictionalist theory of thoughts, whence it too is inadmissible on pain of contradiction.

15. THE PHENOMENON OF ‘THOUGHTINSERTION’: NO OBJECTION Another kind of challenge that I anticipate is this: it may be urged that there is a good sense in which, paradoxical though it may sound, it simply isn’t true that all and only my thoughts necessarily have the same subject—me. For some individuals who are suffering from certain psychopathological

Why My Body is Not Me

145

conditions report that other people’s thoughts are somehow intruding into their own minds—a phenomenon known as ‘thought-insertion’.14 Of course, it is surely never literally true that someone else is thinking a thought in my mind, through some sort of process of telepathy. The phenomenon of thought-insertion is, after all, a delusional one. But doesn’t the fact that the delusion can even arise tell us something important about how we conceive of the relationship between thoughts and thinkers? Doesn’t it show that it at least makes sense to dissociate the subject of a thought—the person whose thought it is, or who is thinking the thought—from the person in whose consciousness the thought arises? And if so, doesn’t that imply that the thesis that all and only ‘my’ thoughts must have me as their subject is not quite as self-evidently true as might be supposed? For if a thought could, at least in principle, arise in my consciousness but have another person as its subject or thinker, then there would be a perfectly good sense in which this thought would be one of ‘my’ thoughts—in virtue of manifesting itself in my consciousness—and yet not be a thought of mine, that is, not be thought by me and therefore be individuated by me as its subject. Well, I confess that I can really make no clear sense of this suggestion. I think we have to say that someone suffering from the delusion of thoughtinsertion really is not just mistaken about their condition but radically confused concerning it. A thought which manifests itself in my consciousness is ipso facto a thought of mine, which is thought by me as its subject. I suspect that what sufferers from this delusion are trying to report, in an incoherent fashion, is a lack of conscious control that they feel they have over their trains of thought. We all experience this from time to time, when thoughts ‘pop’ into our consciousness unbidden and often unwanted, having no apparent connection with the train of thought that we are endeavoring to pursue. We might, for instance, be trying to plan a holiday trip when, quite ‘out of the blue’, a thought about a sick relative springs into our mind and disrupts our train of thought. It seems likely that the delusion of thoughtinsertion is an exaggerated and much more debilitating version of this kind of everyday experience.15

16. THE UNITY OF CONSCIOUSNESS: AN IRRELEVANCE Yet another clarificatory point I should make is that when I speak, in connection with the unity argument, of the unified way in which all my thoughts depend upon me as their unique subject, I am not in any way appealing to the philosophical notion of the so-called ‘unity of consciousness’. The thesis of the unity of consciousness is a phenomenological one, alleging that a subject’s states of consciousness at any given time are ‘unified’, in the sense that they all belong to one ‘stream’ of consciousness, rather than being split into separate ‘parallel’ streams. The implication is that if a subject, S, is aware of one content of consciousness, X, and simultaneously aware of

146

E. Jonathan Lowe

another content of consciousness, Y, then S must also be aware of X and Y together, as a unified content of consciousness. For example, according to this thesis, it would be impossible for S to be aware of a certain pain in his foot and simultaneously be aware of a certain taste in his mouth, with being aware of that pain conjoined with that taste. Whether this thesis is true seems to me to be very much an open question. Obviously, it is difficult if not impossible to test the thesis directly, by means of introspection, for one could not notice that one was not aware of the pain and the taste together, but only separately, since the putative act of noticing that one was aware of them only separately would imply that one was in fact aware of them together. However, just because the thesis cannot be directly falsified through introspection doesn’t mean that it must be true—although, of course, its unfalsifiability by these means may go a long way toward explaining why it may be an intuitively compelling thesis. I strongly suspect, indeed, that it is false.16 But, in any case, it is irrelevant to my present concerns, because I make no appeal to it in support of my contention that all of my thoughts depend upon me as their subject. My argument for the latter thesis, which is a metaphysical one, appeals entirely to considerations concerning the individuation of thoughts, as I made clear earlier. This argument will consequently still stand even if the thesis of the unity of consciousness needs to be discarded.

17. A FINAL SUMMING UP Now it is time for me to sum up my fi ndings. I think that there are compelling arguments in favor of the real existence of selves or persons, conceived as subjects of thought, and also compelling arguments for regarding selves as constituting a fundamental ontological category of things, distinct from and irreducible to anything bodily, whether that be an entire living organism or animal, or a part of such an animal, or even a bodily system of some sort. It is simply incoherent to regard selves as fictions created by the brains of certain animals. They are perfectly real and, as far as I can see, ontologically basic. They are what we are, in the most fundamental sense. This does not mean that we are something wholly immaterial or non-physical, in the sense of being capable of existing quite apart from bodies and their exemplifications of physical properties, as Descartes believed. But it does mean that we should reject as utterly extravagant and indefensible the views of those physicalists who say that we are ‘nothing but’ animals of a certain kind, or ‘nothing but’ our brains. A rejection of Cartesian dualism does not compel us to embrace such a physicalist position as the only tenable alternative that is consistent with the fi ndings of empirical science. An emegentist view of the self, in the sense explained earlier, is not only consistent with those fi ndings but, it seems, the only coherent option open to us from a metaphysical point of view.17

Why My Body is Not Me

147

NOTES 1. See Descartes (1984, II), Second Meditation. For a recent study of the plural form of the question, see Olson (2007). 2. For more on ontological categories in general, see Lowe (2006a). 3. But see further Lowe (2008). 4. For more on individuation in general, in both of these senses, see Lowe (2003). 5. See further Lowe (2001). 6. See Lowe (2009) and also Baker (2000: 122–123). 7. For an earlier and much briefer version of the argument, see Lowe (2008: 95–99). The present chapter is designed to strengthen that version in various respects. 8. Such philosophers include, prominently, Eric T. Olson: see Olson (1997). 9. For an important and influential instance of this line of thought, see van Inwagen (1981). 10. Compare Unger (2006: ch. 7). 11. Compare Beckermann (2009). 12. But see further Lowe (2008). 13. See further Lowe (1996: 27–30). 14. See, for example, Stephens and Graham (2000). 15. See further Lowe (2006b). 16. See again Lowe (2006b). 17. I am grateful for comments received when earlier versions of this chapter were delivered at Davidson College and at Washington University, St Louis.

REFERENCES Baker, L. R. (2000). Persons and Bodies: A Constitution View. Cambridge: Cambridge University Press. Beckermann, A. (2009). What is property dualism? In B. P. McLaughlin, A. Beckermann, & S. Walter (Eds.), The Oxford Handbook of Philosophy of Mind (pp. 152–172). Oxford: Oxford University Press. Descartes, R. (1984). The Philosophical Writings of Descartes. Trans. J. Cottingham, R. Stoothoff, & D. Murdoch. Cambridge: Cambridge University Press. Lowe, E. J. (1996). Subjects of Experience. Cambridge: Cambridge University Press. . (2001). Identity, composition, and the simplicity of the self. In K. J. Corcoran (Ed.), Soul, Body, and Survival: Essays on the Metaphysics of Human Persons (pp. 139–158). Ithaca, NY: Cornell University Press. . (2003). Individuation. In M. J. Loux and D. W. Zimmermann (Eds.), The Oxford Handbook of Metaphysics (pp. 75–95). Oxford: Oxford University Press. . (2006a). The Four-Category Ontology: A Metaphysical Foundation for Natural Science. Oxford: Oxford University Press. . (2006b). Can the self disintegrate? Personal identity, psychopathology, and disunities of Consciousness. In J. C. Hughes, S. J. Louw, & S. R. Sabat (Eds.), Dementia: Mind, Meaning, and the Person (pp. 89–103). Oxford: Oxford University Press. . (2008). Personal Agency: The Metaphysics of Mind and Action. Oxford: Oxford University Press.

148 E. Jonathan Lowe . (2009). Dualism. In B. P. McLaughlin, A. Beckermann, & S. Walter (Eds.), The Oxford Handbook of Philosophy of Mind (pp. 66–84). Oxford: Oxford University Press. Olson, E. T. (1997). The Human Animal: Personal Identity without Psychology. New York: Oxford University Press. . (2007). What Are We? A Study in Personal Ontology. New York: Oxford University Press. Stephens, G. L., & Graham, G. (2000). When Self-Consciousness Breaks: Alien Voices and Inserted Thoughts. Cambridge, MA: MIT Press. Unger, P. (2006). All the Power in the World. New York: Oxford University Press. van Inwagen, P. (1981). The doctrine of arbitrary undetached parts. Pacific Philosophical Quarterly, 62, 123–137.

7

What About the Emergence of Consciousness Deserves Puzzlement? Martine Nida-Rümelin

1. INTRODUCTION At some point in the evolution of life on Earth some specific pieces of matter got arranged in a way that led to the occurrence of consciousness. At some ‘moment’ deep in the past the fi rst animals must have started experiencing. One may speculate that the fi rst experience ever had on our planet was nothing more than a vague feeling of being comfortable or uncomfortable; maybe it was a feeling of warmth or a feeling of hunger. The emergence of phenomenal consciousness in the evolution of life is astonishing. Upon reflection one will wonder how this radical change could possibly occur. The same may be said about the emergence of consciousness in an individual life. Upon reflection one will realize that something puzzling, amazing, startling, or bewildering happens when an animal or a human organism physically changes in a way that gives rise to the occurrence of phenomenal consciousness.1 What has been said in the preceding few lines will appear obvious to many and is hard to deny. But philosophers are known for denying the obvious. A few decades ago a majority of analytical philosophers interested in issues about the mind would have denied that they share that natural intuition. Many would then have reacted saying that they simply do not understand what some people fi nd puzzling about the occurrence of consciousness. But the intellectual climate has changed. It is now widely acknowledged that the occurrence of consciousness on a physical basis appears to be puzzling and seems mysterious. The ongoing discussion about the explanatory gap triggered by Joseph Levine’s seminal paper (Levine, 1983) and the debate about what David Chalmers calls the harder problem of consciousness (see Chalmers, 1996 and Levine, 2001) illustrate that change. Many contemporary materialists still insist that there really is nothing mysterious about the occurrence of consciousness. They admit, however, that we are all naturally inclined to be under a cognitive illusion that tends to persist even in the light of convictions to the contrary: the illusion that there is a deep puzzle about the emergence of consciousness in a material world. 2 Many of those presently involved in the discussion about the ontological status

150

Martine Nida-Rümelin

of consciousness pursue what has been labeled ‘the phenomenal concept strategy’. They try to develop a materialist account of special concepts we have of phenomenal states (phenomenal concepts) in order to explain why consciousness appears mysterious to human thinkers although it really is a normal physical phenomenon the occurrence of which on a physical basis does not deserve deep puzzlement.3 The descriptive claim that humans are cognitively so equipped that they tend to see a mystery in the occurrence of consciousness is widely accepted. The related normative claim, however, is seldom discussed and often explicitly rejected: the claim that the occurrence of consciousness in a material world deserves puzzlement or amazement. According to this normative claim, it is a mistake (so to speak) not to be puzzled when considering the occurrence of consciousness, which may be due to a lack of understanding. The claim may be put like this: a thinker with an adequate understanding of what it is for an individual to be phenomenally conscious will thereby see that the occurrence of phenomenal consciousness on a physical basis deserves puzzlement and will normally experience that intellectual emotion upon reflection on the issue. In the present chapter, I hope to give some motivation for the normative claim that the emergence of consciousness deserves puzzlement. It is not my main purpose, however, to defend that claim. Rather, I invite the reader to presuppose the normative claim for the moment and to reflect upon the following question: what is it exactly about consciousness that makes it adequate to wonder about its very existence? In other words, what is—in that case—the proper object of amazement?

2. PRELIMINARY REMARK: ASTONISHMENT AND FASCINATION The question in the center of this paper is about the object of astonishment, puzzlement, or amazement. These intellectual emotions are typically triggered by a lack of understanding or by happenings that contradict our rational expectation. They are for that reason closely related to surprise. The type of intellectual emotion at issue must be clearly distinguished from another kind of intellectual emotion called ‘fascination’. Obviously there is reason to be fascinated by the phenomenon of consciousness. Consciousness, human consciousness in particular, is fascinating by its creative and rich features illustrated by works of literature, art, and music; by scientific theorizing and philosophical reflection; or, in a different manner, by the strange concreteness and detailed vividness of the stream of consciousness present to a subject in normal dreams, and more impressively even in lucid dreaming or under the influence of psychedelic drugs. No reasonable person will deny that these phenomena of consciousness deserve the reaction of fascination. Materialists who deny that there is reason to be puzzled by the

Puzzlement of Emergence of Conciousness 151 existence of consciousness have no reason to deny the fascinating character of consciousness phenomena. The difference between being fascinated and being puzzled may be illustrated by the case of the ‘flowers’ of frost on the window on a cold winter day. One may be touched by their beauty and fascinated by their rich structure even when—after learning about the mechanisms underlying their formation—one has no temptation to be puzzled about their existence. Given enough physical knowledge, puzzlement, or amazement about the existence of frost patterns is inappropriate and will normally disappear. Contrary to this, physical knowledge normally has no bearing and at any rate should have no bearing on fascination. The rich structure of frost patterns remains fascinating even when the relevant causal mechanisms are well understood. The physicalist can and will see the case of consciousness in analogy to the case of frost flowers. She will not deny the fascinating character of consciousness and yet insist that—as in the case of frost patterns—there is no mystery about its existence and so no rational basis for puzzlement. It is important to be clear about the difference between fascination and puzzlement (amazement, bewilderment, wondering) in the present context. For the reasons just given, the two normative claims (a) that consciousness phenomena are fascinating (deserve fascination) and (b) that their existence is puzzling (deserves puzzlement) are distinct claims and neither of them implies the other. Also, and for the present inquiry even more importantly, the object of the intellectual emotion, when it occurs in the context of reflection on consciousness, may well be and normally is different. The fascinating character of consciousness is due to the richness of its specific content in specific cases and to its apparent creativity. In typical cases, when we are fascinated by consciousness phenomena, we are fascinated by their particular features, and the object of the intellectual emotion is not the existence of consciousness as such. The latter is the object, however, of the kind of puzzlement relevant in the present context. It is puzzlement about the mere fact that consciousness occurred at all, that nature somehow brought it into existence. Puzzlement about the existence of consciousness is puzzlement about the existence of something that all consciousness phenomena, even the simplest ones, have in common. It is not triggered by the complexity or richness of consciousness in highly developed conscious beings. Given the difference in the object between typical cases of fascination with consciousness on the one hand and the relevant kind of puzzlement about consciousness on the other, confusing fascination with puzzlement would lead one astray in one’s reflection on the question here at issue.

3. THE THEORETICAL INTEREST OF THE QUESTION Why should we be interested in gaining a clear understanding of the proper object of puzzlement or amazement in the case of puzzlement (or amazement)

152

Martine Nida-Rümelin

about the existence of consciousness? An obvious reply is this: only then do we have a clear understanding of what an explanation of the occurrence of consciousness would have to amount to. Only then will we be able to judge whether a proposed explanation of the occurrence of consciousness fully explains what should be explained. But there is an objection against the theoretical value of the present enterprise that might immediately come to the reader’s mind. We know the supposed proper object of amazement under the following preliminary description: it is the emergence of consciousness on a physical basis. So it seems that we can search for a more detailed and deeper understanding of the proper object of amazement simply by reflecting upon what it is for consciousness to come into being. And we surely have a clear understanding of the latter once we have a satisfactory theory about what being conscious essentially consists in. So it seems that we can address the question here asked simply by doing what philosophers of mind are normally engaged in: by developing and defending some account of the nature of consciousness. It follows that there is no point in asking the question about the proper object of puzzlement. We can address that issue more directly by reflecting upon the nature of consciousness. This objection, however, overlooks that the two philosophical activities—searching for a satisfying theory about the nature of consciousness and searching for a satisfying description of the proper object of amazement in our puzzlement—are guided by different constraints and different intuitions. The fi rst activity, for instance, will be guided by the following constraint: the account should be compatible with accepted background assumptions about the functioning of biological organisms. Contrary to this, the second activity should not be influenced in the same way by theoretical background assumptions. Its aim is to give an adequate account of what it is that appears to deserve puzzlement in the occurrence of consciousness. Intuitions about what deserves puzzlement are important in that context and put constraints on potential answers. These constraints, however, play no central role in the fi rst philosophical enterprise. An example may illustrate the point: it is no convincing objection against a functionalist account of consciousness when proposed within the fi rst intellectual enterprise that the instantiation of functional properties on a given physical basis does not even appear to be puzzling or amazing. But a functional account of consciousness cannot be used in an acceptable account of the proper object of amazement precisely for that reason. We cannot see why consciousness should appear to be mysterious if it is nothing but the instantiation of certain functional properties. A different conceptualization of consciousness must be used in an adequate account of the proper object of puzzlement. The proper object of amazement must be described, within the second enterprise, using a conceptualization that gives rise to a problem about consciousness. Although the two intellectual activities are different and guided by different intuitions and constraints, they are not independent. The result of

Puzzlement of Emergence of Conciousness 153 the second is relevant for the fi rst. Let us fi rst consider the relevance for a philosopher in search of a materialist theory about the nature of consciousness who accepts that the emergence of consciousness on a physical basis appears to be mysterious. She will have to accept the following constraint on her theory: the best description of the apparently proper object of amazement and the description proposed by a materialist theory about the nature of consciousness must converge in the following sense: the account of consciousness given in the fi rst description and the one given in the proposed materialist theory must refer to the same phenomenon. The materialist must be able to show that the referents are identical. Any materialist theory of consciousness must be supplemented by an argument showing that the phenomenon described by the theory really is the phenomenon we were puzzled about in the fi rst place. Otherwise one will object that the materialist has simply changed the subject. In order to answer this potential objection and to supplement her theory in the required way, the materialist needs an adequate account of what appears to deserve amazement in the emergence of consciousness. The dualist will be interested in the issue about the proper object of amazement for a different reason. Contrary to the materialist, the dualist does not believe that our puzzlement about the emergence of consciousness is the result of some cognitive mistake due to the architecture of the human mind. Rather, according to the dualist, the puzzlement results from an adequate understanding of what being conscious consists in. The dualist therefore will have to explain what it is about the nature of consciousness that makes its occurrence on a physical basis mysterious in a way that deserves astonishment, puzzlement, or amazement. It is in this context that the dualist will need a satisfactory account of the proper object of puzzlement.

4. THE ANSWER TO THE QUESTION PRESUPPOSED IN THE ONGOING DISCUSSION The question about the proper object of amazement is seldom explicitly addressed in the ongoing debate. It will however be easy to agree on the following close relation between this question and issues that are explicitly addressed in the debate about the explanatory gap. According to the explanatory gap thesis, the emergence of consciousness on a physical basis does not have and might not ever have a satisfactory explanation. We may assume as generally accepted: the phenomenon which is in need of explanation according to the controversial explanatory gap thesis coincides with the proper or apparently proper object of amazement. Presupposing this close link, it is clear what the answer to the present question implicitly accepted in the ongoing discussion looks like. A common assumption— accepted by defenders and enemies of the explanatory gap thesis as well as materialists and dualists—seems to be this: the phenomenon to be

154

Martine Nida-Rümelin

explained is the instantiation of phenomenal properties or qualia. What we wish to understand is how and why nature brings it about that qualia are instantiated in certain organisms. Two explananda are distinguished in the debate: that there are any qualia at all instantiated in a given animal and that there are certain specific qualia instantiated in a given organism under certain physical conditions. Transferring this picture to the question about the proper object of amazement, we are led to the following claim: the (apparently) proper object of amazement is the instantiation of qualia in biological organisms. What we wonder about and what is or appears mysterious is the fact that qualia are instantiated in human beings and other animals. To understand the implicit claim just arrived at, it is important to remember what qualia are supposed to be according to mainstream opinion. According to the standard view, qualia are properties of experiences, and experiences are neurological processes. Furthermore, qualia are directly given in experience—the instantiation of qualia is phenomenally present to the experiencing being. Using the plausible presupposition formulated earlier, the standard view implies the following claim about the (apparently) proper object of puzzlement: it is the fact that certain physical processes in human and other brains instantiate qualia. According to this view, the (apparently) proper object of puzzlement is the fact that physical processes instantiate qualitative properties which are phenomenally present to the experiencing subject. I will argue in this chapter that this account of the (apparently) proper object of amazement gives us a distorted picture of what it is about the occurrence of consciousness that deserves or seems to deserve astonishment. Let us call this account implicit in the debate ‘the standard view about the (apparently) proper object of puzzlement’ or simply ‘the standard view’.

5. A PROBLEM FOR THE STANDARD VIEW It is common for materialists to deny the existence of qualia in the commonly accepted sense alluded to previously. I agree with the result but for different reasons. There are no qualia, I argue, if the two conditions mentioned previously characterize what it is for a property to be a quale.4 There are no qualia in that sense since there are no properties that satisfy both of these conditions: C1: Qualia are properties of neurological processes. C2: The instantiation of qualia is phenomenally present to the subject in the relevant experience. The argument can be summarized like this: the only properties the instantiation of which is present to the subject in experience are either properties

Puzzlement of Emergence of Conciousness 155 things appear to have or properties of the experiencing subject. Neither properties things appear to have nor properties of experiencing subjects are properties of neurological processes. Therefore, no property that satisfies C2 is a property that satisfies C1; so there are no qualia in the sense specified. The argument can and must be based on concrete examples. Let us ask, for that purpose, what properties are phenomenally present to a subject who sees an object which appears red to her (the subject has an experience belonging to the type commonly referred to by ‘red experiences’). In this phenomenal kind of experience, an object appears to the subject to have a certain surface property. The object appears to be red. If the experience is veridical, then the object instantiates the property it appears to have in the experience. Properties things appear to have are potentially instantiated by perceived objects; they are not instantiated by neurological processes, and they do not seem to be instantiated in neurological processes. In having a ‘red’ experience we are not presented with an experience which appears to be red, nor are we presented with a neurological process that appears to be red. In seeing something as red we are not under the impression that something going on within our body has the property of being red. If there are any entities in the world which actually are red, then these entities are external objects.5 The only reason we have to assume that there are any red objects in the real world is that it certainly appears in visual experience as if there were. But obviously there is no parallel basis in phenomenology for the assumption that inner processes (‘experiences’, neurological processes) are red. So we can conclude: instantiations of the property of being red are phenomenally present to the experiencing subject in visual perception (the property of being red satisfies condition C2). But the property of being red does not satisfy condition C1. The same may be said, or so I claim, about any property something appears to have in perceptual experience. Before I consider a second candidate for being a quale, I would like to address a possible objection to the fi rst part of my argument. Someone may say—in the spirit of what has been called the projection theory of color—that when things appear red to us we erroneously ‘project’ a mental property of mental entities onto the physical world. What really is red (or what really has some corresponding qualitative property of redness*) is something mental, some ‘mental object’ for instance in ‘the visual field’. In visual experience, however, we are under the illusion that the mental property at issue (red*) or some corresponding qualitative property (‘red’) is instantiated in the perceived objects themselves. I mention the projection theory of colors not in order to discuss it but to avoid a possible misunderstanding. What has already been said about the property of being red is compatible with the projection theory. Even according to the projection theory it is not the experience itself (or a physiological process the materialist identifies with the experience) which is red or red* but rather something given in the experience to the experiencing subject, some object or apparent object the subject is presented with in having the

156 Martine Nida-Rümelin experience. The projection theory may have some intuitive plausibility. If the reader feels attracted by some variant of it, she should not thereby feel motivated to reject the previous claim. It is only when we confuse the experience with what is given in the experience that we are tempted to think that the experience itself instantiates redness (or any corresponding directly given qualitative property). Let us now turn to the second candidate mentioned before for being a quale in the sense specified by the two foregoing conditions. We need to search for a property which does in some sense fulfi ll the second condition. Returning to the example of seeing something which appears red to the subject a second plausible candidate for a property the instantiation of which is phenomenally present in having the experience will come to mind: it is the property itself of having the relevant kind of experience, the property of seeing something as red, the property of having a ‘red’ experience. A person who has a red experience is in a specific sense thereby aware of having a ‘red’ experience. It would be misleading to say that the person is aware of the fact that she is having a red experience. Awareness of a fact seems to presuppose some conceptualization of the fact at issue. However, in being aware—in the relevant sense—of seeing something as red, one need to have a concept neither of oneself nor of the phenomenal property of seeing something as red. It may be less misleading to say that the subject is aware of being in the relevant state (for instance of seeing something as red). We should accept, or so I propose, that the following is true of certain properties of subjects: instantiating the property is to be aware (in a sense to be specified) of instantiating the property. We may say that this is true for all phenomenal properties if we define what it is for a property to be a phenomenal property in the following way: Defi nition: A property P is a phenomenal property if and only if having or not having that property at a given moment constitutes a difference in the phenomenology of the overall state of the subject concerned. The instantiation of a phenomenal property in that sense is necessarily ‘present’, we may say, in the overall phenomenology of the subject’s present state (if it wasn’t present in that sense, it wouldn’t satisfy the preceding defi nition).6 The instantiation of any phenomenal property in the sense defi ned makes a difference for the subject—the instantiation partially characterizes what it is like for the subject to be in her present state. In that basic phenomenal sense of awareness, the following is necessarily true: when a subject instantiates a phenomenal property, then the subject is aware of instantiating that property, or, in other words, her own instantiation of that property is phenomenally present to the subject in the moment at issue. Examples of phenomenal properties in the sense just given are properties of subjects. Seeing a red apple, hearing the melody of a violin, having a

Puzzlement of Emergence of Conciousness 157 headache, and feeling sad about something are uncontroversial examples. Controversial examples are being struck by the thought that one has forgotten one’s key in the apartment, understanding a sentence in a certain way, and experiencing oneself as the causal origin in one’s actions.7 To make it clear that these properties are properties of subjects and to distinguish them from so-called phenomenal properties of experiences, let us call these properties s-phenomenal properties. As argued earlier, s-phenomenal properties do fulfill the second condition C2 (in a legitimate sense of ‘phenomenally present’ which however needs some explanation). But s-phenomenal properties are properties of subjects; they are not properties of experiences. So they do not fulfill C1. The only candidates for fulfillment of C2 are properties things appear to have (like being red) and s-phenomenal properties (like seeing something as red). Neither properties of the fi rst kind nor properties of the second kind fulfill C1. So—in the sense specified by C1 and C2—there are no qualia. The argument depends on the assumption that there is no further kind or properties that might fulfi ll C2. I must admit that I cannot see any. But the reader might still think that there are properties of experiences the instantiation of which is phenomenally present to the subject concerned. A different way to get to the same negative result (to the denial of the existence of qualia in the relevant sense) is to ask what precisely these properties of experiences that are supposed to be phenomenally present might consist in. To simplify things let us assume, as is commonly done in the debate not only by materialists but even by property dualists, that experiences are neurological processes. Let us call the physiological process which is identical (according to that proposal) to a concrete ‘red’ experience the process R. According to the idea I wish to reject, R has a phenomenal property which is (a) responsible for its membership in the kind ‘red experiences’ and (b) phenomenally present to the subject having the experience. Let us first see which property fulfills (a) and then check if the property we found also fulfills (b). The fi rst question may be put like this: what is the property which is constitutive for the fact that R is adequately classified as a ‘red experience’? It seems clear that the adequate answer must be given along the following lines: R is the physiological basis responsible for the fact that the subject concerned sees something as red.8 The exact account is not important here. What is crucial is this: it seems clear upon reflection that we can only define the relevant property of the physical process R indirectly and by reference to the corresponding s-phenomenal property of the subject concerned. The complicated property of R (of being the basis for the instantiation of a certain s-phenomenal property), however, does not appear to be a plausible candidate for being phenomenally present to a subject when having a ‘red’ experience. In having a red experience the subject is phenomenally aware of redness and of experiencing redness. But there does not seem to be any basis in phenomenology for the assumption that the subject is, in addition, phenomenally presented with the property mentioned before which is

158

Martine Nida-Rümelin

instantiated by R. This complicated property is, however, the only plausible candidate for fulfillment of condition (a); it is the only plausible candidate for the property that makes it the case that the process belongs to a certain ‘phenomenal’ kind. But the instantiation of that property is not phenomenally present. So we arrive at the same result as before: there are no qualia in the relevant sense.9 The negative result obtained makes it plausible to conclude: the (apparently) proper object of amazement when we wonder about the existence of consciousness should not be described as puzzlement about the instantiation of qualia in neurological processes. The standard view should be rejected. The reader may object at this point that the last conclusion has not been sufficiently justified. Even if there are no qualia in the relevant sense, it may still cognitively appear to us as if there where qualia, and in that case our puzzlement might still be adequately described as being about the supposed occurrence of qualia. But this hypothesis has little plausibility. If it was the instantiation of qualia in the sense specified which appears puzzling, then once we have seriously abandoned the idea that there are qualia in this sense, the apparent mystery should disappear. But it does not. Thinking about the issue one will easily realize: even if we leave so-called qualia out of the picture, the mystery about s-phenomenal properties persists. There is no need to talk about qualia when we describe what appears astonishing. How can it be, we might say, that—on the basis of some arrangement of matter—some subject has visual experiences, feels pains, suffers from depression, wonders about philosophical issues, experiences itself as active, and so on?10 When we wonder about these cases, we wonder about the instantiation of s-phenomenal properties. It seems safe to assume: rational puzzlement about the occurrence of consciousness is puzzlement about the first instantiation of s-phenomenal properties. Still, the following question remains: what feature of s-phenomenal properties makes their instantiation and in particular their first instantiation on a physical basis seem to deserve puzzlement?

6. TENTATIVE ANSWERS AND A PROPOSAL Let us come back to the fi rst animals capable of experiencing on our planet. Let us assume that they were small animals living in the ocean which developed at some point the capacity to enjoy a feeling of warmth. The occurrence of these fi rst feelings constitutes, I suppose (and I assume that this is intuitively evident to most of us), a radical and fundamental change which deserves astonishment. The question I am interested in may be put like this: what is it about this property—the phenomenal property of enjoying a feeling of warmth—that makes its fi rst instantiation as a result of some physical arrangement appear mysterious upon reflection? A fi rst answer can be readily excluded. It is not the specific phenomenal character at issue. Had it been a feeling of cold or of hunger or of lust or of fear, this would not have

Puzzlement of Emergence of Conciousness 159 made any difference. Our puzzlement does not depend on which specific s-phenomenal property is instantiated in the case considered. We can test the intuition considering the case of a fetus. Let us assume that we know of a child, Anna, what her very fi rst experience was: it was an uncomfortable and frightening loud noise when her mother went to a heavy metal concert during pregnancy. The thought that there is someone who starts feeling something, on the basis of some arrangement of matter, is puzzling, and the fact considered in that thought appears mysterious. But the mystery obviously does not depend on the specific kind of experience considered. Anna’s fi rst experience is puzzling independently of whether it is a noise or a melody or a pain or a sudden light. The beginning of ‘a stream of consciousness’ as the result of biological processes is puzzling independently of the subjective character of the experiences with which it starts. A second answer one might propose starts with the idea that it is something all s-phenomenal properties have in common that makes the instantiation of s-phenomenal properties on a physical basis appear mysterious. One common feature of all s-phenomenal properties is this (see the definition formulated earlier): the instantiation of s-phenomenal properties (and only of these properties) necessarily involves awareness of their instantiation. No individual can instantiate an s-phenomenal property without thereby being aware (in the relevant primitive sense) of instantiating that property. It might be proposed that what makes the instantiation of s-phenomenal properties on a physical basis puzzling is that we lack an understanding of how the relevant kind of awareness could possibly occur in a physical world.11 This answer, I believe, is correct. But it can be made more explicit. What is it about awareness in the relevant sense (awareness of being in a certain s-phenomenal state) that makes its occurrence on a physical basis puzzling? The intuitive plausibility of the answer that I would like to propose will become specifically clear when we consider the first instantiation of s-phenomenal properties on our planet or in an individual’s life. The fi rst instantiation of s-phenomenal properties is phenomenally present to someone. The instantiation of s-phenomenal properties requires the existence of an experiencing subject, of someone to whom anything can be phenomenally present. We may assume that at some moment prior to the fi rst instantiation of s-phenomenal properties by a given subject S, S did not exist, or at least, if S existed, S was not yet a subject of experience. At some moment before the fi rst feeling of warmth experienced by S (if this is the fi rst experience ever had by S), S must have come into existence. Let me say it in a more cautious way which allows for the materialist claim that the subject to whom something is given is nothing but the biological organism12: at some moment prior to the fi rst feeling experienced by S, either S must have come into existence, or—if S existed already as an organism without consciousness—S must have turned into a subject of experience. My hypothesis is this: it is not the instantiation of the fi rst s-phenomenal property as such which deserves amazement, rather, what deserves puzzlement and what

160

Martine Nida-Rümelin

I take to be ‘behind’ our astonishment is what the fi rst instantiation of s-phenomenal properties requires: it requires that—on the basis of some physical arrangement—a subject of experience has ‘popped’ into existence (or an organism has undergone a radical change, it has changed into a subject of experience). The change that deserves puzzlement according to this proposal when consciousness arises in the case of a particular organism, is this: after the change, there is someone, someone capable of being aware of its own properties (‘aware’ in the basic phenomenal sense of awareness); before the change no-one was there to whom something could have been phenomenally given. According to this proposal, what deserves puzzlement about the emergence of consciousness on a physical basis is the beginning of the existence of a subject of experience. A parallel claim applies to astonishment about the instantiation of s-phenomenal properties in general. When we wonder about how it can possibly be the case that s-phenomenal properties are instantiated in some given adult animal, we really wonder about the existence of the subject concerned. We wonder about the fact that there is ‘someone there’ to whom something is given in the experience. It is easier to overlook that point when considering consciousness in a subject that already existed for a while. Upon reflection, however, we can see that even in these cases the existence of a subject capable of being aware in the phenomenal sense is the prior and prominent object of puzzlement.

7. ANSWERING AN OBJECTION Some readers might be tempted to react as follows. Subjects of experience are nothing but individuals capable of instantiating s-phenomenal properties. If so, then there is no difference between the two possible claims about the proper object of amazement: (a) it is the fi rst instantiation of s-phenomenal properties, and (b) it is the beginning of the existence of subjects of experience. In my view, both descriptions are all right but the latter gives a deeper analysis and is more informative. This may be seen as follows. One may use the description given in (b) in order to explain why (a) is a good answer but not vice versa. When given the fi rst answer (a), one may still ask: what is it about s-phenomenal properties that renders their instantiation puzzling? A partial answer to this further question can be given by reference to (b): the instantiation of s-phenomenal properties requires the existence of a subject, and we lack any satisfying understanding of how they come into existence on the basis of some arrangement of matter. When given the answer (b), one may still insist on asking in a similar manner: what is it about subjects of experience which makes it difficult or impossible to understand how some arrangement of matter may lead to their existence? In this case, however, no adequate answer can be given using the fi rst description (a). Answer (a) does not help to gain a better understanding of answer (b).

Puzzlement of Emergence of Conciousness 161 8. FINAL REMARK The thesis presented in this chapter may be formulated like this: it is the coming into being of subjects of experience, of individuals capable of instantiating s-phenomenal properties, that deserves puzzlement when we consider the emergence of consciousness in nature. If this is accepted then the common focus in contemporary discussions on phenomenal properties construed as properties of experiences appears quite unfortunate. Working with the standard view of what is puzzling about consciousness, one leaves the subject of experience out of the picture and thereby misses the crucial point. Progress in philosophical theorizing of consciousness might require a shift of attention. It might require one to withdraw one’s attention from socalled qualia and to focus on what is conceptually and ontologically more fundamental: individuals capable of phenomenal awareness, experiencing subjects, individuals to whom something can be phenomenally present. Emergentist theories of consciousness endorse the intuition that something fundamentally new occurs in nature when consciousness arises. It has however been a persisting problem for emergentists to explain the sense of ‘novelty’ they have in mind in that context. Emergentists need to explain the relevant sense of ‘fundamentally new’, and they need to say what it is about consciousness that renders its fi rst occurrence ‘fundamentally new’ in that sense. Some ideas here presented can be seen as preliminary work for a solution to this problem. In the spirit of what has been said we may start a solution with the following proposals: the novelty when consciousness arises is—fi rst of all—the coming into being of subjects of experience. Related to this, the novelty is, in addition, the fi rst instantiation of a new quite puzzling kind of properties, s-phenomenal properties. They are puzzling for two reasons: their instantiation necessarily involves awareness of their instantiation and thereby ontologically requires the existence of someone to whom something is phenomenally present.13

NOTES 1. In this paragraph I use the term ‘emergence’ in its non-technical everyday sense as a variant for ‘occurrence’. 2. See for instance Papineau (2002). 3. Compare for the current debate about phenomenal concepts Balog (2009). 4. I would like to add that I have no objection against a different notion of qualia which refers to experiential objects (like smells and melodies). 5. For a defense of the view that, strictly speaking, color experience is never veridical, see Maund (2006). 6. The defi nition is not meant as a reductive defi nition, as a reductive defi nition it would be viciously circular. I suppose that we have a quite clear intuitive grasp of what it is to constitute a difference for the phenomenology of the overall state a subject fi nds itself in at a given moment. It is for that reason that the defi nition, even if not reductive, might still be helpful.

162

Martine Nida-Rümelin

7. The first two examples are taken from Siewert (1999: ch. 8), where he forcefully defends the claim that it is something like to think; for agentive phenomenology (third example) compare Horgan (2007) and Nida-Rümelin (2007a). 8. There is of course an issue about what ‘basis’ means in that context. According to the physicalist, it cannot be a ‘causal’ basis. She will have to talk of metaphysical constitution. 9. In the second argument (contrary to the fi rst) I used an additional assumption about qualia which is however commonly accepted: that qualia are the properties constitutive for membership in certain phenomenal kinds. The idea that we are phenomenally presented with instantiations of qualitative properties in inner processes is related to a mistaken model of phenomenal awareness sometimes called the perceptual model of phenomenal awareness; for more about this issue, see Nida-Rümelin (2007b). 10. I agree with Siewert (1999) that thinking a thought has phenomenal character and with Horgan (2007) that there is a specific phenomenology of acting or being active (compare Nida-Rümelin, 2007a). 11. Levine (2007) locates the fundamental problem in a way similar to this idea when he focuses on the acquaintance a subject has with the referent of phenomenal concepts. 12. I reject the view that the subject is identical to the corresponding organism, but I wish to formulate my description of the proper object of amazement in a way that does not presuppose my ontological preferences. 13. For a sketch of a dualist version of emergentism see M. Nida-Rümelin (2006).

REFERENCES Balog, K. (2009). Phenomenal concepts. In B. McLaughlin, A. Beckermann, & S. Walter (Eds.), Oxford Handbook in the Philosophy of Mind. New York: Oxford University Press, 292–312. Chalmers, D. (1996). The Conscious Mind. Oxford: Oxford University Press. Horgan, T. (2007). Mental Causation and the Agent Exclusion Problem Erkenntnis, 67(2) 183–200. Levine, J. (1983). Materialism and qualia: The explanatory gap. Pacific Philosophical Quarterly, 64, 354–361. . (2001). Purple Haze: The Puzzle of Consciousness. Oxford: Oxford University Press. . (2007). Phenomenal concepts and the materialist constraint. In T. Alter & S. Walter (Eds.), Phenomenal Concepts and Phenomenal Knowledge. Oxford: Oxford University Press. Maund, B. (2006). The illusory theory of colors: An anti-realist theory. Dialectica, 24, 245–268. Nida-Rümelin, M. (2006). Dualist emergentism. In B. McLaughlin & J. Cohen (Eds.), Contemporary Debates in Philosophy of Mind. Oxford: Blackwell. . (2007a). Doings and subject causation. Erkenntnis, 67(2) 255–272. . (2007b). Transparency of experience and the perceptual model of phenomenal awareness. Philosophical Perspectives, 21(1), 429–455. Papineau, D. (2002). Thinking about Consciousness. Oxford: Oxford University Press. Siewert, C. (1999). The Significance of Consciousness. New York: Oxford University Press.

8

The Emergence of Rational Souls Uwe Meixner

This chapter describes what the emergence of rational souls consists in, explicating the concepts of emergence, soul, and rationality. On the basis of intuitive, yet defended, premises, it argues the existence of rational souls. In a more indirect approach, it points out which basic feature the world must have in order to allow the existence and emergence of rational souls, it gives an argument to the conclusion that the world has this feature in fact, and it fi nally shows that if the world has the said feature, then the emergence of rational souls is a natural expectation.

1. EMERGENCE Emergence is, among other things, the coming into existence of something that has never existed before. However, in order to speak of emergence, what emerges does not have to be just new; it must be ontologically new: it must be the fi rst exemplar of an ontological kind that has never been exemplified before. Thus, my coming into existence at t 2 (my only coming into existence ever) was not an instance of emergence, although I had never existed before t 2 . And the coming into existence at t1 of the (or of a) fi rst human being was not an instance of emergence, either—although this human being had never existed before t1, and although the kind human being had never been exemplified before t1; for human being is not an ontological kind, just a biological one. But the coming into existence at t 0 of the (or of a) fi rst living being was indeed an instance of emergence: that living being had never existed before t 0, and it was the fi rst exemplar of an ontological kind that had never been exemplified before t 0: the (or a) fi rst living being was not just new, it was ontologically new. This characterization of emergence implies that in asserting the emergence of a rational soul, one is presupposing that rational soul is an ontological kind. But that rational soul is an ontological kind seems plausible. It even seems plausible that rational soul is an irreducible ontological kind—which, note, does not already imply that there are rational souls (it merely means: if there are rational souls, then they are irreducible entities).

164

Uwe Meixner

I will call emergence which is such that the ontological kind involved in it is irreducible radical emergence. Thus, the emergence of the (or a) fi rst rational soul—if it did emerge—was an instance of radical emergence. The emergence of the (or a) fi rst living being, however, was very likely not an instance of radical emergence, since living being is an ontological kind— indeed—but apparently not an irreducible one. (Today living beings, to the extent that they are just living beings, are almost unanimously regarded as being purely physical things—though they were not regarded this way at all times of the philosophical past.) The preceding remarks on emergence can be summed up and made precise by a sequence of formal defi nitions: Suppose X comes into existence at time t: D1 The coming into existence of X at t is an instance of emergence with respect to K if, and only if, (a) X never existed before t and (b) K is an ontological kind that was never exemplified before t, and X exemplifies K at t. D2 The coming into existence of X at t is an instance of radical emergence with respect to K if, and only if, it is an instance of emergence with respect to K, and K is an irreducible ontological kind. D3 The coming into existence of X at t is an instance of emergence (simpliciter) if, and only if, it is true for some K that the coming into existence of X at t is an instance of emergence with respect to K. D4 The coming into existence of X at t is an instance of radical emergence (simpliciter) if, and only if, it is true for some K that the coming into existence of X at t is an instance of radical emergence with respect to K. If the coming into existence of X at t is an instance of emergence with respect to K and if there is no coming into existence of anything else at t that is an instance of emergence with respect to K, then X can be described as the first [exemplar of the ontological kind] K, and the coming into existence of X at t is a case of singular emergence with respect to K. If besides the coming into existence of X at t also the coming into existence of something else at t is an instance of emergence with respect to K, then X can still be described as a first [exemplar of the ontological kind] K; but the coming into existence of X at t is not a case of singular emergence with respect to K: it is part of a case of plural emergence with respect to K. Since plural emergence with respect to K requires exact simultaneity of the several (at least two) comings-into-existence participating in it, plural emergence with respect to K is bound to be a very rare thing, no matter which ontological kind K we are looking at. Perhaps there are no (true) cases of plural emergence at all. Nevertheless, plural emergence (like causal overdetermination) can certainly not be excluded on a priori grounds. According to the preceding definitions, emergence is—relatively or simpliciter—a property of comings-into-existence. But, not infrequently,

The Emergence of Rational Souls 165 emergence is also ascribed to individuals that are not event-like entities: there are assertions in singular of the form “the fi rst NP emerged” (where “NP” stands for a noun-phrase)—for example: “The fi rst living being emerged”—and there are assertions in plural of the form “NPs emerged”— for example: “Living beings emerged”. Such assertions are derivative, analogical ways of speaking. “The fi rst NP emerged” just means the same as “There is a time t and an X which is such that X came into existence at t, and the coming into existence of X at t is an instance of emergence with respect to NP [where NP is the kind that corresponds to the noun-phrase NP], and for every Y that came into existence at t and that is different from X: the coming into existence of Y at t is not an instance of emergence with respect to NP”. And “NPs emerged”, in one interpretation (the lectio facilior), just means the same as “There is an X and a time t such that X came into existence at t, and the coming into existence of X at t is an instance of emergence with respect to [the kind] NP”. Accordingly, the plural in the statement “NPs emerged” is not to be taken referentially seriously: one might as well have said, “A [or some] NP emerged”. Note that, according to the present interpretation, “NPs emerged” is logically equivalent not only to “A NP emerged” but also to “[The kind] NP emerged”. And note that this interpretation is precisely the one in which the statement “Rational souls emerged” underlies the title of this chapter, “The Emergence of Rational Souls”. Hence I might logically equivalently (and without ambiguity) have called this chapter “The Emergence of a Rational Soul” or “The Emergence of Rational Soul [the kind]”. But these alternative titles would not have been as catchy as its present title. In another interpretation of “NPs emerged” (the lectio difficilior), the phrase means the same as “There is a time t and an X such that X came into existence at t, and the coming into existence of X at t is an instance of emergence with respect to NP, and there is a Y different from X that also came into existence at t and the coming into existence of Y at t is also an instance of emergence with respect to NP”. In this interpretation the plural in “NPs emerged” is taken seriously, and “NPs emerged” is logically equivalent to “Several NPs emerged”. Note fi nally that statements of the following forms are pairwise logically equivalent: The fi rst NP emerged. NPs emerged1 [lectio facilior]. NPs emerged 2 [lectio difficilior].

NP emerged singularly. NP emerged. NP emerged plurally.

2. SOULS Presuming that rational soul is an irreducible ontological kind, what are the characteristics of that kind? In answering this question, I shall fi rst look at

166

Uwe Meixner

the concept of soul in and by itself, leaving rationality, for the moment, out of consideration. The notion of soul is a very old one, with a very long tradition in the history of ideas. Given this tradition, very many philosophical, religious, theological, psychological, and folk-psychological ideas, perspectives, and connotations have accumulated around that notion. I herewith present the definition of a sober, strictly philosophical concept of soul; it is inspired by the tradition, but at the same time highly selective with regard to the riches of that tradition: D5 X is at t a soul of Y if, and only if, (1) Y is an animal that is conscious at t, (2) X is a non-physical individual without inherent temporal dimension, (3) X is a subject of the consciousness of Y at t, and (4) if Y is up and about in the time-stretch following t, then X causes some movement1 of Y after t in the light of the consciousness of Y at t. According to D5, the concept of soul—or better: soul-of—is a relational concept, a rather special form of time-dependent being-of (other, more commonplace forms of time-dependent being-of are wife-of, home-of, father-of, son-of, etc.). Note also that according to D5 the concept of soul is identified with the concept of active soul (see condition (4)); this should be kept in mind. There is a normal way of X being at a time a soul of Y (but D5 does not propose a priori that it is the only way). This is the normal way: The normal time-dependent -relationship (a) For some time t: Y is an animal that is conscious at t. (b) For every time t: if Y is an animal that is conscious at t, then X is at t a soul of Y, and nothing else but X is at t a soul of Y. (c) There is no Y´ distinct from Y such that at any time t X is a soul of Y´ at t. If X and Y fulfill the condition of normalcy that is specified by the conjunction of these three statements, then one can address X as the soul of Y (this is based on (a) and (b)) and as the soul of Y alone (this is based on (a), (b), and (c)). And vice versa: if one can address X as the soul of Y and also as the soul of Y alone, then X and Y fulfill the condition of normalcy that is specified by the conjunction of those three statements. (I leave it to the reader to consider which non-normal animal-soul-relationships are not excluded by D5, and which of those may plausibly be assumed to actually occur in the light of psychopathological evidence.) Now, the simple sense in which the word “soul” is meant in the title of this essay is this: D6 X is a soul (simpliciter) if, and only if, there is an animal Y such that X is the soul of Y.

The Emergence of Rational Souls 167 If X is at time t a soul of Y, then there is a connection between X and Y in virtue of which X is of Y at t. What is the nature of that connection? Generally speaking, it is a connection between a non-physical individual and a physical one; this follows on the basis of D5 and the fact that being an animal entails being a physical individual. In part, the connection in question is effected by consciousness, and in part, it is effected by causal power. The two aspects of the connection are closely related to each other—not only on the basis of the obvious fact that each of the two connects the same thing, X, with Y, but also on the basis of a functional relationship between them (indicated in D5 by the words “causes [ . . . ] in the light of the consciousness”). I shall fi rst consider the consciousness-related part of the connection of X with Y. If Y is an animal that is conscious at t, then there is such a thing as the consciousness of Y at t, in other words: the totality of those experiences that have both the index Y and the index t. The index Y becomes (not immediately, but mediately) attached to that totality by the fact that it is the brain of Y that produces the experiences in the totality (and forms them into a homogeneous whole); the index t becomes attached to that totality by the fact that the experiences in the totality all occur at time t (for this, one need not necessarily assume that t has no extension). Thus, the consciousness of Y at t is the totality of all experiences produced by the brain of Y that occur at t. But, note, this does not make Y or the brain of Y a subject of the consciousness of Y at t. Rather, such a subject is itself a product of the brain of Y, which is embedded in, or better: implied by, the (brain-produced) consciousness of Y at the given time t. That a subject of the consciousness of Y at t is neither Y nor the brain of Y is confi rmed by elementary phenomenology (even without reference to evidence that shows beyond reasonable doubt that a subject of the consciousness of Y at t is a dependent product of the brain of Y, and hence cannot be identical with either Y or its brain): I, for example, am not merely a, but the subject of the present consciousness of this animal, UM; but I am neither UM nor its brain. Why? Because there is at present, that is: at t*, a certain, narrowly circumscribed location L in space from which I see the world. If I am located anywhere in space at t*, then I am located in L at t* (this is what phenomenology rather convincingly tells me; for now, at t*, my right toe is three feet, in this particular direction, away from L—and away from me, it seems). Suppose now that I am located somewhere in space at t*; hence, according to the conditional just asserted—the conditional in italics: CI—, I am located in L at t*, and hence I am neither this animal nor its brain (because neither of the two is located in L at t*). Suppose then, on the contrary, that I am not located anywhere in space at t*; hence, again, I am neither this animal nor its brain (because both are located somewhere in space at t*). In fact, if CI is true (and I believe it is true), then I am distinct from every physical individual. This is seen by the following argument: Since I exist at t*, [1] I am distinct from every physical individual that does not exist at t*.

168

Uwe Meixner

And [2] if I am nowhere in space at t*, then I am distinct from every physical individual that exists at t* (because every such individual is somewhere in space at t*). Finally, [3] if I am somewhere in space at t*, then I am in L at t* (according to CI), and therefore, again, I am distinct from every physical individual that exists at t* (because every such individual is either not in L at t*, or is in L at t* but is not a suitable candidate for being me: a brain cell or a small collection of brain cells of UM may be in L at t* but would certainly not be me). Taking [1], [2], and [3] in conjunction, what was to be shown is a logical consequence of that conjunction: I am distinct from every physical individual. And therefore also the subject of the present consciousness of UM—the subject of the consciousness of UM at t*—is distinct from every physical individual (that subject being me). Finally, since I certainly am an individual, it follows that I—the subject of the consciousness of UM at t*—am a non-physical individual. In parallel to what was said previously about the normal way of X being at a time a soul of Y, I note, at this point, that there is also a normal way of Z being a subject of the consciousness of Y at a given time. This is this normal way: The normal time-dependent relationship (a´) For some time t: Y is an animal that is conscious at t. (b´) For every time t: if Y is an animal that is conscious at t, then Z is a subject of the consciousness of Y at t, and nothing else but Z is a subject of the consciousness of Y at t. (c´) There is no Y´ distinct from Y such that at any time t Z is a subject of the consciousness of Y´ at t. If Z and Y fulfill the condition of normalcy that is specified by the conjunction of the preceding statements, then one can address Z simply as the subject (of the consciousness) of Y, and of Y alone. In fact, this condition of normalcy is fulfilled by the animal UM (for Y) and myself (for Z). Hence it follows that I am the subject (of the consciousness) of UM. Am I also the soul of UM? According to the previously stated necessary and sufficient condition for the normal time-dependent -relationship, the truth of the following three statements is necessary and sufficient for me being the soul of UM: (a*) For some time t: UM is an animal that is conscious at t. (b*) For every time t: if UM is an animal that is conscious at t, then I am at t a soul of UM, and nothing else but me is at t a soul of UM. (c*) There is no Y´ distinct from UM such that at any time t I am a soul of Y´ at t.

The Emergence of Rational Souls 169 The statements (a*) and (c*) are true (beyond reasonable doubt). All that remains to be shown for establishing me as the soul of UM is (b*). Assume, then, that UM is an animal that is conscious at time t. In order to demonstrate (b*), we must derive from this assumption the following five statements (but in doing so, we may make use not only of the assumption but also of propositions established on independent grounds): (1*) UM is an animal that is conscious at t, (2*) I am a non-physical individual without inherent temporal dimension, (3*) I am a subject of the consciousness of UM at t, and (4*) if UM is up and about in the time-stretch following t, then I cause some movement of UM after t in the light of the consciousness of UM at t. (5*) For every X that is not me: (1*)X/I, 2 or (2*)X/I, or (3*)X/I, or (4*)X/I is not true of X. For from these five statements, it follows, according to D5, that I am at t a soul of UM and that nothing else but me is at t a soul of UM—which is what has to be derived from the assumption in order to demonstrate (b*). Now, (1*) is a trivial consequence of the assumption. And since the previously stated condition (b´) is already known to be true of UM and me (putting UM for Y, and me for Z), we—given the assumption—also get (3*) and (5*). This is obvious with regard to (3*), and with regard to (5*), merely consider that the assumption and (b´) (for UM and me) entail the following: for every X that is not me: (3*)X/I is not true of X. Moreover, it has already been shown that I am a non-physical individual, hence we already have part of what is stated in (2*). There remain two questions that are still open: (I) Is it true that I am without inherent temporal dimension? (The answer “yes” is required to be correct for establishing (2*).) (II) Given any time t at which UM is an animal that is conscious [and, of course, there are in fact many such times t], is it true, then, that I cause some movement of UM after t in the light of the consciousness of UM at t, if UM is up and about in the time-stretch following t [as UM normally is]? (The answer “yes” is required to be correct for establishing (4*).) In order to give the present section a clear and impressive conclusion, I maintain that the correct answer to both of these questions is “yes” (but my reasons for having this position are presented in the next section and in the section after it). If I am right, then it follows that besides (1*), (3*), and (5*) also (2*) and (4*) have been (rather trivially) derived from the assumption,

170 Uwe Meixner and hence that the truth of (b*) has been demonstrated, and that, therefore, it is quite true: I am the soul of UM (since (a*) and (c*), which are also necessary for this conclusion, have already been established).

3. SOULS AND CAUSATION Common sense is all in favor of giving the answer “yes” both to question (I) and to question (II) posed in the previous section. Given that UM is an animal that is conscious at time t, common sense tells me that, if UM is up and about in the time-stretch following t, there is no circumstance that could keep me—the subject of the consciousness of UM—from causing some movement of UM after t in the light of the consciousness of UM at t. After all, without the almost immediately manifest effects of my guidance, UM would almost immediately be not up and about after t. And common sense also tells me that I do not have an inherent temporal dimension, since I experience myself as not having any temporal parts and as not being a temporal part of anything (not even a point-like temporal part). But common sense does not seem to be a good advisor with regard to the questions (I) and (II)—because common sense does not seem to present a defensible position with regard to them. How so? The relevant worries can be summed up in two arguments: Argument 1: If question (I) is answered with “yes”, then it is presumed that I am not an event, nor anything event-like (for example, a temporally specified state of affairs). But then I am not capable under any circumstances of causing anything, since only events or event-like entities are capable of causing something. Now, this negative result makes it (rationally) impossible to answer question (II) with “yes”; it must be answered with “no”. Conversely, if one insists on the answer “yes” to question (II), then question (I) must be answered with “no”. Clearly, one cannot answer both questions with “yes” (contrary to what is suggested by common sense). Argument 2: Previously, the result has been reached—by fairly commonsensical considerations—that I am a non-physical individual. But if this is correct, then I am not capable under any circumstances of causing anything, since only physical entities are capable of causing something. This negative result makes it impossible to answer question (II) with “yes”. A unified effective response to both these arguments consists in showing the viability—the coherence, or even better: the existence—of non-physical agent-causation. In this section and the next I will try to demonstrate the existence of non-physical agent-causation, which attempt, if successful,

The Emergence of Rational Souls 171 takes away the threat Arguments 1 and 2 pose to assuming that questions (I) and (II) can be commonsensically and correctly answered with “yes”. As a preface to the considerations that follow, I put on notice that causation— whether of the event-causal or the agent-causal type—is here always taken to be sufficient causation, according to which conception of causation a cause, qua cause, is sufficient for bringing about its effect (whether or not the cause raises the effect’s probability, and whether or not the cause is a condition without which the effect would not have come about). Note that causal sufficiency need not be an affair of the cause in and by itself: though a cause is here taken to be, qua cause, sufficient for its effect, a cause (in this sense) may be—and commonly is—sufficient for its effect merely in a relative way: on the basis of sufficiency-enabling circumstances (for example, the well-functioning of the mechanical structures involved, this well-functioning not being counted as part of the cause). Given the general conception of causation employed in this essay, the following is a highly plausible (but contingent a posteriori) assumption: Uniqueness of path of physical event-causation For any physical event X: if there is a physical event-causal chain that ends with X (on the effect side, not the cause side), then all physical event-causal chains that end with X can be integrated into a single physical event-causal chain that ends with X. One consequence of this is that any two physical events Y and X that are connected by a physical event-causal chain (with Y causing X) are also connected by a unique integrative physical event-causal chain: namely, that causal chain that begins with Y and ends with X and is a section of the (unique) complete integrative physical event-causal chain that ends with X. Another consequence of the preceding principle is that causal over-determination (in the interesting sense) is out of the question for physical eventcausation (contingently so and a posteriori, since the invoked principle, which is the sole ground for this result, is itself contingent and a posteriori). In what follows, if there is mention of physical event-causal chains, always integrative physical event-causal chains are intended (either complete ones or sections of complete ones). Now, agent-causation has three possible varieties: it is either purely direct, or purely indirect, or neither of the two. In purely direct agent-causation, an agent causes an event X, and there is no event-causal chain3 which ends with X and the beginning of which is caused by the agent; in purely indirect agent-causation, an agent causes an event X, but only via an event-causal chain which ends with X and the beginning of which is caused by the agent. It is in principle possible that agent-causation is neither purely direct nor purely indirect; however, it seems to me, in view of its hardly being realized in reality, this possibility can be safely neglected. The same distinction that

172

Uwe Meixner

has just been made with regard to agent-causation can, of course, also be made with regard to event-causation. How non-physical agent-causation—in both (remaining) varieties: purely indirect and purely direct—can plausibly enter into the economy of a living conscious animal can be seen in the following way (using UM as a mere example): Suppose there occurs a certain voluntary movement E of UM. E is a physical event, and there is a (single integrative) physical event-causal chain running backwards in time from E. That chain, going backwards, leads to a physical event E´ in the brain of UM. E´ causes E (purely indirectly). Now, if E´ itself were the (counting in the direction of time, not inversely to it) last link in a physical event-causal chain that does not originate in the brain of UM, then E could certainly not be regarded as a voluntary movement of UM—contradicting the preceding assumption that E is a voluntary movement of UM. Therefore, either (1) E´ itself has no physical event-cause, or (2) there is a physical event-causal chain that runs backwards in time from E´ and that originates in the brain of UM, which means: the chain has a (counting in the direction of time, not inversely to it) first link in the brain of UM beyond which it cannot be further prolonged (backwards in time): a physical event E´´—in the brain of UM—that has no physical event-cause. In both cases, we have a physical event E* in the brain of UM that has no physical event-cause but that event-causes E, the voluntary movement of UM. (It has already been concluded that E´ causes E; and in case (2), E´´ causes E because E´´ causes E´, which in turn causes E.) Now, if E* had no cause, then E* would be a chance-event, and hence the event E, being causally based on E*, could not be regarded as a voluntary movement of UM—contradicting the earlier assumption that E is a voluntary movement of UM. Therefore, E* has a cause (although it has no physical event-cause), which cause is, transitively, also a cause of E. But who or what is that cause? The interpretation of self-experience, of self-phenomenology knows of only one answer to this question, notwithstanding the fact that it has been fed, and also has digested, the neurophysiological fi ndings concerning the role of an animal’s brain in the causation of the voluntary movements of the animal: I—that is, the non-physical individual that is the subject of the consciousness of UM (as we already know) and that does not have a temporal dimension (there is no indication of it in self-experience)—am a purely direct agent-cause of the UM-brain-event E* (of which fact I am only very indirectly aware); and by causing E* purely directly, I am, moreover, a purely indirect agent-cause of the voluntary UM-movement E, in virtue of the physical event-causal chain that begins with E* and ends with E. In fact, I cause E in the light of the consciousness of UM at a time shortly before the inception of E (say, E is the withdrawal of the left hand of UM from this metal pipe, after I hand-felt the pipe to

The Emergence of Rational Souls 173 be getting warmer and warmer rather quickly and anticipated its being burning hot to the hand very soon). There are alternatives, of course, to this neurophysiologically informed, self-interpretational view of the causation of E* and E. But how plausible are they? One might, for example, assume that E*—though not caused by any physical event—is caused by a physical agent (hence not by me, since I am a non-physical entity). However, there do not seem to be any physical agents that can plausibly fi ll that role. But is not the brain of UM a likely candidate for fi lling it, in fact, the best candidate? It is true that E* happens in the brain of UM, but E* is certainly not caused by the brain of UM (as little as the movement of an engine’s piston, which movement moves the engine, is caused by the engine). E* is simply something that happens to (in) the brain of UM. And there is also a more fundamental problem with the idea that the brain might be a physical agent-cause of E*: a physical non-event Y can cause an event X only in virtue of some physical event Z that (i) essentially involves Y and (ii) causes X. But we have already found that no physical event causes E*. Hence the brain of UM (though plausibly a physical non-event) cannot cause E*. By the same token, no part of the brain of UM can cause E*, nor, for that matter, can E* be caused by UM itself (which is very plausibly a physical non-event). Alternatively, one might assume that E*—though not caused by any physical event—is caused by a non-physical event (hence, again, not by me, since, having no temporal dimension, I am a non-event). However, there do not seem to be any non-physical events that can plausibly fi ll that role. This has become apparent only recently, in the wake of the experiments of the neurophysiologist Benjamin Libet. The only likely candidate for being a non-physical event that causes E* would be an E-directed decision-experience in the consciousness of UM that occurs before the inception of E*: an experience in which I—the subject of the consciousness of UM—experience myself as effectively (not just premeditatively4) deciding to bring about E. However, the Libet experiments quite unequivocally show that there is no such decision-experience: the decision-experience that does occur follows E*, though it is still ahead of E. There is no getting around this, it seems to me, and it is fatal for the view that E* is caused by a non-physical event. But it presents no problem for the view that E* is caused by a non-physical agent: me; it seems only natural that I become aware of my effectively (not just premeditatively) made decision to bring about E merely after I have effectively made it: by my causing E* (at the very time of E*5) in order to bring about E (via the neurophysiological chain of command). All that seems to stand in the way of accepting the existence of nonphysical agent-causation is the “How could . . . ?” attitude: “How could

174 Uwe Meixner a cause be non-physical?”, “How could a cause be not an event?”, “How could a cause be not even event-like?”. The argument implicit in rhetorical questions of this type is, of course, that something—namely, that which is addressed in the questions—does not exist because one does not understand how it could exist. Not a very cogent way of arguing, it seems to me—and especially uncogent if there is positive evidence for the existence of that regarding which one does not understand how it is possible. Non-physical causation has been attacked on the grounds that it supposedly violates the laws of physics, or, what many erroneously believe to be the same thing, the principle of the causal closure of the physical. The laws of physics, at least, are certainly not violated by non-physical causation; the causal closure of the physical, in turn, is not a law or principle of physics, but a rather shaky metaphysical assumption and certainly far from being a non-negotiable rational obstacle to accepting non-physical causation.6 Agent-causation has been attacked on the grounds that it is absolutely incomprehensible, “mysterious” in the disparaging, philosophical sense of the word.7 Much can be done to alleviate this impression.8 And ever since Hume incontestably showed what perhaps should have been obvious to philosophers from the start—that event-causation is very unlike the relation of logical consequence—one can rationally choose to fi nd something incomprehensible in event-causation, too. Essentially, event-causation is no better off than agent-causation: there is some rational opaqueness to both—which is no good reason to hold that either of the two does not really exist. The real problem for the existence of non-physical agent-causation is not its alleged incomprehensibility or its alleged confl ict with physics. It is the problem that there might be no use for it, that it might be superfluous. True: if something is superfluous, then this does not logically entail that it does not exist. Yet, it must be admitted that the superfluity of non-physical agent-causation, if shown to be a fact, would cast doubt on its existence (we need not invoke Occam’s Razor for this). In this section of my essay, I have argued for the existence of non-physical agent-causation via arguing (1) that I—already shown to be a non-physical individual—have no temporal dimension, hence that I am a substance and thus fit for being an agent, and (2) that I in fact purely directly agent-cause the UM-brain-event E*, and purely indirectly, on the basis of E* (and in the light of the antecedent consciousness of UM), the UM-movement E, thereby, by arguing so, rendering crucial support to common sense in the completion of the argument, started in the previous section, that aimed to establish me as the soul of UM. My argument in this present section crucially relied on the premise that E is a voluntary movement of UM, and hence it crucially relied on the presupposition that there are voluntary movements of animals. Suppose there were no voluntary movements of UM or any other animal. If this supposition were true, my argument, as stated, would collapse.

The Emergence of Rational Souls 175 4. RATIONALITY Can it be doubted that there are voluntary movements of animals? Can it be doubted that at least some of UM’s movements are voluntary? It cannot be doubted—at least not by me (who is the subject of the consciousness of UM)—that some of UM’s movements feel voluntary. But perhaps none of them is voluntary. How can this be? Easily. It is the case if the following is true: The coming about of any animal movement is a necessary consequence of the laws of physics and of the physical events that happened before the fi rst animals came into existence. This—call it “animal determinism”—is just a consequence of the doctrine of universal physical determinism, but it has the advantage that one can believe in it even if one considers universal physical determinism to be false (as is strongly suggested by quantum mechanics). If animal determinism is true (and a fortiori if universal physical determinism is true), then there are no voluntary movements of animals (though for some mysterious reason some of these movements seem voluntary: to the outside observer, and to the subject of the consciousness of the performing animal). Thus: if animal determinism is true, then my basis for arguing for the existence of non-physical agent-causation is gone. And more than that: if animal determinism is true, then non-physical agent-causation is superfluous, and therefore non-existent (“therefore” indeed!—but under the assumption of animal determinism, one is irresistibly drawn to this conclusion, though it does not logically follow). But if there is no nonphysical agent-causation, then there are no souls (simpliciter), as defi ned by D6 (taking into consideration what precedes that defi nition and assuming that any animal is at some time t conscious, and up and about in the time-stretch following t). Therefore, if souls emerged (and a fortiori if rational souls emerged), then the world is more or less bound to be in such a way that animal determinism (and a fortiori universal physical determinism) is not true. Do we have any basis for assuming that animal determinism is not true? Yes, indeed, there is a basis for assuming that it is not true: this basis simply consists in the fact that some animals have conscious interests. The fact of conscious interests is not logically incompatible with animal determinism, but it fits ill with it, and animal determinism has nothing to offer to explain it. If animal determinism is true, conscious interests are superfluous, and therefore (ought to be) non-existent. Under the assumption of animal determinism, one is irresistibly drawn to this conclusion. However, in this case, one cannot rationally accept the conclusion after all: the existence of conscious interests is just too plain obvious. This does not bode well for animal determinism.

176

Uwe Meixner

Conscious interests are embedded in the consciousness that is produced by the brain of any animal. In the indirect sense that can be extracted from the preceding sentence, any conscious animal has conscious interests; but the proximate owner of an animal’s conscious interests is not the animal, but the subject of the animal’s consciousness. Thus, in the case of UM, it is me who is the proximate owner of UM’s conscious interests. Now, why should this be, if I could bring about nothing in the physical world that would contribute to pursuing these interests (typically relating to the physical world)—as must be the case if animal determinism were true? Conscious interests are a part of any consciousness and make it a locus of (at least) minimal rationality: in the light of conscious information about what is the case, however incomplete and distorted, and in the light of conscious interests, however simple, the animal’s subject of consciousness is fitting means to ends—at least in a rudimentary fashion (it need not be remarkably deliberate or remarkably intelligent, nor need it be accompanied by self-reflection or conceptualization).9 This is the essence of rationality. If having minimal rationality in the sense just described were all that is being required for an animal to be an animal rationale—as I would advocate— then human animals would, of course, be far from being the only animalia rationalia; rather, being a rational animal would then coincide with being a conscious animal. In whatever way one might ultimately wish to decide this merely conceptual question (i.e., what is to be understood by the word “rational” when applied to animals, including human animals), even minimal rationality makes biological sense only if the subject of consciousness of an animal can act on it and make, in the light of it, a difference in the physical world that would not have come about without that subject’s action. And minimal rationality does make biological sense (it can hardly be denied). Again, this does not bode well for animal determinism.

5. WHY THE EMERGENCE OF RATIONAL SOULS IS A NATURAL EXPECTATION Suppose animal determinism is false (as I have argued in the previous section), in fact, false on a broad scale (why should there be only a few instances of its falsification?): the coming about of many movements of many animals is not a necessary consequence of the laws of physics and of the physical events that happened before the first animals came into existence. In other words, there are many and widely distributed animal movements that are not subject to ancient physical predetermination. Consider the complete integrative physical event-causal chains10 that lead up to these movements. For defi niteness, consider a particular animal, AN, many of whose movements are not subject to ancient physical predetermination: they are the notanciently-descended AN-movements, in short: the NAD-AN-movements.

The Emergence of Rational Souls 177 It contributes significantly to the survival chances of AN—by dramatically increasing AN’s adaptiveness—if most of the complete integrative physical event-causal chains that lead up to the NAD-AN-movements are short and begin in a central place within AN, where the beginnings of many of these chains are subject to survival-directed central control. This is why brains—in particular, AN’s brain—evolved. But what does it mean that many beginnings—in the brain of AN—of complete integrative physical event-causal chains that lead up to NAD-AN-movements are subject to survival-directed central control? First, it means that these beginnings—events in AN’s brain—have a single cause, which is not a physical event-cause (otherwise, they would not be beginnings of complete integrative physical event-causal chains); second, it means that that single cause is neither blind nor without wishes nor without a sense of power: it wishes to survive in the situation it is being informed about (normally correctly) and knows (“in its bones”) what it can do in this situation (normally correctly). Its own survival is the survival of AN. Thus, there is a biological, evolutionary argument for the existence of a single entity that causally controls many of the event-causal origins (the event-causal first causes) in the brain of AN of NAD-AN-movements, most of which have, as was noted, an event-causal origin in the brain of AN. This evolutionary argument functions in the same way as all such arguments do: it exhibits how a certain feature of living beings (though not usually a feature whose very existence is controversial) fits naturally into the economy of survival and hence, given the right circumstances, may be expected to appear and be perpetuated in the course of biological evolution. It is strongly suggested by this argument that the causal controller it points to is a non-physical conscious agent. Now, I do not know about AN, but in the case of UM I know this causal controller personally: it is me, the rationally active non-physical substance that is the subject of the consciousness of UM, and the soul of UM. If rational souls exist, then, very plausibly, there has been an emergence of rational souls (in the sense explained in the fi rst section of this essay)—since, doubtless, there is a time at which at least one rational soul has existed, and a time at which no rational soul has (as yet) existed, and therefore, very plausibly, a fi rst time of the exemplification (by something newly existent) of the ontological kind rational soul. This chapter should not end without addressing two objections (not of the usual kind). It might be objected that the emergence of rational souls—qua non-physical substances—amounts to a natural creation ex nihilo of non-matter by matter, and that such a natural creation ex nihilo is simply incredible.11 Indeed, as things have been described earlier, the coming into the world of souls can be regarded as an entirely natural occurrence (that is, one need not invoke for its explanation any special intervention by God): the laws of nature do allow the existence of souls; given the right circumstances, souls come into existence, and stay in

178

Uwe Meixner

existence because they constitute a survival-asset for animals. In fact, souls simply co-evolved with brains: with the physical organs of animals whose very function it is to serve as the causal vehicles of souls, the nonphysical organs of animals.12 Once one has become accustomed to the idea that nature has two sides (a physical and a non-physical one), one does not see why the previously described occurrences of natural history should deserve the designation of creatio ex nihilo (of “new stuff” to boot).13 It might be objected that the agent-causality of souls cannot be distinguished from the workings of chance (and “Why assume agent-causality and souls then?” is the implied skeptical question). This objection is proposed—with a new twist—by Peter van Inwagen.14 A brief response: If, under indeterminism, the objective probability that UM goes through door A is 0.5, and the objective probability that UM goes through door B is also 0.5 (prior to UM going through either door), then nothing I can do can change that probability; but this does not mean that UM’s going through door A is, or is indistinguishable from, a chance-event, in the sense of its having no (sufficient) cause but coming about nonetheless. Of course that event has a cause; in fact, it has an agent-cause: me; and if I so choose, my rational agency will become apparent over, say, 1000 repetitions of my door-choice in a non-random pattern of frequency—which pattern, whatever pattern it is (it might be a very artful one), will yet not violate what is required by the objective A-B-probability in the long run.

NOTES 1. Note that motionless rests are here also counted among the movements (as limiting cases). 2. The expression (N*)X/I results from the expression (N*) by replacing “I” (everywhere in (N*)) by “X”. If “I” does not occur in (N*), then (N*)X/I = (N*). 3. The shortest such chain has two links: the event which is the beginning of the chain and the event which is the end of the chain, the fi rst event causing the second. In general, being-a-cause-of is the relation which connects each link of a causal chain to the links that follow it in the chain in the direction of time (i.e., from earlier to later). 4. If I resolve that eight hours from now I shall jump into the pool, then I have made that decision merely premeditatively, not effectively. If after the eight hours have passed, I say “now” and therewith jump into the pool, then I have made another decision effectively (which, however, fulfi lls, so to speak, the promise of the earlier decision). 5. In agent-causation, the time of causation (if, indeed, one insists on requiring such a thing) is always the time of the effect. 6. I examine the causal closure of the physical and defend non-physical causation—of the event-causal and/or the agent-causal sort—in several of my publications; see Meixner (2004, 2006, 2008, 2009). 7. For a typical disparagement of agent-causation along these lines, see Dennett (2003:100).

The Emergence of Rational Souls 179 8. My own efforts in this regard can be found in Meixner (2004: ch. 8 and 9, 2008). 9. For this role of consciousness in the economy of animal survival, compare James (1950: 140–142). 10. For this concept, see Section 3 in this chapter. 11. See for this objection O’Connor (2002: 344). 12. Note that William James regarded consciousness itself (in which, according to the present view, souls are embedded) as a causally efficacious organ—a “selecting agency” of the animal, helping it in the struggle for existence; see James (1950:138–139). 13. See also Meixner (2004: 314–323) and Corradini (2008: 205–206). 14. See van Inwagen (2002: 168–175).

REFERENCES Corradini, A. (2008). Emergent dualism. In A. Antonietti, A. Corradini, & E. J. Lowe (Eds.), Psycho-Physical Dualism Today (pp. 185–209). Lanham, MD: Lexington. Dennett, D. C. (2003). Freedom Evolves. London: Allen Lane (Penguin). James, W. (1950). The Principles of Psychology. 2 vols., New York: Dover. Meixner, U. (2004). The Two Sides of Being. A Reassessment of Psycho-Physical Dualism. Paderborn: Mentis. . (2006). Consciousness and freedom. In A. Corradini, S. Galvan, & E. J. Lowe (Eds.), Analytic Philosophy Without Naturalism (pp. 183–196). London: Routledge. . (2008). New perspectives for a dualistic conception of mental causation. Journal of Consciousness Studies, 15, 17–38. . (2009). Three indications for the existence of God in causal metaphysics. International Journal for Philosophy of Religion, 66, 33–46. O’Connor, T. (2002). Libertarian views: dualist and agent-causal theories. In R. Kane (Ed.), The Oxford Handbook of Free Will (pp. 337–355). New York: Oxford University Press. Van Inwagen, P. (2002). Free will remains a mystery. In R. Kane (Ed.), The Oxford Handbook of Free Will (pp. 158–177). New York: Oxford University Press.

9

Are Deliberations and Decisions Emergent, if Free? Achim Stephan

1. INTRODUCTION Traditionally, the debate on whether or not free will is compatible with determinism did not connect to the debate on the mind-body problem. This has changed only recently, when several distinguished neuroscientists have claimed that the sub-personal processes of the brain, and not we, as persons, are “responsible” for who we are and what we do. Hence, it has become an issue how processes at the level of the brain relate to those at the personal level, and an account of how they are related might modify our understanding of human actions and decisions. My goal in this chapter is to shed some further light on this important debate by considering three different positions with regard to the problem of free will, namely libertarianism, compatibilism, and hard determinism, and to discuss how these positions relate to mind-body accounts and to the problem of mental causation. These aspects are often neglected by those who have contributed to the free will debate in the past. A rare exception is Timothy O’Connor, who carefully examined how the personal and the sub-personal level relate to each other with respect to putatively free decisions and deliberations (see particularly 2000: ch. 6). In particular, I will address the arguments of Wolf Singer, a neuroscientist, and two philosophers, Ansgar Beckermann and Geert Keil, each of who have been highly influential in the contemporary debate in Germany. Although they are representatives for completely different positions in the free will debate, what they each have in common is that their positions are compatible with physicalism in one way or another. Thus, Wolf Singer opts for a thorough determinist picture of the brain, which leaves no room for free decisions. He is a hard determinist. In sharp contrast, Geert Keil opts for an indeterminist perspective of the world that leaves room for genuine free decisions. He is a libertarian, however, without being committed to substantival dualism. In this respect, his position is similar to the one argued for by Robert Kane (2002b). In contrast to both Singer and Keil, Ansgar Beckermann takes the side of compatibilism. He develops a notion of free will that allows one to accept both free decisions and determinism.

Are Deliberations and Decisions Emergent, if Free?

181

Beckermann, Keil, and Singer also agree on three conditions that should be met for treating an action or decision as genuinely free. First, the choice that is made in a certain situation must depend on the agent. This criterion is often called the principle of origination: Beckermann dubs it the “authorship condition” (2005: 111), and Keil refers to it under the label “initial authorship” (“Ersturheberschaft”; 2007b: ch. 5.5). Second, the choice that is made in a certain situation should be open for rational considerations. This criterion is often called the principle of intelligibility. For Keil freedom involves the agent’s capability to deliberate before deciding and acting, but he also allows that some of our actual free actions may not be thought through in the right way (2007b: ch. 5.4).1 Beckermann states that our decisions are free, “if and only if, they rest on processes that can be influenced by rational arguments and considerations” (2005: 120). Singer, too, acknowledges the principle of intelligibility citing the common sense view: “We judge decisions as being free, which are based on [ . . . ] the rational deliberation of contents accessible to consciousness” (cf. 2005: 156; my translation). The third condition requires that a decision that de facto is made in a certain situation could have been a different one. This criterion is often called the principle of alternative possibilities. While Keil refers to it under exactly this label (2007b: ch. 5.1), Beckermann calls it the “could-have-done-or-chosen-otherwise condition” (2005: 111). Singer does not explicitly refer to the principle of alternative possibilities; however, he seems to implicitly invoke it when he denies that in given situations we could decide differently than we in fact do (2006).

2. THE PERSONAL LEVEL AND THE BRAIN LEVEL Let us now consider how the three authors address the free will problem. Beckermann develops his position from the personal perspective; in his focus are the agent’s deliberations and decisions. If those are sensitive to arguments and rational considerations, they are classified as free. Here is an example to illustrate his position. Think of yourself as having decided to withdraw an article you promised a year before to the editors of a collection of essays on the free will problem. You don’t find the time to carefully elaborate your ideas, and you think: “Better no publication than a lousy one”. Now you receive an email from the editors, in which they explain the importance of your contribution for the volume and that you will be allowed more time to finish the article. You think everything through again and find a time slot that allows you to finish your paper to your satisfaction. Eventually you withdraw your withdrawal. A decision like the one reported meets Beckermann’s standard for free decisions. You could have done otherwise, namely stuck to the first withdrawal, you are the author of the decision, it was up to you to revise the first decision, and you have been open to new arguments and considerations—which were given in this instance by the editors.

182 Achim Stephan But, or so you may ask, how can my conscious considerations influence what I do or decide, when all neuronal processes underlying these considerations are deterministic processes? Beckermann has found a tricky formulation to circumvent challenges like this typical one. Being a naturalist and willing to accept thorough determination at the neuronal level, Beckermann sees only two alternatives: “either it is the case that not all decisions rest on natural processes or there are natural processes that can be influenced by considerations and arguments” (2005: 121). He himself, clearly, opts for the second option. Thus, Beckermann reconciles our personal-level talk about arguments and considerations with our sub-personal level talk about neuronal processes by suggesting that our decisions rest on processes that can be influenced by arguments. What he assumes is that the processes underlying our decisions are neuro-physiological ones. Therefore each act of deliberation that results in a decision rests on some corresponding physical process or state. However, resting on physiological processes does not make decisions less sensitive to arguments, according to Beckermann. It only shows that brain mechanisms can take care of semantics, critical thinking, and rational considerations. Contrary to Beckermann’s compatibilism are the incompatibilist positions as held by Wolf Singer and Geert Keil, which, of course, are also contrary to each other. Singer, for example, claims that our distinction between free and non-free decisions, which seems to rest essentially on the idea that we have conscious access to the motives that lead to seemingly free decisions and actions, has no basis in fact if looked at from the perspective of the brain. There are, of course, behaviors we perform without having consciously thought about, and there are other behaviors where we consciously weigh arguments and deliberate about them to arrive at certain decisions. According to Singer, then, there is no real difference from the perspective of the brain that allows us to fundamentally distinguish between these two types of behaviors—with all the implied (social and evaluative) consequences. From the perspective of the brain, Singer cannot see a decisive difference between processes that are accompanied with conscious access to the motives we refer to in supposed free decisions and those which are not. According to Singer, one total brain state determines the next one; some of these processes reach the conscious level; most don’t. Therefore, from the perspective of the brain there would not be a real difference between the deliberative processes and the more automatic, non-deliberative processes. Since the “solutions” which are found at the brain level do not differ, in principle, in those cases where conscious deliberations are involved from cases where no conscious consideration is involved, Singer concludes that the distinction we draw on the basis of our own experiences (from a “participant’s perspective”) has no robust basis in fact. Hence, Singer opts for free will illusionism or free will eliminativism. It is, however, not at all settled how much weight we should give to bottom-up-type arguments, such as is delivered by Singer. In order to support

Are Deliberations and Decisions Emergent, if Free?

183

the claim that there is no difference in the brain processes in question, Singer only refers to “general principles” that govern the underlying brain processes. At issue are, however, remarkably different macro behaviors. Let me illustrate the situation by an example taken from chemistry: if we refer to “general principles” only, the same general quantum principles are at work in water and ice, and still there can’t be a greater difference between water and ice at their macro level. Have you ever tried to walk on water? In that case we would not claim that the well-established difference in their macro properties is illusory but would ask for an explanation that accounts for the different dispositions of water and ice with respect to their micro structure. Similarly, the task for the neuroscientist might be better understood as to provide us with an account for the different kinds of human behaviors. Here, too, the issue should not be to claim that the established distinction between free and non-free choices is spurious. To claim this Singer would need arguments which show that there is no counterpart at all on the neuro-physiological side that could account for the difference between so-called free and non-free decisions and actions—a difference we regularly notice at the personal level. The arguments Singer provides really show that conscious processes play a less significant role in human behavior than the common sense view allows and also many philosophers have assumed. There is in fact neuroscientific evidence that our decisions rest heavily on processes that proceed in specific areas of the brain which are not open to conscious access (e.g., the influences from the limbic system or from glands that release neurotransmitters). Thus, it depends on whom Singer really wants to argue against. Is it someone who thinks that our conscious decisions rely only on what we can consciously deliberate about? Or, is it someone who thinks that besides the factors that unbeknownst to us play an important role in shaping our behavior, conscious deliberations have also an influence on what we will decide? It seems that Singer mainly argues against the fi rst option. In contrast to Wolf Singer, Geert Keil takes the manifest image we have of ourselves as his starting point. And this image depicts us as free agents who really decide sometimes between different alternatives. Keil does not try to establish a libertarian position that is impossible to reconcile with science, not at all. Rather, he points to consequences we would have to accept at the neuro-physiological level: if there are non-deterministic processes at the personal level, then there should exist corresponding non-deterministic processes at the subpersonal level. We can illustrate his ideas with Figure 9.1. What we see here is a branching in the possible courses of subsequent events right after the one referred to by the pair (m1, p1). Here, the agent is depicted, according to the libertarian, in the situation of decision, in which she can chose at least some (here two) courses of action. Keil stresses that from a diachronic perspective the free agent has alternatives to decide and act: she can proceed either with what is called here (m 2 , p2) or with (m*2 , p*2). However, Keil also underlines that even free agents do not have

184 Achim Stephan

Figure 9.1 Keil’s libertarianism.

alternatives from a synchronic perspective. No agent can proceed with, say, what might be referred to by (m 2 , p*2) or by (m*2 , p2); that is, no agent is free to decide or act differently from what she does given the neuro-physiological processes (the neuronal correlates of these decisions and actions). Accordingly, Keil seems to acknowledge that mental events are physically realized while stressing that this realization relationship has nothing to do with diachronic determinism, and as such is not limiting our freedom: In fact, the can-do-otherwise of the libertarian is not a can-do-otherwise vis-à-vis an actual physical event, this would be absurd, but rather it is a can-do-otherwise by a given previous history. [ . . . ] Why should the fact that mental processes are physically realized, i.e. that something is happening in my brain while I am deliberating or wanting, endanger my freedom? Those who see a contradiction herein base their freedom indeed on dualism. (Keil 2007a: 277–278; my translation) Although the authors discussed thus far differ considerably with respect to their answers to the free will problem, they all agree that there is no variance in synchronous respects: neither Geert Keil, nor Ansgar Beckermann, nor Wolf Singer thinks that the very same physical event could go along with different decisions.

Are Deliberations and Decisions Emergent, if Free?

185

3. PHILOSOPHY OF MIND MEETS THE PROBLEM OF FREE WILL In the last section we saw that all authors we discussed thus far would probably subscribe to what I have elsewhere called the thesis of synchronic determination, according to which there can be no difference in a system’s systemic properties without some difference in the properties or in the arrangement of its parts (see, e.g., Stephan, 2005: ch. 3.4). When learning, however, that even one’s seemingly free decisions are synchronically determined by corresponding brain processes, one might feel a certain tension: how could an agent be treated as the author of her decisions and doings if what she does is micro-determined by brain processes? Ansgar Beckermann explicitly takes up these worries when he says: “If in biological creatures all decisions are based on neuronal processes— and it is exactly this that neuroscience seems to show—how then should these processes be influenced by rational arguments and considerations?” Such an objection, however, is not cogent for Beckermann. He sees only the two possibilities I quoted earlier: “Either it is the case that not all decisions are based upon neuronal processes, or it is the case that neuronal processes exist that can be influenced by considerations and arguments. [ . . . ] The fact that something is a neuronal process, does not exclude that the very same process is a process of consideration” (2006: 301–302; my translation). What we encounter in the apparent tension is an old problem for the philosophy of mind, although not often discussed within the debate about the problem of free will: the so-called “qua”-problem of mental causation: How could we cause what we do by a conscious decision qua it’s being a conscious decision? In Kane’s comprehensive 638-pages Handbook of Free Will (2002a) we fi nd, for example, exactly two sentences and one footnote on “mental causation”. Moreover, those philosophers who have contributed to the debate on free will have also been silent about the synchronous relationship of the personal level to the sub-personal level. But it is exactly this relationship that has come into focus through the contributions by the neuroscientists. Within the philosophy of mind, this relationship has been discussed particularly by reductionists and emergentists in terms of reductive explanation (for details, see Richardson and Stephan, 2009). But what does it mean for a property to be or not to be reductively explainable? Generally, we ask for reductive explanations when we want to understand why and how a certain entity instantiates a certain property, in fact a property that is only attributed to the system as a whole. The aim of each reductive explanation is to explain (or predict) a system’s having its dispositions and properties solely by reference to its components, their properties, arrangement, and interactions. For a reductive explanation to be successful, the property to be reduced must fi rst be functionally construable or reconstruable; second, it must be shown that the specified functional role

186

Achim Stephan

is filled by the system’s parts and their mutual interactions; and third, the behavior of the system’s parts must follow from the behavior they show in isolation or in simpler systems than the system in question (cf. Stephan 2006: 488–489). Particularly, Kim and Levine have explicitly argued for the fi rst step where we have to work the concept of the system level property “into shape” for reduction. Kim calls this the “priming procedure” in which we must construe, or re-construe the property to be reduced relationally or extrinsically (see 1998: 98). So, what we are looking for are functional characterizations of the properties to be reduced. Usually we refer to these properties by concepts that classify properties at the system level, where specific patterns bring the instantiation of a property to our notice. To propose such conceptual preparations has the aim to allow conceptual transitions from the level of components to the level of systems. If, however, the conceptual “priming procedure” fails, the corresponding reductive explanation fails, too. And if reductive explanations for some systemic property fail, and fail in principle, then the property targeted for reduction is irreducible and thus emergent in a strong sense. But it is not easy to get a grasp on the problem of free will from a synchronous perspective. It will not suffice to simply consider possible candidates for free decisions and actions, which would be some manifest human behaviors. Suppose we succeed in reductively explaining some seemingly free human behaviors by fi nding the neuro-physiological correlates that play exactly the causal role specified of them. Would we then have also succeeded in reductively explaining that they are truly free decisions or actions? No, not so. Note that there is no agreement about whether any particular behavior truly has the property of being a free decision or free action. Accordingly, there is no agreement about how we should construct or re-construct the causal role of the alleged property of being a free decision or action. Since we have no agreed starting point, we have to discuss the different positions concerning the problem of free will, that is, libertarianism, hard determinism, and compatibilism, separately. The causal role we are looking for depends, of course, considerably on what somebody thinks is essential for a decision or action to be free. Let us start with the libertarian position as characterized by Geert Keil. According to him, in the moment of making a free decision we have true options: as agents we have the capacity to decide to do this or that. Thus, besides the manifest action, which is presumed to be a free one, the causal role should comprise one or more alternative behavioral outputs. Now compare the presumed disposition of free agents to do this or that with the dispositions of an ice cube. The disposition of an ice cube to melt in a glass of single malt whisky, say, is synchronically determined by its microstructure, and it also can be reductively explained by reference to its microstructure. Whether or not it will melt depends, however, on the

Are Deliberations and Decisions Emergent, if Free?

187

environment. Analogously, it is reductively explainable that an ice cube that in fact melted in a glass of single malt whisky had a structure that would have stayed stable if it had remained in the fridge. For the libertarian, however, the situation is completely different. She does not say that under (perhaps slightly) different circumstances human agents are capable of doing something different from what they in fact do; rather, the libertarian claims that under exactly the same circumstances agents are sometimes capable of doing different things. For an ice cube this would mean that it sometimes is able to melt or not to melt under the very same circumstances, which seems quite strange for an ice cube. But it is also difficult to conceive of this in the case of human actions and decisions. Even in situations, in which we feel ourselves as free deciders, it is not at all certain whether we really have the ability to do alternative things under the very same circumstances. It might be nothing but an illusion. If, on the other hand, we in fact have the ability to do this or that under the very same circumstances, it would be an ability that escapes reductive explanation: how could a (brain) mechanism admit two (or more) different outcomes under the very same conditions? We do not know of any indeterministic mechanism. But if there were a genuine random generator in the brain, which could allow for different outcomes under the very same circumstances, we could not reductively explain its workings. So we get: either there exist libertarian free decisions and actions or not. If they exist, they escape reductive explanations, in principle. Hence, if there is such a property, being a (libertarian) free decision is an emergent property. Now, consider the “Singer”-type hard determinist. For her, neither mental causation nor genuine free decisions do really exist. Those behaviors that are thought to be fair candidates for free actions and decisions are synchronously determined by neuro-physiological processes and should be reductively explainable. What, interestingly enough, might remain a problem for the hard determinist is to reductively explain the subjective feelings agents have when they experience themselves as if performing free decisions. To reductively explain this kind of experience translates more or less to the problem of phenomenal qualities: sometimes, at least, we experience ourselves as the authors of our actions—as free agents. But these phenomenal experiences are themselves good candidates for emergent properties (cf. Stephan, 2004, 2006). And in that respect, the subjective feeling of being a free agent might be a strong emergent property, too. Then, what about the compatibilist? For her, what reductive explanations should provide us with is an account of how the processes that underlie our seemingly free actions and decisions are sensitive to arguments and rational deliberations. Thus, the issue is not to provide a mechanism for two alternative world runs, but for how we could literally be said to weigh arguments in a rational way in order to draw informed decisions when these behaviors are synchronically determined by brain processes. The compatibilist has no problem to accept that the same circumstances always

188

Achim Stephan

lead to the same results. She only wants to make sure that we as free agents are open to rational considerations and further arguments. In this case, the problem of free will transforms into the problem of mental causation. The traditional answer is to opt for the identity theory in one way or other.2

4. FINAL REMARKS The human brain is an extremely complex system. Probably it is the most complex system we know of. It is constantly undergoing modifications and changes. No single brain state (or state of an organism) that might correspond to a seemingly free decision will occur again; and if it could, we would not be able to notice that it did. There is no chance to know. Therefore, there is no and there will be no experiment that could help to decide between libertarians, hard determinists, or compatibilists. Although both compatibilists and hard determinists might reach the conclusion that “decision procedures” can be reductively explained in principle, this endeavor might be completely utopian. If, as it is in dynamical systems, small changes can make a huge difference—we will never be able to foretell from the brain’s bottom up perspective what decision we will eventually come up with. The question, thus, is how much weight we are ready to give to the phenomenological perspective (or manifest image), on the one hand, and how much weight we are ready to give to the neuroscientific perspective, on the other. We experience ourselves as the authors of our decisions since we have the impression that it is we ourselves who reason, that it is we ourselves who consider pros and cons, and that it is we who decide and act on the basis of weighing reasons. Shall we go with this phenomenology as does Geert Keil, or shall we go with the neuroscientist’s generalizations as suggested by Wolf Singer? I am not sure where the compatibilist ends up; she seems to try both: marry the manifest with the scientific image of ourselves. But she, too, or so it seems, has to give up essential ideas about ourselves . . . What we should be ready to agree on is that there are far more factors contributing to what we think, decide, and do than we might have thought before: even than we can consciously be aware of and eventually influence. This however, does not solve the problem of free will. It will remain unsolved. Under these circumstances, it looks a bit strange to me how farreaching conclusions (e.g., ones that suggest deep changes in the legal system) are so easily drawn by some of our leading neuroscientists.

NOTES 1. A standard objection against libertarians maintains that it would be irrational if someone could decide differently under the very same conditions (i.e., the same considerations). Keil replies that we always have the possibility to think things through again (2007b: 110–115; cf. also 2009: 88–92).

Are Deliberations and Decisions Emergent, if Free?

189

2. Until recently, Beckermann was silent about how to explain the mechanisms that underlie what we phenomenologically perceive as being conscious considerations of arguments. But in his new book, entitled Gehirn, Ich, Freiheit, he takes care of this problem (cf. 2008: 121–132).

REFERENCES Beckermann, A. (2005). Free will in a natural order of the world. In C. Nimtz & A. Beckermann (Eds.), Philosophy—Science—Scientific Philosophy. Main Lectures and Colloquia of GAP.5 (pp. 111–126). Paderborn: Mentis. . (2006). Neuronale Determiniertheit und Freiheit. In K. Köchy & D. Stederoth (Eds.), Willensfreiheit als interdisziplinäres Problem (pp. 289–304). Freiburg-München: Alber. . (2008). Gehirn, Ich, Freiheit. Neurowissenschaften und Menschenbild. Paderborn: Mentis. Kane, R. (2002a). The Oxford Handbook of Free Will. Oxford: Oxford University Press. . (2002b). Some neglected pathways in the free will labyrinth. In R. Kane (Ed.), The Oxford Handbook of Free Will, (pp. 406–437). Oxford: Oxford University Press. Keil, G. (2007a). Mythen über die libertarische Freiheitsauffassung. In D. Ganten, V. Gerhardt, & J. Nida-Rümelin (Eds.), Die Naturgeschichte der Freiheit (pp. 281–305). Berlin, New York: de Gruyter. . (2007b). Willensfreiheit. Berlin, New York: de Gruyter. . (2009). Replik: Freiheit, die ich meine. Erwägen, Wissen, Ethik, 20, 75–94. Kim, J. (1998). Mind in a Physical World. Cambridge, MA: MIT Press. O’Connor, T. (2000). Persons & Causes: the Metaphysics of Free Will. Oxford: Oxford University Press. Richardson, R. C., & Stephan, A. (2009). Reductionism (anti-reductionism, reductive explanation). In M. Binder, N. Hirokawa, & U. Windhorst (Eds.), Encyclopedia of Neuroscience (pp. 3395–3398). Heidelberg: Springer. doi: 10.1007/978–3-540–29678–2_4991. Singer, W. (2005). Selbsterfahrung und neurobiologische Fremdbeschreibung. Zwei konfl iktträchtige Erkenntnisquellen. In H. Schmidinger & C. Sedmak (Eds.), Der Mensch—ein freies Wesen? (pp. 135–160). Darmstadt: Wissenschaftliche Buchgesellschaft. . [im Gespräch mit Markus C. Schulte v. Drach] (2006). Hirnforschung und Philosophie: “Der freie Wille ist nur ein gutes Gefühl”. sueddeutsche. de—Ressort-Wissen. Retrieved from http://www.sueddeutsche.de/wissen/ artikel/113/74039/ Accessed: March 25, 2007. Stephan, A. (2004). Phänomenale Eigenschaften, Phänomenale Begriffe und die Grenzen Reduktiver Erklärung. In W. Hogrebe with J. Bromand (Eds.), Grenzen und Grenzüberschreitungen. XIX. Deutscher Kongress für Philosophie, Bonn, 23–27 September 2002. Vorträge und Kolloquien (pp. 404–416). Berlin: Akademie Verlag. . (2005). Emergenz. Von der Unvorhersagbarkeit zur Selbstorganisation (2nd ed.). Paderborn: Mentis Verlag. (1st ed., 1999. Dresden: Dresden University Press). . (2006). The dual role of ‘emergence’ in the philosophy of mind and in cognitive science. Synthese, 151, 485–498.

10 Is Emergentism Refuted by the Neurosciences? The Case of Free Will Mario De Caro

Emergentists have to defend their highly controversial view on several fronts, the most harshly fought of which is that between the emergentists and orthodox scientific naturalists. In the last years, in particular, an eliminativist strategy has become increasingly popular with orthodox scientific naturalists, which—if correct and generalized—could potentially annihilate emergentism (as well as all other forms of non-reductive naturalism). This chapter will discuss an interesting example of this strategy and propose some possible replies on behalf of the emergentists.

1. THREE EMERGENTIST WARS Emergentism has been characterized in different ways and comes in different versions.1 In general, however, it can be defi ned as the view according to which, at certain non-basic ontological levels of natural reality, properties ‘emerge’ (i.e., they have genuine causal powers and are ontologically and explanatorily irreducible to, and cannot be predicted on the basis of, the properties of the more basic, less complex ontological levels—on whose instantiation, however, the emergent properties still depend for their occurrence). 2 In general, the prototypical example of emergence concerns mental properties, insofar as they are supposed to emerge from physical properties. Sometimes, however, emergentists claim that in nature, besides mental properties, there are other cases of higher-level properties that ‘emerge’ from lower-level properties (the alleged emergence of the biological from the chemical-physical level is an example of that), (cf. Dupré, 1993). Emergentists explicitly refuse both ontological monism (since they endorse pluralism of properties) and explanatorily monism (since they claim that, in order to account for the emergent properties, the conceptual framework with which we explain the more basic ontological levels is in principle insufficient). At the same time, emergentists refuse the supernaturalistic idea that there exist upper-level properties—in particular, mental properties—that are completely independent of lower-level properties for their instantiation.

Is Emergentism Refuted by the Neurosciences?

191

In order to assess the ontological credentials of emergentism, the conceptual link between the notion of ‘emergence’ and that of ‘supervenience’ is crucial (cf. Howell, 2009). More specifically, several authors, such as Crane (2007), have argued that ‘global supervenience’ is a necessary, but not sufficient condition of emergentism (it is not sufficient, since in itself global supervenience is compatible also with some strikingly non-emergentist views of the mental, such as reductionism and eliminationism). 3 Dupré (2001: ch. 7, Forthcoming), Humphreys (1997), and O’Connor (2000), however, have criticized this view and developed some interesting emergentist proposals that do not interpret the emergence of mental properties on physical properties as a form of supervenience at all. Be that as it may, in advocating a view that tries to solve the mindbody problem in purportedly naturalistic but not reductionistic terms, the emergentists have to fight three different philosophical wars, which have very different rationales, stakes, and implications. In the fi rst war the emergentists are opposed by orthodox scientific naturalists, who, in a monistic spirit, claim that mental concepts can be reduced (ontologically or at least explanatorily) to physical concepts, if not eliminated altogether, at least in principle.4 The second emergentist war is fought against the supernaturalists, who believe that the mental does not need the physical for its instantiation, and consequently do not attribute to the notion of emergence any interesting explanatory role. 5 Finally, in the third war the emergentists are opposed by other non-reductive naturalists who—although sympathetic with the view that the mental, on the one hand, is ineliminable and irreducible to the physical and, on the other hand, requires the physical for its own instantiation—do not appeal to the notion of emergence, which they consider too murky, explanatorily empty or conceptually groundless.6 The conflict between the emergentists and the orthodox scientific naturalists is, by far, the most widespread. According to the orthodox scientific naturalist view, the mental does not represent an exception to the natural order of the world in any interesting sense—since this order, in principle, could be exhaustively described by the natural sciences as we now know them or by a recognizable extension of them. However, scientific naturalists disagree with each other as to why the mental is no exception to the natural order. According to some of them, this is because mental phenomena are reducible to physical phenomena, from an ontological point of view; according to others, this happens because they can be eliminated altogether from the correct description of the world. But it has to be said that nowadays the eliminationist version of scientific naturalism is gaining much credit, especially because of the support it allegedly receives from the neurosciences. If elimativism is correct, emergentism as well as all the other views that aim at shaping liberal naturalistic views of the mind are dead ends. In order to understand, fi rst, how eliminativism is supposed to gain ground by appealing to the neurosciences and, second, how the supporters of emergentism

192 Mario De Caro could respond to this form argumentation, it is useful to consider a concrete, recent application of the eliminativist strategy—one that concerns a key notion of the non-reductionist view of the mind: human freedom.

2. FREEDOM, NATURALISM, AND EMERGENTISM Jennifer Hornsby has defi ned an agent as ‘a person (or other being) who is the subject when there is action’ (2005: 18). From this follows the question, what does it means to be the subject of an action? According to Hornsby, being an agent implies ‘(i) possessing a capacity to choose between options and (ii) being able to do what one chooses’ (2005: 18). However, this definition, if certainly acceptable, opens the door on the most controversial philosophical question—that concerning free will. In order to be an agent, Hornsby’s defi nition implies, free will is required. But how exactly should the concept of free will be interpreted? And is it correctly applicable to humans? With regard to this discussion, the goals of the different versions of naturalism are somehow different. The orthodox scientific naturalists’ goal is to offer an adequate treatment of the concept of free will in scientific terms (i.e., one showing that this concept is either reducible to the conceptual apparatus of the natural sciences or illusory). Considering Hornsby’s defi nition of agency, if the goal of the orthodox scientific naturalist were to be reached (i.e., if the concept of free will were to be ‘naturalized’), one would have to conclude that, for centuries, philosophers have been attributing too much importance to the idea that humans, besides being natural objects, are also agents. Liberal naturalists of all sorts, including the emergentists, have a very different goal. They want to vindicate a full-blooded concept of freedom—and thereby the idea of agency as a whole—by proving that that concept is legitimate, ineliminable, and irreducible to the concepts of the natural sciences. Today many scientific naturalists are trying to use canonical naturalization strategies to domesticate the concept of freedom. These strategies can be divided into two main groups, depending on the claims that their advocates defend: a.

The concept of freedom (whether factually eliminable or not) is illusory. b. The causal powers traditionally called ‘free will’ are not illusory, but they can be fully explained in scientifically acceptable terms.7 At the opposite side of the naturalistic spectrum, on the liberal side, several interesting attempts have recently been made in order to rescue full-blooded conceptions of freedom. For example, Timothy O’Connor (2000) has advanced a proposal, which is of great interest here, since it

Is Emergentism Refuted by the Neurosciences?

193

conjugates agent causation with an emergentist view of the agents’ place in the natural world. In O’Connor’s view, the agents’ causal capacities depend, for their existence, on the microstructural properties of the brain. However, agents also have features—including the capacity of instantiating ‘top-down causation’, which for O’Connor is essential for free will— that are irreducible to the causal powers of the microstructural level. In this view, having the properties that subserve an agent-causal capacity is not something that directly produces an effect. Rather, this condition enables the agent to determine an effect—within a circumscribed range of possibilities, of course. The crucial question, then, is whether such a capacity, when exercised, is freely determined by the agent (later one may also ask how and when this capacity can in fact be exercised). This proposal is, of course, speculative. However, according to O’Connor, a correct conceptual analysis can prove that, if free will has to exist, agents must have special causal powers rooted in an indeterministic context; and the existence of these powers is up to empirical science to determine. In this light—by appealing to philosophers of science such as Nancy Cartwright, to physicists such as Ilya Prigogine, and to some innovative research in biology—O’Connor states that “contemporary scientific knowledge is sufficiently incomplete to not rule out an emergentist picture of some factors within some highly organized phenomena”.8 In this light, therefore, “the question of emergence may be settled only in the end game where completed theories are compared” (2000: 115). This sounds fair. Still, of course, it is not enough for the emergentists to argue that their view may be true and that only a complete scientific account of the world could be able to assess this pretence. They also have to argue that their views are better than the alternatives when measured against the evidence and the arguments that are available to us now. Therefore—in the absence of a complete scientific account of the world—O’Connor and the other advocates of the libertarian-emergentist view of free will have to face two main problems.9 The fi rst problem is conceptual in character, since they have to refute the traditional antilibertarian argument according to which indeterminism, far from being a necessary condition of freedom, only generates randomness.10 The second problem is empirical in character, because these authors have to prove that—in the light of the evidence available to us now—their accounts of freedom are preferable to the competing ones. This empirical problem can in turn be divided in two parts. The fi rst (a concrete example of which will be discussed in the next paragraph) presents emergentists like O’Connor with the critical task of proving that attempts made to reduce or eliminate the concept of free will by scientific-naturalist-oriented thinkers do not work. The second part of the empirical problem facing the emergentists (which will not be discussed in this chapter) is that of proving that the agents’ allegedly necessary causal powers should be explained in terms that, while non-reductionistic, still are naturalistically acceptable.

194 Mario De Caro 3. IS FREEDOM AN ILLUSION? As said, one of the tasks that O’Connor and the other emergentists have to face is to show that the attempts at reducing or eliminating the concept of free will do not succeed. In this perspective, it is interesting to assess the value of a very recent but already influential neurophilosophical proposal, whose ambition is to show that the traditional view of freedom is illusory. Since Benjamin Libet’s pioneering research (Libet, 1985), much neurophilosophical research has been dedicated to the problem of free will, generally in a reductionist or an eliminativist spirit.11 Not infrequently, however, this kind of research has been spoilt, both at the conceptual and at the methodological level, by false steps or real blunders. An interesting example of this kind of approach, and of its limits, has recently been offered by an article, meaningfully entitled “Unconscious Determinants of Free Decisions in the Human Brain” (Soon et al., 2008). It may surely be granted that the experiment discussed in this article has offered an interesting contribution to the debate concerning the neurological bases of human decisional processes. However, as we will see, differently from what is boldly claimed in its title, it can certainly be doubted that it throws any new light on the question of whether humans can really make free decisions. Soon et al.’s article is based on an experiment in which the subjects were asked to relax and then, “when they felt the urge to do so, they were to freely decide between one of two buttons, operated by the left and right index fi ngers, and press it immediately”. In the meantime, the subjects were asked to fi xate on the center of a computer screen where a stream of letters was running (in particular, they had to notice which letter was on the screen when they made their decisions). At the same time, the subjects’ brain activity was measured through a fMRI. According to Soon et al., the results obtained with this technique are significantly more impressive than the ones that Benjamin Libet had obtained with his famous pioneering experiments on volitional processes. In order to see that, it is useful to look at the abstract of their article: There has been a long controversy as to whether subjectively “free” decisions are determined by brain activity ahead of time. We found that the outcome of a decision can be encoded in brain activity of prefrontal and parietal cortex up to 10 s before it enters awareness. This delay presumably reflects the operation of a network of high-level control areas that begin to prepare an upcoming decision long before it enters awareness. (Soon et al., 2008: 543) As is well-known, Libet came to the conclusion that an unconscious electrical activity called ‘readiness potential’ precedes the awareness of the related volition to act of about 300 milliseconds, suggesting that it may play a relevant causal role in the production of the volition (a conclusion that gave

Is Emergentism Refuted by the Neurosciences?

195

hope to the project of accounting for decisional processes by only considering the corresponding neural level). In Soon et al.’s experiment, however, the interval between the unconscious neural processes that allegedly ‘encoded’ the outcome of the decision and the awareness of that decision was about thirty times longer than that claimed by Libet. Besides these amazing quantitative differences, there are at least two more reasons for thinking that the experiment described in this article improves upon Libet’s research. First, where for measuring the activity of the brain Libet only had the possibility to appeal to an electroencephalogram or EEG (which is the recording of the electrical activity produced by the firing of neurons), Soon et al. appealed to the much more sophisticated functional Magnetic Resonance Imaging or fMRI (which measures the haemodynamic response related to neural activity). Second, and more importantly, Soon et al.’s experiment was set so that a difficult objection that affected Libet’s experiment could not be repeated. From the beginning of the experiment, in fact, Libet had told his subjects which action they had to perform (i.e., flexing a finger); so the task the subjects were asked to perform was only that of deciding when to execute that action. In this way, however, Libet’s interpretation of his own experiment, according to which the conscious decision of flexing the finger was preceded, and arguably caused, by unconscious processes—that is, by the readiness potential—could be objected. This was because one could point to the moment in which the subjects had consciously decided to agree to perform the experiment as a relevant cause of the unconscious start of the readiness potential. Libet’s experiment, indeed did not give any reason to infer that that conscious decision was preceded by another unconscious process (one could, of course, speculate that that was the case, but no evidence for this claim was offered by the experiment). In Soon et al.’s experiment, however, the subjects were not told to make their own decisions at the onset of the experiment, but only that they would have to make it later, at the moment in which one of the two buttons needed to be pressed. In agreeing to participate in the experiment, therefore, the subjects only made a meta-decision that was different from the decision that, for the sake of the experiment, they had to make later. So this experiment was protected from the previously discussed objection.12 An important thing that should be noticed in the abstract of Soon et al.’s article is that the word “free” is put in scare quotes. This is a very interesting typographical feature, since it clearly signals that, in the discussion on free will, the authors sympathize for a particular version of ‘incompatibilism’— that is, the view according to which causal determination and freedom are incompatible. As is well-known, incompatibilism comes in two main different versions. According to the fi rst, called ‘libertarianism’, humans are free and—as long as they act and decide freely—they are not determined (since, in fact, there are indeterministic gaps in the course of the causal processes that generate actions).13 The second version of incompatibilism, often called ‘illusionism’, is the opposite of libertarianism. According to

196 Mario De Caro this view, we do not enjoy free will since our decisions and actions are causally determined—and perhaps, as suggested by Soon et al., are even predictable. In the latter perspective, of course, it makes very much sense to write the adjective “free” in scare quotes (whereas it does not make sense if one endorses the view that free will exists). So in their article, Soon et al. clearly advocate the illusionist version of incompatibilism—that is, they endorse the view that the subjective feeling of freedom does not have any objective correlate, that human free will is nothing more than an illusion (cf. Smilansky, 2000). However, if the main purpose of the article is clear, its line of argumentation is vulnerable to several objections. Some of these objections are well-known, since they have been raised also against other articles and books discussing free will in the light of neurophysiology, beginning with Libet’s seminal work. In this light, for example, it can be noted that the experiment by Soon et al. presupposes two extremely controversial assumptions: i) Neurological processes, on the one hand, and conscious processes, on the other hand, can be measured in analogous ways and put in a precise correlation; ii) All mental events that occur in the subjects and are relevant for Soon et al.’s experiment (i.e., deciding which button to press, becoming aware of the moment in which one makes that decision, fi xing the screen of a computer, being aware of the letter that is on the screen when one makes the decision) are simultaneous. One could also repeat against Soon et al. a criticism that Daniel Dennett raised against Libet (Dennett, 2003). According to Dennett, Libet assumed an obsolete Cartesian vision of the mind, according to which human conscious activities are performed in front of an elusive ‘I’, who witnesses and reports them. Furthermore, recently some general doubts have been raised with regard to the methodology of pattern recognition to which these kinds of experiments appeal in order to individuate the causal correlations that link the neural processes with specific mental activities and actions (Vul et al., 2009). As said, however, these are standard objections, frequently raised against experiments similar to that of Soon et al. However, even if one discarded those objections, and granted that Soon et al.’s experiment is methodologically sound, other specific objections could be raised against the interpretation of the experimental data offered in the article. First of all, as we have seen, the authors interpret their experiment as a clear indication that free will is illusory.14 It is very doubtful, however, that this is the case. In fact, even if one grants that the experiment is methodologically sound, it can be argued that all the traditional views of freedom (at least in some of their versions) can be reconciled with the fi ndings of this experiment.

Is Emergentism Refuted by the Neurosciences?

197

To begin with, certainly it is legitimate to use the evidence described in this article in support of the view that free will is an illusion, as done by Soon et al. In order to prove that this is the correct view, however, two arguments are necessary, one empirical and one conceptual. First, one has to show that our decisions are causally determined; second, one has to demonstrate that freedom is irreconcilable with causal determination. The article can be seen as an attempt to corroborate the former assumption;15 but not a word is said in favor of the latter. That assumption, however, should not be taken for granted, since many philosophers have challenged it with arguments that certainly are not obviously wrong.16 More interestingly, however, the fi ndings of Soon et al. are compatible also with the two major views of freedom, ‘compatibilism’ and ‘libertarianism’. In general, compatibilism—probably the most popular view about free will among Anglophone philosophers—states that acting freely amounts to nothing more than performing the actions that one intended, desired, or willed to perform, disregarding whether one’s intentions, desires, or wills were causally determined or not. However, according to some compatibilists (including Leibniz, Hume, J. S. Mill, and many contemporary philosophers), the agent’s intentions, desires, and wills have to be causally determined if freedom is to be possible at all. In the perspective of this view, often called ‘supercompatibilism’, then, it is not just that freedom is compatible with determinism, but actually requires it.17 So supercompatibilists could welcome Soon et al.’s experiment and argue, on its basis, that our decisions are determined and because of that we may well be free (whereas a regular compatibilist, according to whom freedom is simply compatible with determinism but does not require it, may consider this experiment simply irrelevant with regard to the free will issue). Moreover, and even more surprisingly, the findings of Soon et al. can even be seen as consistent with libertarianism. This is because the advocates of this view could appeal to the fact that, in this experiment, the accuracy of the experimenters’ predictions about which button the subjects will press is in the order of 60 percent. Undoubtedly this is a statistically meaningful figure, and it certainly would be interesting to determine why it holds; the 40 percent gap, however, leaves open the possibility for libertarians to argue that that inaccuracy not only is due to our (perhaps contingent) epistemic limitations but is also a meaningful consequence of the objectively indeterministic causal structure of the world.18 It has to be noted that Soon et al.’s experiment adds something interesting to the huge literature concerning the inaccuracy of the conscious reports of our own mental lives. However, as proven by the title and abstract of their article, the authors seem much more interested in dealing with the free will problem—and this is, perhaps, because they assume that the evident lack of awareness with which we perform certain actions implies that those actions are not performed freely. As to this assumption, however, it is important to notice that there are cases in which we interpret certain actions of ours

198

Mario De Caro

as free even if they are performed at unawares. An example, which just happened to this writer, can clarify this point. I was cooking pasta and just before adding salt to the boiling water, my cellular phone rang. I answered and talked for a couple of minutes; then I hung up and added the salt to the water. Too bad, the pasta came out salty. The reason, of course, was that I had already salted it while I was talking on the phone. I certainly feel that, when I salted the water for the fi rst time, while talking on the phone, I performed that action freely (because adding salt to the water was exactly what I wanted to do and I had the precise intention to carry out that action immediately). Subjectively, then, my salting the water was a free action; still, I had no awareness of my having performed it. In general, it happens frequently that we are not aware of performing actions that we would consider free, and this is particularly true of actions that are not particularly important for us or that we have performed many times in our lives, such as putting salt in the pasta water or—as is the case in Soon et al.’s experiment—pressing a button. In one word, the actions that we judge to be free do not necessarily require awareness. Then, by the logical principle of contraposition, the performance of an action without awareness does not imply that that action is unfree. Therefore, experiments like Libet’s or Soon et al.’s—which are certainly relevant for the discussion about the limitations of awareness in human decisional processes—do not support any particular conclusion in the discussion of the free will question (let alone that they are crucial). However, Soon et al.’s interpretation of their own experiment is also exposed to more powerful objections. First, as seen, the main claim of their article is that some, and possibly all, actions that appear to be free from our perspective (i.e., actions that are subjectively free) are not free at all objectively. It can be argued, however, that the experiment presented in the article does not concern free decisions at all—and this not only in the objective sense, but even in the subjective sense. In order to see why those decisions are not free, even in the merely subjective sense, it is enough to look at the description of the instructions given to the subjects, as they are described in the article. In fact, the subjects were requested “to freely decide between one of two buttons”, when “they felt the urge to do so” (my italics). Feeling the urge to do something, however, is neither a necessary nor a sufficient condition of a subjectively free decision.19 It is not a necessary condition, since, in the vast majority cases in which agents believe they have freely performed an action, they have felt no urge to perform that action. (You have not felt the urge to read the phrase you have just read; but, at least subjectively, you would not consider that action unfree.) But feeling the urge to perform an action is not a sufficient condition either, since very frequently, when we do something after having felt the urge to do that, far from feeling that we have done it freely, we feel that we have been constrained. This happens, for example, when one feels the urge to sneeze or yawn in front of an interlocutor or when an akratic person goes for the next, very unhealthy

Is Emergentism Refuted by the Neurosciences?

199

glass of gin or, more spectacularly, when a kleptomaniac steals because of her neurotic condition. Normally, in these cases we would not say that the agent who has sneezed, yawned, drunk or stolen has acted freely, even if he or she has felt the urge to do so. (Actually, in terms of attributing freedom, the stronger the urge an agent feels to do something in cases such as these, the weaker their ability to act freely actually is.) And this proves that the set of actions that are preceded by subjectively free decisions has (at most) a small overlap with the set of actions that are preceded by an urging need to perform them. Undoubtedly, then, the latter actions should not be taken as prototypical examples of the former—as is instead done in the article we are discussing here. There is no reason, then, to think that Soon et al.’s “free decisions” are really free, not even in the merely subjective sense. What is worse, however, is that the alleged decisions discussed by Soon et al. are not even genuine decisions. This is because—as uncontroversially assumed by both decision theory and common sense—one can talk of a decision only when, in choosing between alternatives in a situation of uncertainty, an agent considers her preferences. In the context of Soon et al.’s experiment, however, the subjects are selected exactly because they do not have any preference between pressing the right or left button. In cases like that, in which nothing relevant for the agents is at stake by defi nition, it is extremely plausible that the subjects press either button automatically, paying no attention to which “decision” could be better for them—exactly because no genuine preference is at stake and there is no best or worst option at all. The upshot of this, then, is that Soon et al.’s “free decisions”, besides not being free, are not even genuine decisions. It seems safe to conclude, therefore, that the main claim of this article—which can be taken as a good example of a very common strategy for disproving our intuitive belief in freedom—is substantially ungrounded. Arguably, objections similar to the ones raised here could be moved against many of the other fashionable attempts to appeal to the neurosciences to reduce or eliminate the concept of freedom and, more generally, all mentalistic concepts. 20 In general, at least for what we know now, neurophilosophy does not seem to offer any convincing reasons to think that emergentism (and, in general, the forms of non-reductive, liberal naturalism) are erroneous. Of course, it is fair to add that this is no proof that those views are correct. It is up to their proponents, then, to find convincing arguments for showing that they are. 21

NOTES 1. Cf. Bedau and Humphreys (2008), Clayton and Davies (2006), Crane (2001, 2004), Humphreys (1997), Kim (1999, 2006), O’Connor and Wong (2006), Stephan (1999).

200 Mario De Caro 2. It should be noticed that sometimes this view is presented in terms of emerging substances, instead of emerging properties, but this is not very common anymore (see O’Connor & Wong, 2006). The kind of dependence of the emergent properties on the properties of the less-complex levels of reality is complicated to spell out; certainly, however, it is crucial in order to distinguish emergentism from some supernaturalistic views such as Cartesianism. Crane (2007: 198) cashed out the notion of emergence in terms of supervenience, when he wrote that “emergent properties of macroscopic objects are dependent on the properties of their parts, in such a way that there is no variation in the object’s macro-properties without variation in its parts’ micro-properties”. As noted later in this chapter, however, some emergentists deny that the notion of supervenience is of any help here. 3. Global supervenience concerning the mental is the thesis according to which two physically identical worlds cannot be different in any of their mental properties. On this issue, cf. O’Connor and Wong (2006). 4. See, for example, Kim (1999, 2006). A manifesto of the anti-monistic spirit of emergentism is Dupré (2004). 5. Supernaturalism is defended, for example, in Goetz and Taliaferro (2008). 6. Forms of naturalism that are, at the same time, non-reductive and non-emergentistic have been developed, for example, by P. F. Strawson (1985) and J. McDowell (2004). It is worth noticing that on the third emergentist front a civil war is fought, because the emergentists and the other non-reductionist naturalists share several fundamental assumptions, including the idea that a middle way can, and should, be shaped between orthodox scientific naturalism (which attempts at reducing or eliminating all agential notions) and supernaturalism. For some proposals in this direction, cf. De Caro (2010) and De Caro and Macarthur (2004, Forthcoming). 7. Another group that one could consider is composed by the so-called “mysterians” (which include Noam Chomsky, Peter van Inwagen, and Colin McGinn), according to whom the problem of freedom is unsolvable, at least for human beings, because of their biologically grounded cognitive limitations. 8. O’Connor (2000: 114–115). O’Connor does not, but with regard to this issue, one could also mention Dupré (1993). 9. Another view of this kind is proposed by Duprè (2001: ch. 7). 10. Today, following a suggestion by Peter van Inwagen (1983), this argument (already known, in its basics, to medieval philosophers as well as to Hobbes and Hume) is frequently called “the Mind argument”, after the British journal with the same name that has published many articles that developed it. 11. An informed summa against the many naturalization projects that appeal to the neurosciences as their rationale is Bennett and Hacker (2003). 12. Perhaps, however, that objection can be reformulated so that it can also work against Soon et al.’s experiment (2008). One could argue, in fact, that the subjects of that experiment know from the beginning which choice they will have to make later—and a very basic one, indeed, such as choosing with which fi nger to press a button. Therefore it is not absurd to speculate that, subconsciously, the subjects make their decision about which fi nger to use before the start of the neural processes that allegedly “encode” the outcome of that decision, and that it is this subconscious decision that causes the activation of the neural processes. In this light, the reported objection against Libet can be reformulated also against Soon et al.’s experiment, since the subconscious decision of which fi nger to use is clearly caused by the preliminary conscious decision that the subjects make when they agree to do what they are told by the scientists. In this reading of the experiment, therefore, the remote cause of the fi nal decision is still a conscious one.

Is Emergentism Refuted by the Neurosciences?

201

13. For a fi rst orientation of libertarianism, cf. Kane (2001: part vi). 14. In this sense, Soon et al. differ from Libet, who notoriously claimed that humans at least have the power to veto the actions that otherwise they would be unconsciously moved to perform. 15. One could note, by the way, that in order to argue, on a scientific basis, that our decisions and actions are causally determined, there is no real need for new neuro-physiological data, since for a long time we have had enough evidence of that from other fields—in particular from general relativity (which, according to the vast majority of scientists, holds in the mesoscopic and macroscopic world, notwithstanding the indeterminism of quantum mechanics). This does not mean, of course, that in itself the claim that there is something such as neuro-physiological determination of our decisions is not interesting or informative. It only means that if one has the specific goal of proving the illusionist view about free will, the most interesting task is to argue for the conceptual assumption that freedom is irreconcilable with causal determination, since arguably the required empirical evidence is already out there (even if, of course, the possibility of arguing that the physical evidence in this sense is either not conclusive or irrelevant is open to the advocates of libertarianism). However, as said, Soon et al.’s article does not even mention the fact that the conceptual assumption of the irreconcilability of free will and causal determination is necessary to prove the illusionist view. 16. For a fi rst orientation, see Kane (2000: part iv). 17. This is because in these authors’ opinion, the lack of determination would only generate randomness, which is obviously incompatible with freedom. It can be noticed that this is, in a nutshell, the Mind argument mentioned earlier, n. 10. 18. Kane (1996, 2005) and Searle (2004) develop two interesting causal-indeterministic views of free will (according to Kane (2005), in the human brain there may be chaotic processes that “magnify quantum indeterminacies in the fi rings of individual neurons”; 134). As to this view, Flanagan (2002, 121) has written that “There is work nowadays in chaos and complexity theories and in self-organizing dynamical systems theory that suggests that the human nervous system operates, at least sometimes, in ontologically indeterministic ways”. He adds, however, that it is unclear whether such indeterminism is ontologically based or depends on our cognitive limitations. 19. Bennett and Hacker (2003: 228–231) raised a similar criticism against Libet’s experiment. 20. An enlightening analysis of what explanatory potential the neurosciences may have with regard to the free will issue is developed in Roskies (2006). See also Mele (2007). 21. My thank to the audiences of Milan (Cattolica and Statale), Paris “Jean Nicod”, Locarno, Codisco (Messina), Florence, Padua, Holy Cross (Rome), where I gave papers connected with the present chapter. I am also grateful to Michele Di Francesco, Andrea Lavazza, and Stephen White for discussions on the issues treated here.

REFERENCES Bedau, M.A., & Humphreys, P. (Eds.) (2008). Emergence: Contemporary Readings in Philosophy and Science. Cambridge MA: MIT Press. Bennett, M. R., & Hacker, P. M. S. (2003). Philosophical Foundations of Neuroscience. Blackwell: Oxford.

202 Mario De Caro Clayton, P., & Davies, P. (2006). The Re-Emergence of Emergence. Oxford: Oxford University Press. Crane, T. (2001). The significance of emergence. In C. Gillett & B. Loewer (Eds.), Physicalism and Its Discontents (pp. 207–224). Cambridge: Cambridge University Press. . (2004). Summary of Elements of Mind and replies to critics. Croatian Journal of Philosophy, 4, 223–240. . (2007). Cosmic hermeneutics vs. emergence. The challenge of the explanatory gap. In M. Beaney, C. Penco, & M. Vignolo (Eds.), Explaining the Mental. Naturalist and Non-Naturalist Approaches to Mental Acts and Processes (pp.192–206). Newcastle: Cambridge Scholars Publishing. De Caro, M. (2010). Varieties of naturalism. In G. Bealer & R. Koons (Eds.), Waining of Materialism (pp. 365–374). Oxford: Oxford University Press. De Caro, M., & Macarthur, D. (Eds.) (2004). Naturalism in Question. Cambridge. MA: Harvard University Press. . (Forthcoming). Naturalism and Normativity. New York: Columbia University Press. Dennett, D. (2003). Freedom Evolves. New York: Viking Press. Dupré, J. (1993). The Disorder of Things. Cambridge, MA: Harvard University Press. . (2001). Human Nature and the Limits of Science. Oxford: Oxford University Press. . (2004). The miracle of monism. In M. De Caro & D. McArthur (Eds.), Naturalism in Question (pp. 36–58). Cambridge. MA: Harvard University Press. . (Forthcoming). How to be naturalistic without being simplistic in the study of human nature. In M. De Caro & D, Macarthur (Eds.), Naturalism and Normativity. New York: Columbia University Press. Flanagan, O. (2002). The Problem of the Soul. Two Visions of Mind and How to Reconcile Them. New York: Basic Books. Goetz, S., & Taliaferro, C. (2008). Naturalism. Grand Rapids, MI: Eerdmans. Hornsby, J. (2005). Agent. In T. Honderich (Ed.), The Oxford Companion to Philosophy (p. 18). Oxford: Oxford University Press. Howell, R. J. (2009). Emergentism and supervenience physicalism. Australasian Journal of Philosophy, 87, 83–98. Humphreys, P. (1997). Emergence, not supervenience. Philosophy of Science, 64, S337–S345. Kane, R. (1996). The Significance of Free Will. Oxford: Oxford University Press. (Ed.). (2001). The Oxford Handbook of Free Will. Oxford: Oxford University Press. . (2005). A Contemporary Introduction to Free Will. Oxford: Oxford University Press. Kim, J. (1999). Making sense of emergence. Philosophical Studies, 95, 3–36. . (2006). Being realistic about emergence. In P. Clayton & P. Davies (Eds.), The Re-Emergence of Emergence (pp. 189–202). Oxford: Oxford University Press. Libet, B. (1985). Unconscious cerebral initiative and the role of conscious will in voluntary action. Behavioural and Brain Sciences, 8, 529–566. McDowell, J. (2004). Mind and World. Cambridge, MA: Harvard University Press. Mele, A. (2007). Free will: Action theory meets neuroscience. In C. Lumer & S. Nannini (Eds.), Intentionality, Deliberation and Autonomy: The Action-Theoretic Basis of Practical Philosophy (pp. 257–272). Aldershot: Ashgate. O’Connor, T. (2000). Persons and Causes: The Metaphysics of Free Will. New York: Oxford University Press.

Is Emergentism Refuted by the Neurosciences?

203

O’Connor, T., & Wong, H. Y. (2006). Emergent properties. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy. Retrieved from http://plato.stanford. edu/entries/properties-emergent/. Accessed: July 25, 2009. Roskies, A. (2006). Neuroscientific challenges to free will and responsibility. Trends in Cognitive Sciences, 10, 419–423. Searle, J. (2004). Freedom and Neurobiology: Refl ections on Free Will, Language, and Political Power New York: Columbia University Press. Smilansky, S. (2000). Free Will and Illusion. New York: Oxford University Press. Soon, C. S., Brass, M., Heinze, H.-J. & Haynes J. D. (2008). Unconscious determinants of free decisions in the human brain. Nature Neuroscience, 11, 543–545. Strawson, P. F. (1985). Skepticism and Naturalism. New York: Columbia University Press. Stephan, A. (1999). Varieties of emergentism. Evolution and Cognition, 5, 49–59. Van Inwagen, P. (1983). An Essay on Free Will. Oxford: Oxford University Press. Vul, E., Harris, C., Winkielman, P., & Pashler, H. (2009). Puzzingly high correlations of fMRI studies of emotion, personality, and social cognition. Perspectives on Psychological Science, 4, 274–290.

Part III

Physics, Mathematics, and the Special Sciences

Part III Introduction Antonella Corradini and Timothy O’Connor

The concept of emergence appears to have application to a wide variety of phenomena in nearly every scientific domain. Developing a single precise, and fully useful characterization of the abstract concept is made difficult, if not impossible, as a result. However, in every disciplinary domain, ostensibly emergent phenomena are correlated with two fundamental aspects of scientific knowledge, both of which have been taken into account in Parts I and II. On the one hand, emergence pertains to the ontology of the objects constituting the interpretation model of the scientific theory; on the other hand, emergence has implications for the way in which one secures epistemic access to an ontologically emergent domain. A system is usually deemed as emergent, with respect to other systems or subsystems by which it is constituted, if its structure and the dynamical laws that characterize it are not derivable from the structure and laws of the lower systems or subsystems. Epistemic access to the ontology of an emergent system would require adequate cognitive capabilities, ones that must necessarily encompass concepts such as evidence, intuition, information, and perception, and one might well contend that one or more of these capacities are ontologically emergent. Both the ontological and the epistemic point of view about emergence are well represented in the contributions of Part III. The fi rst chapter in Part III, by Patrick McGivern and Alexander Rueger, deals with the nature of emergence in physics. The authors start from the assumption of a relatively non-controversial set of pre-theoretical features of their target emergent phenomena and seek to make them more precise on the basis of their mathematical characterization within physics. McGivern and Rueger take their examples from classical physics and concentrate fi rst upon the behavior of an undamped oscillating system, as an example of diachronic emergence. Secondly, they consider the treatment of steady-state heat conduction in a one-dimensional rod as an example of synchronic emergence. Both examples show that even at a very basic level there are (in one clear sense) novel and irreducible phenomena, phenomena which are irreducible to the extent that they are not only novel but qualitatively novel. Novelty is characterized in terms of topological differences between the representations of a system’s behavior before and after a control parameter

208

Antonella Corradini and Timothy O’Connor

has reached or exceeded a certain critical value. In assessing the chosen examples of emergence, the authors focus upon three main philosophical issues: whether emergence is to be understood ontologically or epistemically; whether emergent phenomena supervene on lower-level phenomena; and what the nature of emergent causation is. While the authors maintain that emergence should be viewed as an ontological feature of the physical world, they restrict supervenience to only synchronic emergence. As regards the third issue, McGivern and Rueger, in accordance with their ontological understanding of emergence, acknowledge the existence of new causal powers that guarantee the reality of the emergent macro-phenomena. Moreover, they accept downward causation as typical of emergent causation, provided that its acceptance neither is incoherent nor violates supervenience. The three chapters that are included in the middle of Part III deal from different viewpoints with a series of topics concerning epistemic emergence. In his contribution regarding the emergent character of the intuition of truth in mathematics, Sergio Galvan examines the feasibility of applying the category of emergence to the hierarchy of formal systems and to the hierarchy of the evidence systems which are their correlates. This hierarchy is a direct consequence of the phenomenon of incompleteness, in its various forms that result from Gödel’s theorems. Hence the contribution explores the meaning of the most significant form of incompleteness resulting from Gödel’s theorems, namely the omega-incompleteness of arithmetic theories, and examines some methods able to remedy this form of incompleteness. In particular, two ways are explored to overcome the incompleteness of primitive recursive arithmetic (PRA): the fi rst consists in the use of an explicit Tarskian theory of truth and the second is based on a notion of truth as non-fi nitary evidentiability. In both cases, the determination of the theory’s consistency always requires the emergent character of truth from derivability; in the fi rst case, due to the non-fi nitary extensional character of the model-theoretical notion of truth (inasmuch as the objectual domain is understood in an actualistic sense) and, in the second, because of the abstract intensional character of the soundness proof within a nonconservative extension of PRA. To operate according to principles based on forms of evidence that are abstract and open to novelty is very different from operating in accordance with the computational logic of a machine. The chapter shows how the mind follows its own paths in discovering the truth and the modalities of its justification, this being due to the fact that this undertaking cannot be restricted to the use of only concrete contents and their fi nite combinations. Galvan argues that it follows that a cognitive system endowed with the capability of intending the abstract is characterized by a relationship to the object which is not one of reception but of presence. The object is known insofar as it is present to the system and not as a mere source of stimuli elaborated by the system. This relationship shows the mark of the emergent novelty, which is the mark of the emergence of

Part III Introduction 209 consciousness, which in its turn is a pre-condition of the presence of an object to the cognitive system. An analogous point of view is argued for in Arturo Carsetti’s chapter, where attention focuses on the cognitive implications of the limitation results of theories, due not only to the Gödelian theorems but fi rst of all to the non-categoricity results. The author examines the meaning of the noncategoricity of fi rst-order languages in order to consider its consequences for the modalities involved in development of the cognitive apparatus and of the construction both of knowledge and self-knowledge. Carsetti begins by considering the fact that any fi rst-order arithmetic theory admits of nonstandard models. It is known that no nonstandard model can contain the set of the standard numbers. But how is this possible, if the domain of a nonstandard model is an extension of the standard domain? This is possible because a semantic model is not a piece of reality on its own, but it is cut out from reality through the cognitive instruments offered by language. Now, in the fi rst-order language it is not possible to defi ne the property of being a standard number, which means that, from the point of view of our cognitive capabilities as restricted within fi rst-order language, the set of the standard numbers is not the unique model of the language. Instead, there are many domains, standard as well as nonstandard, which are linguistically indistinguishable. Hence, it is necessary to employ a more complex language than the fi rst-order language, which yields the capability of grasping the meaning of a standard number. This language is the second-order language, which, as is known, is categorical. In this way, the emergence of the second-order upon the fi rst-order becomes evident. This is confi rmed by the fact that at the second-order level—due to Henkin’s theorem—there are also nonstandard models. However, the existence of these models is allowed only within the so-called non-full Henkinian models, that is, the models which do not include in the property-domain the very property of being a standard number. As we have seen, epistemic emergence is a plausible phenomenon, since the higher level is characterized by a set of properties or abstract objects, whose epistemic access requires new cognitive capabilities relative to those explainable through a causal theory of knowledge. But epistemic emergence can be understood in a weaker sense, as merely the point of view that guides the knowledge process. These are the most interesting forms of emergence from the point of view of the psychological sciences, which are discussed in Alessandro Antonietti’s chapter. According to the author, analyses carried out by psychologists of the difference between a set of sounds and a melody or between a set of sounds and a chord do not show that the melody represents an objectively emergent reality with respect to the set of the sounds (which are in their turn emergent relative to the corresponding vibrations of the bodies which transmit the sound itself). Instead, they bring to the fore the emergence of the subjective reality constituted by the particular way in which the cognitive apparatus of the subject perceives

210

Antonella Corradini and Timothy O’Connor

this set of sounds. The reception of these sounds is in fact determined by the emotive resonance they arouse and on which their evocative function depends. But these components are not part of the reality of the sounds itself, but of the way in which they appear to our subjectivity. The importance of subjectivity is decisive in creating the phenomenon of emergence, and this is one of the reasons why the author intends to include psychology in the conceptual frame of emergentism. At the beginning of his chapter, Antonietti lists some important representatives of the psychological sciences whose thought contains some traces of emergentistic ideas (Wilhelm Wundt, Wolfgang Köhler, Kurt Lewin, Jean Piaget, Lev Vygotskij, and Aleksandr Luria). He then aims at applying to the psychological context some features usually attributed to emergent phenomena, such as ontological heterogeneity, bi-directional causality, property-asymmetry, and epistemological distinction. The author chooses to concentrate upon the latter two characteristics, which he lets interact with the further distinction between synchronic and diachronic emergence. Kanizsa’s triangle, which is described on pages 270–273, is an example of synchronic emergence (whereby the concept of synchronicity is not equivalent to that of staticity). Looking for forms of diachronic emergence, the author presents a series of experiments carried out by his research team and pertinent to the domain of analogical thinking as applied to problem solving. Antonietti underlines how the experiments have refuted the theories according to which experimental subjects solve their analogical problems through an incremental process. Analogies, in fact, come to mind suddenly and not through a cumulative process. The same result is obtained with computer-based problems-solving setting. Moreover, further experiments have shown that the new qualitative aspect, proper of the emergent processes, must be traced back to awareness which, in turn, cannot derive from a mere summation mechanism. The fi nal conclusions of the chapter are that, in the psychological field, true emergent phenomena do exist and that their existence puts a limit to wholesale reductionism and, as a consequence, justifies the use of irreducible, genuinely psychological explanations. The last essay of this section, by Antonella Corradini, deals with issues such as the unity of science, the autonomy of the special sciences, reductionism, physicalism, and the role that the concept of emergence plays in the debate about these topics. Discussion of these themes is a basic task for emergentism, as Broad reminds us in the introduction to his (1925) book, whose main purpose was to analyze “the relation or lack of relation between the various sciences”. The debate takes place mainly at the epistemological level, but its implications for the ontological point of view are evident. Moreover, among all the special sciences, the chapter pays particular attention to psychology. The author develops her point of view through critical examination of three significant perspectives held in contemporary epistemological discussion. Thus, according to Jerry Fodor, three theses are entailed by reductionism: the generality of physics, token physicalism,

Part III Introduction 211 and reductionism itself (i.e., the idea that every natural kind predicate of a special science is related to a natural kind predicate of physics). Fodor maintains that, in order to safeguard the autonomy of the special sciences, the reductionist thesis should be given up, as a consequence of the validity of the multiple realization thesis. Besides the generality of physics, only token physicalism is needed to guarantee both the autonomy of the special sciences and the unity of science on a physicalistic basis. However, reacting to Fodor’s thesis, Jaegwon Kim points out that adoption of token physicalism leads to consequences which are undesirable for the supporters of the autonomy of the special sciences. Moving from assumptions also shared by non-reductive physicalists, Kim argues that reductionism comes up again through “local reductions” and that, as a consequence, sciences such as psychology are devoid of any disciplinary unity. In the author’s view, Kim’s conclusions show that, in order to safeguard the autonomy of the special sciences, token-physicalism needs to be abandoned along with reductionism. In the context of present-day philosophy of science, John Dupré’s perspective is taken as an example of a position which gives up both of these conditions along with the unity of science thesis, as traditionally understood. The alternative to physicalism and to reductionism is an epistemological and ontological pluralism, according to which the different domains and levels of reality display autonomous characteristics and autonomous causal powers. But how should these latter be conceived? Does downward causation fi nd its place in the picture? The author’s aim in the fi nal part of her chapter is to show that Dupré’s allegiance to a liberalized form of empiricism is incompatible with an autonomous form of mental causation as well as with the most typical characteristics of the human being as a personal agent. The conclusion is drawn that in the particular case of psychology as the science of the mental, the last of Fodor’s conditions, the generality of physics, should also be rejected. Instead, a strong form of emergent property-dualism should, as a minimum, be accepted.

11 Emergence in Physics Patrick McGivern and Alexander Rueger

1. INTRODUCTION Discussions of emergence in physics1 usually adopt one of two general approaches. Starting from intuitive desiderata a concept of emergence should satisfy, one can (i) develop a philosophical account of emergence that incorporates and makes precise the initial intuitions and then try to fi nd illustrations for this model in scientific practice; alternatively, one can (ii) study cases in science where the intuitive desiderata seem to be satisfied and then develop a philosophical model of emergence that subsumes those cases. Given these different approaches, disagreements over which phenomena count as emergent are likely, even if everybody were to agree on the pre-theoretical desiderata. Differing views about the latter, of course, compound the confl icts. Whether emergent phenomena, for instance, supervene on ‘base level’ phenomena or whether the ‘completeness’ of physics at the base level should be satisfied are questions which different philosophical views will answer in different ways. Here, we will follow the second strategy. We will start with a relatively uncontroversial set of pre-theoretical features of emergent phenomena and make them more precise by characterizing them in terms of features that we fi nd in the mathematical treatment of certain problems in physics. We can then check whether having made our intuitions more precise has resulted in an account of emergence that allows us to answer philosophically motivated questions. Unsurprisingly, many of these questions can be answered only if we adopt further assumptions that aren’t directly entailed by the accounts we extract from physics; for instance, assumptions about the nature of properties and causation will need to be made. Our aim is to make it clear what these assumptions are, how they can be justified in the context of our examples, and what their consequences are for several prominent questions about emergence. Thus, our main focus will not be on detailed arguments in favor of certain assumptions and against others, but instead on the nature of the assumptions needed to interpret our examples in various possible ways (although we indicate our own preferences).

214

Patrick McGivern and Alexander Rueger

The examples we choose from classical physics may at fi rst not appear to be striking enough to count as cases of emergence, especially when compared with the spectacular cases that are often mentioned in the literature, such as entangled quantum systems, phase transitions, or certain effects in solid-state physics. Our modest choice of examples, however, has its advantages: the technical details of the cases are less daunting, and although we do not have space to elaborate the point in detail, some of the basic features we discuss play important roles in the more complex cases as well. In assessing our examples, we will focus on three philosophical questions about emergence: (i) Is emergence ‘merely epistemic’ or are there examples of ‘ontological’ emergence in physics? (ii) Do emergent phenomena supervene on the ‘base’ phenomena from which they emerge? (iii) What is the nature of emergent causation? The rough idea behind the fi rst question is a distinction between situations where the appearance of emergent phenomena depends on our way of describing a system (e.g., in a fi ne-grained or coarse-grained way) and situations where the occurrence of emergence is independent of how the system is represented (for instance, by demonstrating that emergent phenomena are characterized by special causal powers). 2 Many emergentists have hoped to show that there are indeed cases where emergence has to be understood ontologically. Since in general our claims about ontological issues in physics must be based on features of physical theories, drawing the distinction between epistemic and ontological forms of emergence is not as straightforward as simply noting whether a particular feature is a characteristic of theories or not: Answering this fi rst question will require some further assumptions about what features of theories are ontologically significant, and when. For question (ii), the assumption that emergent phenomena supervene on their base phenomena is widespread, although there are dissenting views (e.g., Humphreys, 1997). We do not assume that supervenience is a requirement for emergence, but instead investigate whether it is found in each of the cases we examine. Question (iii) is the most elaborate of the three. There are three closely related worries about causation and emergence. The fi rst concerns whether or not emergent properties have distinctive causal powers relative to the causal powers of their base properties. Here the key problem is to distinguish between the sort of causal novelty associated with emergence and the ‘regular’ novelty associated with mere ‘resultants’. The second worry concerns whether or not emergent causal efficacy is even coherent: for instance, Kim (1999) argues that emergent causation is unavoidably linked

Emergence in Physics

215

with ‘downward’ causation, and that downward causation is incoherent. The third worry concerns the idea that whatever causal efficacy is found among emergent properties, base level properties will always compete with or exclude those emergent properties as genuine causes. We will examine each of these questions in detail. However, fi rst we will clarify the pretheoretic concept of emergence.

2. THE CONCEPT OF EMERGENCE Emergence is a relational concept: emergent phenomena are always emergent relative to some ‘base’ or reference phenomena from which they emerge. Emergence is also a contrastive concept: emergence involves phenomena that are ‘emergents’ relative to their base phenomena rather than ‘resultants’. An account of emergence should be able to accommodate both of these features, for instance by explaining how emergents are to be distinguished from resultants. We can distinguish between two general types of emergence, diachronic emergence, where an earlier state of the system, over time, gives rise to a later state which is classified as emergent with respect to the former, and synchronic emergence, where the reference phenomena coexist with putatively emergent phenomena (Rueger, 2000a). Classical accounts of emergence typically contain both types, often without distinguishing them. Again, an account of emergence should be able to accommodate both. Given that emergence is a relational concept, what are the relata? Does emergence involve the emergence of entities, properties, laws, behavior, or some combination of these? We take it that emergence must involve emergent behavior of some sort, since it is only through a system’s behavior that we ever have reason to suppose that there are any entities, properties, or laws at all. Hence, for the most part, we will talk of emergent behavior. However, since it is also common to speak of emergent properties, at times we will do this as well. In these cases, such properties can be understood in the sense of being the property of having a particular sort of behavior. With these clarifications in mind, the core criteria we fi nd associated with emergence are those of non-reducibility and novelty: emergent phenomena are in some way irreducible to and novel with respect to their base phenomena, whereas ‘resultant’ phenomena are reducible and/or non-novel. In fact, we see these two criteria as two sides of the same coin: emergent phenomena are typically taken to be not only novel but in some way ‘qualitatively’ novel, and talk of irreducibility often seems intended to capture just this distinctive feature. 3 With this basic concept of emergence in mind, we will now turn to specific examples in order to draw out the details.

216 Patrick McGivern and Alexander Rueger 3. DIACHRONIC EMERGENCE Consider a damped oscillating system with an equation of motion of this form:

which characterizes the three forces operating in the system: the inertial force (first term), the restoring force (kx), and the damping force . The solution of the equation (given sufficient initial conditions)—the integration of the equation over some time interval—describes the behavior of this system, that is, the distribution of properties (here: positions) of the system over time:

a series of oscillations with frequency ω and phase δ and with a gradually decaying amplitude A(t). With increasing time, the trajectory of the system in phase space will spiral down into the origin, the point where the motion comes to a rest. The system has a ‘focal point attractor’, a final state to which all trajectories lead, irrespective of which initial conditions they started from (Figure 11.1, left). Now imagine we decrease the damping in the system. This will not change the qualitative nature of the phase space portrait; it will only take longer for the system to arrive at the point attractor. This is true until we completely eliminate the damping and change the system into an ideal harmonic oscillator. At this point the attractor disappears from the phase space portrait and we see a ‘qualitatively different’ behavior of the system, regardless of how much time passes: the trajectory has turned into a closed curve (an ellipsis, or a series of concentric ellipses if we consider different initial conditions), also called a ‘center’. The system never comes to a state of rest and keeps oscillating. We have arrived at an undamped harmonic oscillator:

with solutions x0(t) = A cos(ωt - δ) in which the amplitude doesn’t gradually decay (Figure 11.1, right). This example illustrates a simple phenomenon that satisfies our criteria for emergence: the behavior of the undamped oscillator is both irreducible to and qualitatively novel with respect to that of the damped oscillator. Here’s why.

Emergence in Physics

217

Figure 11.1 Representative phase space trajectories for damped oscillator and undamped oscillator.

(i) Non-reducibility There’s no doubt that there’s a connection between the equations of motion for the damped and undamped system: the equations for undamped system are the result of setting the damping parameter in the damped equations to zero. So in some sense, there is a reduction between the two. But this isn’t the sense of reduction we’re interested in here. Rather than asking whether or not the equations of motion for the undamped systems can be derived from those of the damped system, we want to know whether the damped and the undamped systems exhibit the same type of behavior and, if not, in what sense their behaviors differ. Hence, what we are interested in are not the equations of the systems but their solutions which describe the behavior. For example, as damping decreases—but before it reaches zero—the system’s behavior will change: with lower damping, if left untouched, it would take longer to reach its equilibrium state than it would with higher damping. But these superficially different behaviors can still be seen as behaviors of the same type since the systems eventually wind up in the same state. Once damping reaches zero, though, the system’s longterm behavior becomes significantly different: as long as it is undisturbed, such a system will never settle into a stationary state. We can characterize this difference between the damped and the undamped systems in terms of the relation of limit reduction (Batterman, 1995; Nickles, 1973; Rueger, 2000b). For a successful limit reduction, the solutions of the damped equations would need to go over smoothly into the solutions of the undamped equations in the limit of vanishing damping, analogous to the way the solutions of certain equations in Special Relativity Theory go over into the relevant solutions of Newtonian mechanics in the limit v/c → 0. Intuitively, what we are asking when we ask about reduction in this sense is whether the behavior of the damped system becomes more and more like that of the undamped system as damping is continuously reduced. For limit reduction to be successful, we would need to be able to

218 Patrick McGivern and Alexander Rueger show that we could make the behavior of the damped system arbitrarily close to that of the undamped system by sufficiently reducing damping. More precisely, if we choose a measure of the distance between the damped and undamped solutions, ε> 0, then for a successful reduction, the damped solutions would have to stay within this ε-neighborhood of the undamped solutions, with decreasing damping (c → 0) and increasing time. However, limit reduction is not successful in this case: the behaviors of the damped and undamped oscillators are not connected by a continuous limit but rather by what is called a singular limit. This shows that the transition from the behavior of the damped system to that of the undamped one is characterized by a discontinuity. This irreducibility is reflected formally in the fact that the two limit operations relevant to the two systems—the limit of infi nite time represented in the phase space portrait (focal point attractor vs. center) and the limit of damping → 0 do not commute: Eq. 1 where

= x0(t). The limit on the right-hand side is 0 (the focal point),

while the limit on the left is not defi ned (since the system keeps oscillating for infi nite time).

(ii) Novelty In the damped system, any change in damping will lead to a change in the system’s behavior. However, there is a formal sense in which the behaviors of any pair of damped systems are similar: given the phase space portraits of any two damped systems, we can always find a smooth mapping from the one space into the other that preserves the phase space trajectories. Such a mapping will deform the trajectories without changing their topological features. In the case of the undamped system, however, this is not possible: there is no way to deform the trajectories of any damped system into those of the undamped system without ‘cutting’ the trajectories. More precisely, there is no homeomorphism, no one-to-one mapping continuous in both directions, between the phase space of the damped oscillator and the phase space of the undamped oscillator, that converts the spiral trajectories of the former into the elliptical trajectories of the latter. Such a mapping would connect the two portraits in a way that always mapped neighboring points in the one onto neighboring points in the other: the two portraits would then be said to be topologically equivalent, and the behavior of the one system could be seen as merely a quantitative variation on the other. But in the damped/undamped case, the two portraits are topologically inequivalent, and in this sense, the behavior of the undamped system is qualitatively novel with respect to that of the damped system. This is a standardly applied notion in dynamical systems theory which characterizes qualitative changes in the behavior

Emergence in Physics

219

of a system.4 We can illustrate this inequivalence as follows. In diagram 1 (see Figure 11.2), the horizontal arrows represent the evolution of behavior described by the phase space portraits of the damped and undamped oscillators, respectively. The vertical arrows represent a mapping h between the two phase space portraits. Topological inequivalence is indicated by the fact that no such mapping is homeomorphic on both the left- and the right-hand sides of the diagram: for instance, any mapping that succeeds at always mapping neighboring points to neighboring points on the left-hand side will inevitably map neighboring points to non-neighboring points on the right-hand side.5 We propose to adopt this way of understanding novel behavior in the context of emergence: novelty of behavior is to be characterized in terms of topological differences between the representations of a system’s behavior before and after a control parameter reaches or crosses a critical value (here: damping = 0). Merely quantitative differences (e.g., a shorter period of oscillation when damping is diminished) are not sufficient for the sort of novelty associated with emergence, as has often been pointed out. That a body of 10 kg can behave differently than a body of 1 g is not reason enough to call the heavier body’s behavior qualitatively different. Compare diagram 1 with a diagram that represents the non-commutation of the limits in Eq. 1 (Figure 11.3). That there is no appropriate map to be found for the ‘closure’ of diagram 1 corresponds to what we pointed out earlier in the discussion about non-reducibility: the limits t

Figure 11.2 Diagram 1.

Figure 11.3 Diagram 2.

220

Patrick McGivern and Alexander Rueger

→ ∞ and c → 0 do not commute. This suggests that there is a connection between our notions of non-reducibility and novelty. But they are not identical concepts. Though the limit notion of reduction implies a sense of novelty of behavior that can be characterized topologically, this is not the same concept that we defi ned with diagram 1. In the failure of limit reduction, the transition from the behavior of the damped system to that of the undamped one is characterized by a discontinuity. This discontinuity manifests itself formally when we study the topological space in which the solutions of the equations of motion figure as points. Our notion of reduction as uniform convergence of one solution to the other in the limit of vanishing damping imposes the ‘topology of uniform convergence’ on the space of solutions, a fairly ‘fi ne’ topology in the sense that it excludes many sequences of functions from the class of converging sequences. The occurrence of a singular limit, as in our example, means that we have to change the topology of the space of solutions so as to be able to deal with cases like the transition from damped to undamped behavior. We’ll see later, in the case of synchronic emergence, that there are ways of ‘regularizing’ the discontinuous limit behavior so that a uniformly valid approximation relation between ‘old’ and ‘new’ behavior can be achieved. Since these techniques, however, do not restore uniform convergence, they effectively introduce a new topology on the space of solutions. Since changes in the topology of the space in which we represent the behavior of a system are usually characterized as qualitative changes, the topological perspective on non-reducibility suggests an explication of ‘novel’ or ‘qualitatively different’ behavior in topological terms. How is this topological notion of novelty related to topological inequivalence? The latter concept was explicated earlier in terms of topological inequivalence of the families of trajectories in phase space, a purely topological or qualitative notion. To characterize non-reducibility, the discontinuous transition of old to new behavior, as a breakdown of uniform convergence in the space of solutions, we need a topological space that is also equipped with a norm or metric—a measure of distance between points. Hence this notion of novelty is not purely topological. In the case of synchronic emergence we’ll see a further application of this concept later. Perhaps this example seems too simple to illustrate an instance of emergent behavior. Where is all the ‘complexity’ of the system that one traditionally thought of as a condition for emergence? We seem to have a case in which the system actually becomes simpler. In response we have to say that the traditional fi xation on emergence as associated with complexity is a mirage. If the spectacular cases of emergence in phase transitions, solidstate physics, and quantum theory are bona fide instances of emergence, then they fulfi ll this role because they share the relevant features—novelty of behavior and non-reducibility in our sense—with our simple example (as we’ll indicate in the fi nal section).

Emergence in Physics

221

3.1. PHILOSOPHICAL QUESTIONS ABOUT DIACHRONIC EMERGENCE

(i) Epistemic vs. Ontological Emergence Is our example a case of diachronic emergence in the ontological or the epistemic sense? Recall that that distinction is supposed to turn on whether the appearance of whatever characteristics are indicative of emergence can be attributed to a shift in perspective in our description of a system. Ontological emergence requires a real change in the system, whereas epistemic emergence requires only a change in how the system appears from a different perspective. Since the emergent behavior of the undamped oscillator is brought about by a change in an actual parameter in the system (the damping), it seems most natural to view this as a case of ontological emergence. One way to understand this case as an instance of epistemic emergence would be to claim that the change in damping represents a change in the conditions of idealization, for instance as our interests shift to time intervals over which the effects of damping can be considered negligible. Note that though it may seem that undamped oscillatory behavior must be regarded as an idealization, the example of the undamped oscillator can be replaced by a more ‘realistic’ (though more complicated) system like the van der Pol oscillator. In these systems, damping does not have to be reduced to zero for novel behavior to occur: when the damping reaches a critical value, the system develops a new attractor, a limit cycle (see Rueger, 2000a). We return to the issue of idealizations briefly later in a more general context.

(ii) Supervenience Supervenience is commonly understood as a synchronic relation. In the diachronic case it would therefore seem inappropriate to ask whether the later state of the system supervenes on the earlier state. We could defi ne a diachronic supervenience relation and ask whether the state of the system at one (earlier) time necessitates its state at another (later) time, or whether the later state could vary without a change in the earlier state. The answer seems to be ‘no’: understood diachronically, supervenience fails in our example. The earlier, damped behavior alone doesn’t necessarily give rise to the later, undamped behavior: for this to occur, the damping must be eliminated. And the later, undamped, behavior could itself vary—for instance, reverting to damped behavior if damping is reintroduced—without any change in the earlier, damped behavior. But this is not surprising and has no bearing on the question as it is usually asked, namely, the question whether some property of the system supervenes on other properties of the system at the same time.

222

Patrick McGivern and Alexander Rueger

(iii) Causation Does the diachronically emergent system have distinct causal powers relative to the base system? In one sense this question seems trivial in the diachronic case: after all, the systems are distinct, they occur at different times, and the difference in causal powers seems no more mysterious than the difference in causal powers between a chicken and an egg. However, there is another way of asking about causal powers which leads to more interesting results in this case. Recalling the contrast between ‘emergents’ and ‘resultants’, we can investigate the difference in relationship between diachronic emergents and their base properties and diachronic resultants and their base properties. In the case of the oscillators, an example of ‘resultant’ behavior, relative to a base level of damped behavior, would be some later damped behavior with a reduced level of damping. As desired, such behavior would both reduce to the base level behavior and be non-novel with respect to that behavior, in our senses of these terms. Since the resultant system will exhibit behavior that is distinct from (though not qualitatively distinct from) that of the base system, we expect there to be causal powers associated with this resultant system that are distinct from those associated with the base system, just as we expect there to be distinct causal powers associated with the emergent system. However, the non-commutativity illustrated in diagram 2 (see Figure 11.3) shows a way in which those causal powers are themselves of different types. We can regard the horizontal arrows as indicating causal processes associated with the temporal evolution of the system, and the vertical arrows as indicating causal processes associated with the change in damping. Because the diagram does not commute (that is, x0(t′) cannot be reached starting from xc(t) and first going to xc(t′)) , we see that these two sorts of causes really are different in the sense that the order in which they are applied is not arbitrary. (If we considered a case where the diagram did commute, the two causes would be equivalent.) We can describe this distinction in terms of Dretske’s (1988) distinction between ‘structuring’ and ‘triggering’ causes, where the causal relations represented by the horizontal arrows correspond to triggering causes, and those represented by the vertical arrows correspond to structuring causes. In the case of non-reducibility we have a real difference between structuring and triggering causes: the emergent state of the system (c = 0) can bring about later states that the earlier state (c > 0) could not trigger. Only the structuring kind of causation opens up the possibility of (diachronic) emergence. 4. SYNCHRONIC EMERGENCE The classic case of synchronic emergence involves phenomena occurring simultaneously on macro and micro levels. To illustrate this, consider the treatment of steady state heat conduction in a one-dimensional rod of length

Emergence in Physics

223

L.6 This system is described in terms of its temperature T(x) and its thermal conductivity k(x) which both vary in dependence on the spatial variable x. We assume that at both ends of the rod (x = 0 and x = L) the temperature is held constant at some values. Suppose now that we take the rod, at the micro level, to have a discrete, ‘atomistic’ constitution, that is, we stipulate that the system consists of individual atoms, separated by empty space—a periodic lattice with a period of length P = εL, with ε <<1. The microscopic conductivity k(x) will then be a rapidly oscillating function of position: high around the location of each atom, low in the inter-atomic spaces. This behavior of the conductivity is indicated by writing k as a function of x and x / ε. The dependence on x / ε manifests itself as rapid variations because , the derivative of k, is large for small ε. The temperature distribution at the micro level is thus described by Eq. 2

Note that any solution of (2), that is, an integration of (2) over the length of the rod, will represent the property of having such and such a distribution of micro-level temperature. To investigate synchronic emergence, we need to compare the solutions for Eq. 2 with the solutions describing the rod from the macro-level perspective. At the macro level, we assume that the rod’s structure is continuous rather than atomic, and hence that conductivity is steady rather than rapidly varying. The description of temperature distribution associated with the macro level will thus be different from that associated with the micro level, but this difference alone isn’t enough to show that the one is irreducible to or emergent with respect to the other. To test for reducibility, we need to show that the discrete description of the rod on the micro scale, indicated by the spacing P between individual atoms of the rod, converges to a continuous description at a larger scale, characterized by the macroscopic length L of the system, as the ratio ε = P/L → 0. This ‘continuum limit’ reflects the intuitive requirement that the macroscopic representation smoothes out the details at the micro level. We therefore seek an expansion of the solutions of Eq. 2 in terms of the small parameter ε = P/L and expect to obtain, in the limit ε → 0, the solution of the sought-for macroscopic equation, T 0(x): Eq. 3

...

A successful relation of this sort would show that the macro description reduces to the micro one: even though the descriptions are different, the macro one could be seen as a direct consequence of the micro.

224

Patrick McGivern and Alexander Rueger

It turns out, though, that letting the parameter ε go to zero results in a singular limit. In general—that is, unless is chosen in special ways— the solution T(x) will therefore not converge uniformly to T 0 (x) in the limit ε → 0. That is, we have , for some x T 0 (x) is not reducible to T(x). In topological terms, the topological space in which the solutions of the micro equation live (for ε → 0) cannot be characterized by the topology of uniform convergence. The singular limit, however, can be tamed if we explicitly introduce two length scales in the micro description, the macroscopic scale x and a microscopic scale y = x/ε.7 We fi rst replace Eq. 3 with an expansion of temperature T in terms of both x and y: Eq. 4

...

We then substitute the right-hand side of this expansion for T in Eq. 2 (the micro theory) and attempt to solve for the various Ti: the result should give us a good approximation of the exact solution, T(x, y). The advantage of Eq. 4 over Eq. 3 is that by explicitly distinguishing between the two scales, we can impose constraints on the Ti that guarantee that the series is ‘asymptotic’: the higher-order terms don’t become larger than the lower-order terms as ε → 0. Note that the series in Eq. 4 is not convergent: adding more terms to the expansion does not necessarily give us a better approximation of the exact solution to Eq. 2 (or, more precisely, the multi-scaled version of Eq. 2), but truncating the series after a few terms will give us a good approximation of T(x, y). Imposing these constraints has interesting results. The leading term in the approximation—T 0 (x, y)—turns out not to depend on the microscopic variable, y. Instead, it depends only on the macroscopic variable, x. Thus, this term represents a purely macroscopic quantity. Furthermore, it turns out that the constraint that allows us to force the expansion to remain asymptotically valid—the ‘solvability condition’—is precisely the macroscopic heat conduction equation we are trying to recover from the micro Eq. 2:

where K(x) is the ‘effective’ macro conductivity—a sort of average over the micro conductivity but not the simple arithmetic mean of the k(x, y) over the length of the rod that one might have expected. It’s important to stress that we don’t simply assume that the leading term in the expansion is constrained by this equation: that constraint falls out of the effort to keep the later terms in the expansion from diverging too quickly, and thus ruining

Emergence in Physics

225

the approximation. The macro equation, thus, arises as a constraint that has to be imposed on the micro-level description of the behavior of our system. With these constraints in place, Eq. 4 can be used to give a close approximation of the exact solution of Eq. 2. Asymptotic expansions typically give good approximations after only a few terms: in fact, the leading term alone, T 0 (x), gives an empirically adequate, and mathematically justified, approximation of the exact solution. Nevertheless, there is still no reduction of one description to the other: the approximation of the micro solution is well behaved, but it does not uniformly converge on the macro solution. As discussed earlier, irreducibility and qualitative novelty can be seen as two closely related ways of formally characterizing the core intuition that emergence phenomena are different in kind from their ‘base level’ or ‘resultant’ phenomena. The rod example satisfies this criterion of emergence. Furthermore, it is an illustration of a fairly general phenomenon in physics: whenever a problem is characterized by two scales of very different magnitude, relating the behavior at one scale to that at the other will typically involve a singular limit.

4.1. PHILOSOPHICAL QUESTIONS ABOUT SYNCHRONIC EMERGENCE

(i) Epistemic vs. Ontological Emergence Since the limit ε→0 does not represent a change in the system (in contrast to the parameter change in the diachronic case) one might think that only the epistemic interpretation is available: the irreducibility of the macro behavior to the micro is due merely to the irreducibility of one type of description to another—nothing about the system itself has been shown to be irreducible or emergent. Furthermore, the multi-scaling method we described does eventually lead to a sort of derivation of the macro-level description from the micro, and we could interpret this more sophisticated relationship as a demonstration that the difference between macro and micro is merely perspectival. However, an ontological interpretation is also available, and in many ways this interpretation is more natural. We saw that although the macro behavior of the system is not reducible to the micro behavior a quantitative approximation of the macro behavior can be achieved if we explicitly distinguish a macro scale from a micro scale in the description of the system. What we get, T(x,y), is an approximation of the system’s micro-scale behavior that includes its macro behavior as one component—the leading term in the expansion of Eq. 3. If we take this result ontologically seriously,8 we are led to the conclusion that macro and micro behavior of the system are not entirely distinct; their relation is one of part to whole. That’s what a literal reading of the expansion equation

226

Patrick McGivern and Alexander Rueger

(Eq. 3) indicates: T 0 (x) is part of T(x, y).9 The behavior of the heated rod, the distribution of temperature over the length of the system, consists of several components, one of which is the purely macro behavior T 0 (x). On this view, the operation of taking the limit ε → 0 does not represent a process of changing the system, but instead a process of paring down the system’s detailed micro-level behavior to its core component: the smooth macro-level behavior that many slightly different systems can share. The claim that a system’s macro-scale behavior could be part of its micro-scale behavior might at fi rst glance seem absurd: if anything, we might think, the parthood claim should be the other way around, since it is smaller (micro) things that are part of larger (macro) things. When we consider the parts of entities, we normally expect parts to be smaller than, and spatially contained within, whatever ‘whole’ they compose. Call this familiar sense of parthood spatial parthood. Obviously, the macro behavior is not a spatial part of the system’s behavior. But the spatial sense is not the only way to think about parthood. There are a variety of senses of ‘parthood’ that do not imply a particular spatial relation between parts and wholes. One sense that is appropriate for our case has been frequently used in the philosophy of mind. Shoemaker and others have suggested that the ‘realization’ relationship between a supervenient property and its realizer should be understood as a type of parthood relation between properties: realized properties are (non-spatial) parts of their realizers.10 This idea is then explicated in terms of a subset relation between causal powers: the causal powers of realized properties form a subset of those of their realizing properties. In our case, the corresponding claim needs to be that the causal powers of the macro behavior are a subset of those of the micro (or mixed) behavior.11

(ii) Supervenience In the case of synchronic emergence, one might expect emergent properties to supervene on their base properties (though this is controversial). When we discussed the diachronic example, we noted that the question about supervenience doesn’t seem to be well posed. The question in the synchronic case is whether there could be any variations in the macro description without any changes in the micro description. Since the series for T(x, y) does not converge, one might think that this question is again ill posed. But this is not the case. Once we have chosen the series of coefficients of T 0 (x), T 1(x. y), and so on, in Eq. 4 (that is, in our case, the series ε 0, ε1, . . . ), the asymptotic expansion of T(x, y) is uniquely determined (cf. Holmes, 1995: 11). So, despite the fact that this expansion does not converge, its leading term (representing the macro property) cannot vary without variation in the function (representing the micro property) which the series asymptotically approximates. And while expansions of the same function in terms of other series of coefficients may well have different leading terms—thus perhaps

Emergence in Physics

227

threatening macro-micro supervenience—there is an important constraint that is used to select one expansion series as the most appropriate for a given problem. That constraint is that the series of coefficients should give us a good asymptotic approximation of the function with as few terms of the expansion as possible. Given this constraint, the choice of coefficients— and hence the terms of the expansion—is determined, and supervenience of the macro on the micro is secured.

(iii) Causation How do the causal powers associated with the macro-level description of the rod relate to those associated with the micro level? If the heated rod case is formulated as a dynamic problem (with a time variable and introducing two time scales in addition to the spatial scales: see Frisch, 1995: 227), we can set up a diagram analogous to diagram 2 in the preceding discussion (Figure 11.4). In contrast to the diachronic oscillator case, however, it is now inappropriate to interpret the vertical arrows as causal relations; the change in the parameter in the rod example doesn’t represent a change in the system itself. But from the fact that the diagram does not commute we can see that the horizontal causal relations, representing micro and macro causation, are not equivalent in the sense that there is no homeomorphism that would map the micro onto the macro on both the left- and the right-hand sides.12 The remaining question about emergent causation concerns whether or not the causal efficacy of the macro can be distinguished from that of the micro in a way that alleviates any worries about the two ‘competing’, or the micro ‘excluding’ the macro as a genuine cause. Here we can avail ourselves of an argument of Yablo (1992: 434 ff.). Yablo defends the causal efficacy of higher-level properties and rejects the claim that these properties are causally preempted by the lower-level properties on which they supervene. For this purpose, he develops a notion of causation that imposes two requirements on property instances (or events) so that property instance C causes instance E: (i) C has to be sufficient for the occurrence of E. This condition, if it is satisfied by C, is also satisfied by any C*

Figure 11.4 Diagram 3.

228

Patrick McGivern and Alexander Rueger

such that C is a part of C*. To rule out C* as causing E, Yablo imposes that (ii) C is not only sufficient but also required for E. If C is required for E, then any C** that is a part of C will not be sufficient for E. Genuine causes, in other words, have to be commensurate or proportional to their effects, that is, the cause has to be sufficient and required for the effect. Some property instances are sufficient to bring about an effect E, but, intuitively, they contain ‘too much’ detail that is not required for causing E; other property instances, although causally relevant for E, contain ‘not enough’ detail to cause E. The cause of E strikes the balance and is therefore called proportional to the effect. Applied to our case of macro and micro behavior, the proportionality requirement tells us that, depending on the effect we are interested in, T(x, y) will sometimes be the cause, but for other choices of effect we’ll have to select T 0 (x) in order to satisfy proportionality of cause and effect. The worry about T 0 (x) not having a causal role distinguishable from that of T(x, y) can now be countered by pointing out T 0 (x)’s causal efficacy: for certain effects, T 0 (x) cannot be replaced by T(x, y) as the genuine cause. Thus we see the non-uniform limit which leads from T(x, y) to T 0 (x) as a procedure for isolating that part of the microscopic behavior that is causally efficacious at the macro level (i.e., the part which can satisfy the proportionality requirement for macro causes). As we have seen, in our example the limit can be taken (approximately) successfully only if we introduce an independent macro scale besides the micro scale. It is tempting to characterize this method as stripping away or eliminating causally extraneous (here: microscopic) detail. But this mustn’t be misunderstood: the continuum limit does not change the system as given by the microscopic description of the rod. The limit operation rather isolates a part of the behavior of the given system. This perspective is relevant to the question (mentioned earlier) of whether the macro description is just an ‘idealization’ which must not be taken ontologically seriously. It should be clear now that there is an ambiguity in the notion of idealization on which this objection relies. The macro description is idealized compared to the micro description because the former ‘leaves out’ certain causal powers included in the micro characterization. This is one sense of idealization. Another sense is that, therefore, the macro description must be false. But this second meaning of idealization is not equivalent to the fi rst. According to the argument in this section, we can say that the macro description is true of the macro phenomena whose reality is secured by their having different causal powers than the micro base. Does emergent causation necessarily involve ‘downward’ causation? Within the mathematical framework we used, downward causation would be illustrated, presumably, by the effect of a higher-level constraint on the lower level. In the heated rod case, imposing the solvability condition on the solution of the micro equation might count as a relevant higher-level constraint. Such interpretations have precedents (e.g., Bishop, 2008; Sperry,

Emergence in Physics

229

1986). To the extent that these interpretations can escape the charge of incoherence against synchronic reflexive downward causation (Kim, 1999), they are tenable. If the way in which they diffuse the incoherence objection relies on a violation of supervenience, however, we don’t see how they could be compatible with our example.

5. RAMIFICATIONS We said earlier that the relatively simple examples with which we illustrated our understanding of emergence are representative of the crucial features that underlie the more spectacular cases of emergence in physics. What all of these cases share is the occurrence of a singular limit which is responsible for the non-reducible and novel behavior observed.13 Phase transitions are often regarded as candidates for emergent behavior, that is, transitions from an unordered to an ordered state (or vice versa) when a system parameter (like temperature) is changed. Examples include the transition to ferromagnetism or when a gas changes its state to a liquid. In all these cases, the behavior of the ordered phase seems intuitively novel with respect to that of the unordered phase. Since a system parameter is being changed, the examples are analogous to our discussion of diachronic emergence in the oscillator case. To represent such phase transitions mathematically, the system has to be studied in the ‘thermodynamic limit’, the limit in which the volume of the system goes to infi nity while the density is kept constant. This limit is singular. The occurrence of classical behavior at a macro level from a quantum mechanical micro level is another standard candidate for emergence. Again, the differences are intuitively striking enough to classify the classical behavior as novel with respect to the quantum mechanical base. In our classification, this would be a case of synchronic emergence, and, indeed, the mathematical treatment of the relation between the levels involves taking the ‘classical limit’ h→0 of the Schrödinger operator h2/2m(∂2/∂x2). This limit again turns out to be singular. Space does not permit a detailed comparison of our view with other proposals in the recent literature. Batterman’s (2001) account of emergence is formally closely related to our view, although he uses the more advanced techniques of the renormalization group to analyze the extracting procedure that we phrased in terms of singular limits and multi-scale analysis.14 Batterman, however, shies away from giving an interpretation of his results in terms of causal powers, and thus his view should probably be classified as epistemic. Similar approaches can be found in Primas (1983, 1998) and various works by Atmanspacher (e.g., Bishop & Atmanspacher, 2006). A purely ontological account in terms of a ‘fusion’ operation between properties is given by Humphreys (1997, 2008), a theory that seems to be applicable only to quantum phenomena. A sophisticated metaphysical model

230 Patrick McGivern and Alexander Rueger of diachronic emergence that includes downward causation is presented by O’Connor and Wong (2005). Both, Humphreys and O’Connor, require violations of supervenience in their proposals. The relation of our view to Wimsatt’s (1997) suggestion of understanding emergence as ‘violations of aggregativity’ is sketched in Rueger (2006).

NOTES 1. ‘Emergence in physics’ is here understood as a discussion of the question whether there are emergent physical phenomena, not an investigation into the possibility of non-physical phenomena emerging out of physical ones. 2. Humphreys (2008) draws a tripartite distinction between ‘inferential’, ‘conceptual’, and ‘ontological’ conceptions of emergence. While the inferential and conceptual varieties seem close to what we call the epistemic conception, the fact that Humphreys does not view his categories as mutually exclusive makes this difficult to decide. 3. There are several other criteria that have often been associated with emergence, such as their inexplicability or unpredictability from knowledge of the base phenomena alone. However, criteria of this sort seem to be biased toward a purely epistemic sort of emergence, whereas we want our pre-theoretic account to be neutral insofar as the question of an epistemic or ontological interpretation is concerned. 4. For a more precise defi nition see, e.g., Arnold (1983: 89–91). 5. The diagram also shows that the undamped system is structurally unstable: small changes in the parameter (d) turn the system into a topologically inequivalent one. See Rueger (2000b). 6. See, for instance, Holmes (1995: 224ff.) and Rueger (2006). The example can be modified to contain a time variable and so describe a process: see Frisch (1995: 226–228). 7. An analogous procedure can be applied in the damped oscillator case to handle the singular limit: by introducing two time scales, a ‘fast’ one to describe the oscillations of the system and a ‘slow’ one to characterize the decay of the amplitude, we can produce an asymptotic approximation of the undamped behavior even though this approximation does not converge uniformly. 8. Arguments for why we should do so are discussed later. 9. It might seem arbitrary to interpret an equation like T(x) = T0(x) + εT1(x) + . . . as representing a part-whole relation in which T(x) is the whole. The equation itself is symmetric in the sense that we can just as well write T0(x) = T(x)—εT1(x) + . . . so that it looks as if T0(x) might be the whole and T(x) a component. But there is no arbitrariness here. Although the equations are symmetric, the asymmetry required for our interpretation is introduced by the perturbation approach itself. We are looking for a representation of the system’s (total) behavior, an appropriate solution of the equations of motion, which is T(x). T0(x), by contrast, solves the equations of motion only approximately, at the lowest order (e.g., ε0) of the perturbation theory; the complete solution is T(x), and therefore we are justified in interpreting T(x) as the whole and T0(x) as a component. 10. See Shoemaker (2001: 78 ff.) (with further references). Compare also Yablo (1992). One of the attractions of the subset view is that, as Lewis (1991) notes, the subset relation satisfies the standard axioms of mereology.

Emergence in Physics

231

11. The subset view of realization has also been used to characterize the difference between properties that are reducible and those that are not (cf. Wilson, 1999). In cases of successful reduction, the causal powers of the realized property form an improper subset of the powers associated with the realizing property. In cases where reduction fails, the powers of the realized property are a proper subset of the set of powers of the realizing property. 12. Lest this be interpreted as a violation of supervenience of the macro on the micro, remember that the vertical arrows do not represent the supervenience relation. 13. More examples can be found in Primas (1998). 14. These techniques are related: see Goldenfeld (1992: 318–329) (with further literature).

REFERENCES Arnold, V. I. (1983). Geometrical Methods in the Theory of Ordinary Differential Equations. New York: Springer. Batterman, R. (1995). Theories between theories. Synthese, 103, 171–201. . (2001). The Devil in the Details. Oxford: Oxford University Press. Bishop, R. (2008). Downward causation in fluid convection. Synthese, 160, 229– 248. Bishop, R., & Atmanspacher, H. (2006). Contextual emergence in the description of properties. Foundations of Physics, 36, 1753–1777. Dretske, F. (1988). Explaining Behavior: Reasons in a World of Causes. Cambridge. MA: MIT Press. Frisch, U. (1995). Turbulence. Cambridge: Cambridge University Press. Goldenfeld, N. (1992). Lectures on Phase Transitions and the Renormalization Theory. Reading, MA: Perseus Books. Holmes, M. H. (1995). Introduction to Perturbation Methods. New York: Springer. Humphreys, P. (1997). How properties emerge. Philosophy of Science, 64, 1–17. . (2008). Computational and conceptual emergence. Philosophy of Science, 75, 584–594. Kim, J. (1999). Making sense of emergence. Philosophical Studies, 95, 3–36. Lewis, D. (1991). Parts of Classes. Cambridge, MA: Blackwell. Nickles, T. (1973). Two concepts of intertheoretic reduction. Journal of Philosophy, 70, 181–201. O’Connor, T., & Wong, H. Y. (2005). The metaphysics of emergence. Nous, 39, 658–678. Primas, H. (1983). Chemistry, Quantum Mechanics, and Reductionism. New York: Springer. . (1998). Emergence in exact natural sciences. Acta Polytechnica Scandinavia, 91, 83–98. Rueger, A. (2000a). Physical emergence, diachronic and synchronic. Synthese, 124, 297–322. . (2000b). Robust supervenience and emergence. Philosophy of Science, 67, 466–489. . (2006). Functional reduction and emergence in physics. Synthese, 151, 335–346. Shoemaker, S. (2001). Realization and mental causation. In C. Gillett & B. Loewer (Eds.), Physicalism and Its Discontents (pp. 74–98). Cambridge: Cambridge University Press.

232

Patrick McGivern and Alexander Rueger

Sperry, R. (1986). Macro- versus micro-determinism. Philosophy of Science, 53, 265–270. Wilson, J. (1999). How superduper does a physicalist supervenience need to be? Philosophical Quarterly, 49, 33–52. Wimsatt, W. (1997). Aggregation: Reductive heuristics for fi nding emergence. Philosophy of Science, 64, S372–S384. Yablo, S. (1992). Cause and essence. Synthese, 93, 403–449.

12 The Emergence of the Intuition of Truth in Mathematical Thought Sergio Galvan

1. INTRODUCTION The concept of emergence has a twofold meaning: the fi rst ontological and the second epistemic. (1) In the ontological domain, a system is said to be emergent with respect to the subsystems of which it is constituted when the laws that govern its structure (including the laws of development that characterize it) cannot be derived from the laws of its subsystems. An emergent system is therefore endowed with an essentially irreducible structure, and description of this structure is assigned to an ontological theory irreducible to the ontological theory able to describe the subsystems of the system in question. (2) In the epistemic domain, the concept is applied to the evidence systems at the basis of justification of particular theories: an evidence system is said to be emergent with respect to other evidence systems when the theories justifiable on the basis of the fi rst system are irreducible to the theories justifiable on the basis of the other systems considered. An emergent evidence system is therefore one which gives access to information that is essentially new with respect to the information yielded by the systems from which it is emergent. Consequently, theories based on this information are irreducible to theories based on information of another type. The two concepts of emergence are linked by the fact that systems of evidence which are such to allow access to emergent structures are epistemically emergent with respect to those necessary to access less complex ontological systems. This chapter examines the feasibility of applying the category of emergence to the hierarchy of formal systems and to the hierarchy of the evidence systems which are their correlates. This hierarchy is a direct consequence of the phenomenon of incompleteness, in its various forms, that result from Gödel’s theorems. Hence the intention is to explore the meaning of the most significant form of incompleteness resulting from Gödel’s theorems, namely the omega-incompleteness of arithmetic theories, and to examine some methods—in particular, the elaboration of a truth theory—which are able to remedy this form of incompleteness. These two themes are addressed respectively by the fi rst and the second sections of the chapter. The fi rst section analyzes the omega-incompleteness of arithmetic theories, starting

234 Sergio Galvan from the primitive recursive arithmetic (PRA) system. The result of this analysis will be the identification of two fundamental reasons for incompleteness, both of which involve the irreducibility of truth to provability. However, the analysis will also show that the reason for this irreducibility is not that there are truths extraneous to the proof process; instead, there are truths relative to the model of a theory T not provable in that same theory but which can be proved in theories T’ more powerful than T. In this sense, the truth of a proposition relative to (the standard model of) T can be defi ned as the provability of that proposition in a theory more powerful than T: that is, a theory T’ which is a non-conservative extension of T. Moving to a more powerful theory, therefore, makes it possible to overcome the phenomenon of incompleteness. But on what does the greater power of the superior theory T’ depend? Does it depend on the fact that the truth theory relative to the language of theory T is formalized in T’? Or is it sufficient that the reflection principle relative to the formulas of T be derivable in T’? And in this latter case, what is the basis for this principle? In regard to these questions, the chapter will address the debate between Ketland (and Shapiro) on the one hand, and Tennant on the other, concerning the so-called ‘conservativeness argument’. While Ketland maintains that in order to obtain the consistency of PRA, for example, one must move to a higher theory wherein it is possible to formalize a truth theory for PRA from which its consistency can be derived, for Tennant it is sufficient to use the truth of the uniform reflection principle restricted to the PR formulas. It will be important to establish whether justification of the reflection principle necessarily entails assumption of the existence of the standard model. We shall see that this is unnecessary if the theory in question is PRA (i.e., the theory that formalizes the system of finitist evidence). The third section of the chapter, in fact, will make use of three distinct notions of the truth of an universal proposition: a notion of truth based on an infinitary model; a notion of truth understood as finitary evidentiability; and a notion of truth understood as non-finitary abstract evidentiability. In the case of PRA, the truth of its consistency can be established by resorting both to an explicit (Tarskian) theory of truth (which formalizes the first notion of truth) and to a notion of truth as non-finitary evidentiability. However, determination of a theory’s consistency always requires the emergent character of truth on derivability. As regards the technical aspects of the formulation and the demonstration of the theses contained in the chapter, it is worth informing the reader that the formal language employed in my foregoing demonstration will not be explained here, but can be found in Smorynski (1977). 2. OMEGA-INCOMPLETENESS OF A THEORY T ≥ PRA The subject of this section is the omega-incompleteness of a formal theory which seeks to formalize fi nitist arithmetic. PRA (i.e., primitive recursive

The Emergence of the Intuition of Truth 235 arithmetic) is normally considered to be the theory that formalizes fi nitist arithmetic.1 But the arguments which follow also hold if one assumes other theories such as PA (i.e., Peano arithmetic) as formalizing fi nitist arithmetic. I shall seek to show that omega-incompleteness entails an irreducible distinction between truth and provability. Consider the following three statements (where T can be considered coincident with PRA or an extension of it): – (a) (∀n)(T-⊢α(n-)), i.e., T-⊢α(0) and T-⊢α(1) and . . . . (b) T-⊢∀xPrT (¯α(x )¯) (c) T-⊢∀xα(x) Firstly, (a)⇒(c) expresses the usual property of omega-completeness (in short, omega-3), and its falsity is well-known. The formalization of (a)⇒(c) also enables one to show that omega-3 entails the inconsistency of T. In fact, the formalization of (a)⇒(c) is omega-3

. ∀xPrT(¯α(x )¯) → PrT(¯∀xα(x)¯)

Now, for a specific α we have . ∀xPrT(¯¬ProvT( x , ¯

¯)¯) → PrT(¯∀x¬ProvT(x,¯ ¯)¯)

hence . ∀xPrT(¯¬ProvT(x ,¯ ¯)¯) → PrT(¯ConsT ¯) def. ConsT . ∀xPrT(¯¬ProvT(x ,¯ ¯)¯) → ¬ConsT by G2 ¬ConsT by Feferman’s lemma (cf. Smorynski, 1977: 847) The non-validity of omega-3 shows immediately that the derivability predicate does not behave like the truth predicate. If Tr(α(0)) and – Tr(α(1)) . . . then Tr(∀xα(x)), whilst the derivability of α(n- ) for all n, (∀n) (T-⊢α(n-)), does not guarantee the derivability of ∀xα(x). If we say truth consists in derivability in T, then we cannot say that Tr(∀xα(x)) even if Tr(α(0)), – Tr(α(1)), etc. But why is it not possible to pass from (a) to (c)? The passage from (a) to (c) would require two critical steps which keep the extremes (a) and (c) detached. Each of these critical steps is a reason for omega-incompleteness. I begin with the fi rst, which consists in the fact that it is not generally the case that (a) implies (b) (i.e., omega-1). Let (a) be the starting point. We want to see why it is not always possible to reach (b). To show why this is not the case, I consider the moves by which one usually goes from (a) to (b).

236

Sergio Galvan

(a), i.e. (∀n)(T-⊢α(n- )), is an abbreviation for the following meta-theoretical infi nite conjunction: – – T-⊢α(0) and T-⊢α(1) and T-⊢α(2) and . . . Now, from the usual perspective of proof theory, a theory is constructed in order to obtain all the propositions that are true in the standard interpreta– – tion of the theory. But if (∀n)α(n- ) (i.e., α(0) and α(1) and α(2) and . . . ) is true in the standard interpretation of T, then also ∀xα(x) is true, at least in the standard interpretation of T. It is therefore to be expected that also T-⊢∀xα(x) (i.e., the theorem relative to ∀xα(x)) follows from the infi nite conjunction of the theorems relative to each of the numerical examples (∀n) (T-⊢α(n-)). How can this infi nitary relation be translated into a fi nitary relation of derivability? The fi rst step is to translate the meta-theoretical infi nite conjunction into a fi nite expression of the theory. Again, how can this be done? The usual arithmetical practice in cases like this is to find a proof of α(k) (for a certain k) which does not depend on the specific nature of k but only on the fact that k is a numeral. If the proof satisfies this requirement, it coincides with a particular exemplification of a uniform structural scheme which is invariant in the proof of the other cases, with the sole difference that other numerals take the place of k. As is well-known, this is the notion of prototype-proof proposed by Herbrand. Now, how can a prototype-proof be translated within the context of a purely formal standard language? The translation is performed by identifying a term t(n) which describes uniformly for all n the proof of α(n) in T and by proving this in T. For T to be able to do this, however, the proof in T must be carried out with respect to the open code for the closure of α(x) under substitution of numerals. In symbolic terms, this requires establishing the following: . T-⊢∀xProvT(t(x),¯α(x )¯) Whence . (b) T-⊢∀xPrT(¯α(x )¯) Note the essential presence of the variable x in the preceding formula. This guarantees that t(x) is the description of the invariant proof schema underlying the proof for each single numeral. If this were not a free variable, the empty structure of the schema would not be expressible in T. T would thus express only instances of the schema, and this would entail the impossibility of the fi nitary translation of (a). Yet the passage from (a) to (b) is not always guaranteed. It is possible that the theory T does not know (∀n)α(n-), even though it does know that α – – holds individually for each numeral: α(0), α(1), α(2), . . . The non-validity

The Emergence of the Intuition of Truth 237 of omega-1 expresses the general non-validity of (a)⇒(b). The formalization of (a)⇒(b), in fact, is omega-1

. . ∀xPrT(¯α( x )¯) → PrT(¯∀xPr (¯α( x )¯)¯) T

which is incompatible with the scheme of uniform omega-consistency restricted to the PR formulas (see Galvan, 1994). Hence, with respect to the context fi xed by a numerical theory T, it is not always possible to prove in T that a particular property is provable in T for all numerals. Sometimes, the existence of the proof for each individual numeral may not be brought to evidence. To conclude: this is the fi rst reason that explains the phenomenon of omega-incompleteness. It is not possible for (∀n)(T-⊢α(n- )) to imply T-⊢∀xα(x) (that is, (a)⇒(c)), because this . would also imply T-⊢∀xPrT(¯α(x )¯) (that is, (a)⇒(b)). . But let us suppose for a moment that T-⊢∀xPrT(¯α( x )¯) holds. Does this necessarily guarantee also T-⊢∀xα(x)? No, it does not, because this is the step when the second reason for omega-incompleteness comes into play, . and the second critical juncture arises. To assume that T-⊢∀xPrT(¯α( x )¯) implies T-⊢∀xα(x) is in fact to hold that (b)⇒(c), which can be formalized as follows: omega-2

. PrT(¯∀xPrT(¯α(x )¯)¯) → PrT(¯∀xα(x)¯)

And yet, as previously, considering a specific α, we have . PrT(¯∀xPrT(¯¬ProvT(x ,¯ ¯)¯)¯) → PrT(¯∀x¬ProvT(x,¯ ¯)¯) then . PrT(¯∀xPrT(¯¬ProvT(x ,¯ ¯)¯)¯) → PrT(¯ConsT ¯) def. ConsT . PrT(¯∀xPrT(¯¬ProvT(x ,¯ ¯)¯)¯) → ¬ConsT by G2 ¬ConsT by Feferman’s lemma and D1 The result thus obtained is the same as follows from assuming the validity of omega-3 (the derivation is also the same, with the sole difference that D1 is applied to Feferman’s lemma): the assumption of omega-2 implies the non-consistency of T. To conclude: the non-validity of omega-2 tells us that even if T-⊢∀xPrT(¯α( . x)¯), that is, even if it is provable in T that all the numerical instances of α(x) are derivable in T (and it is not only true that all the numerical instances of α(x) are derivable in T), ∀xα(x) is not provable in T. In other words, the non-validity of omega-2 tells us that—at least regarding formulae α(x) like ¬ProvT(x,¯ ¯)—the fact that the truth that all the numerical instances of α(x) are derivable in T does not guarantee the derivability of ∀xα(x) in T depends not on the underivability of that truth in T but on the fact that

238

Sergio Galvan

the derivability of that truth is not sufficient to guarantee the derivability of ∀xα(x) as well. We may now ask why also omega-2 fails. If, at least in certain cases, incompleteness is due not to the fact that T is unaware that all the numerical cases of α(x) are demonstrable but to the fact that this does not enable the theory to be aware of the truth of ∀xα(x), what is the reason for this inability? The fact is that the theory’s knowledge is closed only under the formal relation of logical consequence. However, the truth of (∀n)α(n-) necessarily implies the truth of ∀xα(x) only in the standard model, and it is well-known that a fi rst-order numerical theory is unable to characterize the standard model of natural numbers. For this reason, the theory knows (∀n) α(n-) without knowing ∀xα(x). All three forms of omega-incompleteness express the distance between truth and provability. Omega-1-incompleteness confi rms the distance and increases it by extending it or deepening it. Here the difference resides not in the fact – – that although T-⊢α(0) and T-⊢α(1 ) and T-⊢α(2 ) and . . . , T-⊬∀xα(x)), – – but in the fact that although T-⊢α(0) and T-⊢α(1 ) and T-⊢α(2 ) and . . . , . T-⊬∀xPrT(¯α(x )¯), which means that the truth of the infi nite conjunction of the derivability assertions for each of the numerical instances is not substitutable by the derivability in T of the fi nitary assertion which expresses that conjunction. The gap between truth and derivability is increased even further in this case by the fact that truth—which cannot be replaced with derivability—has a syntactic content (it concerns, that is to say, facts of derivability). Moreover, particular forms of formulae α(x) determining omega-1-incompleteness have the complexity of PRformulae—that is, they are decidable formulae. The truth of the infi nite conjunction therefore cannot be disputed even from a constructivist point of view. It is already established by the way in which the n-th case must be decided, although the time of the decision may be distant. (In other words, the proofs potentially already exist although they have not yet been actualized.) Yet, although the series has already been determined, it is not possible to prove the statement that describes it in fi nitary terms. Omega-2-incompleteness manifests another aspect of the irreducibility of truth to derivability. This consists in the fact that certain truths—for example, the truth of ∀xα(x) with respect to the standard model of arithmetic—are not derivable from the axioms of the theory because they are not their logical consequences. Moreover, there are no axioms able to restrict the structures to the standard ones—like the previously mentioned standard model of natural numbers—so that the said truths could be transformed into logical consequences. Omega-3-incompleteness immediately shows that the derivability predi– cate does not behave like the truth predicate. If Tr(α(0)) and Tr(α(1)) . . . , then Tr(∀xα(x)), while the derivability of α(n) for all n, (∀n)(T-⊢α(n)), does not guarantee the derivability of ∀xα(x).

The Emergence of the Intuition of Truth 239 As has been said, omega-3-incompleteness is implicated by forms of omega-1- or omega-2-incompleteness. The third section considers some significant forms of omega-2-incompleteness. It will do so in the context of analysis of the so-called ‘conservativeness argument’. This analysis will evidence the gap between the primitive arithmetic PRA system and the ΙΣ2 system, in which it is possible to develop a truth theory to prove the formulae in respect to which PRA is omega-incomplete. This conclusion fi nds in favor of the proponents of the conservativeness argument, and in particular of Ketland’s refl ective argument. However, further analysis in the fourth section will grant partial validity to the semantic argument put forward by Tennant in opposition to the reflective one propounded by Ketland.

3. TRUTH THEORY AND DEALING WITH OMEGA-INCOMPLETENESS

3.1. Truth Theory and the Conservativeness Argument As is well-known, Ketland and Shapiro have recently propounded the socalled conservativeness argument.2 This states that a deflationist conception of the truth is false because a system T to which the relative truth theory is added allows derivation of the consistency of T. It is therefore a non-conservative extension of T. In fact, if an extension is non-conservative, the truth predicate from which it is generated cannot be devoid of substantial meaning.

3.2. Reflective Argument Ketland’s reasoning in favor of this argument is based on the assertion that the truth predicate can be used to show that all the theorems of T are correct, that is, that the following general reflection schema RFN(T) is valid: . ∀xPr (¯α( x )¯) → ∀xα(x) T

According to Ketland, indeed, this principle can only be justified with a truth theory based on the existence of a model able to satisfy the axioms of T. On the basis of the correctness of the logic of T, in fact, it can be proved that the truth of the axioms is transferable to the theorems, so that by performing an induction on the length of the derivation it can be shown that also the theorems are true to the same extent as the axioms. In short, it is possible to prove the following reflection principle: . ∀xPr (¯α(x )¯) → Tr T(¯∀xα(x)¯) T

from which, by exploiting Tarski’s schema: Tr T(¯α¯) ↔ α, one obtains the normal reflection schema RFN(T):

240 Sergio Galvan . ∀xPr (¯α(x )¯) → ∀xα(x) T

It is for this reason that Ketland calls his argument refl ective (2005: 75–80); for what is essential in this argument is the use of the truth predicate to show the validity of the principle. Tennant disputes Ketland’s claim that it is necessary to invoke a truth theory in order to justify the reflection principle and consequently the consistency of T. He points out that it is not necessary to assume the reflection principle without restriction. It is sufficient to assume it is restricted to PR-formulae (in symbols, RFN PR(T)). From the reflection principle restricted to PR-formulae, in fact, one can obtain the following instance: . ∀xPrPRA(¯¬ProvPRA(x ,¯ ¯)¯) → ∀x¬ProvPRA(x,¯ ¯) from which is possible to derive the consistency of T using Feferman’s lemma. Tennant maintains that this restricted principle is the formalization of the so-called semantic argument, on the basis of which, according to Tennant’s reading of the incompleteness phenomenon, it is possible to obtain the truth of Gödel’s formula and, by equivalence, the assertion of consistency. He states the argument thus: “G is a universally quantified sentence (as it happens, one of Goldbach type, that is, a universal quantification of a primitive recursive predicate). Every numerical instance of that predicate is provable in the system S. (This claim requires a subargument exploiting Gödel-numbering and the representability in S of recursive properties.) Proof in S guarantees truth. Hence every numerical instance of G is true. So, since G is simply the universal quantification over those numerical instances, it too must be true.” (see Tennant, 2002: 556)

3.3. Semantic Argument The semantic argument is now set out step by step in order to determine its points of contact with the RFN PR(T) principle. Let α(x) be the PR-formula of which the Gödelian γ is the generalization: that is, γ ≡ ∀xα(x). The argument starts from the fact that every numerical instance of the predicateα(x) is provable in the system T: – – T-⊢α(0) and T-⊢α(1) and T-⊢α(2) and . . . But the argument assumes that provability in T guarantees truth. Therefore, – – TrT(¯α(0)¯) and TrT(¯α(1)¯) and TrT(¯α(2)¯) and . . .

The Emergence of the Intuition of Truth 241 But then also ∀xα(x) must be true by virtue of the defi nition of numerical truth. Clearly, the formalization of the argument is therefore expressible through the following principle: . . ∀x[Pr (¯α(x )¯) → Tr T(¯α(x )¯)] T

from which by distribution one obtains . . ∀x[Pr (¯α(x )¯)] → ∀x[Tr(¯α(x )¯)] T

But then, because of the Tarski’s bi-conditional, one finally has . ∀xPr (¯α(x )¯) → ∀xα(x) T

The semantic argument is therefore faithfully expressed through the RFN PR(T) principle. However, the faithfulness of the expression is not the same thing as its justifiability. Ketland insists on this aspect in his reply to Tennant (2005: 76). In effect, there are two essential steps in the argument which are not self-evident. Justifying them requires at least an implicit truth theory. The fi rst step is the derivation of the truth of the individual numerical instances. Such derivation is justified only on the condition that T is correct in regard to the closed PR-formulae. This is a minimal requirement founded on a still fi nitarian truth theory. The second step is decidedly more demanding, in that it involves an infi nitarian notion of truth. This is the move from the truth of each numeral example (which are infi nite) to the truth of the universal proposition, and which is justified only by the satisfiability of the truth-defi nition clause of a universal formula on the standard structure of natural numbers. Hence the semantic argument owes its correctness to the possibility of declaring the truth of a universal numerical proposition starting from the truth of its instances. It therefore cannot do without a theory, at least implicit, of numerical truth.

4. SEMANTIC ARGUMENT, TRUTH THEORY, AND THE OMEGA-RULE

4.1. Semantic Argument and Reflection Principle It was said earlier that both the reflective and the semantic arguments require recourse to a truth theory. In this regard, the foregoing discussion, including the treatment of omega-completeness, seemingly supports the views of Ketland and Shapiro. Nevertheless, more detailed analysis shows that the way in which Ketland’s reflective argument appeals to the truth is different from the way in which the semantic argument does so. This difference depends on how the truth is conceived in the former

242 Sergio Galvan case and how it is conceived in the latter. For that matter, a certain difference has already been pointed out when it was said that the semantic argument invokes at least an implicit notion of truth with respect to the standard model, whilst the reflective argument requires an explicit theory of the truth. The discriminating factor is the fact that an explicit truth theory is a theory wherein it is possible to derive the existence of a model (i.e., a structure and its interpretation) with respect to which the propositions of theory T are true. In the case of numerical theories this model is an infi nitary model, in the precise sense that the support structure on which it is based consists of an actual infi nity of numerical elements. The truth of a universal proposition thus comes to coincide with the truth of all the numerical instances, which, in their turn, are true because each of the numerical elements to which they refer is existent and, moreover, endowed with the property signified by the proposition. By contrast, an implicit truth theory does not require the existence of the actual model because, given that it is implicit, it is also susceptible to a constructivist interpretation. In this case, the truth of a universal proposition comes to coincide with the presence of a procedure able to produce the individual instances one after the other. It is therefore now necessary to determine the conditions under which it is legitimate to interpret Tennant’s semantic argument in constructivist terms. Radically, it will be assumed that the constructivist point of view is emblematically represented by fi nitism. Unlike constructivism in general, fi nitism imposes precise conditions upon the procedures to generate numerical instances. These procedures must be concrete procedures—that is, of a primitive recursive nature—and this requires that they be formalizable within the system PRA. Constructivism is instead more liberal, in the sense that it allows for procedures of more abstract nature—to the point, indeed, that they are not formalizable. We shall see later that the semantic argument can have advantages if the constructivity constraints are relaxed to become non-fi nitist. Let us therefore focus on the PRA system and seek to apply the semantic argument to it. We have seen what PRA lacks for its consistency to be internally provable. PRA is unable to derive the following instance of the reflection principle restricted to PR-formulae. . ∀xPrPRA(¯¬ProvPRA(x ,¯ ¯)¯) → ∀x¬ProvPRA(x,¯ ¯) There are two ways to bridge the gap created by the non-derivability of the principle: (1) defi ne a concept of truth with which it is possible to derive the correctness schema corresponding to the principle; (2) extend PRA to use of the omega-rule. The former method corresponds to the reflective argument. The latter is a sort of formalization of the semantic argument. Both methods presuppose the reliable use of non-fi nitary abstract concepts. But, what is the meaning of non-fi nitary abstract concepts? Before continuing

The Emergence of the Intuition of Truth 243 with the discussion, brief illustration of Gödel’s theory of abstract concepts will be helpful.

4.2. A Note About Gödel’s Theory of Non-finitarian Abstract Concepts First to be illustrated are the meanings of the terms ‘abstract’ and ‘infi nitarian’ in light of Gödel’s theory of abstract concepts. This will be useful in gaining full understanding of the reflective and semantic arguments. In his 1951 lecture and in the two versions of his essay of 1953/59 in honor of Carnap, Gödel furnishes a series of examples of non-fi nitarian abstract concepts. For example, on page 318 of the Gibbs Lecture, at note 27, Gödel declares: Examples of such abstract concepts are, for example, ‘set’, ‘function of integers’, ‘demonstrable’ (the latter in the non-formalistic sense of ‘knowable to be true’), ‘derivable’ etc. or fi nally ‘there is’, referring to all possible combinations of symbols. Similar concepts are put forward in the other two texts. Gödel writes: Abstract concepts, e.g., are: ‘proof’ and ‘function’, if these terms are understood in their original ‘contensive’ meaning, i. e., if ‘proof’ does not mean a sequence of expressions satisfying certain formal conditions, but a sequence of thoughts convincing a sound mind, and if ‘function’ does not mean an expression of the formalism, but an understandable and precise rule associating mathematical objects with mathematical objects (in the simplest case integers with integers). An example for a transfi nite (i.e., non-constructive) concept is ‘there exists’, if this phrase means object existence irrespective of actual producibility. (1953/59, III: 341, n. 20), He fi nally adds in the same note: ‘Infi nite set’ is an abstract or transfinite concept according to whether potential or actual infi nity is meant. The abstract and transfi nite concepts together form the class of ‘non fi nitary’ concepts. Gödel therefore insists on two aspects (or two categories) of non-finitarian concepts. He distinguishes the aspect whereby many of these concepts are transfi nite from the aspect whereby they are abstract. The transfi nite character is the non-fi nitarian extensional element, and it substantially coincides with the existence of the actual infi nite. The other aspect instead coincides with the intensional character of such concepts (where function is understood not as a formal entity but as a rule, proof as a sequence of thoughts, and infi nite as potentially infinite). This second aspect is

244

Sergio Galvan

particularly important because it highlights the non-perceptual meaning of conceptual terms (in their generality and therefore also of transfi nite ones). But Gödel emphasizes it for another reason as well: concepts are the bearers of relations which don’t depend on the fact that they hold among the respective instances; that is, in the strong sense they are universal. This is the second emergent aspect of conceptual intuition with respect to perceptive intuition, and Gödel’s view on it is unequivocal. In “Is Mathematics Syntax of Language?” (1953/59: V, 359), Gödel writes: The similarity between mathematical intuition and a physical sense is very striking. | It is arbitrary to consider ‘This is red’ an immediate datum, but not so to consider the proposition expressing modus ponens or complete induction (or perhaps some simpler propositions from which the latter follows). For the difference, as far as it is relevant here, consists solely in the fact that in the fi rst case a relationship between a concept and a particular object is perceived, while in the second case it is a relationship between concepts. In short, infi nitarian intuition is the intuition of concepts and the relations among concepts, where the concepts are bearers of relations independently of the fact that they hold for the respective instances.

4.3. Derivation of Truth Through the Construction of a Model We may now turn to an analysis of the two methods that serve to demonstrate the consistency of PRA. The fi rst method consists in the formal construction of a model that satisfies PRA. Proof of the existence of this model is only possible in ΙΣ2 . But ΙΣ2 is the system characterized by the principle of induction up to the ΙΣ2-formulae. It is therefore a more complex and powerful system than the ΙΣ1 system which corresponds to (in the sense that it is a conservative extension to the Π2-formulas of) PRA. In accordance with Ketland and Shapiro’s reflective argument, therefore, the truth theory relative to PRA formulated in ΙΣ2 is effectively a non-conservative extension of PRA. An important ontological consequence follows from the non-conservativeness of PRA plus the truth theory for PRA (with the details of the system within which it is formalizable) with respect to PRA. What is the ontology underlying the model for PRA derivable in ΙΣ2? Whereas fi nitist truths (i.e., the propositions derivable in PRA) do not require that the model of the natural numbers be given in actu (it is not necessary for the natural numbers to be conceived as entities actually existing, but simply as potential, indefinitely incrementable, entities), the model for PRA (derivable in ΙΣ2) that guarantees the truth of the reflection principle RFN PR(PRA) is an infi nitarian model. That is to say, it is a model based on a support constituted by an actual infi nity of the abstract entities that are natural numbers.

The Emergence of the Intuition of Truth 245 This derives from the abstractness of the formula that defines the truth predicate and on which induction in ΙΣ2 is performed. This is a Σ2-formula (to be precise, a Δ2-formula), that is, a formula characterized by a prefi x consisting of a universal quantifier followed by an existential quantifier. Now, because the truth predicate has the complexity of a Σ2-formula, it has an infi nitarian content. In fact, the quantifiers of the inductive formula vary on the infi nite set (in the actual sense) of natural numbers, so that the truth of a formula of such complexity presupposes the existence of the actual infi nity of natural numbers, just as the truth of a universal formula presupposes the actual truth of all the generable number examples: It is not correct to argue that induction only involves the numbers from 0 to n; the property of n being established may be a formula with bound variables that are thought of as ranging over all numbers. That is, the induction principle assumes that the natural number system is given. (See Nelson, 1986: 1.) It is evident the model comprises the fi rst aspect of the abstractness of nonfi nitarian concepts: the extensional aspect consisting in the actual existence of the domain of infi nite objects. (i) Objection. It might be objected that the model underlying the consistency of PRA is already contained in the procedure for constructing the set of the primitive recursive functions. This is because the procedure guarantees the existence of all the functions obtained by means of the substitution schema and that of recursion from the initial functions. The procedure also ensures that these are all calculable. (ii) Reply. This is obviously true. However, the function of ΙΣ2 is exactly that of formally ensuring the existence of such functions (truth of the respective defi ning equations), and their calculability, without resorting to the intuitive model of the standard numbers—which is instead implicit in the procedure for the intuitive construction of the set of primitive recursive functions restricted to numerals. In effect, the claim that the procedure of constructing the primitive recursive functions ensures the existence of the model neglects to consider that the entire argument rests on the assumption that each stage in the construction is constituted by a finite number of steps, and that the fi niteness is itself contained in the calculation process. Precisely because the functions are defined by recursion, and because at each stage of the recursion the value can be constructed with a fi nite number of steps, this value exists for every primitive recursive function. In other words, the notion of fi niteness allows the construction of the standard sequence of natural numbers constituting the basis for the existence of a model of the theory, and which is therefore the test of its consistency. The ΙΣ2-induction serves to achieve the same result, but without using the notion of fi niteness, which is formally not expressible. The greater power of ΙΣ2 makes it possible to derive in purely formal terms the existence of a model

246 Sergio Galvan able to ensure the correctness of PRA, while the correctness is obtained by means of a sort of intuitive model if one remains at the formal level of PRA and appeals to the fi niteness of the primitive recursive constructs. In effect, this aspect coincides with the second component of the abstractness of non-fi nitarian concepts: the component whereby the infi nite appears in the form of universal abstractness (and not as extensional abstractness, i.e., as the actually given set of natural numbers). This is the aspect present in the second method to which the semantic argument can be related.

4.4. Evidence of Truth, Omega-rule, and the Semantic Argument The second method consists in the use of a particular instance of the omega-rule. It is intuitively formulatable in the following form: (∀n)(¬ProvPRA(n-,¯ ¯)) ∀x¬ProvPRA(x,¯ ¯) that is, (∀n)(¬ProvPRA(n-,¯ ¯)) | –∀x¬ProvPRA(x,¯ ¯) Now, how can the use of the omega-rule be justified? Or in other words, how can the derivability sign be interpreted? (i) Firstly, | – cannot be taken to be the derivability sign in PRA (or in any recursively enumerable formal theory). In fact, formalizing within PRA, one would obtain: . PRA- | –∀xPrPRA(¯¬ProvPRA(x ,¯ ¯)¯) → PrPRA(¯∀x¬ProvPRA’ (x,¯ ¯)¯) which would be in contrast with G2: as was seen in Section 1, the scheme of the assertion would be equivalent to the inconsistency of PRA. Therefore, this interpretation is to be rejected. (ii) The omega-rule can be justified only if the following condition holds: If (∀n) it is evident that (¬ProvPRA(n- ,¯ ¯)); hence it is evident that ∀x¬ProvPRA(x,¯ ¯). Of course, we can take it for granted that the antecedent holds because we know that (∀n)(PRA- ⊢ ¬ProvPRA(n- ,¯⊥¯)), and the derivability in PRA corresponds to the fi nitist evidence. How, therefore, can it be possible that ∀x¬ProvPRA(x,¯ ¯) is evident? It is possible only if the domain of interpretation of the variable x coincides with the domain of interpretation of the numeral terms. But the numeral terms are one-to-one interpretable onto the set of the standard natural numbers. It follows that the consequent is

The Emergence of the Intuition of Truth 247 evident only if the domain of interpretation is circumscribed to the standard model; that is, only if our mind is able to keep the meaning of standard natural number fi xed and consider it as a criterion for inclusion in the model. On the other hand, a standard natural number is an abstract element of an abstract omega-sequence: that is, it is an element produced by the fi nite iteration of the abstract operation of succession. It is now clear why the semantic argument applied to PRA (but a fortiori to more powerful theories) invokes a constructivist notion of abstract and non-fi nitarian truth. This is a constructivist notion because it coincides with the epistemic notion of evidentiability. However, it cannot be fi nitist evidentiability (i.e., founded on procedures pertaining to PRA) because fi nitist evidentiability coincides with derivability in PRA. Finally, this outcome highlights a marked difference between the model’s approach and this one. Prima facie, in fact, the impression is that this is a situation similar to the one analyzed in Section 3.2, where the gap due to the underivability of the reflection principle was bridged by the shift to a non-conservative extension of PRA. As in ΙΣ2, a notion of truth can be formulated which is able to close the gap at a rigorously formal level. One might therefore be induced to believe that in the second approach, too, one may hypothesize a formal theory able to formalize the notion of standard number and thus obtain the consistency of PRA. Yet this is not the case, for the abstract notion of fi niteness is not formally defi nable. As is wellknown, a compact logical arithmetic formal system is not categorical. In other words, it is incapable of circumscribing the domain of its models to the standard model alone. The overall consequence is that the basis of the second approach to consistency is not the existence of a model (the existence of the actual totality of natural numbers), that is, a realistic ontological hypothesis, but the reliability of the notion of finiteness, which, in its turn, and in consequence of its soundness, is able to generate the standard model. The second approach is therefore based on a cognitive hypothesis concerning the mind’s capacity to grasp intensional relationships which are not strictly finitist, but abstract and universal. From this a constructivist ontological interpretation may then ensue.

5. CONCLUSIONS On drawing together the threads of the foregoing discussion, we may say that the following conclusions have been reached. We can distinguish between two fundamental theses which ramify into a series of presuppositions and further arguments: (a) Any system T’ able to represent the truth theory relative to system T is emergent with respect to T in that it is a non-conservative extension of T. Consequently, the system of evidence that justifies T’ is emergent

248

Sergio Galvan

with respect to the system of evidence at the basis of T. This essentially depends on the fact that the ontology underlying T’, to which one has access, is more complex than that underlying T. The fi rst part of the thesis rests on the irreducibility of the notion of truth to that of derivability. The second part rests on the fact that truth in T coincides with derivability in T’, where T’ is characterized by an ontology richer than that of T. This thesis links with Ketland and Shapiro’s reflexive argument. (b) A theory PRA’ = PRA + RFNPR(PRA) is able to demonstrate the consistency of PRA. RFNPR(PRA) can be justified by means of a truth theory for PRA, and in this case there ensue the consequences of the previous thesis, where assumption of the truth theory entails an increase in ontological commitment. However, it can also be justified constructively by appealing to the evidence that the primitive recursive relations are true of any (universal abstract intensional) numeral and are therefore (extensionally) true of all of them. In this case, acceptance of RFNPR(PRA) does not depend on an increase in ontological commitment (in the realistic sense), but on an increase in gnoseological commitment consisting in acceptance of the reliability of the passage from assertion for the abstract concept of numeral (for any universal) to assertion for all numerals. This acceptance presupposes, in its turn, the willingness to recognize the validity of nonfinitist abstract evidence. This thesis is linked with Tennant’s semantic argument.

NOTES 1. Cf. Galvan (1992: 117–126), Simpson (1999: 373–374, 381–382), Smorynski (1985: 16–25). 2. See Field (1999), Ketland (1999), Shapiro (1998), Tennant (2002).

BIBLIOGRAPHY Azzouni, J. (1999). Comments on Shapiro. Journal of Philosophy, 96, 541–544. Boolos, G. (1998). Logic, Logic, and Logic. Cambridge, MA: Harvard University Press. Buss, S. R. (1998). First-order proof theory of arithmetic. In S. R Buss (Ed.), Handbook of Proof Theory (pp. 79–147). Amsterdam: Elsevier Science B.V. Feferman, S. (1991). Reflecting on incompleteness. Journal of Symbolic Logic, 56, 1–49. Field, H. (1999). Deflating the conservativeness argument. Journal of Philosophy, 96, 533–540. . (2008). Saving Truth from Paradox. Oxford: Oxford University Press. Galvan, S. (1983). Teoria formale dei numeri naturali. Milano: Franco Angeli. . (1992). Introduzione ai teoremi di incompletezza. Milano: Franco Angeli . (1994). A note on the ω−incompleteness formalization. Studia Logica, 53, 389–396.

The Emergence of the Intuition of Truth 249 . (2004). Gödel e il modello computazionale della mente. Rivista di Filosofi a Neoscolastica, 96, 145–174. . (2007). Sistemi dell’aritmetica da Q a PA. Milano: I.S.U. Università Cattolica. Gödel, K. (1951), Some basic therorems on the foundations of mathematics and their implications. Gibbs lecture. In Gödel (1995: 304–323). . (1953/59-III), Is mathematics syntax of language? In Gödel (1995: 334– 356). . (1953/59-V), Is mathematics syntax of language? In Gödel (1995: 356– 362). . (1958). On a hitherto unutilized extension of the fi nitary standpoint (1958). Translation from: Über eine bisher noch nicht benützte Erweiterung des fi niten Standpunktes. In Gödel (1990: 240–251). . (1972). On an extension of fi nitary mathematics which has not yet been used (1972). Translation from: Über eine bisher noch nicht benützte Erweiterung des fi niten Standpunktes. In Gödel (1990: 271–280). . (1986). Collected Works, Vol. I, Publications 1929–1936. S. Feferman, J. W. Dawson Jr., S. C. Kleene, G. H. Moore, R. M. Solovay, & J. van Heijenoort (Eds.). New York-Oxford: Oxford University Press. . (1990). Collected Works, Vol. II, Publications 1938–1974. S. Feferman, J. W. Dawson Jr., S. C. Kleene, G. H. Moore, R. M. Solovay, & J. van Heijenoort (Eds.). New York-Oxford: Oxford University Press. . (1995). Collected Works; Vol. III, Unpublished Essays and Lectures. S. Feferman, J. W. Dawson Jr., W. Goldfarb, C. Parsons, & R. M. Solovay (Eds.). New York-Oxford: Oxford University Press. Hàjek, P., & Pudlàk, P. (1993). Metamathematics of First-Order Arithmetic. Berlin: Springer. Halbach, V. (1996). Axiomatische Wahrheitstheorien. Berlin: Akademie Verlag. Hilbert, D., & Bernays, P. (1968). Grundlagen der Mathematik I (2nd ed.). BerlinHeidelberg: Springer-Verlag. (1st ed., 1934). . (1970). Grundlagen der Mathematik II (2nd ed.). Berlin-Heidelberg: Springer-Verlag. (1st ed., 1939). Ketland, J. (1999). Deflationism and Tarski’s paradise. Mind, 108, 69–94. . (2005). Deflationism and the Gödelian phenomena. Reply to Tennant. Mind, 114, 75–88. Mancosu, P. (1996). Philosophy of Mathematics and Mathematical Practice in the Seventeenth Century. Oxford: Oxford University Press. . (Ed.). (1998). From Brouwer to Hilbert. The Debate on the Foundations of Mathematics in the 1920s. Oxford: Oxford University Press. Nelson, E. (1986). Predicative Arithmetic. Princeton, NJ: Princeton University Press. Parsons, C. (1995). Platonism and mathematical intuition in Kurt Gödel’s thought. The Bulletin of Symbolic Logic, 1, 4–74. . (1998). Intuition and the abstract. In M. Stamm. (Ed.), Philosophie in synthetischer Absicht. Sinthesis in Mind (pp. 155–187). Stuttgart: Klett-Cotta. . (2000). Reason and intuition. Synthese, 125, 299–315. Schirn, M. (Ed.) (1998). The Philosophy of Mathematics Today. Oxford: Clarendon Press. Schirn, M., & Niebergall, K. G. (2005). Finitism = PRA? On a thesis of W. W. Tait. Reports on Mathematical Logic, 39, 3–26. Shapiro, S. (1998). Proof and truth: Through thick and thin. Journal of Philosophy, 95, 493–521. . (2004). Deflation and conservation. In V. Halbach & L. Horsten (Eds.), Principles of Truth. Frankfurt: Ontos Verlag.

250

Sergio Galvan

Simpson, S. G. (1999). Subsystems of Second Order Arithmetic. Heidelberg-New York-Berlin: Springer. Smorynski, C. (1977). The incompleteness theorems. In J. Barwise (Ed.), Handbook of Mathematical Logic. Amsterdam: North Holland. Tait, W. W. (1981). Finitism. The Journal of Philosophy, 78, 524–556. . (2005). The Provenance of Pure Reason: Essays in the Philosophy of Mathematics and its History. Oxford: Oxford University Press. Tennant, N. (2002). Deflationism and the Gödel-phenomena. Mind, 111, 551– 582. . (2005). Deflationism and the Gödel-phenomena: Reply to Ketland. Mind, 114, 89–96. Tieszen, R. (2005). Phenomenology, Logic, and the Philosophy of Mathematics. Cambridge: Cambridge University Press. Van Heijenoort, J. (1967). From Frege to Gödel. A Source Book in Mathematical Logic, 1879–1931. Cambridge, MA-London: Harvard University Press. Zach, R. (2006). Hilbert’s program then and now. In D. Jacquette (Ed.), Philosophy of Logic (Handbook of the Philosophy of Science, Vol. 5, pp. 411–447). Amsterdam: Elsevier.

13 The Emergence of Mind at the Co-Evolutive Level Arturo Carsetti

From an informational point of view, cognition (as well as life) can be characterized in terms of a concrete answer to three diffi cult questions: “how is information generated?”, “how is information transmitted?”, and “how is information assimilated?”. With respect to this last interrogative, we have immediately to realize that the assimilation-process of external information implies the existence of specifi c forms of determination at the neural level as well as the continuous development of a specific cognitive synthesis. Actually, information relative to the system stimulus is not a simple amount of neutral sense-data to be ordered, it is linked to the “unfolding” of the selective action proper to the optical sieve, it articulates through the imposition of a whole web of constraints, possibly determining alternative channels at, for example, the level of internal trajectories. Depth information grafts itself on (and is triggered by) recurrent cycles of a self-organizing activity characterized by the formation and a continuous compositio of multi-level attractors. The possibility of the development of new systems of pattern recognition, of new modules of reading will depend on the extent to which new successful “garlands” of the functional patterns presented by the optical sieve are established at the neural level in an adequate way (i.e., in accordance with the unfolding of a meaning-oriented and functional grid “surfacing” by forms). In this sense, the aforementioned self-organizing activity articulates as the real support for the effective emergence of an autonomous cognitive system and its consciousness. It embodies as a living map, articulating in an informational landscape, capable of crossing the semantic territories in order to permit the very meaning to mirror itself at the surface level and develop as truth in action. Insofar as an “I” manages to close the “garlands” successfully, and imprison the thread of meaning, thereby harmonizing with the ongoing “multiplication” of mental processes at the visual level, it can posit itself as an adequate instrument for the “vision-reflection” on behalf of the original Source of itself, for its self-generating and “reflecting” as Natura naturata, a Nature which the very units (monads) of multiplication will actually be able to read and see through the eyes of mind.

252

Arturo Carsetti

When we take into consideration, for instance, visual cognition, we can easily realize that vision is the end result of a construction realized in the conditions of experience. It is “direct” and organic in nature because it is the product of neither simple mental associations nor reversible reasoning, but, primarily, the “harmonic” and targeted articulation of specific attractors at different embedded levels. The resulting texture is experienced at the conscious level by means of self-reflection; we actually sense that it cannot be reduced to anything else, but is primary and self-constituting. We see visual objects; they have no independent existence in themselves but cannot be broken down into elementary data. Grasping the information at the visual level means managing to hear, as it were, inner speech. It means fi rst of all capturing and “playing” each time, in an inner generative language, through progressive assimilation, selection, and real metamorphosis (albeit partially and roughly) and according to “genealogical” modules (as envisaged by Husserl), the articulation of the complex semantic apparatus which works at the deep level and moulds and subtends, in a mediate way, the presentation of the functional patterns at the level of the optical sieve. What must be ensured, then, is that meaning can be extended like a thread within the fi le in accordance with the engraving of a coordinated sequel of “garlands”; only on the strength of this construction can an “I” posit itself together with a sieve: a sieve in particular related to the world which is becoming visible. In this sense, the world which then comes to “dance” before my eyes is impregnated with meaning. The “I” which perceives it realizes itself as the fi xed point of the garlands with respect to the “capturing” of the thread inside the fi le and the genealogically modulated articulation of the file which manages to express its invariance and become “vision” (visual thinking as it emerges through the successive identification of the observers), anchoring its generativity at a deep semantic dimension. The model-telos can shape itself as such and succeed in opening the eyes of the mind in proportion to its ability to permit the categorial to anchor itself to (and be filled by) intuition (which is not, however, static, but emerges as linked to a continuous process of metamorphosis). And it is exactly in relation to the adequate constitution of the channel that a sieve can effectively articulate and cogently realize its selective work at the informational level. This can only happen if the two original selective forces (drivers) meet (the force linked to the full expression of the original incompressibility, on the one hand, and the force linked to the meaning in action, on the other) and a telos shapes itself autonomously so as to offer itself as guide and support for the task of both capturing and “ring-threading”. It is the (anchoring) rhythm-scanning of the labyrinth by the thread of meaning which allows for the opening of the eyes, and it is the truth, then, which determines and possesses them. Hence the construction of an “I” as a fi xed point: the “I” of those eyes (an “I” which perceives and which exists in proportion to its ability to perceive). What they see is generativity in action, its surfacing rhythm being dictated intuitively. What this also produces, however, is a

The Emergence of Mind at the Co-Evolutive Level 253 file that is incarnated in a body that posits itself as “my” body, or more precisely, as the body of “my” mind: hence the progressive outlining of a meaning, “my” meaning which is gradually pervaded by life. Vision as emergence aims fi rst of all to grasp (and “play”) the paths and the modalities that determine the selective (and coupled) action expressed by meaning, the modalities specifically concerning the revelation (from within) of the aforementioned semantic apparatus at the surface level according to different and successive phases of generality. These paths and modalities thus manage to “speak” through my own fibers: I see by gestalten that home my interiority and determine the external vision. It is exactly through a similar self-organizing process, characterized by the presence of a doubleselection mechanism, that the mind can partially manage to perceive (and assimilate) depth information in an objective way. The extent to which the model succeeds, albeit partially, in encapsulating the secret cipher of this articulation through a specific chain of programs determines the model’s ability to see with the eyes of the mind as well as the successive irruption of new patterns of creativity. To assimilate and see, the system must fi rst “think” internally (at the productive level) the secret structures of the possible, and then posit itself as a channel (through the precise indication of forms of potential coagulum) for the process of opening and unfolding of depth information. This process then works itself gradually into the system’s fibers, via possible selection (by enumeration), in accordance with the coagulum possibilities and the meaningful connections offered successively by the system itself at the co-evolutive level. Herein we can recognize the progressive construction of a living map relative to the actual realization of a specific embodiment process. The revelation and channeling procedures thus emerge as an essential and integrant part of a larger and coupled process of self-organization. In connection with this process we can ascertain the successive edification of an I-subject conceived as a progressively wrought work of abstraction, unification, and emergence. The fi xed points which manage to articulate themselves within this channel, at the level of the trajectories of neural dynamics, represent the real bases on which the “I” can reflect and progressively constitute itself. The I-subject can thus perceive to the extent in which the single visual perceptions are the end result of a coupled process which, through selection, fi nally leads the original Source to articulate and present itself as true invariance and contemporarily as “harmony” within (and through) the architectures of reflection, imagination, computation, and vision, at the level of the progressive constitution of a body and “its” intelligence: the body of “my” mind. These perceptions are (partially) veridical, direct, and irreducible. They exist not in themselves, but, on the contrary, for the “I”, but simultaneously constitute the primary departurepoint for every successive form of reasoning perpetrated by the observer. As an observer I shall thus witness Natura naturata since I have connected functional forms at the semantic level in accordance with a successful and

254

Arturo Carsetti

coherent “score”. The emergence of the “I” is indissolubly linked to the successful coordination of the fi lters and manifests itself through the vision in action. It is precisely through a coupled process of self-organization of the kind that it will finally be possible to manage to define specific procedures of reconstruction and representation within the system, whereby the system will be able to identify a given object within its context, together with its Sinn. The system will thus be able to perceive the visual object as immersed within its surroundings, as a self-sustaining reality, and, at the same time, feel it living and acting objectively within its own fibers. In this way it will be finally possible for the mind to perceive (and assimilate) depth information according to the truth (albeit partially). As Kanizsa maintained, the world perceived at the visual level is constituted not by objects or static forms but by processes appearing “imbued” with meaning. The line per se does not exist: only the line which enters, goes behind, divides, and so on: a line evolving according to a precise holistic context, in comparison with which function and meaning are indissolubly interlinked. The static line is in actual fact the result of a dynamic compensation of forces. Just as the meaning of words is connected with a universe of highly dynamic functions and functional processes which operate syntheses, cancellations, integrations, and so on (a universe which can only be described in terms of symbolic dynamics), in the same way, at the level of vision we must continuously unravel and construct schemata, simulate and assimilate, make ourselves available for selection by the coordinated information penetrating from external reality, and, in particular, continuously adjust our action in accordance with the internal selection mechanisms through a precise “journey” into the regions of intensionality. *** Given a structure Ψ = n≥1, G ∈ O.C.> and a second-order language L’, we can distinguish many kinds of relations. For instance, we can distinguish: (a) fi rst- and second-order relations on the universes of the structure; (b) relations into the universes of the structure; (c) defi nable relations of the structure using a given language, and so on. It is important to underline that these kinds of relations which we have just referred to are not always restricted to the category of relations among individuals. In other words, not all of them are first-order relations: in this way we can realize that hidden in the structure, but defi nable with the second-order language in use, there exist some relations that do not appear as relations among individuals but are utilized in order to defi ne fi rst-order relations. On the other hand, we know that in the universes of any secondorder structure Ψ there are only relations among individuals; when the structure is standard, all the relations among individuals are in the universes of the structure. As M. Manzano (1996) correctly remarks, in standard structures all the n-ary fi rst-order relations on Ψ are into Ψ.

The Emergence of Mind at the Co-Evolutive Level 255 In this sense, when we are faced with a standard structure the ur-elements are fi xed and we cannot “inspect”, with respect to the inner relational growth of the structure, the successive unfolding of some specific depth dimensions different from the simple dimension relative to the full exploitation of the “surface power” of the structure itself. Things are different when we take into consideration structures with non-full relational universes. In order to understand the secret nature of this kind of passage it is useful to examine more carefully the problematics concerning the definition of non-standard models. As is well-known, Skolem discovered the existence of non-standard models of arithmetic in the thirties. At the end of the forties Henkin utilized non-standard structures in order to prove his famous weak completeness theorem for the theory of types and, at the same time, outlined a non-standard model of N2 . In order to present the modalities of construction of this kind of nonstandard model, let us, fi rst of all, show how to build a non-standard model of the fi rst-order theory of Peano arithmetic (N1): a very well-known model, which results non-isomorphic with the structure < ℕ, 0, S > (i.e., with the intended model of N1 in the natural numbers). The construction can also be carried out for the enlarged model <ℕ, 0, S, +, · >. Consider the theory N1* which results from N1 by adding the individual constant “a” together with the following infi nite sequence of axioms, one for each natural number: a≠0 a ≠ S0 a ≠ SS0 ... It is easy to show that the infi nite set of axioms of N1* will be consistent if N1 is consistent. Now, by Gödel’s incompleteness theorem, any consistent set of fi rst-order formulas has a model. But, the intended interpretation of N1 in the natural numbers cannot be a model of N1*. Actually, any model Q of N1* must be a model of N1 and, at the same time, a model of the new formulas a ≠ 0, a ≠ S0, a ≠ SS0. Therefore, the universe of Q contains nonstandard numbers. We know that for every infi nite cardinal, there are at least 2ℵ° non-isomorphic models of N1 of that cardinality. Those models of N1 that are isomorphic with the intended model of N1 are its standard models. All other models are non-standard models. Now, let us imagine building a particular extension of our non-standard model of the fi rst-order theory of natural numbers, Q, that is a secondorder structure capable of presenting itself as a model of N2 . Let us call this structure Q’. It is easy to show that if the structure Q’ were, as required, a model of N2 , it must be non-standard in the second-order sense (i.e., such that each Dn ⊆ P Dn and Dm ≠ P Dm for at least one m ≥ 1). Actually, in the universe of Q’ there are non-standard numbers. This means that the set of standard numbers is not in the universe of unary relations of Q’. Thus, the structure Q’ is non-standard in the second-order sense.

256

Arturo Carsetti

When we are in second-order logic, but we make essential reference to non-standard interpretations and allow structures with non-full relational universes, quantification only applies for the sets and relations that are present in the structure. In the general structures of Henkin, for instance, we put in the universes all sets and relations that are parametrically definable in the structure by second-order formulas. In this sense, it is not surprising that the set of standard numbers is not defi nable by a second-order formula in a structure having non-standard numbers. If we indicate with P. Def. (Ψ, L’) the set of all parametrically Ψ-defi nable relations on individuals using the language L’, we can say directly that a given frame Ψ is a general structure iff Dn = P Dn ∩ P. Def. (Ψ, L’). What it is important to stress again is the fact that hidden in the structure some specific relations exist, some “rules” (second-order relations) that cannot be defi ned as relations among individuals, but are utilized to defi ne fi rst-order relations (i.e., relations among individuals). As a result, we obtain a particular structure where the n-ary relation universe is a proper subset of the power set of the n-ary Cartesian product of the universe of individuals. So, whereas in the standard structures the notion of subset is fi xed and an n-ary relation variable refers to any n-ary relation on the universe of individuals, in the non-standard structures, on the contrary, the notion of subset is explicitly given with respect to each model. Thus, in the case of general structures the concept of subset appears directly related to the definition of a particular kind of constructible universe, a universe that we can explore utilizing, for instance, the suggestions offered by Skolem (cf. his attempt to introduce the notion of propositional function axiomatically) or by Gödel (cf. Gödel’s notion of constructible universe). From a more general and philosophical point of view, we can say that at the level of general structures, the relations among individuals appears as submitted to a bunch of constraints, specifications, and rules having a relational character, a bunch that is relative to the model which we refer to and that acts “from the outside” on the successive configurations of the fi rst-order relations. In other words, as we have just remarked, in the universes of any second-order frame Ψ there are only relations among individuals, but it is no longer true that all the n-ary fi rst-order relations on Ψ are into Ψ. These hidden relations, these particular “constraints” play a central role with respect to the genesis of our models. In particular, let us remark that as a consequence of the action performed by these constraints, the function played by the individuals living in the original universe becomes more and more complex. We are no longer faced with a form of mono-dimensional relational growth starting from a given set of individuals and successively exploring all the possible relations among individuals, according to a preestablished surface unfolding of the relational texture. Actually, besides this kind of mono-dimensional growth, further growth dimensions reveal themselves at the second-order level; specific types of development that spring from the successive articulation of the original growth in accordance with a

The Emergence of Mind at the Co-Evolutive Level 257 specific dialectics. Such a dialectics precisely concerns the interplay existing between the fi rst-order characterization of the universe of individuals and the whole field of relations and constraints acting on this universe at the second-order level. As a result of the action of the rules lying at the secondorder level, new dimensions of growth, new dynamic relational textures appear. Contemporarily the original universe of individuals changes, new elements grow up, and the role and nature of the ancient elements undergo a radical transformation. In this sense, the identification of new growth dimensions necessarily articulates through the successive construction of new substrata. The aforesaid dialectics reveals itself as linked to the utilization of specific conceptual tools: limitation procedures, identification of fi xed points, processes of self-reflection and self-representation, invention of new frames by “fusion” of previously established structures, coagulum functions, and so on. Moreover, as we shall see, we have to recognize the presence of specific patterns of selection and differentiation. Discovery procedures, construction processes, coagulum functions, selective pressures act as a chorus of functions in unison in order to shape the varying texture of mental constructions. At the level of this chorus (if successful) omnis determinatio est negatio. The plot of limitation procedures and cancellations of relations progressively constitutes itself as the gridiron of an intellectual order capable of allowing for the successive “production” of specific gestalten. If we are able to recognize and follow the secret path of this order, we can fi nally manage to illuminate the “good” structures and to “read” (and “play”) the progressive embodiment of that Sinn that selectively determines the real constitution of the events. In this sense, in accordance for instance with Németi’s opinion, standard semantics is not logically adequate because it does not include all logically possible worlds as models. On the contrary, in Henkin’s general semantics many “hidden” possibilities are progressively taken into consideration as possible models. We can have, for instance, models with or without GCH (generalized continuum hypothesis). Things are really different in the case of standard semantics. This argument can be extended in a significant way. Actually, according to Gödel’s incompleteness theorem it is possible to prove that a precise link does exist between non-standard models and formally undecidable propositions. On the other hand, we have just seen how it is possible to outline, according to Henkin’s results, a model containing a non-standard number system which will satisfy all of the Peano postulates, as well as any preassigned set of further axioms. We only have to introduce a new primitive “a” and add to the given set of axioms the infi nite list of formulas, a ≠ 0, a ≠ S0, a ≠ SS0 . . . By adding a non-denumerable number of primitive constants biξ together with all formulas biξ1 ≠ bi ξ2 for ξ1 ≠ ξ2 , we may even build models for which the Peano axioms are valid and which contain a number system having any

258 Arturo Carsetti given cardinal. This kind of theoretical construction shows, as we have just said, that no mathematical axiom system can be categorical, unless it constrains its universe of elements to have some specific fi nite cardinal number. From a general point of view, we know that, in accordance with Mostowski’s results, Gödel’s famous undecidable proposition can be simply considered as a proposition that characterizes the class of natural numbers. If we refer this proposition to a system of non-standard numbers, it will be no longer valid. In this way, we can realize that, along our exploration, we are really driving specific “conceptual” stakes into the ground of an unknown territory and that this exploration articulates in a co-evolutive landscape. At the level of this particular landscape constructing and discovering appear as dialectically interrelated. It is precisely by means of this exploration process that we can ascertain that a formal system can admit models with a universe of individuals that does not have the order type of natural numbers. Henkin explicitly quotes, as an example, a simple result, every non-standard denumerable model for the Peano axioms has the order type: ω + (ω* + ω)η, where η is the type of the rationals. If we make essential reference to non-standard structures, then the set of validities is considerably reduced. At the same time, if we interpret validity as being true in all general models, completeness, Löwenheim-Skolem theorems and other well-known theorems can be proved as in the case of firstorder logic. As a matter of fact, the set of validities will coincide with the set of sentences derivable in a calculus, which is an extension of the fi rstorder calculus. However, this kind of reduction will reveal itself as successful only if we are able to explore the non-standard realm in an intelligent and “creative” way and if the arising differentiation processes articulate in accordance with precise coherence patterns and stability factors. We have just remarked, for instance, that Gödel’s incompleteness theorem concerns a sentence of Z (where Z is a formal system obtained by combining Peano’s axioms for the natural numbers with the logic of type theory as developed in Principia Mathematica) which says of the sentence itself that it is not provable in Z. However, the existence of such a sentence can be identified only because we are able to arithmetize meta-mathematics (i.e., to replace assertions about a formal system by equivalent number-theoretic statements and express these statements within the formal system). In this sense, as we have said before, limitation theorems show that that particular reality (or “essence”) represented by “arithmetical truth” is not exhausted in a purely syntactical concept of provability. From a more general point of view, we can directly affi rm that in Z we cannot defi ne the notion of truth for the system itself. In other words, by constructing a system and then treating the system as an object of our study, we create some new problems, which can be formulated but cannot be answered in the given system. Thus, the limitation procedures permit us to identify the boundaries of our intellectual constructions, to characterize, for instance,

The Emergence of Mind at the Co-Evolutive Level 259 as we have just remarked, the class of natural numbers. They permit us to “see”, once given a specific representation system W, that if W is normal, then every predicate H (the predicates, in this particular case, can be thought of as names of properties of numbers) has a fi xed point. They also permit us, for example, to identify an unlimited series of new arithmetic axioms, in the form of Gödel sentences, that one can add to the ancient axioms. Then, we can use this new system of axioms in order to solve problems that were previously undecidable. We are faced with a particular form of mental “exploration” that, if successful, embodies in an effective construction constraining the paths of our intellectual activity. This exploration concerns the identification of new worlds, of new patterns of relations, the very characterization of new universes of individuals. We shall have, as a consequence, the progressive unfolding of an articulated process of cancellation of previously established relations and the birth of new development “languages” that are grafted on the original relational growth. As we have said before, this type of mental exploration articulates at the second-order level: it can be reduced however (if successful) at the level of many-sorted fi rst-order logic, by means of wellknown logical procedures. In a nutshell, the nucleus of this kind of reduction consists in explicitly showing in many-sorted structures what is implicitly given in second-order or in type theory. According to Post’s famous thesis, any law we become completely conscious of can be mechanically constructed. So, we add to the many-sorted language membership relation symbols and to the manysorted structures membership relations as relation constants. Throughout this reduction process, we simply consider that a second-order structure (or a type theory structure) is basically a peculiar many-sorted structure, since it has several domains. In short, we prove fi rst of all that Henkin semantics and many-sorted fi rst-order semantics are pretty much the same. Then, via Henkin semantics, we establish a form of reduction of second-order semantics to fi rst-order semantics. Second-order logic with the Henkin semantics is, in general terms, a many-sorted logic. However, we immediately have to emphasize that this kind of reduction does not imply that the secret “reasons” that guide, from within the mental activity, the progressive unfolding of the processes of exploration and invention can be reduced to a fi rst-order mechanism or to a set of pre-established rules. In this sense, there must be proofs that are not fully formalizable at a given stage in our mental experience, but that are “evident” to us at that stage on the basis of particular arrangements of limitation procedures, of the successive identification of fi xed points, of the utilization of abstract concepts, of the exploration of new universes of individuals, and so on. In other words, there are, for instance, proofs of Con (PA) (primitive arithmetic) that require abstract concepts as well as the necessary construction of new elements; concepts, for instance, that are not immediately available to concrete intuition (Hilbert’s concrete intuition as restricted to finite

260 Arturo Carsetti sign-configurations). We need, in general, not just rules, but rules capable of changing the previously established rules. In Gödel’s consistency proof, for example, we can directly see that the theory of primitive recursive functionals requires the abstract concept of a “computable function of type t”. According to Gödel, utilizing mathematical reason we are capable of outlining and, at the same time, discovering specific abstract relations that live at the second-order level and that we utilize and explore at that stage in order to defi ne fi rst-order relations. We are faced with a particular “presentation” of the Fregean Sinn, a presentation that selectively constrains the paths of our reasoning in a significant way. So, abstract and non-finitary intellectual constructions are used to formulate the syntactical rules. Once again, this is for many aspects a simple consequence of incompleteness results: mental constructions cannot be exhausted in formal concepts and purely syntactical methods. We have, in general, to utilize more and more abstract concepts in order to solve lower-level problems. The utilization at the semantic level of abstract concepts, the possibility of referring to the sense of symbols and not only to their combinatorial properties, the possibility of picking up the deep information living in mathematical structures open up new horizons with respect to our understanding of the ultimate nature of mental processes. We are actually dealing with a kind of categorial intuition (or rational perception) that does not concern simple data (relative to the inspectable evidence), but complex conceptual constructions. And we know that, in Husserlian terms, meaning “shapes” the forms creatively. However, we must immediately remark that categorial intuition appears to embody in a realm that is far beyond the limits of Gödel’s primitive suggestions, in particular of his primitive Platonist approach. Actually, at the level of the articulation of mental constructions, we are faced with the existence of precise forms of co-evolution. On the one hand, we can recognize, at the level of the aforementioned process of inventive exploration, not only the presence of forms of self-reflection but also the progressive unfolding of specific fusion and integration functions; on the other hand, we fi nd that the Sinn that embodies in specific and articulated rational intuitions guides and shapes, in a selective way, the paths of the exploration. It appears to determine, by means of the defi nition of precise constraints, the choice of some privileged patterns of functional dependencies, with respect to the entire relational growth. As a result, we can inspect a precise spreading of the development dimensions, a selective cancellation of relations and the rising of specific differentiation processes. We are faced with a new theoretical landscape characterized by the unfolding of a precise co-evolution process, by the articulation, in particular, of specific mental processes submitted to the action of well-specified selective pressures, to a continuous “intervention” of depth information determining the successive appearance, at the surface level, of specific gestalten. This intervention, however, could not take place if we were not able to explore the non-standard realm in the right way, if we were not capable of

The Emergence of Mind at the Co-Evolutive Level 261 outlining adequate non-standard models and continuously comparing our actual native competence with the simulation recipes. Meaning selection is creative because it determines ever-new symbolic functions, ever-new processing units which support the effective articulation of new coherence patterns. And, it is precisely by means of these new patterns that we shall be able to “narrate” our inner transformation, to become aware of our mental development and, at the same time, to ascertain the objective character of the transformation undergone. *** With respect to this frame of reference, Reality presents itself as a set of becoming processes characterized by the presence-irradiation of a specific body of meaning and by an inner (iterative) compositio of generative fluxes having an original character. These processes then gradually articulate through and in a (partially consistent) unifying development warp with internal fluctuations and integrations of generative patterns. It is this functional, self-organizing and “irradiating” warp, in the conditions of “fragmentation” in which it appears and is reflected at the interface level, that the model, through the unfolding of the canalization process, progressively manages to read, reconstruct, and “enumerate” as regards its specific functional aspects and living constraints, ultimately synthesizing and reflecting it into an operating architecture of causal programs. In this way, it is then possible to identify (and inscribe) a whole, complex “score” which will function as the basis for the reading-reconstruction of the aforementioned functional warp. However, to read-identify-represent, the score will necessarily require the contemporary discovery-execution of the underlying harmony. Only the individual capable of representing and tuning the work as living harmony, and the score as silent object, will actually be able to depict himself/herself as “I” and as subject. This individual will then not only be able to observe objects, but will himself/herself be able to see the observing eye, modeling those objects. The “I” able to portray himself/herself as such will be able to rediscover the root of the very act of seeing, positing himself/herself as awareness and as the actual instrument allowing for both the progressive surfacing by forms of the original Source and the semantic “grafting” at the deep level. It is thus through the progressive metamorphosis and the embodiment of the model that new Nature can begin to speak, and Reality can channel itself (in primis as regards the “external” selection), in accordance with its deep dimension, ultimately surfacing and expressing as an activity of synthetic multiplication. It is the face-texture of the effected reconstruction which provides the guidelines for the I’s edification; and indeed the original Source which gradually surfaces reflects itself in the constructed work, thereby allowing the fi nal emergence of an “observer” which reveals fi nally itself as a cognitive agent able to observe the Nature around him/her in

262

Arturo Carsetti

accordance with the truth (i.e., we are actually faced with the very multiplication of the cognitive units). The system is thus able to see according to the truth insofar as it constitutes itself as an “I” and as consciousness (i.e., in proportion to the extent it can “see” (and portray-represent) in perspective its own eye which observes things). This can be understood if we start from a number of simple considerations. The visual process, as stated previously, occurs within a coupled system equipped with self-reflection, in which a precise distinction obtains between vision and thought, although they maintain a constant and indissoluble functional exchange. A system of the kind subsumes the successive outlining of functions which self-regulate, as well as the progressive construction of increasingly forms of real autonomy. That function which self-organizes with its meaning, and which posits itself as emergent, is “experienced” as vision insofar as it manages to establish itself, at the network level, as a specific modulation and integration of biological circuits capable of realizing a partial but adequate engraving of the surface representation of the original Sinn (of the deep processes of “scanning” and unification articulating at the level of the system-body of meaning). Insofar as we execute a coherent “enumeration” of the involved virtual “scannings” we, actually, sacrifice to the meaning feeding thus its very body-system. The resulting picture is of a world characterized by continuous emergence and by a constant composition and restructuring of schemes. In this sense, as we have just said, vision extends within a coupled system characterized by the presence of a double selection; throughout this process, meaning reveals itself (albeit partially) in (and through) the effected emergence. Only in this way can a real assimilation process articulate, on the basis fi rst of all of a coherent construction of possible schemes, falsification acts, and so on. This process can then gradually recognize itself in the realized emergence as an act of vision concerning the emergence itself. In self-organizing emergence, then, we fi nd, simultaneously, a process of assimilation, one of growth, one of “inscription” and one of stabilization and reduction through fi xed points. It is therefore not surprising that, as soon as the assimilation (and the unfolding by unification) of meaning occurs correctly, vision appears veridical. What this particularly presupposes as an essential component of the process is also the articulated presence of defi nite capacities of self-reflection and precise replicationmechanisms at the level of the metamorphosis of the original informational flow. Moreover, it is equally clear that there can be no effective vision, at the level of the model, unless specific elaboration has taken place able to “coagulate” the selective activity. The outline offered by the model serves fi rst of all to propose possible integration schemes able to support and prime the nesting proper to the meaning selection. The model must constitute itself as the real carrier and the witness of the “enumeration” process. At the moment of the realization of the embodiment, new vision by gestalten emerges, and the outline as independent instrument is abandoned because

The Emergence of Mind at the Co-Evolutive Level 263 superseded. In this sense, it is true that at the level of the eyes of mind we have visual cognition, and not intellectual reading. Function and meaning articulate together, but in accordance with the development of a process of adequatio, and not of autonomous and direct creation. I will be unable to think of vision during emergence, but will be able to use it, once realized, to construct further simulation models. Growth, modulation, and successive integration thus exist ‘within and among’ the channels together with specific differentiation processes. In order to see more and more I have to support a better canalization of the original flow and to feed a more coherent “circumscription” activity at the meaning level. Hence specific Forms will reveal themselves as natural forms through the progressive realization of my embodiment: in order to join meaning and canalize the Sinn I have to “fi x” the emerging flow into the genesis relative to the Form, I have to give life to specific prototypes, and I have to recognize myself by identifying previously my role in and through them. Amodal completion, for instance, emerges in this context as a privileged window opened on a microcosm which is largely articulated according to the fibers proper to the architecture of mind. Objects are identified through the qualities elaborated and calculated along and through the channels. I neither colonize nor occupy, to use Freeman’s words: I offer myself as a gridiron and I am selected. What remains on my flesh, the operative selection, is the inscription by means of cancellations and negative engravings of the deep functional patterns according to which the real processes “pulsate”. The simulation model thus constructed permits a more coherent integration and articulation of the channels, laying the foundation for the selforganized synthesis of ever-new neural circuits. Objects, in their quality of being immersed in the real world, then emerge as related to other objects possessing different features, and so on. Through and beyond these interrelations, holistic properties and dimensions then gradually reveal themselves, and I must grasp them in order to see the objects with their meaning, if I am to understand the meaning of things. Apples exist not in isolation, but as objects on a table, on a tree; they are, for instance, in Quine’s words, ‘immersed in red’, a reality I can only grasp by means of a complicated second-order process of analysis, elaboration, and comparison which can thereafter be reduced, through concatenations of horizontal and vertical constraints, specific rules and the successive determination of precise fi xed points, to the fi rst-order level. I thus need constant integration of channels and formal instruments to grasp information of the kind (i.e., to assimilate structural and holistic relations and relative ties in an adequate way). In other words, I will understand the meaning of things only if I am able to give the correct coagulum recipes with a view to being selected so as to grasp and capture not only the superficial aspects of objects in the world, but their mutual relations as they interact in depth—in obedience, for instance, to a specific intensional dimension. Only if I provide the correct coagulum, and select the right languages, will these relations emerge

264

Arturo Carsetti

through the trigger operated by the “selective” procedures proper to depth information. Information about the outside world and the “genealogical” apparatus “feeding” it is thus extended: hence the need for a guide to the fi rst articulation and growth of the mechanisms of vision, the need for an intended unfolding of a particular “thread”: the thread of operational meaning, Ariadne’s thread, primarily. The rhythm-scanning of the original labyrinth operated by Ariadne together with the operations concerning the “replica” in action allow the eyes of mind to open: herein we can recognize the progressive opening of the eyes of the Minotaur led by the hand through the process of metamorphosis. Ariadne is a lesson in how to “think by forms”: how to order and unify (according to a semantic unification and a specific representation process) generative thoughts and functional patterns in order to see, while the Minotaur represents the far-flung multiplicity of channels, the pure creativity in action that progressively canalizes itself as (and through) a model (on the basis of the full unfolding of telos’ activities), thus allowing for the emergence of a specific (and multiple) mental activity. Selected and guided by Ariadne, and beginning mentally to perceive her (at the level of the inner “replica” effected on the basis of the blueprint of the original scanning), he becomes aware of a new process of self-organization articulating within his channels. Besides thinking of Ariadne, he will also be able to see an external world to the extent to which he himself has been successfully selected by means of intended “circumscriptions” and fi lled by intuitions (and he will also be able to see himself as a part of the worldNature: i.e., as an observer). The self-organization of the channels coincides with the successive stages in his metamorphosis, with his own gradual cognitive development and with his very achievement of a form of effective, intellectual autonomy. It is exactly within the secret paths of this process of metamorphosis that we can ascertain the objective articulation, at a fi rst (and deep) level, of the specific procedures proper to rational perception as well as the actual expression of the autonomy proper to the telos.

BIBLIOGRAPHY Atlan, H. (1992). Self-organizing networks: Weak, strong and intentional, the role of their underdetermination. La Nuova Critica, 19–20, 51–71. Carnap, R., & Bar Hillel, Y. (1950). An Outline of a Theory of Semantic Information (Technical Report No. 247). Cambridge, MA: MIT. Carsetti, A. (2000a). Randomness, information and meaningful complexity: Some remarks about the emergence of biological structures. La Nuova Critica, 36, 47–109. . (Ed.) (2000b). Functional Models of Cognition. Self-organizing Dynamics and Semantic Structures in Cognitive Systems. Dordrecht: Kluwer. . (Ed.) (2004). Seeing, Thinking and Knowing. Meaning and Self-Organisation in Visual Cognition and Thought. Dordrecht: Kluwer. Chaitin, G. (1987). Algorithmic Information Theory. Cambridge: Cambridge University Press.

The Emergence of Mind at the Co-Evolutive Level 265 Feferman, S., et al. (Eds.) (1986). Kurt Gödel: Collected Works (Vol. I). Oxford: Clarendon Press. . (1990). Kurt Gödel: Collected Works (Vol. II). Oxford: Clarendon Press. . (1995). Kurt Gödel: Collected Works (Vol. III). Oxford: Clarendon Press. Gaifman, H. (2000). What Goedel’s incompleteness result does and does not show. Journal of Philosophy, 97, 462–470. Grossberg, S. (2000). Linking mind to brain: The mathematics of biological intelligence. Notices of AMS, 1358–1374. Henkin, L. (1950). Completeness in the theory of types. Journal of Symbolic Logic, 15, 81–91. Hintikka, J. (1970). Surface information and depth information. In J. Hintikka, & P. Suppes (Eds.), Information and Inference (pp. 298–330). Dordrecht: Reidel. Kanizsa, G. (1980). Grammatica del vedere. Bologna: Il Mulino. Kohonen, R. (1984). Self-organization and Associative Memories. Berlin: Springer. Kozen, D., et al. (Eds.) (1982). Logic of Programs. Berlin: Springer. Manzano, M. (1996). Extensions of First-Order Logic. Cambridge: Cambridge University Press. Németi, I. (1981). Non-standard dynamic logic. In D. Kozen et al. (Eds.) (1982). Talmy, L. (2000). Toward a Cognitive Semantics. Cambridge: Cambridge University Press. Van Dalen, D. (1983). Logic and Structure. Berlin: Springer. Wang, H. (1974). From Mathematics to Philosophy. New York: Routledge.

14 Emerging Mental Phenomena Implications for Psychological Explanation Alessandro Antonietti

1. EMERGENCE IN PSYCHOLOGY The term “emergence” is seldom used by psychologists. An inquiry carried out in Psych-Info (the largest and mostly used databases of scientific articles and books about psychological issues), considering works published starting from the fi rst decades of the last century and by using “emergence” and “emergentism” as keywords, produced a short list of titles. If we exclude papers which addressed the topic by assuming a philosophical perspective (e.g., McLaughlin, 1999; Saeger, 2006), only a few papers are relevant and can highlight how emergence was conceived by psychologists or with reference to psychology. Furthermore, in Sawyer (2002), which appears to provide us the most relevant contribution to our subject, since it was aimed at analyzing the presence and influence of the concept of “emergence” in scientific psychology along its history, we fi nd that few scholars mentioned are psychologists and few theories reported are genuine psychological theories. Most authors considered in the paper are philosophers (Mill, Lewes, Morgan, Bergson) and contemporary philosophers of mind (Davidson, Fodor, Kim), social scientists (Mead, Durkheim, Parsons, Menger, von Hayek), and present-day sociologists. As a third proof of the poverty of the psychological contribution to emergentism, we note that in Clayton’s (2004) book Mind and Emergence—a book in which both the philosophical and the scientific approaches are taken into account—psychology is not included in the scientific disciplines considered (whereas physics, chemistry, biology, neurosciences, and artificial intelligence are considered), and only a couple of psychologists are quoted. The overall impression that can be derived from these remarks is that the concept of “emergence” is imported by psychologists from the philosophical tradition as an extrinsic concept useful to provide a general viewpoint when they have to imbed their statements within a theoretical framework. In other cases, psychologists explicitly employ the concept of “emergence” when they make reference to evolutionary issues (Blitz, 1992); for instance, when they are speaking about the phylogenesis of mental abilities (e.g., Ramachandran & Hubbard, 2005). With reference to psychological topics,

Emerging Mental Phenomena

267

the term “emergence” is occasionally employed—in a more technical sense—also in association with cognitive development (e.g., Bates et al., 1998), with exceptional abilities such as giftedness (Dai, 2005) or creativity (McDonough, 2002; Sawyer, 1999), and with motivation (Csiksentmihalyi, 1978). A more widespread use of the term “emergence” occurs in psycholinguistics (Ellis, 1998; Ke & Holland, 2006; Lantolf, 2006; MacWinney, 1999; Nelson, 1996). However, these cases fail to help us in discovering a specific meaning of “emergence” within psychology.1 In the aforementioned research fields, in fact, the term “emergence” refers only in a broad sense to what is usually meant as emergentism, by denoting simply the fact that a mental capability or aspect appears as a novel or uncommon thing in comparison with what preceded it. However, such a pessimistic conclusion, based on the lack of original psychological contributions to the debate about emergentism, can be contrasted if we assume another perspective. In fact, it appears that the concept of emergence, even though the terms “emergence” and “emergentism” are not explicitly employed, is shared by some psychologists and psychological traditions (Sawyer, 2002). Wilhelm Wundt proposed the principle of “creative resultants” which states “the fact that in all psychical combinations the product is not a mere sum of the separate elements that compose such combinations, but that it represents a new creation” (quoted in Sawyer, 2002: 7). The principle that the whole is more than the sum of the parts is traditionally attributed to Gestalt psychology. Wolfgang Köhler, the Gestalt psychologist who tried to give the most systematic account of the grounds of the Gestalt theory, maintained that mental high-level structures emerge from the patterns of distribution of electro-physiological forces in the cerebral cortex (Stadler & Kruse, 1994). Kurt Lewin extended such a concept and applied it to social phenomena, by arguing that a group of persons is a dynamical whole in which each member is interconnected to the other ones—so that a change in any component of the group influences the other components— and that a group has distinctive features which depend on the kind of relationships which have been established among the members rather than on the intrinsic qualities of the individuals themselves (in this sense, such features can be seen as emergent holistic properties of the components). According to Sawyer (2002), a sort of emergentism can be detected also in Jean Piaget’s theory of cognitive development. According to the kind of constructivism he defended, a new, higher-order schema—a holistic organization of mental elements—emerges from the lower-order level of intellectual functioning not on the basis of an additive process, but as a qualitative transformation of the key features of the previous schemata. A different kind of emergentism can be identified in the socio-cultural school initiated by Lev Vygotskij and extended by Aleksandr R. Luria. In this perspective two forms of emergentism are suggested. The fi rst one concerns the passing from neural activity to mental life: the brain (lower level) is the basis (both in phylogenetic and ontogenetic sense) of the mind (higher

268

Alessandro Antonietti

level), but, once mental activities appear, they constitute an autonomous level, which merits to be described and explained through psychological, but not neurobiological, concepts. The second form of emergentism concerns the relationships between the individual (lower level) and the society or the culture (higher level). The culture is produced by individuals thanks to the social organization of their life and working activities, but it can not be reduced to the sum of the individual contributions, since it is a functionally integrated whole. In this sense, culture emerges from individual mental states. Furthermore, culture exerts a causal power on the individuals’ minds since the latter ones grow by internalizing the structures embedded in the former. The most recent reference to emergentism that can be identified in psychology is provided by connectionism. According to connectionism (or better, to the original connectionist perspective, the so-called Parallel-Distributed-Processing or PDP models), mental activity is supported by neural nets, that is, conceptual systems composed by individual units (the nodes) interconnected with each other. Activation flows through the net starting from the input units toward the output units and passing through the intermediate (called “hidden”) units. Each unit is characterized by a given level of activation, expressed by a number. A number (weight) defines also the facility/difficulty with which activation can be transferred to the other units through the links connecting the emitting unit to the receiving units. If weights are higher than 1, activation is amplified (facilitation); if weights are negative, activation is reduced (inhibition). After each flow of activation the weights change as a consequence of the gap between the actual and the expected activation of the output units. The attempt is to try to make the activation of the output units closer and closer to the expected values. After weight changes, a new course of activation flows through the net. This recursive process lasts until the expected values are reached. When this happens, it is assumed that the net has learned to carry out a given psychological operation (for instance, to discriminate between two categories of stimuli) or to represent a certain state of the world. In this sense a mental property emerges from the functioning of the net. Knowledge or ability is not localized in specific units; it is distributed in the whole net since its counterpart is the pattern of weights reached by the net. In the last decades other computational systems, devised according to different principles (for instance, the dynamic systems theory), have been proposed: also in these cases, however, the mental states and operations that such systems allow to simulate are conceived as emerging properties of an information-processing device. It is clear, from these references, that different interpretations of emergentism can be identified in psychological theories. Indeed, as Saeger (2005) noted, emergentism is not a unified theory but rather a set of related statements characterized through a variety of distinctions. In order to better analyze how emergentism, even if the explicit term was not used, can be conceptualized in psychology, some points have to be clarified.

Emerging Mental Phenomena

269

2. IN WHAT SENSE ARE PSYCHOLOGICAL PHENOMENA EMERGENT? If we try to consider the different meanings associated to emergentism, we can say that “emergence” occurs when at least one of the following conditions is satisfied (Bedau & Humphreys, 2008). 1. Ontological heterogeneity. A higher-level phenomenon (HLP) is different in its own nature from the lower-level phenomena (LLP) which constitute its grounds. In other words, HLP and LLP belong to different levels of reality, to different ontological domains. It might be the case of the relationships between living organisms (HLP) and non-organic structures and processes (LLP), between the mind (HLP) and the brain activity (LLP), between society or culture (HLP) and individual human beings (LLP). However, since psychology deals with mental phenomena—namely, with a homogeneous range of experiences (in the sense that they belong to the same ontological stuff, that is, mental)—this aspect of emergentism is not the most relevant here. 2. Property asymmetry. A HLP has a kind of properties different from the kind of properties of the LLP, even though both HLP and LLP properties belong to the same ontological realm (for instance, they both are mental properties). An HLP possesses properties which are not shared by the components of the LLP, either as singular elements or as sets of elements. For instance we can say that a melody has emergent properties as compared to the properties of the individual notes which constitute that melody. Each tone has physical properties (pitch, duration, intensity) as well as phenomenal properties. The phenomenal properties can be conceived as “emergent”, in the ontological sense mentioned before, with respect to the physical properties (Antonietti, 2008). However, a different kind of emergence can be identified within the realm of the phenomenal properties, that is, within the same ontological domain pertaining the subjective experience of the tones. A tone has phenomenal properties directly grounded in its physical objective properties (the so-called secondary properties). For instance, timbre is an acoustic property depending on the number and type of harmonics produced by the vibrating object, which contributes to generate the shape of the sound wave. A tone also has tertiary properties, such as tenderness and tension. These properties are linked to the physical features of the tones (for instance, it is unlikely that an acute, loud, and very short tone is perceived as smooth), but such features per se fail to give reason of why such a tone is smooth. In this sense smoothness can be conceived as an emergent property of the sounds. There is however a third kind of emergence. Contour is a property which is shared by a sequence of tones (melody) but not by tones considered separately. The same is true of consonance: a tone can be qualified as consonant or dissonant only if it is associated to other tones (so to constitute a chord). Contour and consonance/dissonance are properties which sounds possess when they are assembled but not when they are in isolation. These

270 Alessandro Antonietti properties are acoustic properties (they belong to the same ontological domain), but in the HLP (melody, chord) they appear to be additional and different properties with respect to the properties of the single elements of the LLP. In this sense they can be meant as emergent properties. The cases of emergence mentioned before merit to be thoroughly analyzed. Obviously, if something has more than one element, it is different from the individual element. As Crane (2001: 184) notes, the fact that the whole is not identical to its parts can be meant in a blatant sense: an iron block which weights 100 kilos has a feature (weighting 100 kilos) that is not possessed by each of the ten components which constitute such a block, each of which weights 10 kilos. In this case no emergence occurs since it is the same property (weight) which has different values (measured along the same dimension, with the same unit, expressed with the same symbolic system, and so on) in the individual and in the aggregated blocks. However, there is a second, less obvious sense that does not represent a case of emergence. Consider the three angles reported in Figure 14.1. If I combine them in a certain way (Figure 14.2), I can obtain a triangle (Figure 14.3). Has such a triangle emergent properties as compared to the three angles? It has a closed shape, whereas the components are open shapes. As a consequence, I can compute the measure of the area of the triangle, whereas this is not possible for each angle. But these properties (being closed, having an area) can be derived from the properties of the components. The same is true for the triangle resulting by rearranging the position of the three parts illustrated in Figure 14.4, so to constitute the Figure 14.5. By contrast, if I locate these parts as shown in Figure 14.6 (the so-called Kanizsa’s triangle), I can perceive two partially overlapping, opposite oriented triangles. Such triangles have properties which are neither possessed by the individual components nor derived from them. For instance, the triangles are placed into a three-dimensional space, one triangle being “put” on the other one, whereas the components lie on a bi-dimensional space. The triangles are different with respect to their brightness, whereas as the space delimited originally by the components has identical color qualities. Thirdly, the triangles are delimited in part by boundaries which do not have a physical consistency. All these properties—three-dimensionality, color differences, not-physical edges—not only fail to be possessed by the individual components, but also can not be derived (analytically, logically) by the components. They are genuine emergent properties, since they are visuo-spatial properties (so belonging to the same ontological domain as the properties of the components: it is not the case of ontological emergence), but they are different from the (visuo-spatial) properties of the constituents, so that an asymmetry between the properties of the parts and the properties of the whole can be identified. Such a kind of asymmetry occurs when the individual properties of the LLP components fail to give reason of the HLP properties, namely, when there is nothing in the properties of the components which can motivate or justify the appearance of the HLP property.

Emerging Mental Phenomena

271

Figure 14.1 A non-emergent phenomenon: Take three separate angles.

Figure 14.2 A non-emergent phenomenon: Combine three angles.

Figure 14.3 A non-emergent phenomenon: Obtain a triangle.

Figure 14.4 Another non-emergent phenomenon: Take three round shapes.

3. Bi-directional causality. The HLP is grounded on the LLP. This implies that the components support, produce, create, etc., the whole: the whole would not exist without the components or, better, that specific whole would not exist without those specific components. In addition, if

272

Alessandro Antonietti

Figure 14.5 Another non-emergent phenomenon: Obtain a triangle.

Figure 14.6 An emergent phenomenon: Obtain Kanizsa’s triangle.

the components (even only one component) change, the whole also changes. In this sense the components have a causal power on the whole (bottom-up causation). On the other side, however, the whole can determine the behavior or the meaning of the components (either of all the components or of an individual component). In this sense, the whole has a causal power on the components (top-down causation). If we come back to the example of the tones and the melody: changing even a single tone influences the musical meaning of the whole melody (bottom-up causation); on the other hand the melody gives sense to a tone embedded in such a melody, namely, assigns a specific expressive value to that note (thus inducing, for example, the interpreter to play that tone in a given way or leading the listener to perceive that tone as possessing a given expressive value) (top-down causation). In the Kanizsa’s triangle, different components (or the same components arranged in a different way) might modify the perception of the triangles (see the difference between Figure 14.5 and Figure 14.6); on the other hand the triangles “impose” a specific interpretation to the components (for instance, in Figure 14.6 the circular shapes are perceived not as circles “deprived” of a sector—as happens when they are located one after the other along the same line (Figure 14.4)—but as full circles lying under the vertices of a superimposed triangle.

Emerging Mental Phenomena

273

4. Epistemological distinction. The HLP needs to be described, analyzed, and/or explained by a kind of concepts, categories or principles which are different from those adopted to describe, analyze, and explain the LLP. To make reference to the previous examples again: when I speak about a melody, I employ a kind of evaluation criteria and develop a kind of comments which I cannot apply to the single tones constituting that melody. The same is true of the Kanizsa’s triangle: I have to use a novel vocabulary (to describe the color qualities, for example) and introduce concepts such as “modal completion”, “perceptual vectors”, “field forces” (to explain the not-physical edges of the triangles), and so on, which are not required by, and are not relevant to, the individual components of the figure.

3. SYNCHRONIC AND DIACHRONIC EMERGENCE IN MENTAL PHENOMENA Hence in psychology we can encounter instances of emergence, namely, phenomena which show properties—usually holistic (in the sense that they encompass a set of separate elements) and unitary (in the sense that they can not be decomposed without missing their distinctive features) properties—which can not be derived by the single components which precede or lie under them. Now we have to deepen the analysis of these instances of emergence. The Kanizsa’s triangle is a case of synchronic emergence in the domain of psychological phenomena, since it concerns the properties of a system at a time. Synchronic emergence does not concern necessarily static phenomena. The perception of the contour of a melody is a case of emergence which occurs by linking together tones which appear one after the other in a given time interval. Another example of non-static, in this case spatial but not temporal, emergence is the so-called “phi-phenomenon” or “auto-kinetic motion”: two adjacent luminous dots are switched on and off intermittently, so that when you perceive the disappearance of the left dot, you perceive the concomitant appearance of the right dot and vice versa. If the duration of the time interval during which dots are switched on is progressively reduced, the situation in which you perceive a unique luminous dot moving from left to right and vice versa is reached. The apparent motion is a perceptual property which is not contained in the two elements (the steady dots) which produce it: it emerges. Both the case of the melody and the case of auto-kinetic motion can be considered as examples of synchronic emergence since, even though they include a temporal (melody) and cinematic (phi-phenomenon) dimension, the whole emerges quite from the beginning of the phenomenon. Another type of emergence is diachronic. It concerns not only phylogenetic transitions (Crane, 2001) or ontogenetic transformations, but the fact that in a time and in a specific stage of mental development, or

274 Alessandro Antonietti during a psychological process, or after a period during which the mind operated in certain way, a new way of mental functioning occurs. This leads to apply operations not applied before, to activate a different kind of resources and abilities, to elaborate a different representation, to reach a different level of awareness, and so on. It is a qualitative change since all what preceded it cannot explain what occurred later. As a case of diachronic emergence let us consider the field of analogical thinking (Antonietti, 2001). This kind of reasoning is based on the transfer of ideas from a familiar situation to a novel situation. This induces one to extend some information or principles from a common set of knowledge and experiences to an unfamiliar domain. This extension leads one to view a new situation from a different perspective producing a creative reorganization, or to interpret it in an original way allowing for the discovery of new meanings. This is a process of which artistic intuition, scientific discovery, and technological invention take advantage. Analogies occur also in everyday thinking. Political and social discourses and judgments, as well as advertising messages, are often based on analogies. Legal decision-making is sometimes grounded on analogical reasoning. Moreover, in ordinary language people make use of analogies when they try to explain something, to express their own thoughts, to communicate their emotions and feelings. In all these situations, mapping a situation onto another permits one to identify a wider or deeper meaning in the former and to fi nd an unsuspected meaning in the latter. A field where analogical thinking is often applied is problem solving. In this to reach the solution of a new problem (“target”) persons retrieve a familiar situation (“source”) which is analogous to that problem and which can suggest a relevant response (Antonietti & Maretti, 2003). Recently, a standardized experimental procedure has been followed to investigate analogical problem-solving. Such a procedure consists in presenting a target problem, preceded by a source describing a situation structurally similar to the target, which has been previously solved by means of a set of strategies which can be applied also to the target problem. Consider the following target problem: in a university laboratory a very expensive light bulb used in some experiments does not work because the fi lament inside is broken. The only way to repair it is to use a laser beam. A highintensity laser beam could repair it but would also break the fragile glass surrounding the fi lament; at a lower intensity the beam would neither break the glass nor repair the fi lament. Individuals were asked to fi nd some procedures to repair the fi lament with a laser beam without damaging the glass. One of the possible solutions consists in sending several weak laser beams from different directions toward the fi lament: in this way the weak beams do not damage the glass and, by simultaneously converging on the fi lament, produce a high-intensity effect which repairs the fi lament. The dispersion-concentration solution of the target problem implies four steps:

Emerging Mental Phenomena

275

– substitution of the high-intensity beam with several low-intensity beams; – circular disposition of the weak beams; – simultaneous irradiation; – summation of the intensity of the beams in the central convergence point. These steps are systematically suggested by this source: “An engineer had to plan the construction of an artificial lake to produce electric energy. To feed the lake he thought to build a unique wide canal collecting water coming from a near valley. However, a mason pointed out that during the flood periods the stream of water flowing along the canal might be too strong and might damage the surrounding areas; by contrast, during the drought periods a unique stream of water might be insufficient to feed the lake. In order to avoid these mishaps, the mason suggested to build, instead of a unique wide canal, four small canals whose total flow was the same as the unique wide canal previously planned. These small canals were placed around the lake so that they conveyed water coming from four different valleys. In this way only small amounts of water could flow in each canal and thus during flood periods dangerous overflowing might not occur. At the same time, the lake was fed by water from various belts, so that also during drought periods it was sufficiently fed.” The artificial lake situation shows close structural relations to the lightbulb situation. In fact, the solution of the latter problem requires that two obstacles are overcome: – if a unique high-intensity beam is sent toward the light bulb, the beam will break the glass; – if a unique low-intensity beam is sent toward the light bulb, it will fail to repair the filament. Correspondingly, in the source, in the fi rst plan there are two obstacles: – during flood periods, a unique wide stream of water may damage the surrounding areas; – during drought periods, a unique small stream may be insufficient to feed the lake. Furthermore, in the target problem, according to the dispersion-concentration solution: – the fi rst obstacle is overcome by substituting the unique strong beam with several weak beams; – the second obstacle is overcome by sending several weak beams toward the light bulb from different directions.

276 Alessandro Antonietti Analogously, in the artificial lake situation: – the fi rst obstacle is overcome by replacing the unique wide stream of water with several small streams; – the second obstacle is overcome by placing the small streams in different directions, so that they can convey water coming from different belts into the lake. When people face the light-bulb problem after having exposed the artificial lake story, they usually try to apply different strategies, which however are unsuccessful. At a given point, someone realizes that the dispersionconcentration plan can be applied to the target problem to reach the solution. The impression is that a new idea emerged in the subject’s mind, that is, that reasoning takes a new direction, qualitatively different from those attempted previously. Is this an actual case of emergence? There is a consensus that analogical problem-solving can be divided into the following phases: – the “encoding” of the source; – the “access” phase, which induces individuals to recall a source relevant to the solution of the target; – the “mapping” or “applying” phase, in which subjects construct orderly correspondences between the elements of the source and those of the target to draw a solution. As far as the mapping phase is concerned, the process is usually conceived as follows: – people grasp the correspondence between an element of the source and of the target; – then they fi nd in the source a relationship between the previous element and a second element; – consequently, they map such a relationship onto the target; – the result is that a second source-target correspondence is found. This process continues in this way until the mapping is completed, namely, all the relevant correspondences are identified. In sum, mapping is conceptualized as an incremental mechanism which yields, through the addition of new similarities to those previously found, the gradual extension of the source-target similarities which at the end, through the complete sourcetarget overlapping, leads one to apply the solution strategy described in the source to the target (Gentner, 1989; Holyoak, 1985; Keane, 1990). This view denies that something new emerges in the problem solver’s mind: the goal is achieved step-by-step through the continuous and progressive increase of the correspondences between the source and the target. Now

Emerging Mental Phenomena

277

the question is: do people actually solve problems by analogy through an incremental process as predicted by some theoretical models?

4. DO DIACHRONIC EMERGENT PSYCHOLOGICAL PHENOMENA EXIST? Introspective reports and thinking aloud data (Wedman, Wedman, & Folger, 1999) failed to support the alleged incremental nature of analogical mapping. The same conclusion can be drawn from diary studies where people where asked to track records of the occurrence of analogical reasoning in everyday life and to identify the way they carried out this form of reasoning: interviewees often reported that analogies came in mind suddenly but not through a progressive process leading them to fi nd more and more correspondences between situations.2 Van Lehn (1998), by analyzing protocols of students engaged in physics exercises whose solutions were suggested by previous analogous examples, found that solvers did not proceed in a sequential, progressive way. In addition, a study (Antonietti, Cantoia, & Cordara, 1999a) showed that source-target correspondences are suddenly realized and people seldom transfer the solution to the target by means of a systematic process; furthermore, individuals, just after having focused their attention on the source, quickly fi nd the solution elements and the analogical response is not reached through the sequence of the solution steps predicted by theoretical models. It was also observed that thinking aloud, a procedure that usually improves problem-solving performance since it “imposes” a well-structured way of reasoning leading individuals to proceed step-by-step, fails to enhance analogical transfer (Antonietti, Cantoia, & Cordara, 1999a): this induces to cast doubts about the alleged not-emergent, incremental nature of analogical problem-solving. Data obtained by means of introspective reports and thinking aloud procedures are consistent with what was found in experiments carried out by means of a computer-based problems-solving setting which allowed us to test on-line how participants are perceiving the source-target correspondences (Antonietti, Cantoia, & Cordara, 1999b). College students were presented with three short stories. The structural story (the analogical source) embedded a solution strategy that was isomorphic to that of the target problem, even though it was contextualized in a different domain. The superficial story described a situation in the same domain as the target but the solution strategy included in this story had no relation to the target. The neutral story had neither structural nor superficial correspondences with the target. Participants were asked to memorize carefully the stories. Then students were given the light-bulb problem as the target problem to solve. The target problem was divided into six parts and was presented step by step. At each step a part of the target was presented onto a computer screen. At the bottom of the screen there were three items: participants had

278

Alessandro Antonietti

to rate on a 1- to 10-point scale the similarity they found between each story previously memorized and the target. After this task the whole lightbulb problem was presented and students were asked to solve it. Mean similarity ratings recorded at each step are reported in Figure 14.7. At the beginning of the rating task, the superficial story was perceived as being closely similar to the target, whereas the structural story—as well as the neutral one—was rated low. However, when the first element of the target problem having an analogical correspondence with the source was provided (fourth step), a significant increase of the similarity scores of the structural story occurred. Then, no significant differences in ratings, as compared to the fourth step, occurred. At the same time, in correspondence with the fourth step, there was a significant decrease of the ratings of the superficial story. This pattern occurred also when, in a second experiment, to avoid the evaluation being influenced by the interaction among the sources, the structural source and the superficial source were isolated. In fact, two new conditions were tested where either the structural or the superficial source was presented between two neutral sources. Also in this case, the fourth step was the critical point: similarity ratings for the structural source suddenly increased and ratings for the superficial source decreased. In a third experiment the computer-assisted rating task was applied to the source instead of the target. Participants first read the light-bulb problem but did not have to solve it. Then they were presented with either the structural or the superficial story. The text of the story was divided into six parts and visualized onto a computer screen step by step. At each step participants rated the similarity between the target and the story. Then, the target problem was presented again and students had to solve it. The same picture of similarity ratings was found (Figure 14.8): analogical mapping monitored through this similarity rating task showed a critical point (fourth step) where people suddenly recognized the correspondence between the structural source and the target. The results of this series of experiments induce to cast further doubts about the alleged incremental nature of the mapping process which is hypothesized by several theoretical models. In these experiments analogical problem-solving appeared to be not a continuous but a insight-like process involving mental acts which allow one to perceive, at a given moment, the overall, though not complete, source-target correspondences.3

5. AWARENESS AS A CASE OF EMERGENT PSYCHOLOGICAL PROPERTY If analogical problem-solving, as a case of diachronic emergent psychological phenomenon, has not an incremental nature, but includes a qualitative different aspect which fails to be possessed by its components, what does characterize it? Awareness seems to be the psychological feature that makes the difference, as proven by the following series of experiments (Anolli et al., 2001).

Emerging Mental Phenomena

279

Figure 14.7 Target-source similarity under the conditions of the fi rst experiment concerning on-line rating.

Figure 14.8 Source-target similarity under the conditions of the third experiment concerning on-line rating.

In these experiments two basic conditions were designed: reminding and hint. In the reminding condition, the representation of the source was activated immediately before the presentation of the target by inducing participants to retrieve the critical part of the source relevant to the solution of the subsequent task, with no cue about the connection of such information to the target. By contrast, in the hint condition, participants, before the presentation of the target, were told that the source presented earlier would be useful to the solution of the target itself. The experimenter presented a booklet in which every page corresponded to a task to be carried out. On the fi rst page of the booklet there was a filler story with no relations to the target. Participants had to read the story and then they had to go to the next page. On the second page participants had to solve a puzzle. On the third page there was a question about the fi ller story. On the fourth page

280 Alessandro Antonietti

Figure 14.9 Percentages of analogical solutions under different conditions in the series of experiments concerning the awareness of source-target correspondence.

participants had to examine visually two pictures in order to spot differences. The fi fth page reported the source (that is, the artificial lake story) to be read. On the sixth page there was an anagram. The subsequent task depended on the particular condition the participants had been assigned. In the control condition participants read the target on the seventh page and had to verbalize all the solutions that they found. In the reminding condition, on the seventh page participants found a question concerning the source. More precisely, they were asked: “What did the mason suggest in the artificial lake story?”; then they went to the eighth page where there was the target. In the hint condition, on the seventh page participants found this sentence: “The mason’s proposal mentioned in the artificial lake story can suggest a way for you to solve the next task”; then they could pass to the eighth page where they found the target. Results of Experiment 1 (Figure 14.9) suggest that the lack of spontaneous transfer in analogical problemsolving does not depend on failures in the activation of the relevant source information. In fact, although people were invited to recall source information, they did not transpose it to the target in a significantly greater proportion than did those in the control group. Rather, only when subjects were aware of the relationship between source and target can they apply the solution strategy embedded in the source to solve the target problem. However, these fi ndings might have depended on an experimental artifact. In the reminding condition participants had to turn the page reporting

Emerging Mental Phenomena

281

the question about the source before reading the target problem: this might have induced participants to forget the representation of the source further on. To circumvent this, in Experiment 2 a concurrent reminding condition was devised, in which the question concerning the source was presented on the same sheet as the target. This should induce participants to activate the representation of the source immediately before the target task so that such information is still activated during the target problem because the cue devised for such activation—namely, the question concerning the source—is always available. Participants were presented with the same tasks in the same order and with the same instructions as in the previous experiment. The previous reminding condition was the same as the reminding condition studied earlier. In the concurrent reminding and hint conditions the text of the problem was preceded, respectively, by the question about the mason’s proposal or by the suggestion about the source-target connection. Results (Figure 14.9) showed that also concurrent reminding was ineffective in suggesting the analogical solution to the target problem. One might argue that this occurred because participants engaged in the solution of the target were deeply involved in such task, so that they paid no attention to the cue printed on the top of the page. To “compel” students to activate the source during the target phase, in Experiment 3 the question about the critical part of the source was asked while subjects were looking for the solution to the target. In this way, to answer the source question, participants had to stop thinking about the target and retrieve the source, which consequently should just have been activated during the attempts to solve the light-bulb problem. In this experiment, for all the reminding conditions the procedure was the same as in the concurrent condition of the previous experiment, with the exception that the question about the mason’s proposal was orally presented by the experimenter while participants were engaged in the target problem-solving task. In the first condition the experimenter asked the question after one minute from the beginning of the task; in the second condition this happened after three minutes and in the third condition after five minutes. The hint condition was the same as in the previous experiments, but the experimenter gave the cue about the source-target connection orally one minute after participants began reading the light-bulb problem. Also this experiment failed to support the notion that the activation of the source without awareness of the source-target relationship facilitates analogical problem-solving. The recall of the relevant source information during the target task was ineffective both when the cue was given at the beginning of the reasoning process and when it was given during or after a longer involvement in the target task. Inviting participants to recall information necessary for the solution of the target problem in all the three distinct moments of the problem-solving process did not increase rates of analogical responses as compared to a control condition where such cue was not given. Previous experiments showed that reminding alone did not induce people to relate the source to the solution of the target. However, it might

282

Alessandro Antonietti

be argued that participants did not reach the analogical response because the reminding instructions were too general and not sufficiently focused on the elements of the source which were relevant to the solution of the target. Furthermore, in the reminding conditions the effort to recall the source information might have interfered with its possible subsequent utilization in solving the target. To obviate this, in Experiment 4 different kinds of reminding were designed: an active general reminding, in which participants were asked to repeat what they could remember about all the previous tasks; an active focal reminding, in which they had to explain the mishaps of the engineer’s plan and the advantages of the mason’s suggestion; a passive focal reminding, in which they had to read a summary of the artificial lake story stressing the aspects useful to solve the light-bulb problem analogically. In the last two conditions participants should pay more attention to the critical elements of the source and, consequently, activate a high-focused memory of the isomorphic aspects. The conjecture that mental work needed to retrieve the source may inhibit its transfer to the target was tested through the comparison between the active and passive reminding conditions. Participants were presented with the same tasks in the same order and with the same instructions as in previous experiments. The seventh page of the booklet reported the following instructions, which were different in each condition. The general reminding group was asked: “Relate what you recall about the content of the tasks you have carried out till now”; in the passive focal reminding condition participants were asked to read again the paragraphs of the source story in which the mason’s proposal was described; in the active focal reminding participants were told: “In the artificial lake story, what mishaps of the engineer’s plan did the mason stress, what did he suggest, and what were the advantages of his proposal?”. In this experiment the activation of the source was focused only onto the elements relevant to the solution of the target. However, even in this case reminding was not effective in improving analogical transfer. The previous series of experiments showed that the solution principle embedded in a source story can be applied to an isomorphic target problem of a different domain only in presence of the awareness of the source-target connection. In fact the mere activation of the source immediately before or during the target task has no effect on enhancing analogical transfer unless it is accompanied by a hint suggesting a search for possible source-target similarities. In these experiments the source and the target were semantically unrelated. Keane (1987) found that, if in the target problem an instrument which is identical or semantically similar to the corresponding element of the source is mentioned, a considerable percentage of participants can retrieve the source before or during the target task. This led to wonder as to whether, without hints, the inclusion of such an element in the source favors the analogical solution of the target problem owing to the pure activation of the source itself. To answer this question a new line of investigation was undertaken. More precisely, laser rays—that is, the instrument

Emerging Mental Phenomena

283

used to solve the light-bulb problem—were also mentioned in the source story. In Experiment 5 such a revised story was presented in a control, reminding, and hint condition. The material was the same as in the previous experiment. However, in the source story, instead of the mason, a technician was mentioned and the following sentence “A laser-ray system was devised to monitor the stream of water into the lake so that it would detect both overflowing and insufficient feeding” was included. In the reminding condition, the sheet preceding the target problem reported this question: “In the artificial lake story, what mishaps of the engineer’s plan did the technician stress, what did he suggest, and what were the advantages of his proposal?”. In the reminding condition the percentage of analogical solvers was not significantly higher than in the control condition, whereas in the hint condition many participants reached the dispersion-concentration solution. Thus, if an explicit hint was not provided, the presence of the same elements (laser rays) both in the lake story and in the light-bulb problem did not lead participants to notice source-target correspondences. It might be argued that in the revised version of the source story employed in the previous experiment the introduction of an element (ray) semantically identical to the corresponding element of the light-bulb problem had no effect because the laser ray played an irrelevant role in the artificial lake story, and, consequently, it was likely to be neglected. Therefore, in Experiment 6 the text of the source was changed again so that the laser rays became more crucial and were linked to the central problem of that situation, namely, the need to prevent both overflowing and insufficient feeding. The following sentence was added to the source story: “A laser-ray system was devised to monitor the stream of water into the lake so that it would detect both overflowing and insufficient feeding. More precisely, a laser beam, set at the highest possible water level, would be inactivated if the water rose any higher. Another laser beam, set at the lowest possible water level, would be activated if the water fell under that level”. Also in this case the reminding condition failed to increase analogical transfer as compared to the control condition. The same was true when, in Experiment 7, the critical part of the source was enlarged as follows: “A laser-ray system was devised to control and monitor the stream of water conveyed into the lake so that it would detect and solve overflowing and insufficient feeding problems. More precisely, a laser-ray emitter was located in correspondence with each canal. This emitter would send rays to a station in the center of the lake signaling how much water was conveyed by the canal at that moment. The station would sum up signals coming from the four canals. On the basis of the resulting whole signal the station would give the order either to reduce the stream of water (if excessive) or to increase it (if insufficient)”. In this version of the artificial lake story the element of the source (laser beam) superficially similar to the corresponding element of the target played a notable function both in the control of the water streams and in the solution of the problem that arises from overflowing

284

Alessandro Antonietti

and insufficient feeding, so that the critical element of the source shared by the target had a remarkable role. However, even if the artificial lake story included an element that was superficially similar to a corresponding element of the target problem and such an element was involved in the dispersion-concentration strategy described in the source, reminding source information very close to the target task did not prompt the transfer of the solution strategy to the light-bulb situation as the explicit hint did. Also the latter experiments showed that no spontaneous transfer occurred as a consequence of the immediately previous activation of the source, even though the source and the target shared identical elements and the salience of these elements was progressively increased so much that they were deeply involved in the solution schema underlying both the source and the target. Holyoak and Koh (1987) argued that retrieval of analogies is based on the summation of activation resulting from multiple shared features. If the sum of activation exceeds some thresholds, the representation is retrieved and it can be used for further processing, such as an explicit source-target mapping. According to these authors, retrieval by summation of activation can provide a general mechanism for flexible access to information in the memory that is related to a novel input. An activation model responsible for analogical problem-solving has been proposed also by Anderson (1993). In this perspective analogical transfer occurs when the strength of the activation of the trace of the source is sufficiently high. In that moment the solution procedure embedded in the source becomes so strong that it can be applied to the target so that search for new original solutions is not needed. A question raised by these models is the following: does an adequate degree of activation of source information really enable analogical transfer? Results from experiments on source access showed that activation is not the critical process. The mere activation is ineffective unless persons realize the source-target connection. This awareness does not seem to derive from a summation mechanism.

6. CONCLUSION The initial remarks suggested the impression that psychology provides a poor contribution, on the theoretical level, to the concept of emergence. However, in the field of psychological inquiry several kinds of emergent phenomena can be identified. Such instances of emergence can be useful both to widen the repertoire of the cases of emergent phenomena and to sharpen the analysis of them. In fact, within the broad concept of emergence we can make fi ne-grained distinctions, so as to identify, as we tried to do across this chapter, different forms of emergence. For example, it is possible to distinguish between synchronic and diachronic emergence and, within the domain of synchronic emergence, between static and non-static emergence, or, within the domain of non-static emergence, between spatial

Emerging Mental Phenomena

285

and temporal kinds of emergence. Besides these rather superficial distinctions, it is possible to deepen the analysis in other directions. For instance, assuming that a whole has an emergent property if it is not possessed by the components, we can distinguish between additional emergent properties and transformative emergent properties. If we consider the case of the tones and of the melody, contour is an additional emergent property of the melody since it does not substitute or change previous properties, but rather adds something further to the series of tones. Instead, the expressiveness of a melody is a transformative emergent property since each single tone has per se, to some extent (namely, in nuce), an individual expressiveness, which however can change dramatically when that tone is embedded in a melody. Another distinction which can be made is between predictable and unpredictable emergent properties. The contour of a melody is not contained in each singular tone, but, if I imagine putting three given tones in sequence one after the other, I can figure out both that the resulting sequence of notes will have a contour and that the melody will have that specific contour (for instance, ascending). Contrariwise, given individual elements of the Kanizsa’s triangle, I cannot anticipate what it will result by arranging them in a certain way (unless I have acquired an expertise by managing previously figures of that kind). This kind of distinctions should allow us to improve the analysis of emergence and to refi ne the bases on which we can identify the presence of an emergent phenomenon. This analysis could also have epistemological implications. In order to describe and explain a mental phenomenon, psychology—as well as other scientific disciplines—is tempted by the fascinating promises of reductionism, namely, by trying to explain a given phenomenon in terms of the workings of its components, assumed to be more basic and common to other kinds of phenomena. By considering the case of analogical problem-solving, we realized how the investigation of such thinking process was addressed to detect some core mechanisms (as the recursive identification of sourcetarget correspondences) which can be invoked in order to maintain that it is a particular case of a general cognitive activity, thus ignoring its specific features. Similar attempts were made with reference to insight (Weisberg & Alba, 1981) or creativity (Weisberg, 1986). Such attempts are not per se questionable. Firstly, they contribute to dissipate the aura of “exoticism”, “mystery”, or “exceptionality” which sometimes surrounds certain phenomena. Secondly, they implement one of the general goals of scientific research, that is, unification: obviously, the possibility to explain a wide range of phenomena with the same principle can be seen as an advancement in our understanding of those phenomena, in terms both of “economy” (the description/explanation becomes more parsimonious and simple) and of “depth” (we are led to go beyond the surface appearances of the phenomenon and to postulate more fundamental grounds of it). A problem rises when the reduction produces a distortion of the phenomena, by leading researchers to neglect some aspects of the phenomenon

286

Alessandro Antonietti

in object, and/or by proposing descriptions/explanations based on alleged causes which conflict with the nature of that phenomenon. As regards the case of analogical problem-solving, we saw that the role of awareness, supported by a long series of experiments, was neglected by the incremental models and that the conjectured progressive mapping of source-target correspondences confl icts with the sudden recognition of the similarity between the source and the target. In this sense, emergentism might be conceived as a conceptual framework useful to stress the limits of reductionism when the latter is not the adequate approach, as in case of phenomena showing discontinuities, qualitative “jumps”, sudden transformations, and so on. In these cases, emergentism can justify the adoption of peculiar vocabularies, methods, and explanatory patterns by psychologists who do not share the enthusiasm toward (improper) reductionism.

NOTES 1. Terms associated with emergentism can be found in positions and statements by neuroscientists who provided contributions to psychology or addressed psychological issues, such as Donald O. Hebb (who labeled his solution of the mind-body problem as “emergentist monism”) and Roger W. Sperry (who attributed the qualification of “emergent interactionism” to his view of the relationships between brain and mind). Emergentist claims were supported also by Ragnar Granit and Gerald M. Edelman and by ethologists such as William H. Thorpe, Donald R. Griffi n, and Konrad Lorenz. 2. Unpublished data reported in the workshop “Analogy as a tool for thinking” at the X International Conference on Thinking “Changing Minds: Toward a Thinking Society”, Harrogate, 15–19 June 2002. 3. A further support to the doubt about the alleged incremental nature of analogical problem-solving is provided by the investigation of individual differences: if analogical problem-solving were a systematic, step-by-step process, people possessing analytical skills and styles should perform better than individual with creative and insight skills and intuitive-holistic style, but this was not the case (Antonietti & Gioletta, 1995).

REFERENCES Anderson, J. R. (1993). Rules of Mind. Hillsdale, NJ: Lawrence Erlbaum Associates. Anolli, L., Antonietti, A., Cantoia, M., & Crisafulli, L. (2001). Accessing source information in analogical problem solving. Quarterly Journal of Experimental Psychology, 54A, 237–261. Antonietti, A. (2001). Analogical Discoveries. Identifying Similarities to Solve Problems. Roma: Carocci. . (2008). Must psychologists be dualists? In A. Antonietti, A. Corradini, & E. J. Lowe (Eds.), Psycho-physical Dualism Today. An Interdisciplinary Approach (pp. 37–67). Lanham, MD: Lexington Books. Antonietti, A., Cantoia, M., & Cordara, G. F. (1999a). (Ir)realtà psicologica di alcuni meccanismi della soluzione di problemi per analogia [Psychological (un)

Emerging Mental Phenomena

287

reality of some mechanisms involved in analogical problem-solving]. Giornale Italiano di Psicologia, 26, 401–414. . (1999b). Riconoscimento di somiglianze nel ragionamento per analogia: un transfer progressivo? [Similarity recognition in analogical reasoning: A progressive transfer?]. Giornale Italiano di Psicologia, 26, 567–590. Antonietti, A., & Gioletta, M. A. (1995). Individual differences in analogical problem solvine. Personality and Individual Differences, 18, 611–619. Antonietti, A., & Maretti, C. (2003). Degrees of similarity in knowledge transfer. In n B. Apolloni, M. Marinaro, & R. Tagliaferri (Eds.), Neural Nets (pp. 338–347). Heidelberg: Springer. Bates, E. A., Elman, J. L., Johnson, M., Karmiloff-Smith, A., Parisi, D., & Plunkett, K. (1998). Innateness and emergentism. In W. Bechtel & G. Graham (Eds.), A Companion to Cognitive Science (pp. 590–601). Malden, MA: Blackwell Publishers. Bedau, M. A., & Humphreys, P. (Eds.) (2008). Emergence: Contemporary Readings in Philosophy and Science. Cambridge, MA: Cambridge University Press. Blitz, D. (1992). Emergent Evolution. Dordrecht: Kluwer. Clayton, P. (2004). Mind and Emergence. From Quantum to Consciousness. Oxford, UK: Oxford University Press. Crane, T. (2001). The emergence of consciousness. In The Emergence of the Mind. Proceedings of the International Symposium (pp. 183–191). Milano: Fondazione Carlo Erba. Csiksentmihalyi, M. (1978). Intrinsic reward and emergent motivation. In M. R. Lepper & D. Greene (Eds.), The Hidden Costs of Reward (pp. 205–216). Hillsdale, NJ: LEA. Dai, D. Y. (2005). Reductionism versus emergentism: A framework for understanding conceptions of giftedness. Roeper Review, 27, 144–151. Ellis, N. C. (1998). Emergentism, connectionism and language learning. Language Learning, 48, 631–664. Gentner, D. (1989). The mechanisms of analogical learning. In S. Vosniadou & A. Ortony (Eds.), Similarity and Analogical Reasoning (pp. 199–241). Cambridge, MA: Cambridge University Press. Holyoak, K. J. (1985). The pragmatics of analogical transfer. In G. H. Bower (Ed.), The Psychology of Learning and Motivation, Vol. 19 (pp. 59–87). San Diego, CA: Academic Press. Holyoak, K. J., & Koh, K. (1987). Surface and structural similarity in analogical transfer. Memory and Cognition, 15, 332–340. Ke, J., & Holland, J. H. (2006). Language origin from an emergent perspective. Applied Linguistics, 27, 691–716. Keane, M. (1987). On retrieving analogies when solving problem. Quarterly Journal of Experimental Psychology, 39A, 29–41. . (1990). Incremental analogizing: theory and model. In K. J. Gilhooly, M. Keane, R. H. Logie, & G. Erdos (Eds.), Lines of Thinking, Vol. 1 (pp. 221–235). New York: Wiley. Lantolf, J. P. (2006). Language emergence. Applied Linguistics, 27, 717–728. MacWhinney, B. (Ed.) (1999). The Emergence of Language. Mahwah, NJ: Lawrence Erlbaum Associates. McDonough, R. (2002). Emergence and creativity: Five degrees of freedom. In T. Dartnall (Ed.), Creativity, Cognition, and Knowledge (pp. 284–320). Westport, CT: Praeger. McLaughlin, B. P. (1999). Emergentism. In R. A. Wilson & F. C. Keil (Eds.), The MIT Encyclopedia of the Cognitive Sciences (pp. 267–269). Cambridge, MA: MIT Press. Nelson, K. (1996). Language in Cognitive Development: Emergence of the Mediated Mind. Cambridge, MA: Cambridge University Press.

288 Alessandro Antonietti Ramachandran, V. S., & Hubbard, E. M. (2005). The emergence of the human mind: Some clues from synesthesia. In L. C. Robertson & N. Sagiv (Eds.), Synesthesia. Perspectives from Cognitive Neuroscience (pp. 147–190). New York: Oxford University Press. Saeger, W. (2005). Emergence and efficacy. In C. E. Erneling & D. M. Johnson (Eds.), The Mind as a Scientific Object: Between Brain and Culture (pp. 176– 190). New York: Oxford University Press. . (2006). Emergence, ephipenomenalism and consciousness. Journal of Consciousness Studies, 13, 21–38. Sawyer, R. K. (1999). The emergence of creativity. Philosophical Psychology, 12, 447–469. . (2002). Emergence in psychology: Lessons from the history of non-reductionist science. Human Development, 45, 2–28. Stadler, M., & Kruse, P. (1994). Gestalt theory and synergetics: From psychophysical isomorphism to holistic emergentism. Philosophical Psychology, 7, 211– 226. Van Lehn, K. (1998). Analogy events: How examples are used during problem solving. Cognitive Science, 22, 347–388. Wedman, J. F., Wedman, J. M., & Folger, T. (1999). Thought processes in analogical problem solving: A preliminary inquiry. Journal of Research and Development in Education, 32, 160–167. Weisberg R.W. (1986). Creativity: Genius and Other Myths. New York: Freeman. Weisberg, R. W., & Alba, J. W. (1981). An examination of the alleged role of “fi xation” in the solution of several “insight” problems. Journal of Experimental Psychology: General, 110, 169–192.

15 How Special Are Special Sciences? Antonella Corradini

1. INTRODUCTION Reflection on the epistemological and ontological status of special sciences is reflection on their degree of autonomy or independence from the basic science of physics. The debate on special sciences is thus closely connected to two other major issues in philosophy of science. The fi rst is reductionism; the second is the thesis of the unity of science. For as long as the program of the unification of science has been considered as an epistemological desideratum, even reduction has been viewed as an epistemological virtue (and vice versa). The best-known contemporary representative of this program is logical positivism, and not simply because of its associated project to compile an Encyclopedia of Unified Science. More important were the considerable efforts made by some of its members to rebuild the system of sciences in its entirety on a physicalistic basis. Constitution systems, such as Carnap’s, obviously leave little room for autonomous special sciences, at least when it comes to the language adopted in scientific theories. The constraints they impose on scientific language penalize the human sciences in particular. Psychology, for example, could be conceived as a scientific discipline only insofar as it was in accordance with the basic methodological principles of behaviorism. Other, more typical dimensions of human sciences were considered to be “remainders of theology”, as Otto Neurath once put it (1931: 407). Now, setting aside the internal developments of logical positivism that—as everybody knows—have led to a partial liberalization of the original reductionistic approach to science, it is worth mentioning that at the same time logical positivism was being developed1 another philosophical movement—emergentism—articulated very different views on these topics. Emergentism, which has its roots in nineteenth-century British philosophy, opposed the methodological ideal of unifying science on a physicalistic basis being then pursued by the Vienna Circle. In the introduction to The Mind and Its Place in Nature, a revision of the Turner Lectures delivered by C. D. Broad at Trinity College, Cambridge, in 1923, the author mentions that the purpose of the lectures was

290

Antonella Corradini

to analyze “the relation or lack of relation between the various sciences” (1925: 3). Broad’s main theoretical concern was the epistemological status of “special sciences” such as chemistry, biology, psychology, and others and their relations to physics. Two main views on these issues were offered at that time: mechanism and vitalism. However, in Broad’s view they were both unsatisfactory. One of the advantages of mechanism was that it introduced “a unity and tidiness into the world which appeals very strongly to our aesthetic interests . . . On such a view the external world has the greatest amount of unity which is conceivable. There is really only one science, and the various “special sciences” are just particular cases of it” (1925: 76). Such advantages, however, were obtained at the price of ignoring all those aspects of the external world—such as the existence of secondary qualities—which mechanism was unable to account for. Simplicity and tidiness were therefore the result of an oversimplification of reality and of privileging one science, that is to say, physics. On Broad’s account, however, vitalism is not a better alternative. On the one hand it was able to safeguard the specificity of life sciences, but, on the other, it could do so only under the condition of postulating, as explanatory hypotheses of the peculiar behaviors of living beings, obscure entities like entelechies (1925: 58 ff). Nobody, however—claimed the emergentists—is compelled to choose between a scientific but reductionistic perspective like mechanism and a non-reductionistic but empirically un-testable theory like vitalism. The main ambition of the emergentists, in fact, was to combine scientificity with acknowledgment of methodological pluralism and the autonomy of special sciences. These desiderata are today still on the agenda of many philosophers of science. In the following part of my chapter I shall seek briefly to reconstruct how the debate on special sciences, reductionism, and the unity of science has proceeded in recent decades.2 I shall specifically focus on questions such as: What is needed to ensure autonomy for the special sciences? Is unity a desirable goal to be achieved in science? Is such unity threatened by the autonomy of special sciences? Moreover, is the thesis of the unity of science equivalent to acceptance of reductionism?

2. FODOR ON SPECIAL SCIENCES One of the most authoritative attempts made in recent decades to foster the autonomy of special sciences has been Jerry Fodor’s essay “Special sciences (or: the disunity of science as a working hypothesis)”, published in 1974. In this essay he aims at dismantling a claim that contemporary philosophers of science have inherited from logical positivism: the equivalence of reductionism and the thesis of the unity of science. I shall now outline the main points of Fodor’s essay. The fi rst point to be mentioned is the interpretation proposed by Fodor of the classic model of reduction, inspection of whose elements may prove

How Special Are Special Sciences?

291

to be a useful starting point for our discussion. According to Fodor, reductionism is in the classic model both an empirical claim and a regulative ideal of scientific practice. Any special science, psychology for example, is accepted as a full-blown science, if and only if its theories can reduce to physics. But what is required for a special science to reduce to physics? Reductionism, says Fodor, entails three theses, two of which are only necessary conditions and one which is also sufficient. The two necessary conditions are as follows. 1. Generality of physics: “All events which fall under the laws of any science are physical events and hence fall under the laws of physics” (1974: 97). This condition also expresses the idea that physics is the basic science. 2. Token physicalism: The bridge laws connecting the predicates of the reduced science to the predicates of the reducing science should be interpreted as expressing contingent event identities (and not identities among properties as in type physicalism). “If the bridge laws express event identities, and if every event that falls under the proper laws of a special science falls under a bridge law, we get the truth of token physicalism” (1974: 99–101). We have not yet reached—in Fodor’s view—the core of reductionism, which is the conjunction of the already mentioned conditions with the third condition. 3. Reductionism: “If reductionism is true, then every natural kind is co-extensive with a physical natural kind” and every natural kind predicate of a special science is related to a natural kind predicate of physics according to the corresponding bridge laws (1974: 102). Fodor’s criticism only concerns this third condition, which in his view is what constitutes classic reductionism. As a consequence, a main point of Fodor’s essay is to show why every natural kind is unlikely to be co-extensive with a physical natural kind. The reason put forward by Fodor is similar to the famous argument fi rst presented by Putnam for functionalism (1967) and relies on the multiple realizability of the higher-level natural kinds. Fodor—who accepts the generality of physics—agrees that, for example, every mental event has a true description in the vocabulary of physics, in virtue of which it falls under the laws of physics. But a physical description which covers all mental events must be “wildly disjunctive”, and a disjunctive predicate can neither be expressive of a physical natural kind nor occur in a proper law of physics. The point of reduction—argues Fodor—is not to fi nd a co-extension of some physical natural kind predicate with each natural kind predicate of a special science but “to explicate the physical mechanisms whereby events conform to the laws of the special sciences” (1974: 107).

292 Antonella Corradini These critical remarks induce Fodor to propose a “liberalized reductionism” which takes the following form. Let Sx ⇔ P1x v P2x v . . . v Pnx be a bridge statement. Then the consequent is not a natural kind predicate, because it is disjunctive, and the bridge statement is not a bridge law for the very reason that the consequent is not a natural kind predicate. A liberalized form of reductionism yields some important advantages for special sciences that classic reductionism ruled out. First, it can account for a widespread characteristic of special science laws, that is their being nonexceptionless due to the disjunctive nature of their physical bases. Secondly, it can also explain why there are special sciences at all, whose structure and validation is by and large independent of their proving in the long run to be parts of physical sciences. To sum up, Fodor’s model shows that the unity of science does not go hand in hand with reductionism. The unity of science is still a desirable epistemological goal, but it can be reached without committing oneself to reductionism. What is required to this end is acceptance of just token physicalism. Indeed, the latter is in the enviable position of being able to reconcile the two issues that reductionism separated from each other, i.e. the unity of science and the autonomy of special sciences. As a matter of fact, token physicalism is able to guarantee both the autonomy of special sciences (because token physicalism is weaker than reductionism) and the unity of science (because this latter is anchored, as in reductionism, in acknowledgment of physics as the basic science).

3. KIM’S CRITICISM OF FODOR’S THESIS Fodor seems to represent a satisfactory equilibrium point between the claim for the autonomy of special sciences and that for the unity of science on a physicalistic basis. Nevertheless, from the outset Fodor’s proposal has been the object of many lively and still continuing debates. I shall follow one line of argument—the one put forward by J. Kim in “Multiple Realization and the Metaphysics of Reduction” (1992)—to investigate what may be wrong in Fodor’s “liberalized” or “non-reductive” physicalism. Rejection of condition 3 (i.e., of reductionism)—argues Kim—may have a somewhat different consequence from showing, as Fodor hoped, that psychology is an autonomous special science: to everybody’s dismay, it may follow from the rejection of reductionism that psychology is a science devoid of any disciplinary unity. Let us go through the different steps of Kim’s argument. The physical realization bases of mental properties, according to Fodor, are disjunctive; hence they cannot be considered as natural kinds; hence reductionism is wrong (or at least very implausible). Kim wonders why disjunctive properties cannot constitute natural kinds. His answer is that this does not depend on disjunctivity itself but on the fact that disjunctivity is in

How Special Are Special Sciences?

293

most cases a sign of heterogeneity and diversity among disjunctive properties. The lack of unity of disjunctive properties prevents us, among other things, from guaranteeing the validity of the inductive projections made on their basis. But what makes natural kinds homogeneous or heterogeneous? Kim argues that it depends on whether the respective microstructures are or are not homogeneous. The fact that the macro-level is determined by the microlevel is part of what Kim calls “the metaphysics of micro-determination” (1992: 14), that he views as a partly empirical and partly metaphysical stance shared not only by reductivists but also by functionalists à la Fodor. Kim’s argument against the multiple realization thesis is expressed in terms of “causal powers” and consists of three steps. The starting point is the “Principle of Causal Individuation of Kinds”, conceived as a principle with wide acceptance by both the reductivist and non-reductivist. This principle states that “kinds in science are individuated on the basis of causal powers; that is, objects and events fall under a kind, or share in a property, insofar as they have similar causal powers” (1992: 17). The second step is the “Causal Inheritance Principle”, according to which: “If mental property M is realized in a system at t in virtue of physical realization base P, the causal powers of this instance of M are identical with the causal powers of P” (1992: 18). This principle is, in Kim’s view, quite uncontroversial, above all because it is implied by “the Physical Realization Thesis”, which is in its turn accepted both by reductionists and most non-reductionists. The principle states that “significant properties of mental states, in particular nomic relationships amongst them, are due to, and explainable in terms of, the properties and causal-nomic connections among their physical substrates” (1992: 14). The third and conclusive step allows Kim to affi rm that if the physical realization thesis and the causal inheritance principle hold, according to the principle of causal individuation of kinds, mental kinds cannot be causally individuated independently from the contribution of physical kinds. Thus if according to Fodor physical kinds are heterogeneous, mental kinds must be equally so. Thus mental kinds are not causal kinds. But if they are not causal kinds, they cannot occur in scientific laws and therefore cannot qualify as scientific kinds. What consequences does this result have for the status of psychology as a special science? As just mentioned, psychological kinds are not scientific kinds because they are not causal. However, Kim adds that that does not imply that psychology is a pseudo-science in the same sense as astrology is, since, in contrast to the latter, it has physical realizations. But, according to Kim’s argument, such realizations must be conceived as “local reductions” which take place when each of the psychological kinds occurring in a psychological theory has a physical realization for a fi xed species or structure. Reptiles’ pains, for example, have a physical realization specific to the species of reptiles, so that the corresponding psychological theory on reptilian

294

Antonella Corradini

pain can be locally reduced to the physical theory regarding that species (in the classic sense of reduction as consisting in deducing laws from the reducing theory and making use of bridge laws). The same can be said about all sorts of species-specific pain or whatever mental state is at stake. However, this leads multiple realization into a reductio ad absurdum. First of all, science is pushed back to a form of reductionism which leaves little room for the autonomy of special sciences. Worse still, the local reductions to which multiple realization allegedly leads, deprive psychology of any disciplinary unity, because, taking pain as an example, no unified psychological theory of pain is possible, but only a conjunction of theories of pain each of which will concern different biological species. Even though Kim hastens to specify that this is not yet eliminativism, it is something near enough. The paradoxical moral of the story is well expressed in the final part of Kim’s essay, where he remarks that psychology remains scientific (because it has physical realizations) though perhaps not a science (1992: 26). Not everybody agrees with Kim’s conclusions, as the debate on multiple realization shows, still less does Fodor himself, who vehemently rejects them (1997).3 However, it is undeniable that Kim, in this as in many other essays, has stressed a specific difficulty of non-reductive physicalism. Fodor removes condition 3, that is to say reductionism, from the classic picture of science, but he does not question physicalism, merely presenting a weaker version of it. If we now look back at Kim’s argument against multiple realization, we must acknowledge that “the metaphysics of micro-reduction” shared by theorists of multiple realization is the pivot on which the whole argument revolves. This raises the strong suspicion that, in order to guarantee the autonomy of special sciences, it is not enough to remove condition 3 from Fodor’s list: we should also give up condition 2, token physicalism, along with the thesis of causal completeness of physics to which it is committed. Indeed, some scholars today maintain that in order to obtain autonomy for special sciences we should pluck up the courage to abandon both reductionism and physicalism and, perhaps, even the very idea of the unity of science. Among such scholars is John Dupré, who in his 1993 book The Disorder of Things. Metaphysical Foundations of the Disunity of Science has outlined an image of science radically alternative to both Fodor’s and Kim’s.

4. DUPRÉ ON DISUNITY OF SCIENCE Dupré’s main points of criticism regard two correlated aspects of physicalism: reductionism and causal completeness. Let us fi rst see his criticism of reductionism. Dupré understands reductionism in a sense quite different and broader than that envisaged by Fodor and Kim. He stresses that the term “reduction” refers to a continuum from derivational to replacement, or elimination,

How Special Are Special Sciences?

295

reductionism. In this perspective, eliminativism is the strongest form of reductionism. Surprisingly enough, the weakest form of reductionism in this picture is supervenience, a concept often employed by supporters of non-reductive physicalism. The reason why Dupré labels supervenience as a reductive notion is that the subvenient basis somehow determines what takes place at the supervenient level. “If a level of organization H supervenes on a lower level L, and if God knew the complete state of things at level L, then she could infer the complete description of the state of things at level H” (1993: 97). The “tyranny” exerted by the lower level is the common core of reductionism, shared even by the weakest version of it, that is to say supervenience. Against reductionism Dupré puts forward arguments from disciplines such as ecology, general and population genetics, folk psychology. His aim is to account for the insufficiency of lower-level descriptions to capture peculiar aspects of higher-level phenomena. In general, he applies to the various fields that he discusses the difference between accounting for the “how” and accounting for the “what” of a phenomenon or a behavior. The function of a reductive explanation is to give account of how things belonging to a certain domain of reality do what they do, whereas such an explanation is unable to grasp what those things will actually do. To capture the “what” of a phenomenon is the task of an explanation which makes use of terms referring to higher-level entities. To take an example from folk psychology, a generalization concerning emotionally laden beliefs (such as “if spouses believe that their partners are unfaithful, they tend to ignore evidence to the contrary”) is unlikely to be completely explained either by relying on neurobiology or by making reference to the internal states of the subjects entertaining them. Rather, the more such beliefs are specified in their contents, and the more they are conveyed by language, the more it will be apparent that their explanation requires factors belonging to the higher, social dimension. Thus a really adequate explanation of such beliefs emerges from an interplay of both the psychological and the social dimension (1993: 151 ff). Let us suppose that Dupré has succeeded in showing that reductionism is untenable. This will have wider consequences due to the fact that reductionism is implied by causal completeness. Causal completeness is the view according to which “for every event there is a complete causal story to account for its occurrence” (1993: 99). This defi nition clarifies why causal completeness implies reductionism: if a certain level of reality is causally complete, then either causal powers at the higher levels can be reduced to those of that level or the higher levels are epiphenomenal. But, if causal completeness implies reductionism and if reductionism does not hold, it follows by modus tollens that not even causal completeness will. Equally, we could cast doubt on reductionism by casting doubt on causal completeness, which is connected with a conception of causation that is questionable. In Dupré’s picture, to speak of causal completeness is tantamount

296

Antonella Corradini

to speaking of causal determinism. The negation of determinism, that is to say indeterminism, can be conceived in two different ways. According to the first one, c indeterministically causes e if c increases the propensity of e to occur. The cause c does not necessitate the effect but only makes it more likely. Causal indeterminism, however, is not a sufficient condition for the causal incompleteness of the different levels of reality. The fact that a causal nexus is indeterministic does not rule out that there is a complete causal truth about every situation. Rather, a sufficient condition for incompleteness is that the distribution of probabilities is not uniform (1993: 201 ff). That is to say, it is necessary that conditions which affect the frequency, or the magnitude, of propensities may occur. But this is plausible only if we assume that behind frequencies there are conditioned dispositions of behavior. Such dispositions are contingent insofar as they can be modified, activated, or inhibited by factors external to the level to which behavior pertains. The propensity of lynxes to eat hares, for example, is not due to an internal property of the single members of the species “lynx”, but to the environment in which the hare population evolves. It is worth noting that in this second form of indeterminism a decisive role is played by the thesis that the distribution of probabilities is not uniform. If, on the contrary, it were uniform, causal indeterminism would refer to something invariant, not subject to external influences and thus expressing a causal regularity which would be very similar to the theory of invariant conjunction.4 If causal completeness is made implausible by an ontology of causal powers, reductionism too will be made implausible by it. The alternative, nonreductionistic view stresses the causal autonomy of the different domains and levels of reality, each of which has its own peculiar and irreducible causal powers. Just as physicalism has been illustrated through the metaphor of the “tyranny” exerted by the physical level, ontological pluralism can be likened to a cooperative undertaking in which every member makes its (causal) contribution to the whole, in a thick intertwining of reciprocal interactions. Now, what bearing does ontological pluralism have on the issue of the autonomy of special sciences and on the related question of the unity of science? In a perspective like Dupré’s, the autonomy of special sciences is made possible by the causal autonomy of their domains of objects, that is to say by rejection of physicalism. However, those ready to abandon physicalism must also be ready to abandon the ontological unity that it guarantees, whose acceptance is common both to the supporters and the opponents of the multiple realization thesis. Not by chance does Dupré’s book title contain words such as “disorder” and “disunity”. Once physicalism is considered to be no longer necessary or adequate, the methodological ideal of the unity of science loses a great deal of its appeal. Not by chance, again, does Dupré declare himself to be against any form of unity of science, including those forms that are not directly tied to reductionistic assumptions. But what does it really mean to plead for scientific disunity?

How Special Are Special Sciences?

297

Will acknowledgment of the autonomy of special sciences perhaps lead to anarchism and, in demarcation issues, to the impossibility in principle of distinguishing between science and pseudo-science? Few scholars would choose this option, and Dupré himself is reluctant to thoroughly assert this thesis. Although he is against a clear-cut demarcation criterion à la Popper5 and stresses the many epistemic virtues that practices considered as nonscientific often display, he nevertheless agrees that “it is vitally important . . . that we develop critical principles for assessing the validity of . . . (putative scientific) claims”, for it would be disturbing if “nothing could be said in explanation of the epistemic superiority of the theory of evolution over the apparently competing claims of creationists” (1993: 242). Dupré’s proposal consists in emphasizing that only the conjoint work of a plurality of criteria can help clarify what it means for a theory to be scientific. To this end he appeals to Wittgenstein’s notion of family resemblance and applies it to criteria of scientificity: instead of looking for a unique and general criterion, it would be better to compile an open list of characteristics, without specifying how many and/or which of them are to be satisfied in order to attribute the title of “science” to a discipline. To sum up, Dupré’s “disorder” and “disunity” would be better called “epistemological pluralism”, a view that amid all the possible positions on this issue represents a middle course between the thesis of the unity of science and anarchism.

5. SOME CRITICAL REMARKS

5.1. The Unity of Science and Autonomy of Special Sciences Epistemological pluralism is both a fact and a methodological desideratum. As for the former aspect, it is evident that the development of science in recent decades has been marked by increasing awareness of its plurality. This plurality does not simply concern different disciplines such as biochemistry, astrophysics, and others, but it is also widespread within a single discipline. If we consider psychology, it would clearly not correspond to the facts if we were to claim that there is a unique set of psychological methods to be applied to all sub-disciplines. As a matter of fact, methods which are valid in, say, neuropsychology, can hardly substitute methods typical of clinical psychology, unless one is ready to subscribe to an old fashioned methodological monism. On the other hand, epistemological pluralism is considered to be a methodological desideratum by all those who believe in the plausibility of ontological pluralism. In clinical psychology, for example, the use of distinctive methods of inquiry is required not only by the function of therapeutic intervention—which is neither descriptive nor explanatory but transformative—but also by the specific domain of objects of the discipline, which focuses on the intentional—conscious as well as unconscious—dimension in the human being.

298

Antonella Corradini

Although opponents of epistemological pluralism are nowadays few and far between, the consequences of epistemological pluralism are not completely clear and generally agreed. As we have just seen, Dupré rejects methodological anarchism and declares himself in favor of a moderate form of pluralism. I agree with him that it is safe to avoid on the one hand the extreme of a physicalistic unity without autonomy of special sciences and on the other hand the extreme of an epistemological pluralism resulting in anarchism. The dialectical relationship between unity of science and autonomy of special sciences obtains in the space within these extremes, so that our task is to find an equilibrium point between both claims. It is accordingly worth asking, however, whether the concept of the unity of science should be completely dismissed, as Dupré and others propose. Is it for example entirely out of place, from a methodological viewpoint, to advocate fulfillment of the criterion of justifiability as the unifying characteristics of science, even in full awareness that such a criterion is to be applied differently according to the different special sciences? And does not perhaps the concept of “family resemblance” imply the existence of a common—even though not exhaustive—core of properties shared by all members of a family? The family of scientific theories is wide and heterogeneous, but it is nevertheless a family, inclusive of those third or fourth cousins we do not like that much and we are not so eager to meet. Methodological unity may prove more useful than some philosophers of science are inclined to think. Dupré maintains for example that in our best conditions we are able to distinguish between good and bad science, but not so easily between science and pseudo-science. If this is so, I would like to ask him whether he would be ready to apply this idea to his own example of the epistemic superiority of the theory of evolution over creationism. Should we concede that creationism is a scientific theory, just a bit “worse” than the theory of evolution? Or should we not prefer to say that creationism is not a science at all, since it does not meet the main criterion of justifiability of its assertions?

5.2. Anti-essentialism and Pluralism Part of the ontological and epistemological pluralism that pervades Dupré’s thought is tied to his criticism of essentialism as regards natural kinds. These should not be considered as complexes of essential properties. As a matter of fact, nothing prevents us from classifying as a natural kind a set of objects displaying homogeneous properties. However, we must be aware that there is no way to know a priori how homogeneous a natural kind is, and to sort out as essential some of the properties belonging to a kind. The answer to the question whether a property belongs or otherwise to a natural kind does not rest on essentialistic but on pluralistic criteria, which are sensitive both to the point of view or the perspective that is being taken and to the peculiar features of the objects in question (1993: 57). It is a consequence of ontological pluralism, for example, that an organism can

How Special Are Special Sciences?

299

be considered, according to the point of view, either as an individual or as a population of cells (1993: 44). The issue that I shall now briefly address is whether we really need antiessentialism in order to have pluralism in science and reality. Dupré produces two general arguments against essentialism. In the fi rst one he aims at showing that—against the traditional view—a real essence is not necessary for a natural kind. In fact, “even if a kind is determined by a real essence, the discovery of such an essence presupposes the discovery of the kind” (1993: 63). On my construal, this argument amounts to the claim that we discover the characteristics of the real essence by discovering the characteristics of the kind: that is to say that knowledge of the essence necessarily presupposes sensory knowledge. But this conclusion appears to be quite harmless. Actually, unless we already assume the validity of anti-essentialism, the argument does not imply that the only knowledge we need is the sensory one and that knowledge of the essence is dispensable. Moreover, it is a main thesis of essentialism that natural laws are a posteriori necessary truths, whose discovery requires decisive contribution from experience (Putnam, 1975). Dupré’s second argument against essentialism focuses on the alleged arbitrariness of the distinction between essential and accidental properties, which is, though, a central aspect of essentialism. Since Dupré’s preferred domain of inquiry, biology, shows how difficult clear-cut distinctions are, it should be maintained that essential properties are to be attributed not to all members of a kind but only to its typical members. But, then, laws applying to such kinds would be probabilistic laws, and it would be a mere matter of statistical frequency to decide which properties are essential and which ones are accidental (1993: 64–66). My comment on this argument is that behind statistical frequencies, which do not account for the distinction between essential and accidental properties, there still is a qualitative dimension, capable of signifying the relevance of such a distinction. Essential properties are structural properties, which shape the peculiar features shared by the members of a certain kind.6 That is to say, essence is what makes a natural kind be what it is. Thus, it is true that the distinction essential/accidental is not always precise, but this does not mean in the least that such a distinction is a purely arbitrary one. These remarks lead us to a discussion of a further thesis held by Dupré: that we should conceive of natural kinds without their essences. But what is a natural kind without its essence? Dupré is confident that “there is certainly no harm in calling a set of objects that are found to have a substantial number of shared properties a natural kind” (1993: 83). In other places in his 1993 book, however, he repeatedly suggests that less importance should be attached to the delineation of kinds and more attention should rather be paid to the identification of properties, dispositions, and forces (1993: 80). Some pages before this he even says that “there is no more to the discovery of a kind than the discovery of the correlations of properties characteristic of the members of a kind” (1993: 61). The ontological status

300

Antonella Corradini

of a non-essentialistic kind is therefore quite spurious. It should be something less than an essence and more than a set of correlated properties. My suspicion is however that without an essential form that structures it, a kind is no more than a set of properties. It is at this point that the necessity of postulating real essences becomes evident. Dupré’s maneuver of detaching natural kinds from real essences is condemned to failure, because it would amount to throwing the baby out with the bathwater. According to Dupré, essentialism attracts a good deal of criticism especially because of its apparent incompatibility with the undeniable pluralism exhibited by biological classificatory frames. However, this incompatibility is plausible only if essentialism is identified with the thesis that there is a unique correct scheme of classification. But essentialism must not coincide with this thesis. Instead, it can admit of plural descriptions of the biological reality, provided that such alternative descriptions derive from the assumption of different viewpoints and from the hypothesis that behind them there are different but not mutually exclusive essential ontologies. In general, much of Dupré’s criticism of essentialism is due to the fact that he views it as one of the facets of a mechanistic metaphysics, which lies at the basis of reductionism and determinism. But, as the more recent debate on natural kinds has shown,7 other forms of essentialism are possible. We have good reasons to be skeptical about essentialistically understood natural kinds if we conceive them as abstract and universal entities grounding exceptionless natural laws. But we do not have to conceive them this way if we shift to a causal power ontology.8 Then the necessity intrinsic to the fundamental natural kinds does not confl ict with the contingency of reality or the plurality of epistemological viewpoints. As a matter of fact, natural kinds are bearers of potentialities whose actualization depends on the existence of conditions which are contingent. Science needs both necessity and contingency. Contingency gives an account of the manifold ways in which reality presents itself, whereas necessity is the source of unity, a sort of unity, however, that peacefully coexists with plurality and heterogeneity. Thus essentialism can be seen as the ontological counterpart to an epistemological pluralism that does not on principle eschew every form of order and unity (Lowe, 2006: 169 ff).

5.3. Downward Causation There are at least two main ideas, in Dupré’s thought, that are very close to British emergentism. The fi rst is the claim that every domain of reality has its own causal powers, exercised in full autonomy from the physical domain. This idea of the causal autonomy of the different domains has a lot in common with the form of causation known as “downward causation”. The term was coined by David Campbell in 1974, but the concept dates back to British emergentism. This attributes to the more complex, emergent levels the possession of specific causal powers to be exerted upon

How Special Are Special Sciences?

301

less-complex levels. Such causal action is efficacious and thus changes the course of phenomena at the lower level, as we can infer from Lloyd Morgan’s statement: “ . . . when some new kind of relatedness is supervenient (say at the level of life), the way in which the physical events which are involved run their course is different in virtue of its presence—different from what it would have been if life had been absent. . . . I shall say that this new manner in which lower events happen . . . depends on the new kind of relatedness which is expressed in that which Mr Alexander speaks of as an emergent quality” (1923: 16). Dupré is not only very sympathetic to this conception, but, in addition, he also maintains that humans are “dense concentrations of causal power”, thus rebutting the widespread conviction that humans are at the margin of a network of causal regularities over which they cannot have any influence whatsoever (Dupré, 2001: 157). The second idea that Dupré shares with British emergentism concerns the common belief that rejection of reductionism does not amount to a rejection of empiricism. Just as British emergentists do not want to amend mechanism by introducing entities external to the natural course of things, so Dupré repeatedly asserts his fi rm commitment to empiricism and his rebuttal of mysterious and metaphysically obscure entities such as nonphysical substances (1993: 101). The aim of my fi nal remarks is to consider how compatible acceptance of downward causation is with the liberalized form of empiricism that Dupré and the British emergentists seek to defend. The intention to address this question arises from some criticisms put forward against downward causation, among others by J. Kim. I believe that the attempt to avert such criticism compels us to thoroughly examine the fundamental assumptions at the basis of downward causation. In his criticism of downward causation, Kim distinguishes fi rst between reflexive synchronic downward causation and reflexive diachronic downward causation.9 The former is in Kim’s view an entirely inconsistent concept, since it would lead to causal circularity. “ . . . how is it possible for the whole to causally affect its constituent parts on which its very existence and nature depend?” (1999: 28). By contrast, Kim allows in principle for diachronic forms of downward causation since they are not circular due to the temporal factor they involve. In fact, the whole W’s having a property M at t causes some micro-component aj to have Q at t* and having Q at t* does not belong to the conditions from which M emerges in W. A fi rst remark concerns Kim’s distinction between synchronic and diachronic variants of downward causation. Actually, it is possible to avoid circularity also in the case where downward causation is synchronic. What is needed is that the property generated by this form of causation be different from the properties of the underlying microstructure causing the emergent property. So Kim is right to say that—in order to avoid circularity—an emergent property must cause a micro-level property that is different from the property which causes the emergent property itself, but he is wrong in

302 Antonella Corradini thinking it possible only by virtue of taking into account the temporal factor which, in fact, is not relevant to the thesis of non-circularity.10 So, the real question to be discussed is whether downward causation must always be accompanied by upward determination: the thesis that the constituents of the underlying microstructure, by entertaining a certain relation with each other, determine the fact that the whole W has emergent property M (1999: 27). Indeed, if M is determined by the underlying microstructure, we can say that the underlying microstructure itself is the cause of emergent property M. In its turn, M causes the new property Q, whence it follows by transitivity that the microstructure itself causes Q. M’s causal role is therefore redundant. Kim’s criticism of downward causation poses a difficult dilemma for us:

1. We accept the “metaphysics of micro-determination”, as Kim calls it, and together with it we also accept that downward causation becomes only a way of describing things, without any ontological import (1999: 33). This option, however, is precluded for those who believe that downward causation belongs to the basic equipment of reality and that it is the source of the autonomy of special sciences. In this case we are obliged to turn to the second horn of the dilemma, and 2. We must deny that the emergent property is determined by the microstructure. This can be viewed as a co-determinant or a necessary condition of the rise of emergent properties, but never as a necessary and sufficient condition of them. The crucial point is that emergent properties must have their own ontological autonomy from physical properties. What consequences can be drawn from this fact? According to Dupré and to many others, humans are “dense concentrations of causal power”. Let us assume that such a causal power is to be understood as an emergent power in the sense just outlined. It must then be stressed that the mental presents us with a series of characteristics, such as subjectivity, consciousness, intentionality, normativity, free agency . . . which are not empirical at all. Actually, mental causation is free causation, structured not according to laws but to teleological rules. If this is the case, I believe that the pluralists’ position is neither compatible with empiricism, even if conceived in the most liberal of all its possible variants, nor with the generality of physics, that is to say condition 1 of Fodor’s list.11 Does this perhaps mean that we must be ready to open the doors of science to strange and ghostly entities? My answer is that downward causation requires the existence of emergent mental properties.12 The fact that they are co-determined by the underlying microstructure means that they are structurally embedded in the natural course of things, although they extend beyond the

How Special Are Special Sciences?

303

causal powers tied to purely empirical properties. Downward causation does not require, on the other hand, the introduction of supernatural entities, that is to say of entities fully detached from the empirical world. This is a price we do not have to pay in order to guarantee the autonomy of special sciences.

NOTES 1. Broad’s The Mind and Its Place in Nature appeared in 1925, Carnap’s Aufbau in 1928, the Wissenschaftliche Weltauffassung by Hahn, Neurath and Carnap in 1929. 2. For recent discussions about these topics see: Andler (2006), Bunge (2003), Harré (2006). 3. On multiple realization see Clapp [2001], Shapiro (2000), Sober (1999). 4. On this topic see also Cartwright (1989, 1999), Cartwright-Dupré (1988). 5. “ . . . since the demise of Popper’s falsificationism there have been no serious rivals for the role of a universal methodological criterion of scientificity” (Dupré, 1993: 233). 6. I interpret “structural” here according to its ontological meaning. 7. Representatives of current-day debate on scientific essentialism are, among others, Bird (2008), Ellis (2001, 2002), Lowe (2006, 2007, 2008), Molnar (2003), Mumford (1998), Walsch (2006). 8. As noted in Section 4, Dupré himself strongly sympathizes with this position. 9. Kim defi nes refl exive downward causation as follows: “Some activity or event involving a whole W is a cause of, or has a causal influence on, the events involving its own micro-constituents.” It is to be distinguished from non-refl exive downward causation, “in which an event involving a whole causes events involving lower-level entities that are not among its constituents” (1999: 26–27). 10. For an exhaustive exposition of this argument see O’Connor and Wong (2005). 11. I surmise that this claim is at least valid for psychology as the science of the mental. 12. It is the subject of debate whether downward causation requires only the existence of emergent properties or also of emergent substances. See on this Corradini (2008).

REFERENCES Andler, D. (2006). Federalism in science—complementarity vs perspectivism: Reply to Harré. Synthese, 151, 519–522. Bird, A. (2008). Lowe on a posteriori essentialism. Analysis, 68, 336–344. Broad, C. D. (1925). The Mind and Its Place in Nature. London: Routledge and Kegan Paul. Bunge, M. (2003). Emergence and Convergence: Qualitative Novelty and the Unity of Knowledge. Toronto: University of Toronto Press. Campbell, D. T. (1974). ‘Downward causation’ in hierarchically organized biological systems. In F. J. Ayala & T. Dobzhanski (Eds.), Studies in the Philosophy of Biology (pp. 179–186). Berkeley and Los Angeles: University of California Press. Carnap, R. (1928). Der logische Aufbau der Welt. Leipzig: Felix Meiner Verlag.

304 Antonella Corradini Cartwright, N. (1989). Nature’s Capacities and Their Measurement. Oxford: Clarendon Press. . (1999). The Dappled World. Cambridge: Cambridge University Press. Cartwright, N., & Dupré, J. (1988). Probability and causality: Why Hume and indeterminism don’t mix. Noûs, 22, 521–536. Clapp, L. (2001). Disjunctive properties: Multiple realizations. The Journal of Philosophy, 98, 111–136. Corradini, A. (2008). Emergent dualism. In A. Antonietti, A. Corradini, & J. E. Lowe (Eds.), Psycho-physical Dualism Today. An Interdisciplinary Approach (pp. 185–209). Lanham, MA: Lexington. Dupré, J. (1993). The Disorder of Things. Metaphysical Foundations of the Disunity of Science. Cambridge, MA: Harvard University Press. . (2001). Human Nature and the Limits of Science. Oxford: Clarendon Press. Ellis, B. (2001). Scientific Essentialism. Cambridge: Cambridge University Press. . (2002). The Philosophy of Nature. A Guide to the New Essentialism. Chesham: Acumen. Fodor, J. A. (1974). Special sciences (or: the disunity of science as a working hypothesis). Synthese, 28, 97–115. . (1997). Special sciences: Still autonomous after all these years. Noûs, 31. Supplement: Philosophical Perspectives, 11, Mind, Causation, and World, 149– 163. Hahn, H., Neurath, O., & Carnap, R. (1929). Wissenschaftliche Weltauffassung. Der Wiener Kreis. Wien: Artur Wolf Verlag. Harré, R. (2006). Resolving the emergence-reduction debate. Synthese, 151, 499– 509. Kim, J. (1992). Multiple realization and the metaphysics of reduction. Philosophy and Phenomenological Research, 52, 1-26. . (1999). Making sense of emergence. Philosophical Studies, 95, 3–36. . (2006). Emergence: core ideas and issues. Synthese, 151, 547–559. Lloyd Morgan, C. (1923). Emergent Evolution. London: Williams & Norgate. Lowe, E. J. (2006). The Four-Category Ontology. A Metaphysical Foundation for Natural Science. Oxford: Clarendon Press. . (2007). A problem for a posteriori essentialism concerning natural kinds. Analysis, 67, 286–292. . (2008). Reply to Bird on a posteriori essentialism. Analysis, 68, 345–347. Molnar, G. (2003). Powers. A Study of Metaphysics. Oxford: Oxford University Press. Mumford, S. (1998). Dispositions. Oxford: Oxford University Press. Neurath, O. (1931). Soziologie im Physikalismus. Erkenntnis, 2, 393–431. O’Connor, T., & Wong, H. Y. (2005). The metaphysics of emergence, Noûs, 39, 659–679. Putnam, H. (1967). Psychological predicates. In W. Capitan & D. Merrill (Eds.), Art, Mind, and Religion (pp. 37–48). Pittsburgh: University of Pittsburgh Press. Reprinted as The nature of mental states. In Putnam (1975: 429–440). . (1975). Mind, Language, and Reality. Cambridge: Cambridge University Press. Shapiro, L. A. (2000). Multiple realizations. The Journal of Philosophy, 97, 635654. Sober, E. (1999). The multiple realizability argument against reductionism. Philosophy of Science, 66, 542–564. Walsch, D. (2006). Evolutionary essentialism. British Journal of the Philosophy of Science, 57, 425–448.

Contributors

Alessandro Antonietti is Full Professor of Cognitive Psychology and Chair of the Department of Psychology at the Catholic University of the Sacred Heart of Milan (Italy). He carried out experimental studies about creativity, problem solving, and decision making and investigated the role played by new media in cognition. He is interested also in theoretical and methodological issues in psychology. Mark A. Bedau is Professor of Philosophy and Humanities at Reed College, Visiting Professor at the European School of Molecular Medicine, and Editor-in-Chief of the journal Artificial Life. He has co-edited Emergence: Contemporary Readings in Philosophy and Science (MIT Press, 2008), Protocells: Bridging Nonliving and Living Matter (MIT Press, 2009), The Ethics of Protocells: Moral and Social Implications of Creating Life in the Laboratory (MIT Press, 2009), and The Nature of Life: Classical and Contemporary Perspectives from Philosophy and Science (Cambridge University Press, 2010). His research interests include emergence, evolution, and adaptation, the nature of life and intelligence, the evolution of technology, machine learning methods and their application to designing and optimizing complex biochemical systems, and the social and ethical implications of creating life from scratch. Arturo Carsetti is Professor of Philosophy of Science at the University of Roma-Tor Vergata, Rome (Italy). His principal fields of interest are epistemology, logic, cognitive science, complexity theory, semantics, evolutionary models, symbolic dynamics, and self-organization theory. He is the author of three volumes and of about 100 articles. He has edited six volumes, among which Functional Models of Cognition (Kluwer, 2000) and Seeing, Thinking and Knowing (Kluwer, 2004). Antonella Corradini is Associate Professor of Philosophy of the Human Sciences at the Catholic University of Milan (Italy). Her main scientific concerns regard the philosophy of psychology, the philosophy of mind and action, the philosophy of biology and the human sciences, and the

306

Contributors

foundations of ethics. Her latest publications in English are Analytic Philosophy Without Naturalism (Routledge, 2006), co-edited with S. Galvan and E. J. Lowe and Psycho-physical Dualism Today: An Interdisciplinary Approach (Lexington Books, 2008), co-edited with A. Antonietti and E. J. Lowe. Mario De Caro is Associate Professor of Moral Philosophy at the University of Roma-Tre (Italy). Besides authoring three books and editing several anthologies in Italian, he has edited Interpretations and Causes (Kluwer, 1999) and, with David Macarthur, Naturalism in Question (Harvard University Press, 2004), Normativity and Naturalism (Columbia University Press, 2010), and Philosophy in an Age of Science, Hilary Putnam’s next collection of essays (Harvard University Press, 2010). Michele Di Francesco teaches Philosophy of Mind and Philosophy of Cognitive Science at the University Vita-Salute San Raffaele of Milan, where he is Dean of the Faculty of Philosophy and Director of the PhD program in philosophy and cognitive science. He is the president of the European Society for Analytical Philosophy and the author of La coscienza (Rome, 2000), Introduzione alla filosofi a della mente (Rome 2002), and guest editor with Patricia Churchland of Neurophilosophy, special issue of the journal Functional Neurology (Volume 22, No. 4, 2007). Sergio Galvan is Full Professor of Logic at the Catholic University of Milan (Italy). He works mainly on Gödel’s theorems, on philosophical logic, and on the ontology of abstract objects. Among his publications are Teoria formale dei numeri naturali (Formal Theory of Natural Numbers) (Milan 1983), A Note on the ω-Incompleteness Formalization (Studia Logica, 1994), The Principle of Deontic Refl exivity and the Kantian Axiom (Logique et Analyse, 2001), and Einführung in die Unvollständigkeitstheoreme (Introduction to the Incompleteness Theorems) (Paderborn, 2006). Carl Gillett is an Associate Professor of Philosophy at Northern Illinois University (USA). He has published extensively on topics such as the nature of realization, multiple realization, reduction, and emergence, in the metaphysics of mind and the metaphysics of science. E. Jonathan Lowe is Professor of Philosophy at Durham University (UK), specializing in metaphysics, philosophy of mind and action, philosophical logic, and early modern philosophy. Books include Kinds of Being (Blackwell, 1989), Locke on Human Understanding (Routledge, 1995), Subjects of Experience (Cambridge University Press, 1996), The Possibility of Metaphysics (Oxford University Press, 1998), A Survey of Metaphysics (Oxford University Press, 2002), The Four-Category Ontology

Contributors 307 (Oxford University Press, 2006), Personal Agency (Oxford University Press, 2008), and More Kinds of Being (Wiley, 2009). Patrick McGivern is a Lecturer in Philosophy at the University of Wollongong (New South Wales, Australia). He works in philosophy of science and philosophy of mind, and his current research focuses on physicsbased characterizations of emergence, reduction, and ‘levels of reality’. His publications include papers in Philosophy of Science and Synthese. Uwe Meixner is Professor of Philosophy at the University of Regensburg (Germany). His main areas of work are metaphysics (including the philosophy of mind), formal ontology, logic, and the history of philosophy. He has published numerous books, notably, Axiomatic Formal Ontology (Springer, 1997), The Two Sides of Being (Mentis, 2004), The Theory of Ontic Modalities (Ontos Verlag, 2006), and Modelling Metaphysics (Ontos Verlag, 2010). He is co-editor of two international philosophy journals: Logical Analysis and History of Philosophy and Metaphysica. Martine Nida-Rümelin is Professor of Philosophy at the University of Fribourg (Switzerland). Her main interest lies in the development of a non-materialist account of consciousness compatible with a scientific approach to phenomena of consciousness. She has been working in this context on phenomenal consciousness, transtemporal identity of experiencing subjects, and the phenomenology of being active. Timothy O’Connor is Professor and Chair of the Department of Philosophy at Indiana University (USA). He has published fi fty articles in metaphysics, philosophy of mind, and philosophy of religion. He is the editor of four other volumes and the author of Persons and Causes (Oxford, 2000) and Theism and Ultimate Explanation (Wiley-Blackwell, 2008). Alexander Rueger, PhD from University of Konstanz (1989), has taught in the Philosophy Department at the University of Alberta since 1993. His main interests are in the philosophy of modern physics (dynamical systems theory, theory reduction, questions of emergence) and in seventeenth- and eighteenth-century aesthetics (especially Kant). He recently published a paper, with P. McGivern, on hierarchies of levels in science (Synthese). Achim Stephan, born in 1955, is Professor of Philosophy of Cognition at the Institute of Cognitive Science of the University of Osnabrück (Germany). He had visiting positions at the VU Amsterdam and at the University of Ulm. Recently, he was a member of two research groups at the Center for Interdisciplinary Research (Bielefeld) on Emotions as Bio-Cultural

308 Contributors Processes and on Embodied Communication in Humans and Machines, respectively. He has written extensively on the topic of emergence (e.g., Emergenz. Von der Unvorhersagbarkeit zur Selbstorganisation, 2007 3rd edition). Georg Theiner is a Killam Postdoctoral Fellow at the University of Alberta (Canada). He earned his PhD in Philosophy and Cognitive Science at Indiana University (2008), where he received an Outstanding Dissertation Award for his thesis From Extended Minds to Group Minds. His main areas of research lie within the philosophy of mind and cognitive science. He is the author of several papers and a forthcoming book on distributed cognition. Hong Yu Wong is a Jacobsen Research Fellow at the Institute of Philosophy and Birkbeck College, University of London. His research interests are in philosophy of mind and psychology, action theory, metaphysics, and aesthetics. He has written on bodily awareness and agency, emergentism, and mental causation and is currently working on a series of papers on the relation between perception and action.

Index

A

B

abstract concepts 259 Gödel on 243, 248 agent 174 non-physical 173 physical 173 agent-causality see agent-causation agent-causation 170–171, 174, 178 non-physical 170, 172–175 physical 173 aggregation conditioned view of 32–33, 34, 40–42, 43 simple view of 37–38, 43 Alexander, S. 67, 76 alternative possibilities, principle of, 181 amazement 149, 150, 153, 159 proper object of 149, 150, 152, 153, 154, 158, 160 analogical thinking 274, 284 anarchism, methodological 297, 298 Anderson, P.W. 25, 32, 44, 59, 61 animal 129, 134, 135–6, 154, 158, 160, 166–168, 172, 176–178 determinism 175–176 movement 176–177 Argote, L. 79, 92 argument from composition 30–31, 36, 42 argument from composition and completeness 38–39, 42 Arnold, V. 230n4 artificial life, 53, 55 Assad, A. 54, 61 astonishment 153, 154, 158, 160 Atmanspacher, H. 229 auto-kinetic motion 273 awareness 159, 160, 278–284

Baas, N.A. 49, 61, 62 Baker, R.L. 76 Balog, K. 161 Barnier, A.J. 97, 98 Batterman, R. 217, 229 Bechtel, W. 85 Beckermann, Ansgar 60n5, 62, 180–182, 184–185, 189 Bedau, M.A. 199n1 Beginning of the existence of a subject of experience 160, 161 Berlekamp, E.R. 60n7, 61 bodies 131, 132 Bennett, M.R. 200n11, 201n19 Bishop, R. 228, 229 Block, N. 82, 84, 88, 100, 102 Boogerd, F.C. 60n4, 61 brain 132–133, 136–138, 144, 167, 172–173, 177–178 Broad, C.D. 8, 19–20, 47, 61, 289– 290 Bruggeman, F.J. 61 Burge, T. 76, 83

C Capcarrère, M.S. 62 Carnap R. 243, 289 Cartesian dualism 143, 146 Cartwright, N. 73, 76, 193 categorial intuition 260 causal completeness of physics, 9, 11, 12, 13, 22, 38, 43, 80, 109n22, 174, 178, 294, 295, 296 efficacy (of mental properties) 66, 68–71, 73–74, 76 exclusion argument 8, 11–14, 21, 22, 50

310

Index

exclusion problem 227, 228 explanatory exclusion principle 11, 21 fundamentalism 50 inheritance 65, 67, 68, 70–73 novel 7, 10, 14, 17–18, 21 pluralism, 71–73, 76 powers 14, 64, 66–72, 74, 143, 214, 222, 226, 227, 293, 295, 296, 300, 302, 303 theory of properties 22 unity (of the world) 70, 73, 74 causality 271–273 causation 171 emergent 214, 222, 227 mental 180, 185, 187–188 non-physical 174 causes, triggering/structuring 222 cellular automata (CA) 47, 53, 56, 60n7 central nervous system (CNS) 134–135, 142 Chaitin, G.J. 60n6, 61 Chalmers, D.J. 50, 61, 83, 103, 106, 107, 149 Chomsky, N. 200n7 Christensen, C. 79, 89 Churchill, J. 80, 88, 102, 103 Clark, A. 75–76, 98, 103, 106, 107 Clayton, P. 199n1 cognition 78, 81–84 animal 90 distributed 79, 89–90 group 78–81, 84–107 Colombo, M. 75–76 coming into existence 163–164 compatibilism 180, 182, 186–188 complexity 220 composition, scientific 27–29 conceivability argument 132 connectionism 26 conscious interests 175–176 consciousness 50f, 60n3, 82–84, 106–107, 149, 150, 151, 152, 167, 179 conceptualization of 152 emergentist theories of 161 functionalist account of 152 materialist theory of 153 nature of 152, 153 occurrence of 152, 154, 158 ontological status of 149–150 conservativeness argument 239ff

Conway, J.H. 61 Corradini, A. 303 cosmic hermeneutics 67, 75 Crane, T. 15–17, 22–23, 66–67, 73, 75, 76, 191, 199n1, 200n2 Crutchfield, J.P. 53, 60n6, 61

D Davies, P. 199n1 De Caro, M. 76, 200n6 decisions, 194–199 Dennett, D. 82, 196 dependence/autonomy (of mental phenomena) 64, 65, 72, 74 depth information 251, 253, 254, 260 Descartes, R. 127–128, 132, 146 determinism, physical 175 hard 180, 186–188 vs indeterminism 296 Di Francesco, M. 75–76, 201n21 divisibility argument 132 downward causation 7, 18, 50, 66, 68, 79, 215, 228, 229, 230, 300, 301, 302, 303 synchronic reflexive 7, 9–10 diachronic reflexive 8, 10–14 downward causation argument 10–11 nomological sufficiency analysis of 8, 12–13 Dretske, F. 222 dualism 64, 65, 66, 73, 76, 153, 157 property dualism 157 Dupré, J. 76–77, 190, 191, 200nn4, 8, 9, 294–304 dynamical systems theory 218

E eliminativism, 294, 295 embodiment process, 253 emergence 80, 84–107, 163–164, 177, 233, 247–248, 252, 253, 254, 261, 262, 263, 264 and authonomy 48f, 56, 57, 59 and dependence 48f degrees of 54, 60n10 diachronic 215, 216, 221, 273, 284 dynamic, 46, 51 epistemic 14, 59, 221, 225, 230n2, 256 hallmarks of 47f, 49, 50, 56 level model 74 nominal 48–51, 57, 58, 59 of consciousness 149, 152, 153, 161

Index ontological 8, 214, 221, 225, 229 patchwork mode 74 pluralism about 48–51 plurality of emergence relations 65–66, 75 radical 164 spatial 284–285 static 284 strong 31–37, 37, 48, 50f, 55, 59, 60n3 definition of 25 possibility of 33, 33–36, 36 synchronic 215, 220, 222, 223, 225, 226, 229, 273, 284 temporal 285 weak 46–63 emergentism 142–143, 146, 190–194, 199, 289, 290, 300, 301 British 300, 301 minimal 68, 69 moderate 65, 69–71, 73–74 radical 65, 69–70, 73–74, 76 scientific 25, 31–37, 39 epiphenomenalism 101–104 essence 128 essentialism 298, 299, 300, 303 experiencing subject 155, 159 explanation, generative 52, 55f, 60n10 reductive 185–188 extended mind 103–105 event 170, 173 non-physical 173 physical 171–173 event-causation 170–174 non-physical 173 physical 171–173 evolutionary argument 177

F Farmer, J.D. 61 fascination 149, 150 Feferman, S. 235, 237, 240, 248 Feferman’s lemma 235, 237 fictionalism (regarding the self) 143–144 Field, H. 248n2, 248 finitary/infinitary procedures 236, 238, 242, 244, 245, 248 Flanagan, O. 201n18 Fodor, J. 44, 59, 61, 73, 76, 77, 82, 88, 96, 100, 290–292, 293, 294, 304 freedom 192–199

311

Freeman, W. 263 free will 180–181, 184–186, 188 Frisch, U. 227, 230n6 functional, patterns 251, 252, 263, 264 properties 152 roles 128–129, 143 functionalism 291

G Galvan, S. 237, 248n1, 248, 249 game of life 53, 60n7 generality of physics, 291, 302 Gestalt psychology 267 Gilbert, M. 79, 95 Gillett, C. 29, 44, 80, 88 Gödel, K. 233, 240, 243, 244, 249, 258, 259, 260 Gödel’s theorem G1 233 Gödel’s theorem G2 235, 237 Goetz, S. 200n5 Goldenfeld, N. 231n14 Goldbach, C. 240 Goldstone, R. 87, 92 Guy, R.K. 61

H Hacker, P.M.S., 200n11, 201n19 Haidt, J. 105 Harré, R. 49, 61 Hayek, F.A. v. 86 Hempel, C. 87 Henkin, L. 259 semantics 259 Herbrand, J. 236 Hinsz, V.B. 79, 89 Hollan, J. 79, 89, 98 Holmes, M. 226, 230n6 Horgan, T. 17, 75, 77, 88, 100, 102, 162 Hornsby, J. 192 Hovda, P. 60n10, 61 Howell, R.J. 191 Hume, David 174, 197 Humphreys, P. 19, 61, 62, 191, 199n1, 214, 229, 230n2 Hutchins, E. 79, 89, 98

I idealization 221, 228 identity 133 incompressibility, of explanation, 51–54, 56, 58, 60n6 individuation 133, 138, 144

312

Index

intellectual emotion 149, 150 intelligibility, principle of 181 intuition 152, 161 ΙΣ2-theory 244–245

J Jackson, F. 51, 62, 75, 77 James, W. 179

K Kane, R. 180, 185, 201nn13, 16, 18 Kanizsa, G. 254 triangle 270–273 Kauffman, S. 25, 32, 44, 87 Kearns, M. 90, 91 Keil, G. 180–184, 186, 188 Kim, J. 8–16, 19–22, 40–41, 44, 50, 54, 55, 62, 66–69, 72, 75, 77, 88, 101, 102, 186, 199n1, 200n4, 214, 229, 292–294, 301, 302, 303, 304 Kitcher, P. 44 Köhler, W. 267

L Larson, J.R. 79, 89 latent properties 70–72 Laughlin, R. 25, 32, 36, 44, 51, 54, 59, 60n13, 62 Lavazza, A. 201n21 laws emergent 8, 9, 18–20 intra-ordinal 19 trans-ordinal 19 Leibniz, G.W. 197 levels, micro/macro 222, 223, 224, 225, 226, 227, 228 Levine, J. 149, 186 Lewin, K. 267 Lewis, D. 13, 82, 88, 93, 94, 230n10 libertarianism, 180, 183–184, 186–188 Libet, Benjamin 173, 195–6, 198 limit, singular 218, 220, 225, 229, 230n7 limit reduction 217 Loewer, B.M. 43, 102, 103 Luria, A.R. 267 Lycan, W.G. 82

M Macarthur, D. 76, 200n6 machresis 36, 43 Manzano, M. 254 Marr, D. 89

materialist(s) 149, 153, 154, 155, 157 Maund, B. 161 Mayer, B. 62 Mayr, E. 59, 62 McDowell, J. 200n6 McGinn, C. 200n7 McLaughlin, B. 19, 21, 51, 54, 62, 66, 75, 77, 102 meaning selection 261, 262 mechanism vs vitalism 290 Mele, A. 201n20 micelles 49, 60n1 micro-causal web 52–56, 58f Mill, J.S. 197 model, concept of 244 Motterlini, M. 75–76 movement, voluntary 172, 174–175 multiple realizability 57, 107n5, 108n7, 109n14, 291–294, 296, 303 mystery 150, 158, 159 mysterious 152, 154, 158, 159

N naturalism, 64, 73, 76, 190–193, 199 Nelson, E. 245, 249 neurological processes 154, 155, 157, 158 neurons 139, 141 Nickels, T. 217 Nida-Rümelin, M. 162 nihilism (regarding the self) 130 Nillson, M. 62 non-causal determination 28, 36 non-standard models 255, 257, 261 novelty 214, 215, 216, 218, 219, 220, 225

O O’Connor, T. 17, 22–23, 50, 62, 75, 77, 80, 88, 102, 103, 180, 191, 192–194, 199n1, 200, 230 O’Gorman, R. 79, 104 Olesen, M.W. 62 omega-1-incompleteness 237–239 omega-2-incompleteness 237–239 omega-3-incompleteness 235–239 omega-incompleteness 234–239 omega-rule 246–247 ontological categories 127, 129, 142, 146 ontological kind 163–165 origination, principle of, 181 overdetermination 12–14, 22

Index P Packard, N.H. 53, 54, 60n8, 61 Papineau, D. 22, 161 parts/wholes 225, 226, 228, 230n9 Peano, G. 255, 257 arithmetic, 255 persons 128, 142 Pettit, P. 51, 62, 79, 83, 100, 102 phase transitions 57, 58, 229 phi phenomenon 273 phenomenal character 158 consciousness 149 properties 154, 156, 157, 158, 161 phenomenally present 154, 155, 156, 157, 159, 161 physical basis 152, 153, 159 processes 154, 157 physicalism 22-23, 64–66, 67, 69–71, 73, 75, 80, 107, 150 non-reductive 16, 22, 65–67, 69, 73, 75, 292, 294, 295 standard 40, 43 strong emergentist, 43 token 291–292, 294 Piaget, J. 267 Pines, D. 59, 62 pluralism 298, 299, 300, 302 epistemological 290, 297, 298, 300 ontological 296, 297, 298 Popper, K. 86 Post, E. 259 PRA-arithmetic 234, 244–248 predictability 285 Prigogine, I 25, 32, 193 Primas, H. 229, 231n13 properties emergent 7–23, 185–187, 269–270, 285 essential 298, 299, 300 vs contingent 299 irreducible 186 of experiences 154, 157, 161 structural vs non-structural 17, 20, 23 Putnam, H. 59, 62, 291, 299, 304 puzzlement 149, 150, 153, 158, 159 proper object of puzzlement 152, 154

Q Qualia 50, 154, 155, 156, 157, 158, 161

313

Quine, W.V.A. 263

R Rasmussen, S. 60n1, 62 rationality, minimal 176 realization 21–23, 29, 88, 96–97, 101–104, 109nn17, 21, 226, 231n11 reducibility/non-reducibility 215, 216, 217, 220, 223, 224, 225, 231n11 reduction, causal 52, 55f context-free vs context-sensitive 46, 47, 54, 59 in principle vs in practice 47, 52, 54, 55f, 58, 59 local, 293, 294 ontological 52, 55f synchronic vs diachronic 52 reductionism 46f, 55, 58, 65–69, 289, 290, 291, 292, 294, 295, 296, 300, 301 vs anti-reductionism 73 compositional 31 hard problems of 65, 69 liberalized 292 ontological 29–31, 38–39, 40, 43 scientific 25, 29–31, 39 reflective argument 239–240 reflection principle 239–242 replacement argument 132 resultants 50, 214, 215, 222, 225 Richardson, R.C. 61, 185 Ronald, E.M.A. 60n2, 62 Roskies, A. 201n Rueger, A. 60n4, 62, 215, 217, 221, 230 nn5–6

S Sawyer, R.K. 86, 87, 98, 100 Schmitt, F. 79, 95 Searle, J. 201n Seeley, T.D. 79, 90 self-body dualism 127–148 self-organization 253, 254, 264 selves 128–129, 142–143, 146 semantic argument 240–241, 246–247 Shapiro, S. 234, 239, 241, 244, 248, 249 Shaw, R.S. 61 Shoemaker, S. 22, 44, 70–72, 75–77, 88, 102, 106, 230n10 Siewert, C. 162

314

Index

Simon, H. 52, 62 Simpson, S., 248, 249 Simulation, computer 51–56, 58f model 263 Singer, W. 180–184, 187–188 Sipper, M. 62 Skolem, T. 255 Smart, J.J.C. 55, 62 Smilansky, S. 196 Smith, A. 86 Smorynski, C. 234, 235, 248n1, 250 Sober, E. 88, 105 Soon C.S. 195–200 soul 131–132, 165–170, 174–175, 177–179 rational 163, 175, 177 special sciences 289, 290, 291, 292, 294, 296, 297, 298, 301, 302, 303 Sperry, R.W. 51, 62, 228 s-phenomenal properties 157, 158, 159, 160, 161 Spurrett, D. 22 standard view 154, 158, 161 Stanley, H.E. 57, 62 Stephan, A. 60n5, 62, 185–187, 199n1 Stevens, C. 97, 98, 99, 100 Strawson, P.F. 200n6 subject 133, 138, 144, 154, 155, 156, 157, 158, 159 of consciousness 167–168, 176–177 of experience 159, 160, 161 substance 174 supervenience 9, 10, 14–23, 67–68, 72, 73, 75, 213, 214, 221, 226, 227, 231n12, 295 of emergent properties on basal properties 14–23 mereological 15–16, 55 strong 15–16 supervenient causal powers 50 Sutton, J. 98 synergy, 54, 59 system stimulus 251 systems 134–135, 142

T Taliaferro, C. 200n5 Tarski, A. 239

Tennant, N. 234, 239, 240, 241, 242, 248, 250 Theiner, G. 98 Tollefsen, D. 79, 83, 95, 104 Thompson, E. 75, 77 thoughts-insertion 144–145 thoughts 130-131, 133, 135, 137, 144 topological (in)equivalence 218, 219, 220, 230n5 transactive memory system 92–97, 103 truth 239, 243, 244–246 truth, evidence of 246–248 Tye, M. 83

U unity argument 133, 139, 142 unity of consciousness, 145–146 unity of science 289, 290, 292, 294, 296, 297, 298 vs disunity of science 290, 294, 296, 297 upward determination 302

V vagueness 137 van Gulick, R. 16, 19 van Inwagen, P. 178, 200nn7, 10 van Vugt, M. 79, 104 Varela, F. 75, 77 Vygotskij, L.S. 267

W Wegner, D.M. 78, 92, 94, 97, 106 Weinberg, S. 25, 31, 44, 46, 57f, 58, 62 Westerhoff, H.V. 61 White, S. 201n21 Wilson, D.S. 79, 104, 105 Wilson, E.O. 79, 104 Wilson, J. 231n11 Wilson, R. 79, 80, 97, 98, 103, 106 Wimsatt, W.C. 54, 60n4, 60n11, 62, 84, 85, 96, 230 Wolfram, S. 60n6, 60n7, 62 Wong, H.Y. 75, 76, 80, 199n1, 200nn2, 3, 230 Wundt, W. 267

Y Yablo, S. 227, 230n10