Wittgenstein: Why Philosophy Is Bound to Err

Die Deutsche Bibliothek - CIP-Einheitsaufnahme Iliescu, Adrian-Paul: Wittgenstein: why philosophy is bound to err / Adrian-Paul Iliescu. - Frankfurt am Main; Berlin; Bern; Bruxelles ; New York; Oxford; Wien : Lang, 2000 ISBN 3-631-35436-3

This book is published with financial support from The Alexander von Humboldt-Stiftung, to which I am particularly grateful for understanding and friendliness.

ISBN 3-631-35436-3 US-ISBN 0-8204-4358-1

© Peter Lang GmbH Europaischer Verlag der Wissenschaften Frankfurt am Main 2000 All rights reserved. All parts of this publication are protected by copyright. Any utilisation outside the strict limits of the copyright law, without the permission of the publisher, is forbidden and liable to prosecution. This applies in particular to reproductions, translations, microfilming, and storage and processing in electronic retrieval systems. Printed in Germany 1 2

4567

Abbreviations

Throughout the book, I have used the usual abbreviations: PU

PB PO

BOM BBB

BB

BPP LSPP OG LCA BFGB Z VB

MS KGW

Philosophische Untersuchungen Philosophische Bemerkungen Philosophische Grammatik Bemerkungen tiber die Grundlagen der Mathematik The Blue and Brown Books Das Blaue Buch (The Gennan Edition of The Blue Book) Bemerkungen tiber die Philosophie der Psychologie Letzte Schriften tiber die Philosophie der Psychologie Ober Gewissheit Lectures and Conversations on Aesthetics, Psychology and Religious Belief Bemerkungen tiber Frazers The Golden Bough Zettel Vermischte Bemerkungen Unpublished manuscripts of Wittgenstein Nietzsche - Kritische Gesamtausgabe

INTRODUCTION Wittgenstein's metaphilosophy can be divided in two main parts. One of them, based upon nummerous analyses, arguments and examples, is devoted to his famous idea that philosophy is and should be a therapy against some characteristic conceptual errors and intellectual 'illnesses' (PU, § 255). The second part, which draws mainly upon some categorical statements and cryptic remarks, is devoted to the idea that philosophy cannot and should not provide explanations, hypotheses and theories, should not aim at discoveries and should not try to produce knowledge. The aim of this book is to reconstruct this second part of Wittgenstein's conception, which is the most provocative but also the most problematic component of his nletaphilosophy; the reconstruction is meant to show that, despite the apparent implausibility and eccentricity of this component, the view contained in it is coherent, founded on solid grounds and penetrating. I do not deal with the 'philosophy as therapy'-part of Wittgenstein's conception for three basic reasons: first, that this component, although not entirely transparent, is quite clearly developed and defended by Wittgenstein himself, so that its main articulations are not hard to grasp; second, that this is perhaps the most familiar part of his thinking, described and analysed in nlany books and articles; third and the most inlportant, that in my view no final and complete understanding of the 'philosophy as therapy'-view can be reached, without a deep analysis of the old German intellectual tradition of assessing and criticizing the limits of language. This tradition, which starts with Lichtenberg, th Herder and Hamann, which continues with many authors· in the 19 century, including such famous names as Fr. Schlegel, Schopenhauer and Nietzsche, and which is represented in our century by many thinkers (including sonle who, like Mauthner, have directly influenced Wittgenstein) has not been fully and properly investigated by the authors who were interested in Wittgensteinean topics. It is my deep conviction that an adequate research upon this tra~ition could produce surprisingly interesting results for the field of 'Wittgenstein Studies': Ludwig Wittgenstein might have not read at all Aristotle (as he hinlself claimed), but he has certainly been familiar not only with Lichtenberg, Schopenhauerand Mauthner (which is well-known), but also with Hamann and Nietzsche, which is in general (and especially in the Anglo-Saxon world) ignored. Obviously, it would have been impossible to deal adequately with such intricate topics in the history of ideas in only one or two chapters of the present book, so I have deliberately ignored the 'philosophy as therapy'-part of Wittgenstein's conception; I could not help, though, giving a hint about the kind

11

of analysis which can be revealing in this respect, so I have added an appendix in which I point out some striking similarities between some language-oriented arguments of Nietzsche and some parallel ones ofWittgenstein. But let me come back to the main topic of this book, which, as I have said, is not 'philosophy as therapy' but rather 'what philosophy should not be' or 'where philosophy is bound to err'. Wittgenstein's ideas on this particular topic have not had a happy destiny; not even among experts in his work, did they benefit from a warm reception. One reason for this unlucky evolution is that most of the 'Wittgensteineans' are scientifically-minded philosophers, who highly appreciate knowledge, discoveries, explanations and theoretical reconstruction (of the kind natural sciences do); it is thus easy to see why they are not happy at all with Wittgenstein's suggestions that 'there is nothing to explain', or 'to discover', and that, in philosophy, theoretical reconstructions are bound to fail. The wide-spread lack of sympathy for such a position which separates philosophy from science and knowledge, explains perhaps why no serious attempt has been made to recapture the internal logic and coherence of Wittgenstein's view of the matter. Another explanation for the fact that (as far as I know) no systematic analysis of this view has been attempted is its extreme difficulty. As already mentioned, most of what Wittgenstein has said on that topic is to be found in scattered, short, counterintuitive or even enigmatic remarks; one has to make a survey of his whole work, to collect relevant aphorisms or suggestions in very different places and to struggle to recover their conceptual (often hard to see) connections, in order to be able to aspire at a coherent reconstruction of his view. This is hard work, implying a lot of interpretation attempts the success of which is never guaranteed and in which one does engage easily. On the other hand, such an enterprise can be fascinating. You can advance step by step, by recovering the invisible connections between different statements made by the author in different manuscripts, or by discovering the 'missing links' between his ideas, and you can have the very pleasant surprise to see that things finally arrange themselves in a coherent and well-articulated picture. This is a particularly satisfactory experience, because it shows how deep a great thinker like Wittgenstein could go with his investigations, and how inadequate is the nonchalance with which certain contemporary authors dismiss his statements as eccentric, idiosynchratic, or narrow-minded; linguists like Chomsky, logicians who see logical reconstruction as the main task of philosophy or contributors who have only a superficial and short-term interest in Wittgenstein l (for experts in 1.

12

A long list of occasional contributors and articles could be given, but I shall restrict myself to only two examples: Dale Jacquette, Metaphilosophy in Wittgenstein's City, International Studies in Philosophy, XXV, 1/1993, and David Henderson, Witt-

Wittgenstein are much more cautious and understanding) feel themselves entitled to decide that 'it simply cannot be as Wittgenstein said' (because what he said is nonsensical, contradictory or at least illegitimate, inadequate, disconfirmed by facts). It is thus a great pleasure to be able to show that 'it could be as Wittgenstein said'; this is, of course, a long-tem1 debate, in which nobody has the authority of pronouncing final sentences, but if that much can be proved, that there is at least one coherent, well-justified, interpretation of Wittgenstein's provocative statements, then the efforts to reconstruct the whole picture behind these statements are not in vain.

* It seems that Wittgenstein has been somehow obsessed by the idea that philosophy (and philosophers) go farther than they should go, and say more than they should say. This is a permanent worry, illustrated not only by his famous confession that the nlain difficulty in philosophy is not to say nlore than you know (BB, p.75), but also by many suggestions he makes: that philosophers see a law in the way a word is used, whereas there is no such law (BB), that they see general, strict, rules in language, where there are none (PU), that they look for definitions (but there are no such definitions and we don't need them - BB and PU), that they see clear-cut distinctions but in most cases language is not based on such dichotomies (BB and PU), that they try to achieve greater generality, which cannot be achieved (PG, IV, § 77), that they think there must be a common elenlent in all particular things designated by the same name (PU) but in fact there isn't, etc. This permanent worry about tresspassing limits, of going farther than facts (experience) entitle us to go, might create the impression that Wittgenstein was a linguistic Humean, an empiricist preoccupied to apply Humean caution principles to the philosophical study of language. On the other hand, his apparently dogmatic statements, like nEs ist nichts zu erkUiren" (PU, § 126) or "Es darf nichts Hypothetisches in unsem Betrachtungen sein" (PU, § 125), might create the feeling that he was a dogmatic thinker, with a rigid view upon what can and what cannot be done in philosophy (a view based, perhaps, on a Kantian-type criticism applied to language - the Kantian element in Wittgenstein's work being often evoked, most notably by authors iike David Pears). I must confess that, at some moments, I was myself under the impression that Wittgenstein's metaphilosophy cannot be accounted for without postulating some genstein's Descriptivist Approach to Understanding: Is There a Place for Explanation in Interpretive Accounts?, Dialectica, vo1.42, No.2, 1988.

13

empiricist (perhaps hidden) commitments ·he had. After several years of investigations into the topic, I do not think any more that Wittgenstein's view can be understood by labelling him as an empiricist or a dogmatic. Its radicality is not rooted in doctrinaire restrictions, but rather in a coherent picture of language and philosophy, as they actually are, and not as they should be, according to common philosophical or non-philosophical prejudices. But this is not to say that Wittgenstein's conception is absolutely presuppositionless and immaculate: it is not, of course, as no other conception is. On the contrary, its internal logic and coherence is based upon some basic presuppositions, the rejection of which leads to its total failure. Thus, the general picture of philosophizing that he proposes will prove well-founded and convincing only for those who accept these premisses. As far as I am concerned, it seems to me that his presuppositions are hard to reject; but I am prepared to admit that this is not impossible, and also that their rejection demolishes the whole construction.. At different points, in the following chapters, I have singled out the relevant presuppositions at work in Wittgenstein's argumentation. But I have the feeling that a short summary of them, from the very beginning, would be very useful. Perhaps the most important supposition of his approach is that there really is a distinction between conceptual matters and empirical matters (and connected to that, the distinction between providing conceptual clarifications and providing informations). This is not a new distinction, and it does not look eccentric at all (as compared with the conclusions that Wittgenstein is going to derive from it) but after Quine's demolishing critique of the analytic-synthetic distinction, it does not appear as fimlly grounded as before. As to myself, I don't think that there are clear proofs about the collapse of the distinction between conceptual and empirical matters, but who so believes need not bother to go on reading: most of the arguments are based on this differentiation, since it is supposed that philosophy is dedicated to conceptual problems while empirical ones are only of scientific interest. A second distinction is between philosophy and science. This used to be an obvious and banal dichotomy, but in these post-modem times many people think that trying to draw such boundaries is nonsensical. This was not Wittgenstein's opinion, and it is not mine either, but who really entertains such ideas won't find anything acceptable in the following arguments. A last specification: Wittgenstein separated philosophy not only fron the natural sciences, but also from logic and mathematics. Philosophy is not a deductive science, it does not draw conclusions (PU, § 599), but rather helps one to see things clearly, to see more and to see better. A third distinction is between describing (on one hand) and explaining or theorizing (on the other). Again, such separations are not fashionable. There is a 14

tendency to argue that, since behind any explanation there is a description and behind any description there are hidden explanations, there is in fact no real boundary to be drawn here. By the way, the same objection is raised against the other two distinctions above: 'there is no empirical item without some conceptual ingredient, and no conceptual ingredient without some empirical element'; or 'there is no scientific inquiry without some philosophical components, and no philosophy without some scientific elements behind it'. The curious thing about that is that precisely Wittgenstein's work is one of the main sources of this tendency to deny the absolute separation between empirical (or perceptual) and conceptual: the authors who have dealt extensively with that - N.R. Hanson, Thomas Kuhn, even Quine - have been strongly inspired by Wittgenstein's analyses. But, of course, although he pointed out subtle connections between empirical and conceptual elements, or between science and philosophy, Wittgenstein never dreamt to claim that there is no distinction any more between them. Such differentiations may be gradual, complex, sophisticated, hard to capture, but they nevertheless exist, they are not simply fictional (this is, at least, a necessary supposition for the arguments in this book). Another presupposition made by Wittgenstein is that philosophical questions are mainly questions of the type 'What is X?': 'What is Time?', 'What is Knowledge?', 'What is Beauty?', 'What is Truth?' etc. This might seem quite restrictive, but it can be shown, I think, that no major philosophical problem can be raised without asking at least one of them. I shall not try to do that here, simply because the central role played by such questions in the Western philosophical tradition is so evident. And it is the same with another, related, supposition, absolutely essential for Wittgenstein's approach, namely that philosophical 'problems' are conceptual and not empirical. A very important presupposition for this approach is the idea that 'essence is expressed by grammar'. Obviously, this is not a metaphysical thesis, but rather a statement of fact: it is n1atter of fact that we do 'deposit' what we consider to be the 'essence' of things in grammar, i.e. in the basic conventions on which our use of words is based. 'Meaning is use', Wittgenstein's famous principle, is also acting as a basic supposition in his arguments about what philosophy should and should not do. Finally, a very important supposition made by Wittgenstein is the 'autonomy of grammar', Le. the idea that grammar (the basic linguistic rules and conventions on which language use is based) is not imposed upon us by something else (be it God, Reality, Natural Laws, or whatever), but exclusively dependent upon linguistic behavior inside a community of speakers: that is, there is nothing "hoher als eben das menschliche Sprachspiel" (PU, § 554), nothing which dictates the way our language games should be played. Rules and conventions are 15

instruments, not items governed by forces or laws which we cannot change. In a certain sense of the word (to be clarified below, in chapters 3 and 4), rules and conventions are arbitrary (PU, § 497). The artificial and conventional character of grammar also used to be a banal idea, but in the last years a certain naturalist orientation pressed for another way of seeing things.

* A very important preliminary remark is the following. Many people are upset or even revolted to see Wittgenstein expressing general ideas of the kind 'philosophy is...' or 'is not', 'philosophy should...' or 'should not'. Some are revolted simply because they believe (as many people do nowadays) that 'anything goes', that any kind of 'text' could belong to philosophy and that, consequently, nobody is entitled to make general judgements on what philosophy is or should be (such judgements having a normative character and implying that not anything can belong to philosophy). Some others, more reflective people, are upset because they think that they have detected a contradiction in Wittgenstein's conception: on one hand, he says that ,it is not necessary at all that a set of items have a common, definitory, feature, in order to be labelled in the same way (or to belong to the same category), which implies that there is not necessary that all philosophical enterprises be of the same kind; on the other hand, he keeps speaking about what philosophy is and should do, as if there actually was a characteristic feature of all philosophical enterprises and it was necessary that any piece of philosophical approach meet that standard. Is this a coherent position? I do not think that Wittgenstein was simply incoherent. In order to reply to such objections, I shall first point out his preoccupation with improper generalizations and then a very important distinction he made. In a lecture he gave quite late, Wittgenstein said (referring to William James): "»The thinking takes place before you talk«. /.../ So James says. But how did he know? It is inductive generalization? How many people had James examined? If it is an hypothesis, that isn't what James wants. He wants to say something essential to thinking'!.2

This fragnlent (like many others, of course) shows first of all that Wittgenstein was alert at the danger of overgeneralizing. Is it plausible that he has made the very mistake he was so cautious of?

2

16

Wittgenstein's Lectures on Philosophical Psychology, 1946-47, ed. by G.T.Geach, Harvester Wheatsheaf, New York, "19,88, p. 8.

I do not think so. To the objection that we hear sometimes ('Wittgenstein says that philosophy is so and so. But how does he know? Is this an inductive generalization? How many cases of philosophy did he analyze, in order to reach his conclusions?'), I think the right answer is the-following. Wittgenstein has denied, indeed, that there must always be a feature which is common to all things in the same category, and which can be found by inductive generalization. He would have probably denied that in the case of philosophical activities too. But he has never denied that things have essential or characteristic features, which make them what they are and differentiate them from other things. Anybody who denies that finds hin1self in the strange situation of not being able to say something about a category of things any n10re: if things have no essential or characteristic features, how on Earth could somebody say something specific about them? In fact, as it is obvious, essential or typical features do exist: the familyresemblances are good examples in point, and Wittgenstein never prohibits talk about that. He only rejects the supposition that such essential traits are to be described as common elements, to be found in all the relevant cases. There is, thus, no contradiction involved here: one is entitled to talk about essential or characteristic elements; what one is not entitled to do, is to speak about common, canonic, elements. The feeling that Wittgenstein is contradicting himself comes precisely from the wrong assummption that, when talking about philosophy, he talks about canonic elements, common to all kinds of philosophical enterprise. In fact, he only speaks about the characteristic or essential ones. Now, the vital questions are: if he does not engage in inductive generalizations (which would perhaps be easier to verify), how does he know what the characteristic or essential elements of philosophy are? how can he be so sure about them? The answer to the former question is that there are two ways to learn about that: either by looking at what philosophers have done (and detecting familyresemblances) or by paying attention to the grammar (to the characteristic conventions) governing the use of the word 'philosophy'. From both these sources one can learn what philosophy is, and then, of course, one is entitled to speak about that. Therefore, there is nothing wrong with making assertions about 'what philosophy is'. The answer to the latter question is that there is, of course, absolutely no guarantee that Wittgenstein's characterization of philosophy must be correct; he cannot be absolutely sure about that, and perhaps nobody can be. But this does not prove that there is something wrong with his intellectual strategy, or that he contradicts himself. In principle, his strategy is unobjectionable, but it is up to his

17

cntlcs to show that he has eventually misinterpreted or mis-characterized philosophy: the burden of proof is on them. Uttering sentences of the type 'philosophy is...' or 'should be...' is, in some sense, unavoidable, if one is to say anything at all about philosophy as a specific kind of intellectual enterprise: it would be quite bizzarre to ban such sentences, on the ground that there is no common element to all sorts of philosophical approaches. Wittgenstein might, of course, be wrong, but if he is, then his error does not consist simply in the fact that he made statenlents on what philosophy is or should be.

* One final remark is needed here. This book is not an introduction in the (meta)philosophy of Ludwig Wittgenstein. It is not nleant to be read by people who are not familiar with his ideas at all, but rather by people who are familiar with them, at least to some extent, and who are also perplexed by then1. It is not a presentation of his conception, but rather a reconstruction of it and, perhaps, an attempt to do justice to it. My feeling has been that, despite the fact that so many people speak and write about Ludwig Wittgenstein's metaphilosophy, not much attention and consideration has been given to it. Potsdam, 1998

1. 'THERE ARE NO PHILOSOPHICAL PROBLEMS'

"Wir sehen im philosophischen Denken Probleme, an Stellen wo keine sind"

(Philosophische Grammatik)

Wittgenstein's notion of a philosophical problem is placed at the very heart of his metaphilosophical conception. Not only that he sometimes goes as far as to saying that philosophy is nothing more than the whole set of philosophical problems, but almost all he says about what philosophy is or should do is based upon his (non-standard) idea of philosophical problem. Despite its importance, this idea is never elucidated systematically by Wittgenstein. As always, he keeps making different remarks on the topic, without bothering to explain plainly his view in order to prevent misunderstandings or misinterpretations. The result is (again, as always, in his later work) that his readers face considerable difficulties when trying to get a general picture of Wittgenstein's conception by putting together all his comments. The main question which is raised by his remarks on this topic is this: are philosophical problems real, genuine, problems, or not? This interrogation becomes so important because one can find in Wittgenstein's work a diversity of pronouncements on the topic, all (or almost all) of which seem to suggest that something is not in order with the so-called 'philosophical problems'. T'hus, one cannot avoid questions like 'what actually is a philosophical problem?', 'is such an item a proper problem or not?' and, in case that the answer was negative, 'why not?'

1.1 What is a philosophical problem? The variety of names that Wittgenstein keeps glvlng to what we call 'philosophical problems' is amazing. He talks about confusions (PU, § 109, 132) and Verwirrungen (PO, Teill, § 141), about misinterpretations (PU, § 111, 120) and misunderstanding (Mif3verstiindnis-PG, Teil I, § 72), about puzzlements or enigmas (BBS, pp. 27,59; I, § 56), about Beunruhigungen (PU, § 111), Unbefriedigungen (BFM, § 85) and Verbluffung (BBB, p. 76). He presents them 19

as the result of a sickness (BFM, § 4, Appendix II; Z, § 382) or of bewitchment PU, § 109; Z, § 690), as examples of mental uneasiness or mental discomfort (as Moore relates in his Wittgenstein's Lectures in 1930-1933, p. 323), or as instances of disorientation (PU, § 123), and so on and so forth. This permanent, almost obsessive, search for the right characterization of philosophical difficulties shows that Wittgenstein was convinced that they are not proper, real, problems at all. The idea that he accepted them as real problems, supported by some authors 1, is unacceptable, not only because it does not fit in at all with his metaphilosophy in general, but first of all because it is sometimes explicitly rejected in his writings: "The very word »problem«, one might say, is misapplied when used for philosophical troubles". 2 Many people who remember Wittgenstein's early interest in mathematics and logic, as well as in the technical sciences, but who forget his later critique of scientism, will probably be tempted to think that he used to identify all problems with the scientific or technical ones, restricting thus the use of the term 'problem' and consequently dismissing the idea that there could be such a thing as a philosophical problem. Since he permanently contrasts philosophy with other fields, nothing would have been easier for Wittgenstein than to state explicitly that problems are to be found only in science or in technical areas of human activity. But he does not do that; on the contrary, he keeps using the expression 'problem' for very different kinds of difficulties, and he even says explicitly that whatever one can tackle is a problem (PS, § 151). This proves that such an interpretation, according to which Wittgenstein used the term in a technical sense, is not adequate. But if he used the expression in an usual sense, and if there are so many different kinds of ('real') problems, why would not philosophical difficulties be genuine problems too? There are, I think, several answers to that question and we have to consider them separately. I shall start with the ones which are not decisive, leaving the most important one for the next chapters. In The Blue Book, Wittgenstein says that the philosopher is not a man who is out of his senses: "Ein Philosoph ist nicht jemand, der nicht bei Sinnen ist" (Das Blaue Buch, in Schriflen, Suhrkamp, 1970, B. 5, p.95). But if philosophers are not mad, then their questions are not absurd, and this creates the impression that the philosopher's problems are sound. Unfortunately, that impression is wrong.

See, for instance, William E. Barnett The Rhetoric of Grammar: Understanding Wittgenstein's Method, in Metaphilosophy, Vol. 21, Nos. 1,2, January, April, 1990,

p.57. 2

20

BBB, p.77. See also PO, Teil I, § 9.

As the context of Wittgenstein's statement clearly shows, what he has in mind is that a philosopher is not a man who suddenly forgets what everybody knows and should know, or who illegitimately ignores obvious facts. A madman is someone who fails to grasp some basic facts about life, the world, others and himself; but the 'philosophical sickness' is not caused by a failure to master some facts. This is why philosophical statements cannot be derived from sentences about facts 3, new facts or empirical discoveries do not solve philosophical enigmas and 'there is no conlmon-sense answer to philosophical questions'! But then, what is wrong with such questions? The first answer is precisely that philosophical difficulties are not rooted in facts, states of affaires, or in 'how things are'; they are not placed in the subject-matter of philosophy (PB, § 2), nor do they appear in ordinary life, in practice: the concepts ('time', 'existence', 'goodness') and the grammar rules which create philosophical perplexities are, in normal, daily, life, absolutely OK (and what is OK in daily life, is always OK, Wittgenstein specifies in Philosophische Grammatik). The phenomena we deal with in philosophy do not strike us as strange in real life (PG, § 120); they only appear as strange when we put them in a strange light, by approaching them philosophically. That is, the strangeness (of philosophical topics) which fascinates us is self-suggested; it does not belong to things, it actually springs from the way we approach these things. Our perplexities spring thus from our own misunderstandings, (PU, § 90) and misinterpretations (PU, § 111) of facts, things, states of affairs and grammar rules, and that is why they are not real problems. A question springing from nlisunderstandings and misinterpretations is not a real problem: when a savage man asks me what powerful ghost pushes forward my car, or when a child asks me why cannot the sun and the moon live in peace together, both shining continuously, these are not proper problems, because they are not rooted in reality, but in a mistaken view of it. And when we do philosophy, we are exactly like savages: "Wir sind, wenn wir philosophieren, wie Wilde, primitive Menschen, die die Ausdrucksweise zivilisierter Menschen horen, sie miBdeuten und nun die seltsamsten SchlUsse aus ihrer Deutung ziehen" (PU, § 194). We are also like children who draw some lines and then ask seriously 'what is this?' That is, in such cases we do not deal with objective difficulties, it is not the intricacy of the world which torments us; we are only misled by our own errors, entangled in our own ways of thinking (PU, § 122-125), disoriented by our own mistaken 'moves'. Exactly as in chess one has to distinguish between a real problem (springing from the complex, intricate, position on the board, the solution to which is hard 3

See LWPP, vol. II, p.45.

21

to find) and a false problem (created by a subjective failure to perceive correctly the situation, by confused reasoning and disoriented thinking of some player, when in fact the position on the board is quite clear and what should be done is obvious), one has also to differentiate between real theoretical problems (created by objective difficulties) and false (philosophical) problems created by personal, subjective, errors in interpreting things. It is precisely because philosophical puzzlements are such false problems, that we have to get rid of them and stop doing philosophy (PU, § 133), to be cured and thus to forget about them as we forget about an old illness (PU, § 133, 255); the job we have to do is one upon ourselves (as Wittgenstein says in MS 213), like in any other case of therapy. Our intellectualist inclinations make us expect that everything should be explained: factual aspects of the world by science, and more general, 'philosophical', aspects by philosophy. But, in fact, we never explain everything, exactly as we never prove everything, because, as Wittgenstein reminds in Uher Gewissheit, not everything is a topic for doubt. If there were strange actual phenomena which required a philosophical explanation, then those phenomena would have represented real philosophical problems. But there are no such phenomena, there are only those 'normal' phenomena which appear strange only because we put them in a strange light by philosophizing; consequently, there are no rea] philosophical problems. Now, at this point of the discussion, we have to consider a possible objection. The above argument of Wittgenstein is based on the premiss that philosophical 'enigmas' do not appear as problems in ordinary life, and that only what would appear as problematic (or strange) in ordinary or 'normal' life would also be a real, genuine, problem for philosophy. But to the idea that philosophical difficulties do not appear in ordinary life, one can be tempted to reply 'So what?', while at the supposition that only things which seem strange in practice are to be accepted as genuinely strange (and thus as really problematic), one could reply by asking: could not philosophy constitute a specific form of life, an autonomous form of life with its own problems, so that their genuine character would not depend upon their presence in ordinary life? In fact, there are some authors who believe that philosophical activity does constitute a genuine form of life,4 which, of course, implies that also its problems are genuine. 4

See, for instance, Dale Jacquette, Metaphilosophy in Wittgenstein's City, in International Studies in Philosophy, XXV, 1, 1993, p. 30: "Another uninteresting answer is that philosophy occupies no territory in Wittgenstein's city because it lacks a grammar, and. does not represent a form of life. I.. j Ostensibly, the claim is simply false, since there are communities of philosophers with their own stated and unstated rules about the correct conduct of philosophical inquiry, standards of professionalism, criteria of what is to count as a legitimate or interesting philosophical pro-

Wittgenstein's answer to that objection is based upon an analysis of the grammar of the word 'problem'. What does the grammar of this word imply? The way in which we use this expression shows that there is an 'internal' connection between 'problem' and 'solution': a problem is something which, in principle, seems to us susceptible of having a solution; where no solution is possible, and we know that, there can be no problem. As Wittgenstein puts it, "Nur wo eine Methode cler Losung ist, ist ein Problem" (PB, § 149). This statement, like so many others made by him, can appear as unacceptably restrictive and dogmatic: after all, why could not problems without solutions exist? A short analysis will show that what made him say that was not dogmatism, but grammatical accuracy. Thus, it must be noticed that we nornlally speak about 'having a problem' only when the following two conditions are fulfilled: (i) we managed to 'grasp' the difficulty we face, i.e. to identify it, to isolate it and to understand it (at least partially); as Wittgenstein suggests, a problem is something sufficiently well-determined as to be 'grasped', in order to 'be handled' - "Was man anfassen kann, ist ein Problenl" (PG, p. 379). Thus, not every difficulty, unfulfilled need or aspiration, does represent a problem; not everything we are after creates a problem: we seek happiness, because we have a vague feeling we could be happy, but this is not, properly speaking, a problenl - it is rather an aspiration. We seek truth, but finding the truth is not always and necessarily a problem, because someone can reach truth through revelation, and having a revelation is not the same thing with solving a problem or with answering a question ("Einer Offenbarung entspricht keine Frage" - PB, § 149). A problenl is, so to say, a 'technical', local and quite well-crystallized difficulty which can be solved in a 'technical' and well-determined manner. (ii) the problem is susceptible of having a solution, appears as soluble. What is in principle insoluble cannot be a problem, because we do not use the expression 'problem' for hopeless, eternal, difficulties. Thus, imnlortality is not, properly speaking, a problem: of course, some people wanted to become immortal, but that was a dream, an aim, or an aspiration, not a problem. We only speak of 'having a problem' when we think that a solution is possible and we are looking for it. That is the 'internal connection' between 'problem' and 'solution'. Of course, there can be cases in which we think a cerblem... ". Obviously, Jacquette makes a confusion between forms of life and professional communities; moreover, he neglects the huge differences between a working professional community like that of mathematicians (which satisfies his description above) and the many (small or large) groups of philosophers, who, contrary to the same description, have no common standards and problems, and who keep dismissing each other's problems, judgements and solutions as meaningless.

21

tain difficulty is a problem and does have a solution, only to discover that there is no possible solution to it. But in such cases we do not say 'True, we do have a problem here, but it is insoluble', but rather 'There is no problem here; there is nothing to be solved, because what we want is simply impossible'. One normally does not try to solve what one knows it cannot be solved, and does not take as a problem a difficulty which one knows it is absolutely hopeless; a problem is thus something one could try to solve: what nobody cares trying to solve (because one knows that there can be no solution) cannot be a problem. To take a step farther, the internal connection between 'problem' and 'solution' leads us to the distinction between 'right' and 'wrong', because a solution is something which can be right (the right solution to the problem) or wrong (the wrong, or the pseudo-solution). A problem is thus something susceptible of having a right solution or a wrong solution. But the distinction between right and wrong leads us to a community in which this distinction is used. As Wittgenstein insists, terms like 'mistake', 'error', 'wrong', 'correct' imply a reference to a community (00, § 156): such terms are not used according to personal feelings on the matter; they are rather used according some inter-subjective standards. But this is precisely what philosophy lacks: in philosophy, there is no substantive . inter-subjective agreement, therefore there is no effective distinction between right and wrong solutions, and therefore there can be no proper problems, since problems imply solutions, solutions imply an effective distinction between correct and incorrect, and such a distinction implies inter-subjective agreement. Deductions like the above one might seem wild, but in fact the argument we have here is not unusual at all; it is even banal. There is a very old complaint about the destructive results which lack of consensus has in philosophy, and a quite familiar contrast between science and philosophy in this respect (let us remember Kant's remarks on the topic, to take just one example); what Wittgenstein says is, after all, that scientific consensus makes problems-solving activity possible (generating agreement on what a good solution to a problenl would be), while philosophical lack of consensus creates a situation in which there can be no proper problems, sinlply because there is no agreement on what is the problem and what a good solution to it would look like. Developing a suggestion once made by Rudolf Haller, in his Questions on Wittgenstein, one could say that philosophers are like a community of people in which no consensus about colours existed; in such a community, there could ~e no problems (of the kind 'finding the colour of X', for instance), because having a problem means looking for a solution, and looking for a solution means knowing what 'a right solution' means, which is not the case, as long as the agreement upon colours was lacking. Philosophers constitute a community of this kind; or,

24

perhaps, they do not constitute a true (a proper) community, because a true community is based on agreement (upon language, definitions and even judgements, as Wittgenstein suggests).

* The above arguments are not unimportant, but the main arguments developed by Wittgenstein against the idea of proper philosophical problems are still to come. As they need developed conceptual support, distinct parts of this chapter will be devoted to them.

1.2 Are philosophical questions meaningful? ~fittgenstein's insistence that philosophical problems are confusions and misunderstandings, rather than genuine questions, is disturbing. One of the main reasons for our uneasiness about this is the thought that philosophical questions must be in order as long as they 'tell' us something and we can grasp reasonably well their 'message'. If we generally manage to understand (at least in the end) philosophical questions, and to discuss them rationally with other people, how on earth could they fail to have any meaning? Was Wittgenstein serious about that, or, perhaps, his claim was only the expression of an allergy to classical philosophy? By asking this, one reaches (I think) an important, but usually neglected, point of Wittgenstein's thinking, and it is at this point that a vital connection between his nletaphilosophy and his theory of meaning becomes apparent. I shall try to show that there are strong arguments for the (apparently wild) claim that all philosophical questions are meaningless, which is much more than a piece of eccentricity coming from a nlan of genius. If we take seriously Wittgenstein's idea that meaning is use, then we have to consider the implication that a sentence is meaningful not simply when it 'says' something which we can (seem to) understand, but rather when it has a definite use on which we agree and the correctness of which can be checked. Consequently, (I claim), when denying the meaningfulness of philosophical questions, Wittgenstein denies neither that they 'tell' us something, nor that some of us can be convinced that they understand them very well; his thesis does not imply that there can be no philosophical discussion, since philosophers simply don't know what they are talking about. Philosophical debate is obviously possible, even if agreement among participants is rare (and they often disagree not only on matters

25

of fact, but also on the meaning which should be attributed to their words and sentences!). What (I think) Wittgenstein actually says, is that the sentences meant to express fundamental philosophical questions do not have a definite, susceptible of correctness-checking, way of using, and consequently, although they 'say' something, what they say is confused and, strictly speaking, meaningless. This idea does not seem paradoxical any more, if we pay attention to the distinction between 'saying something' and 'being strictly speaking meaningful'. (AJ

To begin with banalities, it is obvious that anybody can accidentally hear a sentence which 'says' something, but (being completely unfamiliar with the relevant context) fail to grasp its proper meaning. (Many such examples have been evoked in recent years, so that it is not necessary to add new ones to them). It is even nl0re evident that a question has always an appropriate context of use, and it cannot be sensibly asked outside it (it is only about adults that we ask whether they are married or single, and it does not make sense to ask the same question about five years-old children, and so on and so forth). Now, what happens quite often in philosophy is that some question is asked, although an appropriate context for it is lacking. A case in point would be what Carnap calls 'external questions'. But let us think about the question 'What is the meaning of life?' Normally, the meaning something has for us depends upon the position it occupies in our life, upon its connections with other things that playa major role for us etc. (The meaning of marriage in one's life could be a good example in this respect). But, if meaning is determined by such a position and such connections inside life, then it does not make any sense to ask this question about life itself. 5 The interrogation would only be meaningful when one accepts a larger than life meaning-bestowing order (e.g., a natural or a divine Universal Harmony). This is not always the case, (it is perhaps increasingly seldom the case) but even when this necessary condition is not fulfilled, people are tempted to ask the same question because it seems to 'say' something important to them, and even to touch a fascinating point. The expression 'meaning of life' certainly 'tells' us something, because its component words are meaningful, it is grammatically

5

26

As Hayek says, "if»to have meaning« is to have a place in a order which we share with other people, this order itself cannot have meaning because it cannot have a place in itself" . Rules, Perception and Intelligibility, in F.A. Hayek, Studies in Philosophy, Politics and Economics, The University of Chicago Press, 1967, p. 61.

correct and, moreover, it is analogous to other expressions which undoubtedly make sense. But is this enough? Wittgenstein's conviction seems to have been that 'saying something' is not a sufficient condition (although it may be a necessary one) for meaningfulness. In a less well-known contribution, he discusses the idea of a medicine which prolongs life, so that anyone who takes it has one's life extended with two months. The problem is that this beneficial effect cannot be verified (since we never know how long would one have lived without the help of that medicine), and Wittgenstein suggests that this kind of unverifiability makes the expression 'a medicine which prolongs everyone's life with two months' practically meaningless. 6 Should we conclude that Wittgenstein was committed to some kind of verificationist criterion of nleaning? We certainly don't have to. Another, nlore generous, interpretation would do as well. Let us remember that meaning is use. Consequently, to have a certain meaning is to have a detennined standard use. But the existence of a standard use presupposes the possibility of checking whether an expression has been applied correctly or not. In the example we are talking about, such a possibility does not exist: we simply cannot determine in which cases the expression has been correctly, and in which it has been wrongly applied, because we never really know whether a particular medicine actually prolongs life in the way described. Thus, the difficulty we face is not an epistemological one, created by the limits of verification, but a semantic one, due to our incapacity to establish a determined standard use for this expression, or a 'technique' of applying it. The expression 'says' something, looks like a normal, meaningful, construction, but it fails, in fact, to make sense; and Wittgenstein often insists "daB etwas aussieht, wie ein Satz, den wir verstehen, was doch keinen Sinn ergibt" (PU, §513). In order to see how is this failure possible, we have once again to remember that 'hearing what an expression says' is independent from 'grasping its meaning' or properly understanding it. Wittgenstein stresses this point by insisting that, even in cases in which we understand perfectly well the meaning of the component words of an expression like "I have n friends and n 2 +2n+2=O", and consequently what this sentence 'says', we still don't see immediately if the sentence nlakes sense (PU, §513). Seeing what a sentence 'says' can be an instantaneous and unproblematic act (when the sentence is granlmatical and we know the language); but seeing if the sentence makes really sense is much more difficult. We see what an expression (e.g., the construction of a heptagon) 'says' and we are inclined to think that we can imagine something which it should be applied to and which thus makes it undeniably meaningful, whereas in fact - as many mathema6

Cause and Effect: Intuitive Awareness, in Philosophia, vo1.6, nos. 3-4, 1976, p.423.

27

tical proofs show - we cannot do that (PU, §517). Such cases, Wittgenstein suggests, should make us consider the possibility that sometimes we are simply mistaken in thinking that we understand a question (Ibidem). This happens, for instance, when a philosopher uses (in philosophical purposes) sentences like 'I am here'; he thinks something by it, the sentence 'says' something to him, but he completely overlooks the question of "how, on what occasions" ("wie, bei welcher Gelegenheit") should this sentence be used (PU, §514). Such a philosopher identifies the meaning of the sentence with a thought, with 'what it is said', ignoring that meaning is in fact use, and that meaningfulness presupposes the existence of a well detennined way of using. A similar mistake is being done by someone who understands the sentence 'it's five o'clock here', and is therefore inclined to think that he also understands the sentence 'it's five o'clock on the sun' (PU, §350). This mistake consists in a reduction of meaning to 'what is being said', where 'what is being said' is deduced from the meanings of the component words. Obviously, this reduction is wrong, because understanding what time is on the sun presupposes not only understanding some words, but also a technique, Le. the way in which the time-measuring expressions are applied on the sun (their use there). The distinction between 'saying something' and 'being properly meaningful' (that I insist upon here because, I claim, it is implicit in many arguments, not only in the Philosophische Untersuchungen, but also in other texts) is wellsupported by Wittgenstein's dictum that "Einen Satz verstehen, heiBt, eine Sprache verstehen. Eine Sprache verstehen, heil3t eine Technik verstehen" (PU, §199). This well-known thesis has often been quoted as a typically Wittgensteinean slogan, but - as far as I know - has seldom been used adequately in order to clarify his views on meaning. It will thus be useful, in this respect, to have a look at another fragment in which Wittgenstein speaks about understanding questions. He remarks that we really understand a question like "Kommt die Ziffemfolge 7777 in der Entwicklung von p vor?", not only because it is a (correct) German sentence, but also because we know what it means for a certain number ("415", for instance) to occur in the development of p (PU, §516). And he specifies that "soweit solche ErkHirungen reichen, soweit, kann man sagen, versteht manjene Frage" (Ibidem). This remark about understanding and meaning clearly points out the vital importance of a technique. Any linguistically-competent speaker can see what the sentence mentioned 'says'; but in order to understand its meaning properly, one has to know 'what it means' for a number to occur in a certain sequence and therefore to be familiar with a certain technique. If we want to explain the meaning of this sentence to somebody, we have to explain the way in which we use expressionslike 'sequence occuring in the development of...'; and in order to explain the 28

technique ofusing this expression (when and how it should be applied), we have to explain a certain mathematical technique. Consequently, the understanding one has of this question reaches just so far as these explanations of a technique reach. Thus, grasping the meaning implies grasping the use, and grasping the use implies mastering a technique; seeing what a sentence 'says' is not enough - if the technique of its use is not mastered or is not available, we are not entitled to say that we grasp its meaning. Now, at this particular point of the discussion an important objection might be raised. The very fact (somebody might argue) that we have been able to see what a sentence 'says' is a sufficient proof that we have grasped the technique of use involved and, consequently, its meaning: the danger of meaninglessness is automatically excluded, once we have understood what the sentence 'says', for, were it simply meaningless, how on earth could we understand it? Isn't this whole discussion the unfortunate result of a false alarm? The answer to this last question is No, and the justification of this answer is to be found in the fact that Wittgenstein does not speak about a general meaninglessness (of the kind we face when hearing inarticulate sounds or accidental noises); he rather refers to the kind of meaninglessness which appears in certain language.-games, despite the fact that the expression which fails to make sense in those games can be 'in general' (Le., in other, standard, language-games) meaningful. The kind of meaninglessness which is of interest for him is the meaninglessness of an expression which, although generally makes sense (as it is used in ordinary speech and contexts of use), is nevertheless without meaning when used in other, special (for instance, philosophical) or 'abnormal' contexts. This kind of meaninglessness is possible because every particular instance of use raises new problems and has 'its own laws'; the fact that a sentence is in one particular case meaningful, does not solve the problem of meaningfulness for other cases. As Wittgenstein says, "jede dieser Verwendungen muB eigens beschrieben werden und hat ihre eigene Gesetze. Es niltzt uns nichts, daB wir eine Redeform fertig in unserer gewohnlichen Sprache vorfinden, weil diese Sprache jedes ihrer Worter in den verschiedensten Bedeutungen gebraucht, und, daB wir den Gebrauch des Wortes in einem Fall verstehen, erspart uns nicht die Untersuchung seiner Grammatik in einem andem" (PB, p.307). That Wittgenstein talks only about this particular kind of meaninglessness, it is proven by his remark that "where we say »This makes no sense«, we always

29

mean »This makes no sense in this particular language-game«;"7. And, of course, there is no surprise in this: being mainly interested in the meaninglessness of philosophical utterances, he referred to the special linguistic situations in which philosophers take an ordinary expression from its 'original' language game, and use it in a new (purely philosophical) one. It is precisely in such situations that, although "any expression may make sense", as he said, the danger of meaninglessness appears: simply because "you may think you are using it with sense, when in fact you are not" (op.cit., p.265). In such cases, as philosophers, "we are apt to think that we are using a new system of projection which would give sense to our words, when in fact we are not using a new system at all" (Ibidem); and our sentences "sound as if they »ought to have sense«, when in fact they have none" (op.cit., p.319). For instance: "since we talk of Time 'flowing' as well as of a river 'flowing', we are tempted to think that Time 'flows' in a certain 'direction', as a river does, and that therefore it has sense to suppose that Time might flow in the opposite direction, just as it certainly has sense to suppose that a river might" (Ibidem). Now, this particular kind of meaninglessness can certainly co-exist with an understanding of what the sentence 'says'; we understand what 'Time flows' says, although the sentence is in fact meaningless. The above objection is thus rejected, because it appears that we really can understand what a meaningless sentence 'says'. What happens with philosophical problems seems to be this: some questions are raised which clearly 'say' something (at least to philosophers and philosophically-minded people); we are thus convinced that they are meaningful, but in fact, since there is no well-determined technique of use for them, they fail to make sense. What they 'say' is confused; they are mere Verwirrungen. Wittgenstein describes again and again particular cases in which such philosophical confusions occur and sometimes (PU, §90) presents philosophical work as consisting, in general, in the attempt of clearing them away. The general idea (which is also mentioned in PU, §90) is that philosophers engage themselves in wrong analogies which lead them to misinterpretations of our forms of expression. 8 He gives a lot of examples of misleading analogies: the analogy between mental phenomena and physical processes (PU, §308), between measuring time and measuring spatial distances (BBB, p.26), between mind and a medium (BBB, p.6), between thoughts and sentences (BBB, p. 6-7) or between thinking and a 7 8

30

Wittgenstein's Lectures in 1930-33, in G.E. Moore, Philosophical Papers, George Allen & Unwin, London, Humanities Press, NY, 1959, p.273. In other places Wittgenstein suggests that it is the other way round: language leads us to adopt the wrong analogy. See The Blue Book, BBB, p.48: "here our language is tempting us to draw some misleading analogytt.

process or an activity (PU, §329,330,339; BBB, p.6) etc. His discussion of the case of willing illustrates well the general idea in PU, §90. The mistake we make at the linguistic level is due to a wrong analogy: "Und mein falscher Ausdruck kam daher, daB man sich das Wollen als ein unmittelbares, nichtkausales, Herbeifiihren denken will. Dieser Idee aber liegt eine irreruhrende Analogie zu Grunde" (PU, § 613). This description of the way in which confusions and misunderstandings appear is metaphilosophical1y illuminating, because it shows how a false analogy leads us to questions which 'say' something, but fail in fact to make sense: seeing thinking as an invisible process, we ask where does it take place and what relationship does it have with its companion (speech); seeing time as an invisible process, we ask who created it etc. Such sentences seem, prima facie, to be OK, because their components are meaningful and what they themselves 'say' seems quite similar to what other, meaningful, sentences say. But since in fact what is implied here are false analogies, the questions are actually meaningless, i.e. they have no proper application in the relevant context. Wittgenstein is very (and unusually) explicit upon this point: "Ich mochte antworten: »Es ist ein deutscher Satz; scheinbar ganz in Ordnung, -ehe man namlich mit ibm arbeiten will; er steht mit andem Satzen in einem Zusammenhang, der er uns schwer macht, zu sagen, man wisse eigentlich nicht, was er uns mitteilt (... )«" (PU, §348). That is, since the sentence 'says' something that we (think we) understand, we cannot simply dismiss it as a mere series of noises; but, "Jeder, der nicht durch Philosophieren empfindungslos geworden ist, merkt, daB hier etwas nicht stimmt" (Ibidem). What is wrong here becomes evident as soon as our next objection is answered. We are usually tempted to reply: but the idea that thinking is a hidden process, or that (as in the example discussed in the fragments I am now quoting) the deaf-mutes talk to themselves inwardly in a vocal language, is not at all meaningless; it is only a hypothesis, maybe true, maybe wrong, but which in any case makes sense. Wittgenstein's answer is here the following: a sentence is meaningful only if there are cases in which it can (in principle) be correctly applied; but if a sentence, which in normal circumstances can be applied (being thus meaningful), is used in such a way that (as it happens with philosophical uses) there is no possible (correct) application of it, then we have to agree that it is not a meaningful sentence, but only a series of 'naked words': "»Aber diese Annahme hat doch gewiB einen guten Sinn!« - Ja; diese Worte und dies Bild haben unter gewohnIichen UmsUinden eine uns gelaufige Anwendung. - Nehmen wir aber einen Fall an, in welchem diese Anwendung

31

wegHillt, so werden wir uns nun gleichsam zum ersten Mal der Nacktheit der Worte und des Bildes bewuBt" (PU, §349). Once again, the decisive element is not constituted by the words themselves, but by the kind of use which is being given to them. And if philosophers utter sentences for which no (correct) application is available, i.e. sentences for which an adequate technique of use is lacking, then, since mastering a technique is out of the question, proper use is out of the question too: therefore, there is no proper understanding of such sentences and strictly speaking they have no meaning (although we understand or think we understand what they 'say', in the sense that we understand their C0l11pOnent words and think that we understand them too because they have an analogy with other, familiar, meaningful sentences). It is not clear what are we to say, according to Wittgenstein, in such cases. His question in PU, §517 (quoted above) suggests that, perhaps, we should say that one can be mistaken in thinking that one understands certain questions: one seems, maybe, to understand, although one doesn't. But Wittgenstein is not categorical on this point. And I do not believe that this creates a serious problem for his interpretation, because there always exists the possibility of saying· that in such cases we only have a prima facie understanding (similar to that of one who happens to overhear a sentence and to give some sense to it, although, ignoring the context, entirely misses the point), but not proper understanding (i.e we do not grasp a meaning enabling us to use the sentence correctly ourselves). Exactly like someone who simply overhears some meaningful words, we feel that we understand (in a way) what is said, but we actually fail to, in the sense that we fail to grasp and master an adequate technique of using what has been said. (B) One unpleasant consequence of the interpretation I am proposing here is that it seems to commit us to a distinction between two kinds of understanding (prima facie or proper understanding), and (accordingly) two kinds of nleaning, which, of course, goes against our Occamean preference for simplicity. But there are, I think, also other reasons for our reluctance to accept this way of seeing. Some of them, I would like to argue, are probably wrong. OUT main problem is created by the tendency we have to consider meaningfulness as something internal or, better, intrinsic to expressions, and to neglect the role played by use. In the case of sentences, this tendency leads us to interpret things as if the meaning of the whole sentence were determined entirely and exclusively by the meaning of its component words. We consider these words as intrinsically meaningful, and we make the tacit assumption that, since they are so, the resulting sentence must also be meaningful (which, of course, is to be expec-

32

ted if the game of 'making sense' is played exclusively with 'intrinsicalities', i.e. with what expressions have 'in themselves', independently of the way they are used). In a fashionable logical jargon, this assumption takes a very respectable form, like 'the meaning of a linguistic compound derives from the meanings of its components', which of course implies that the meaning of a whole sentence should be deducible from the meanings of its component words. Then, of course, if 'life' and 'meaning' are meaningful, then so nllist also be 'meaning of life'; if 'created' and 'time' are meaningful, so must be the question 'has time been created?', and so on. In such cases, we completely ignore the importance of use, and this is the source of our philosophical confusions, as Wittgenstein convincingly (I think) shows: for instance, we neglect the fact that we actually use the word 'time' in order to refer to such relations like before, after etc., the consequence being that "people ask »Has Time been created?« although the question »Has before been created?« has absolutely no meaning"9. But this question seems perfectly meaningful to us, since it is analogous to meaningful sentences like 'Has this mountain been created?'. Once again, it is a false analogy which leads to the philosophical confusion: as Wittgenstein put it, "nonsense always arises from fonning symbols analogous to certain uses, where they have no use"lO. Now, a legitinlate question could be the following: how does it happen that a philosophical sentence, although composed of meaningful words and often quite similar to another (meaningful) sentence, fails to have an adequate use? Part of Wittgenstein's answer is that the former lacks the normal context in which the latter is uttered or the appropriate situation which 'determines' the latter; and this is a good opportunity to remind one that the meaning of a sentence does not automatically derive from meaningful components, but must also be somehow 'determined' by a situation: "So wie die Worte »Ich bin hier« nur in gewissen Zusammenhangen Sinn haben, nlcht aber, wenn ich sie Einem sage, der mir gegenUber sitzt und mich klar sieht, - und zwar nicht darum, wei! sie dann ilberflilssig sind, sondem, weil ihr Sinn durch die Situation nicht bestimmt ist, aber so eine Bestimmung braucht"11. Thus, although the construction a of philosophical sentence is often very similar to the construction of an usual one, we use the former in a very different context, where it lacks the kind of use (for a certain situation) which gives meaning to the latter. And the proof that this is the mistake we make is that as soon as 9 Wittgenstein's Lectures in 1930-1933, in G.E. Moore, Philosophical Papers, p.319. lOIn G.E. Moore, op.cit., p.274; see also pp. 304-305. 11 Ludwig Wittgenstein, Ober Gewissheit, (hereafter Va), § 348.

33

we go back to the 'normal' use (and to the customary situation), the opacity of meaning associated with philosophical sentences simply disappears: "Sowie ich aus der philosophischen an eine allUigliche Anwendung des Satzes denke, wird sein Sinn klar und gewohnlich u12 . Thus, Wittgenstein's position appears to be coherent and (as I hope I have managed to show) quite convincing. But that does not mean that no other objections could be raised against it. An important objection, which is sometimes raised also against his views on explaining in philosophy, is that he simply uses some essential terms (meaning, explanation etc.) in a technical way, placing thus too strong, unacceptable, restrictions upon their philosophical use. For instance, many people would probably be ready to agree that wrong analogies could lead to nonsense, but they would still be reluctant to admit that Wittgenstein is right when claiming that a sentence can be meaningless even if it 'tells' us something and we understand it, at least in a way. It might be objected that there is no such thing as 'understanding a sentence in a way, or superficially, but not in another, proper way' - that this distinction is artificially created by imposing some special restrictions on the usual word 'to understand'. The claim could be made that understanding is nothing else than grasping the meaning, so that any understanding of a sentence must be a proof that the sentence does have a meaning; and there are no cases in which (as suggested above) one understands the sentence in a way, but completely misses the point - there are only cases in which we understand it correctly and cases in which we understand it in a wrong way (i.e., we misinterpret it). If so, then, Wittgenstein simply multiplies unnecessarily the entities (in the ontology of mind), by inventing a special kind of understanding, the existence of which is conditioned by the fulfilment of some too strong conditions, and which is declared to be 'proper understanding'. Let me try to answer this objection. It seems that, Occamean idiosyncrasies apart, the objection is based either on seeing language as a calculus, or on seeing it as governed by a tertium non datur-principle in matters of understanding and meaningfulness. Both alternatives can be, I think, rejected and I shall try to show how. The first one is already familiar, since it consists in claiming that the meaning of a sentence is determined (entirely and exclusively) by the meaning of its component words, so that anyone who understands the components can sonlehow 'calculate' the meaning of the conlpound. Thus, Michael Dummett writes: "we de-

12

34

OG, § 347.

rive the meaning of each particular sentence from the meanings of the words that compose it". 13 If we accept this principle - that sentence-meaning can be deduced from the word-meanings - then, of course, Wittgenstein's suggestions seem suspect: any sentence which is (correctly) composed of meaningful words is bound to be meaningful itself, and thus it is hard to see how philosophical sentences could fail to make sense. But, of course, Wittgenstein would reject the principle evoked by Dummett, which is only a consequence of (wrongly) seeing language as a calculus (remember PU, Teil II, p. 181: "Wie setzt sich der Sinn des Satzes »Ich habe ibn noch immer nicht gesehen« aus die Bedeutungen seiner Worter zusammen?"). Dummett's mistake is to be located in a subtle glissando from an acceptable startingpoint to a wrong conclusion. The starting point is simply the acknowledgement of the fact that "we understand a sentence by understanding the words that compose it"14; Wittgenstein has to agree upon this, because, after all, the sentence exists only if its components exists and who cannot understand the component words cannot, of course, realize what it is being spoken about. It is thus undeniable that, in some sense, we understand a sentence !2Y understanding its components or, to put it differently, that understanding the individual words is the first condition for being able to understand the whole sentence. The question is whether this necessary condition is at the same time a sufficient one. Since he speaks about 'deriving', Dummett seems to be convinced that the answer is 'yes'. Not so Wittgenstein, who always insists upon the importance of the context: he would probably say that the meaning of the sentence is determined (not simply by the meaning of its components, but) by the way these components are used in a particular context. One has to be able to understand the component words, no doubt, because otherwise there is no possibility of grasping their use in a certain particular situation; but it is this particular way of using or applying them in a particular context that determines the meaning of the sentence. ('Meaning is use' not only for individual words, but also for a whole sentence!) To put it in a nutshell, understanding the words is only a first step (admittedly essential, but not sufficient); the decisive element is grasping the use which is being given to them in a certain case. Sometimes it might be possible to (simply) deduce the use of the compound from the usual uses (meanings) of its components, if the context in -which the sentence is being uttered has some sort of semantic neutrality, or it is the standard 13 14

Michael Dummett, Truth and Other Enigmas, Harvard University Press, Cambridge, Massachusetts, 1978, p.448. Dummett, Ope cit., p.444.

35

context for that particular sentence; it is in such circumstances that meaning appears to be intrinsic to the words, because the context does no require a special kind of use, but only the one which is already 'associated' with these words and the resulting sentence. But this is by no means always the case, and it is certainly not the case with philosophical sentences, which are used in special contexts, very different from those of ordinary speech. If, then, the meaning of the sentence is not generally deducible from the meanings that its components have l2Yr se, a calculus of meanings does not appear any more as the unique instrument by means of which we grasp the sense of a sentence. Consequently, we are entitled to distinguish between the kind of understanding which is arrived at by simply 'calculating' (i.e., combining in a certain algorithmic way the meanings of the component words) and the kind of understanding which is provided by grasping the way in which a sentence is used in a particular context. The former is a prima facie understanding, in the sense that it helps us see that the sentence 'says' something (and what it apparently says), without necessarily enabling us to grasp its 'true' meaning or to see 'the point'; that is, this kind of initial understanding does not necessarily provide an adequate grasp of what was really meant (by the utterer) when uttering the sentence. Consequently, it does not guarantee that we would be able to answer it in a relevant manner, starting thus an adequate dialogue with the utterer. In cases in which we are unable to do that, it would not be improper to say (as Wittgenstein suggests) that we only believe to have understood the sentence, where in fact we didn't. The proper understanding of a sentence is secured only by a grasp of its use in a determinate context, which enables us to see what was actually meant by it and, consequently, to answer it properly. The proper understanding is, probably, conditioned by the existence of a prima facie understanding, but it is by no means determined by it. Accordingly, we should distinguish between prima facie meaning (which is simply derived from the meanings of the conlponent words and exists even in sentences like 'green ideas sleep furiously'), and proper meaning, which exists only when a coherent technique of use, for particular, reasonably well-determined contexts or situations, is in fact available (which, of course, is not the case with 'green ideas sleep furiously').15

15

36

I am, of course, not trying to suggest that all sentences which lack proper meaning are exactly of the same kind as 'green ideas sleep furiously'. Philosophical sentences obviously differ, in essential respects, from this famous example of nonsense; nevertheless, they may share with it an important feature: lack of an adequate technique of use, allowing inter-subjective agreement on the correctness of their application in determinate contexts.

One can find, I think, in Wittgenstein's work, several arguments which can be used to defend this distinction. Such an argument might be the following: were the sentence-meaning always derivable from the meanings of the component words, we should be able to convey the same sentence-nleaning, in different languages, by using the 'same' component words (Le., by using in a second language individual words which are simply the equivalent of those used in a first one, in order to express the same sentence-meaning). That is, if the component words played the decisive role, then we would have always the same semantic result (at the propositional level), when assembling the same semantic components (Le., equivalent components). But, although this algorithmic method of translating works very well in many (banal) cases - of the kind 'The cat is on the mat'-, it certainly fails in many others, as any translator knows. Wittgenstein had also this fact in mind, for he remarked: "Was ist die korrekte Obersetzung eines englischen Wortspiels in's Deutsch? Vielleicht ein ganz anderes Wortspiel. "16 One could, of course, claim that plays on words are quite irrelevant for our topic, because it is misleading to talk about proper meaning in their case (many such plays on words simply have no cognitive meaning). This is not, I believe, a strong objection, because the same argument can be supported by evoking proverbs, which certainly have a (cognitive) meaning: it is notoriously the case that one can seldom translate a proverb by using in the second language the counterparts of the words which are used in the first one. A second, and stronger, objection could be based on the fact that most proverbs belong to metaphorical discourse, whereas philosophical sentences do not. This point shall be dealt with a bit later. The main argument which supports the distinction between prima facie and proper meaning, as well as this reading of Wittgenstein's view on meaninglessness, is to be found in his remarks about our failure to understand some sentences, despite the fact that we understand perfectly well their component words. A first example is the one given in The Blue Book (BBB, op.cit., p.9), where Wittgenstein stresses the fact that, although we understand 'I feel', 'water', 'in my hand' and 'three feet under the ground', we actually fail to grasp the meaning of the compound 'I feel in my hand that the water is three feet under the ground'. The prima facie meaning is certainly clear, but since an adequate technique of use for this sentence is lacking (we cannot say when and how it should be applied), its proper meaning cannot be grasped. A second example is offered in PU, § 350, where Wittgenstein insists that we don't understand 'it's five o'clock on the sun' by 16

Letzte Schriften tiber die Philosophie der Psychologie, hereafter LSPP, Basil Blackwell, Oxford, 1982, vol.l, § 278.

37

simply understanding the component words or the (apparently) similar sentence 'it's five o'clock here'. Again, it is easy to get a prima facie understanding of it, an understanding which enables us to explain that the sentence speaks about time, about the sun, and that it 'says' that...; but, since we are not familiar at all with the kind of use that this sentence should have (we don't know when and how it should be applied by people living on the sun), we miss, in fact, its proper meaning. When hearing it, we might be sure that we understand what is being said; we can even engage in a talk having this sentence as starting-point. But it is clear that, in some essential way, we don't really know what we are talking about. The conclusion is, once more, that meaningfulness is not a matter of assembling meanings (of individual words) according to some algorithm, but rather a matter of giving an adequate technique of use. What makes this conclusion plausible is the fact that the above-mentioned distinction does not apply exclusively to sentences, but also to other linguistic compounds, the meaning of which cannot be simply deduced from component meanings. For instance, when hearing the French expression 'un escroc sentimental" someone could deduce that it refers to a crook who happens to be particularly sentimental; but this is not the case, and it is only by grasping the particular kind of use which is being given to this expression, that one manages to understand it properly. Quite similar is the case of some apparently paradoxical expressions like 'yes and no', and so on. Now, here we must face what is (as far as I can see) the main objection to this way of seeing things. It might be said that what I have called proper meaning is nothing else than some kind of metaphorical compound meaning, which (obviously but also trivially) cannot be deduced fronl the usual meanings of the components. Thus (it might be argued), in order to understand some sentences proverbs, for instance - and some expressions, like those just mentioned, one has to grasp a metaphor, and not the so-called 'proper meaning' the distinction between the two kinds of understanding and meaning becomes useless and supertluous, as soon as one realizes that what is in fact involved is the well-known difference between nonnal and metaphorical speech. Moreover, realizing this, one is able to reject Wittgenstein's claim about the meaninglessness of philosophical sentences, because these don't belong to the domain of metaphorical speech. Three points can be made in reply to this objection. The first (and the least important) is that it is not sure at all that all philosophical questions and theses do really belong to non-metaphorical speech. On the contrary, many fragments (not only some of the ones written by ancient philosophers, including Plato, but also many written by thinkers like Hegel and most of those belonging to authors like Heidegger) simply cannot be understood otherwise than as pieces of metaphorical thought. But since this is certainly not generally the case with philosophical 38

thinking, and since Wittgenstein attributes meaninglessness to philosophical questions in general, this counterargument is not decisive. My second point is that the distinction between prima facie meaning and proper meaning is applicable (and useful) in many cases in which the distinction between metaphorical and non-metaphorical speech cannot be applied; for instance, in the examples discussed by Wittgenstein: 'it's five o'clock on the sun', 'I feel in my hand that water...' etc. Thus, it simply cannot be the case that every situation in which we understand 'in a way', but we nevertheless miss the point (fail to grasp the proper rneaning) could be explained by evoking a failure to grasp a metaphor. Metaphorical speech cannot be, in general, the key to the problem created by such situations. My third argument is that if, in order to explain these situations, one postulates a distinction between 'normal meaning' (which is grasped) and 'metaphorical meaning' (which is not), while rejecting the distinction between 'prinla facie' meaning and 'proper' meaning as superfluous, absolutely no progress has been made. A similar 'multiplication of entities' appears in both cases. Our Occamean aspirations are not better served by the fonner distinction, than by the latter; and, as said above, some of Wittgenstein's examples cannot be clarified by applying the former, while they can if the latter is applied.

* As already suggested, somebody could be inclined to object against the distinction presented here by using a sort of tertium non datur-principle, and saying that there is no such thing as 'proper' understanding (as opposed to 'prima facie' understanding); the only adequate distinction which applies here - it might be said - is the banal one, between understanding and misunderstanding. Thus, what one would like to call 'prima facie' understanding is in fact either complete lack of understanding (as in the case of 'green ideas sleep furiously') or simple misunderstanding (as in the case of 'it's five o'clock on the sun', in which we misinterpret things by taking time on the sun to be identical with time on earth). Consequently, there are no 'different kinds' of understanding; either one understands a sentence, or one doesn't (misunderstandings are, of course, cases in which one fails to understand) - tertium non datur. The problem with such an objection is that it is based on the refusal to acknowledge clear-cut factual differences. The idea that there is only one kind of understanding and meaning simply flies in the face of well-known, obvious facts. Suppose one utters the sentence: "He wears scarlet". Who does not speak English at all understands, of course, nothing of it. Who speaks English reasonably well, but happens to be unfamiliar with the shade called 'scarlet', obviously can acquire 39

a general idea about what has been said, although one that is incomplete or insufficiently clear for some kinds of dialogue which could start from this sentence. Who knows what 'scarlet' means, but has no idea about gowns and their significance in the academic world of Oxford or Cambridge, has an even clearer idea about what has been said, but can still be unable to grasp some of the meaning involved, and can still fail to engage fruitfully in a certain possible dialogue. It is only someone who knows how this sentence is used in some English colleges that manages to grasp (entirely and clearly) the meaning of it, a performance which can be confirmed by an adequate reply given to that sentence. What this shows is, of course, that there are many different degrees of understanding a sentence. It is therefore not absurd to consider such a distinction like the one proposed here, and it is as natural to talk about 'partial' or 'insufficient' understanding as it is to talk about indefinite or 'insufficient' meaning (as in PU, §99). Moreover, the prima facie understanding is itself gradual, even if, perhaps, proper understanding is not. And, it might be supposed, precisely this possibility of understanding some of the meaning of a sentence opens the way for illusions: one can believe one has grasped the proper meaning of a question because one has understood something of what 'it says'. The last objection I would like to consider is one prompted by the feeling that a Wittgensteinean view of the kind defended here implicitly denies the combinatorial character of language, as well as the existence of some principles which govern our potentially infinite combinations of words. Many people would object to this view by saying that, even if in many cases we could face difficulties when deriving the sentence-meaning from the meanings of component words, it is simply in1possible to account for our huge - unlimited, indeed - capacity of creating and understanding new sentences, otherwise than by admitting that new sentencemeanings are created by combining in new ways the old meanings of our familiar individual words; which, in tum, simply cannot be accounted for, otherwise than by supposing that we somehow master a set of principles which govern this combinatorial process. Thus, o ummett, speaking about the fact that once we have mastered a language we are "able to understand an infinity of sentences of that language", says: "this fact can hardly be explained otherwise than by supposing that each speaker has an implicit grasp of a number of general principles governing the use in sentences of words of that language"l? If giving meaning to sentences is a matter of combining components and of applying principles, then it seems quite strange that, while doing that successfully in so many cases and fields of discourse, we nevertheless fail completely and sy17

40

Dummett, Truth and Other Enigmas, op.cit., p.45].

stematically to do the same thing in philosophy. This looks as if we had a very reliable linguistic tool which worked wonderfully in most cases, while suddenly failing in only one case: philosophical problems. But philosophical problems and topics do not seem to be so different, so unusual, so wild, as to generate such a total failure of language; on the contrary, they sometimes seem to be very near to other (scientific, moral, social and even daily) problems and topics, so that this huge gap appears to be hardly understandable. How could philosophical questions, which are similar and even connected to other meaningful questions, still fail to be meaningful? How could philosophical discourse fail so dramatically, while being apparently analogous to other, successful, kinds of discourse? There seems to be an unacceptable tension between this failure of language in a particular field and the systematic success of its combinatorial techniques in all other fields. Should an acceptance of the Wittgensteinean view compel us to reject the idea that language has a combinatorial character and is governed by some general principles that we apply when combining words? This is a difficult problem, and I shall not try here to give a complete answer to it. My aim will only be to show that (at least a main part ot) the trouble which the Wittgensteinean view seems to create here is in fact caused by a hidden, wrong, supposition. When pointing out the combinatorial character of language and the role played by general principles, authors like Dummett seem to commit themselves to the supposition that, since conlbinatorial procedures and general principles are so successful, the whole success of language is due to them and nothing more is required in order that meaningfulness and understanding be reached. This, I think, is certainly wrong, and the origin of the error is probably an excessively algorithmic view of semantic performance (which is sometimes seen as if it was identical with syntactic performance), a view dominated exclusively by the logical point of view. But, if Wittgenstein managed to prove something, then he certainly proved the truth of the maxim that, in linguistic matters, "Logic is not enough". Of course we manage to produce inummerably many new sentences by combining old words, of course combinatorial procedures playa big role in achieving meanigfulness (for these new sentences); it might even be the case that some general principles are at work here (but, from a Wittgensteinean point of view this is not sure at all). Nevertheless, the success of this creative linguistic enterprise should not be explained (and cannot be, I think) exclusively by evoking the reliability of combinatorial procedures and of general principles. Even if reliable, these are certainly not infallible, and, what is more important, they do not ~rantee and do not, by themselves, secure meaningfulness and success for any semantic performance. When success is present, its presence is not due ex·elusively to them, but also (and always) to other factors, like successful analogy, the existence of the right kind of context or situation etc. Our newly created sen 41

tences are thus meaningful not simply (and only) because combinatorial procedures and general principles have been applied correctly in the cases involved, but also because those cases happened to be sufficiently analogous to other, familiar, cases, so that the procedures and principles used by us could work successfully; and also because the particular contexts in which these new sentences were to be applied were the adequate ones. These supplementary conditions being fulfilled, adequate uses for these sentences were available, and it is the existence of an adequate way of using that guarantees meaningfulness; nothing else can secure this, and (as the examples evoked and many others show) certainly not the combinatorial procedures or the general principles. Now, in the case of philosophical questions, it is precisely this element which lacks: an adequate technique of use, an adequate kind of context, the right kind of underlying analogy or a situation that is sufficiently similar to others (in which analogous sentences can really be applied meaningfully). As Wittgenstein insists, we speak about time in ways similar to the way in which we speak about space, or in which we speak about objects, neglecting the deep disanalogies involved; we speak about meanings exactly as we speak about (hidden) entities or processes; about mind as an aethereal object (BBB, p.47); about redness as an object which could be destroyed (BBB, p.3l); about thinking as an invisible process, etc. Wrong analogies, inadequate contexts, and unsatisfactory techniques of use can thus lead to meaninglessness, even if all the combinatorial procedures and general principles have been correctly applied. Thus, adopting Wittgenstein's view about the meaninglessness of philosophical questions does not imply rejecting the idea that combinatorial procedures are useful and important in the process of producing new meaningful sentences; it only amounts to saying that they are never enough, so that even if a combination of words is (combinatorialIy) correct, an adequate use and consequently an adequate meaning can still be missing. On the other hand, it is certainly true that this view implies that the philosoph ical use of words is, in some important even if not easy to explain sense, abnormal; and Wittgenstein has never made a secret of this: he has always stressed the huge difference between 'normal' and philosophical uses of language 18 . If it can be proved that there is no such difference, and that philosophical uses of language are (in the relevant respects) absolutely normal, then this would be, I think,

18

42

Abnormal uses can occur even in ordinary language - as when, for instance, somebody would ask about a five-year old child whether he is married or single. But this seldom happens, because in ordinary language abnom1al uses are, in general, caused by lack of linguistic competence, This, of course, is not the case with philosophical (abnonnal) uses.

a decisive proof against his view. But the mere existence and usefulness of combinatorial procedures does not provide such a proof.

(C)

By adopting the interpretation I am defending here, one can reconstruct Wittgenstein's thinking in a way in which its internal coherence is displayed. Such a reconstruction would go along the following lines: (i) since philosophical questions are meaningless, there can be no genuine philosophical problems (but only misunderstandings, puzzlements, different kinds of mental discomfort etc.); (ii) if there are no real philosophical problems, there can be no proper philosophical assertions, because "Nur wo ein Problem sein kann, kann etwas behauptet werden ff19 • This is one of the reasons for which philosophy cannot advance theses (see PU, § 128; there is, of course, another main reason, namely that philosophical utterances refer to linguistic conventions, not to the nature of things, as philosophers unfortunately believe - PU, § 104,114. This reason will be dealt with later). (iii) not being able to make assertions, philosophy neither explains, nor deduces anything (PU, § 126) and, consequently, does not construct theories (PU, § 109: "wir dUrfen keinerlei Theorie aufstellen"). (iv) thus, the results of philosophy are not constructive, but destructive: they consists in uncovering pieces of nonsense (PU, § 119) and in destroying houses of cards ("Luftgebaude" - PU, § 118). (v) by playing this destructive role, philosophy proves to be a sort of therapy: philosophical methods are like different therapies (PU, § 133), which is only normal as long as a philosophical question is like an illness (PU, § 255). (vi) this also shows that philosophers do not solve objective difficulties, but only help one to overcome one's own ('subjective') philosophical confusions; that is why Wittgenstein claims that 'the job to be done is... really a job on oneself (MS 213, 407) and that 'philosophy is a tool which is useful only against philosophers and against the philosopher in us' (MS 219, 11). (vii) consequently, as soon as these 'subjective' confusions disappear, philosophical problems themselves disappear completely (PU, § 133). My conclusion is thus that the distinction between prima facie meaning and proper meaning can not only give a foundation to the thesis that philosophical problems are meaningless, but also make Wittgenstein's whole metaphilosophy 19

Philosophische Bemerkungen, § 151.

43

appear very coherent. If, as claimed here, philosophical questions do have prima facie meaning, but they generally lack proper meaning, since there is no coherent, adequate, agreed upon technique of use for them, then the generally negative attitude towards philosophy taken in the Philosophische Untersuchungen is not to be seen as unjustified (as merely a manifestation of an idiosyncrasy, the first symptoms of which were already visible in the Tractatus); and, finally, the interpretation based on this distinction between two kinds of meaning explains the quite surprising and hard to accept fact that Wittgenstein dared to speak in such a general manner about philosophical questions, dismissing them all as meaningless.

1.3 Why do philosophers ask meaningless questions? The case that Wittgenstein makes against philosophical problems is perhaps not convincing enough, but it is by no means a weak or an insubstantial one. He gives many examples of philosophical 'problems' which are wrong· in themselves because they spring from a false analogy: from seeing time as a river, time intervals as similar to spatial distances, future as something hidden, meanings as mental entities, thinking as an invisible, occult process, mind as a queer medium, linguistic use as a kind of object (BBB, p.5), rules as 'rails' which predetermine all correct results and so on and so forth. Such exanlples, of course, cannot be supposed to put an end to all disputes on the topic, but they might convince some people that Wittgenstein was actually right. Now, what I want to consider here is precisely what happens in such a case. Suppose one accepts the main idea that philosophical 'problems' are not problems at all, but rather meaningless questions. Naturally, this cannot simply be the end of the matter. The obvious question that presents itself now is: why are philosophers always so irresistibly tempted to ask hopelessly meaningless questions? It may seem very strange that, for more than two thousand years, resourceful thinkers have kept raising questions which simply didn't make sense, without even noticing this; it may appear as if all philosophers were victims of some sort of illusion that compelled them to engage in a very special kind of useless debate - which is hardly imaginable. Does Wittgenstein has an explanation for this general intellectual mistake? - for ifhe doesn't, then his position becomes extremely vulnerable. Wittgenstein has, I think, some important things to say in this respect and I shall try here to reconstitute his interpretation of the matter, which, as usual, is not explicitly given, as a complete explanation, in any particular place.

44

My starting-point will be his claim that what makes us ask philosophical questions is some kind of n1ental discomfort about our own language. On one hand, language appears to us as something queer (PU, § 93), which has an incomparable essence (PU, § 97) and we feel it to be the profound and major task of philosophy to elucidate its nature; on the other hand, we are always dissatisfied with it, because it does not have the kind of order, precision, definiteness that (we think) it should have. As proved by the objections of Wittgenstein's interlocutor, obviously someone having the usual philosophical prejudices and intellectual inclinations, we can hardly accept its vagueness (PU, § 98,100,101), the lack of well-determined boundaries (PU, § 68-71), the lack of definiteness (PU, § 99), and the imperfections of its rules (PU, § 68,100). All these features appear to us as abnormal, and we are inclined to think that, normally, language should be characterized by clarity, precision, definiteness, or, to put it shortly, by an ideal order materialized in it (PU, § 105). We want to find a formal unity (PU, § 108) and a law in the way every word is used (BBB, p.27); when we fail to discover it, we fear there is something wrong with ordinary linguistic use (BBB, p.27), as if we didn't really know what our words mean (Ibidem). Our disappointment is generated by the fact that we are committed to an ideal of crystalline purity ("Kristalreinheit"), taken from logic (PU, § 107, 108), an ideal that we try to impose upon ordinary language - although the conviction that this ideal should materialize in it is just a prejudice (PU, § 108). Thus, the first mistake of intellectualist thinking and of philosophy is generated by an ideal of crystalline purity the role of which is misunderstood (PU, § 100); it is this ideal which makes us blind, so that we fail to see how language actually works (Ibidem). But what made philosophers, and thinkers in general, succumb to this ideal? What was the origin of this obsession with crystalline order, purity and exactness? The origin is to be found in the logicisn1 and the scientism which dominated their way of thinking. "Philosophers constantly see the method of science before their eyes, and are irresistibly tempted to ask and answer questions in the way science does. This tendency is the real source of metaphysics, and leads the philosopher into complete darkness" (BBB,p.18). It is absolutely remarkable that Wittgenstein is so categorical aboutthe role played by scientism, "the real source of metaphysics". This shows that his rejection of (traditional) philosophical questions as meaningless is not purely idiosyncratic: he had an elaborated view of the matter, and he even suggested a causal explanation for the failure of metaphysics. One is thus encouraged to explore the causal chain which leads to this failure. The first (or the deepest) cause is scientism. Scientism, and particularly a sort of logicist ideal of crystalline order, generates the typically philosophical dissatisfaction with language, which appears to be the second link in 45

this chain. This dissatisfaction is not something which could be simply neglected, for it indicates a 'sharp' conflict between real language and our ideal of order (PU, § 107), and this conflict caBs into question the very basic conventions of our language; consequently, the problems or the disquietudes we face here are deep (PU, § 111) and 'their significance is as great as the importance of our language'. We have thus to concentrate upon these difficulties, the roots of which 'are as deep in us as the forms of our language' (PU, § 111), and which in fact appear as grammatical puzzlements. In order to solve them, the best thing to do would of course be to clarify the grammar of words and the conventions which govern their use. But this is something that philosophers never do: instead of describing the usual and normal way in which words are used (and which is absolutely in order as it is - PU, § 98), instead of realizing that they 'must stick to the subjects of our every-day thinking' (PU, § 106) because they 'are talking about the spatial and temporal phenomenon of language, not about some non-spatial, nontemporal phantasm' (PU, § 108), instead of clarifying the grammar of words in their original language games, they always try to start an investigation into 'the nature of things'; and this is the third link in the chain which leads to meaninglessness and failure. For instance, in the case of time: "in fact it is the grammar of the word 'time' which puzzles us. We are only expressing this puzzlement by asking a slightly misleading question, the question 'What is...?' This question is an utterance of unclarity, of mental discomfort, and it is comparable with the question 'Why?' as children often ask it" (BBB, p.26). What puzzled Augustine, was an apparent contradiction created by "a conflict between two different usages of a word" (Ibidem); but a philosopher's reaction to such a problem is to start looking for definitions like 'time is the motion of celestial bodies' (BBB, p.27) which seem to give us the very essence of time. Unfortunately, one soon discovers that such definitions are unsatisfactory (Ibidem) and the result is that time (or any other thing the essence of which is under scrutiny) now appears as a queer, mysterious thing. Analogously, "when we are worried about the nature of thinking, the puzzlement which we wrongly interpret to be one about the nature of a medium is a puzzlement caused by the mystifying use of our language. This kind of mistake recurs again and again in philosophy; e.g. when we are puzzled about the nature of time, when time seems to us a queer thing. We are most strongly tempted to think that here are things hidden, something we can see from the outside but which we can't look into. And yet nothing ofthe sort is the case" (BBB, p.6). But why are philosophers inclined to start an investigation into 'the nature of things', instead of clarifying grammar and the linguistic conventions? Several suppositions are possible:

46

(i)

they are mislead by the conviction that language is a sort of picture of the world: "Das Denken, die Sprache erscheint tillS nun als das einzigartige Korrelat, Bild, der Welt" (PU, § 96). But if language is the picture of the world, then its vagueness, lack of defmiteness and imprecision must presumably be somehow connected with our vague, indefinite and imprecise view on 'how things really are'; and in order to eliminate these defects, we have first of all to clarifY the actual states of affairs that language is describing. (ii) one sometimes goes even (much) further, reifying the features of some way of representing things, as if they were characteristics of the real world: "Man pradiziert von der Sache, was in der Darstellungsweise liegt" (PU, § 104) "Man glaubt, wieder und wieder der Natur nachzufahren, und fahrt nur der Form entlang, durch die wir sie betrachten" (PU, § 114). The same confusion between real things and the way we represent them makes us think that the difficulties we face are located in the nature of things, not in the grammar of our words, which constitutes a strong reason for investigating this 'nature'. (iii) one is confused by the unifoffil appearance of words (PU, § 11) and, thinking that behind every substantive there must be an entity (BBB, p.5-6), one tends to connect the problems created by words like 'time' with 'hidden features' of the corresponding 'real' entities; and, of course, one hopes that a discovery of these features would eliminate the difficulties. No matter which or how many of these possible explanations are right, the fact is that philosophers wrongly substitute a supposed ontological difficulty to the real, grammatical, one, and this substitution plays a nlajor role in the process which leads philosophy to meaninglessness and failure. We now come to the fourth and, perhaps, the most important link in our chain. After substituting an ontological problem to the original, grammatical, one, philosophers have to find a solution - for instance, to solve 'the problem of time'. Since they do not see this problem as a grammatical one, they abandon any attempt to clarifY the basic conventions governing the use of the word 'time' and concentrate upon the 'real' phenomenon (time): a second substitution (corresponding to the first one) takes place, by which phenomena are substituted to conventions, as the 'real' topic of philosophical investigation. Philosophers want now to talk about 'real' facts which could uncover 'the mystery' or 'the hidden features' of time. But what could they say about 'real' facts, in order to solve 'the problem of time'? They can either speak about the facts which are connected with our nonnal use of the word 'time' (with the language games in which this word is applied); or speak about facts which are not connected with this use (nor with the usual, relevant, language games). If they choose the former and try to advance theses about the facts connected with ordinary use, then they are confined to trivialities, because our ordinary language games (e.g., with the word 'time') are always based on very familiar and banal

47

facts (e.g., 'taking place before or later than', 'not having taken place yet' etc.) - facts on which everybody agrees. And it is precisely this banal character of the relevant facts that makes a philosophical activity of advancing theses so hopeless: "Wollte man Thesen in der Philosophie aufstellen, es kannte nie tiber sie zur Diskussion kommen, weil Aile mit ihnen einverstanden waren" (PU, § 128). In order to avoid trivialities, philosophers have thus to concentrate upon facts which are not connected with the usual language games played with words like 'time', upon some new relevant facts about time. But how are such new facts to be found? Philosophers not being engaged in empirical research, their only source for new relevant elements is purely conceptual: they find new factual elements by making analogies and mental experiments (and sometimes by mere speculation). Thus, the fourth link in the chain which leads to meaninglessness is the attempt to elucidate the nature of things (of time, etc.) and the corresponding linguistic items ('the concept of time', etc.) not by relying upon grammar and ordinary use, but by relying upon various analogies and mental experiments. What is wrong with this? The answer (which occupies an important place in Wittgenstein's thinking) is that in any such attempt philosophers are bound to err, because, in the process, they inevitably start using the basic words ('time', 'thinking', 'meaning') in a new way, in a way shaped not by the usual contexts of use and language games, but (quite arbitrarily) by different new analogies and mental experiments. They transplant, as it were, the relevant words and the corresponding problems in new contexts, in which the familiar conventions of use (and, since meaning is use, of meaningfulness) are not applicable any more, or, at least, in which there is no guarantee that they could be applied in a non-objectionable manner. The only thing that philosophers have in mind, in such cases, is that the relevant words are correlates of what is to be found in reality (PU, § 96), that they refer to 'something rea\', the 'nature' of which is under scrutiny; the underlying idea is that they must be meaningful, because they are meant to grasp something which 'is really out there'. But, according to Wittgenstein, the question is: "Wozu aber sind diese Warter nun zu brauchen? Es fehlt das Sprachspiel, worin sie anzuwenden sind" (PU, § 96). The best example is, once again, time. In order to find the 'secret of time', philosophers start speaking about 'the nature of time', and what they say (not being based upon an examination of the grammar, the nonnal uses or the language games of'time') is inspired by their analogies between time and a river, etc. In the process of developing their new, purely philosophical, way of seeing, a new use of the word 'time' appears - a typical philosophical use, exclusively guided by the analogy made, for which there is no guarantee that the constraints characterizing the usual language game (and securing meaningfulness for 'time') are still maintained. Somebody could suppose that the abandonment of the usual language games is not a danger for mea-

48

ningfulness, because philosophers can easily invent new, but adequate (even more adequate), language games. But, as mentioned above, the philosophical use of words like 'time', in such contexts, is only guided by an analogy; when this analogy is carried on too far, the 'newly created' language game becomes arbitrary - it is not a proper language game, and it does not secure meaningfulness for the words involved. Now, at this point of the discussion, someone could be tempted to raise the following objection: as the argument described above shows, the meaninglessness of philosophical questions (and answers) is due not only to the supposedly inevitable philosophical mistake of abandoning the usual language games which secure meaningfulness, but also to the unfortunate tendency to carry on an analogy too far; but are philosophers bound to make this second mistake (of pushing on their analogies too far)? Why couldn't they avoid this mistake, and make sure thereby that their sentences are meaningful? Can Wittgenstein indicate deeper reasons for the (supposed) recurrence, or even pemlanence, of this error, or is he simply basing his argument on the well-known anti-metaphysical suspicions characterizing certain traditions of thought? It is of course true that Wittgenstein has been strongly influenced by such traditions, especially by the one which starts with Hamann and Lichtenberg and goes to Nietzsche and Mauthner, a tradition according to which language is responsible for the deepest philosophical errors. But it can be shown, I think, that Wittgenstein's view, even if in a way inspired by such traditions, is not theoretically based on their authority, but rather on very sound arguments. In order to understand them, we have to realize fully the situation in which philosophers find themselves. The philosopher (that Wittgenstein talks about) is a man puzzled by some expressions (like 'time') and eager to solve the resulting problems by elucidating 'the nature of things' (e.g., of time). He thinks that there must be some 'hidden features' which, if captured, could explain 'the essence' of things and thus eliminate any puzzle or enigma. But he does not believe that ordinary language, with its familiar conventions, can be of some help (that is why he never pays attention to grammar); his conviction is that something new must be discovered, which would reveal 'the essence' of things, thus enabling us to explain 'how things really are' and to eliminate all the difficulties we face. Therefore, he must push his investigations (based, for instance, on analogies like 'time-river') beyond the limits within which their implications simply coincide with (and are implicitly confirmed by) the conventions of our usual language games - fOf, within these limits, their implications are familiar and banal; there is nothing new, nothing susceptible to uncover 'hidden essential features' in them: as long as the analogy 'time-river' is used only in order to describe relations like 'after' and 'before', which correspond unproblematically to the language games we play with temporal words, no philosophical discovery is possible - in this 49

way, the philosopher can only illustrate our grammatical conventions, but neither elucidate or explain them, nor eliminate the difficulties associated to them. It is only when going further than this, that something new can reasonably be expected. Thus, the philosopher is compelled, by his own assumptions, to go beyond the limits within which the familiar language games for temporal terms guide and help him. Pushing thus further his analogies, he asks questions like 'What is the origin of time?', which are not usually asked in ordinary speech; but raising such questions becomes possible because the analogy made seems to support them (we can ask 'What is the source of this river?' - why couldn't we ask the same thing about time, as long as time is 'like a river'?) There are no language games for temporal terms which would allow such a question and secure its meaningfulness, but there are language games for other terms (which refer to entities with spatial dimensions) which can be imitated. And this is precisely what philosophers do: they use the main words ('time') in a new way, by imitating other language games that they consider analogous; their strategy is "due to the fascination which the analogy between two similar structures in our language can exert upon us" (BBB, p.26). But the result of this strategy is simply nonsense: as Wittgenstein put it, "nonsense always arises from forming symbols analogous to certain uses, where they have no use"20. The analogy seems to suggest that 'What is the origin of time?' makes sense, whereas in fact it doesn't - for 'time' refers to relations like before, after etc., and 'What is the origin of before?' doesn't make any sense. There is no proper use for such sentences (they have a prima facie meaning, but not proper meaning), although philosophers ventilate them. In order to avoid such failures of the philosophical attempt to make sense, it is necessary to eliminate these 'new' ways of using words like 'time'; to go back to the ordinary use of words: "Wir flihren die Worter von ihrer metaphysischen, wieder auf ihre alltagliche VelWendung zuruck" (PU, § 116), and to examine it, in order to explode the misunderstanding - "The thing to do in such cases is always to look how the words in question are actually used in our language" (BBB, p.56). Ifphilosophers do not pay attention to ordinary linguistic use and do not confine their analogies within the limits indicated by the usual language games, they simply have no means to prevent failures of sense, because "no sharp boundary can be drawn round the cases in which we should say that a man was misled by an analogy" (BBB, p.28); analogies can be very useful up to a point, and then suddenly become a source of meaninglessness, for "it is, in most cases, impossible to show an exact point where an analogy begins to mislead us" (BBB, p.28). 20

50

Wittgensteints lectures in 1930-33, in Moore, Philosophical Papers, op.cit., p.274.

Now, what we have here is a very important consequence of the fact that the meaning of a sentence cannot be derived algorithmically from the meaning of its components. Since the meaningfulness of a sentence is not secured by the n1eaningfulness of its components, an analogy which has been pushed on too far can easily mislead us and generate meaninglessness. But analogies simply must be pushed on very far, if they are to lead philosophers to new discoveries susceptible to elucidate the grammar of words like 'time' and to eliminate all difficulties! That is why analogies between different linguistic structures are so dangerous and philosophy should always be "a fight against the fascination which forms of expression exert upon us" (BBB, p.27) or a battle against the bewitchment of our intelligence through language ("ein Kampf gegen die Verhexung unsres Verstandes durch die Mittel unserer Sprache" - PU, § 109): for, while taking us 'beyond the limits' within which the usual language games function, they also take us beyond the limits in which meaningfulness is guaranteed.

51

2. 'THERE IS NOTHING TO EXPLAIN'

"»ErkHirung« nennen wir's; aber »Beschreibung« istes" Nietzsche (Die frohliche Wissenschaft, §112)

Philosophers are still puzzled by Wittgenstein's categorical rejection of explaining and explanations in philosophy. "AIle Erklarung muB fort, und nur Beschreibung an ihre Stelle treten"} and "Da alles offen daliegt, ist auch nichts zu erkUiren "2 are statemen~s difficult to cope with. One reaction to them was to attribute Wittgenstein's rejection of explanations to his 'anti-theoretical' attitude (supposing that explaining is a 'theoretical' enterprise). Another (more sympathetic) reaction was to say that he did not, in fact, reject explanations in general, but only the intrusion of 'scientific' (or causal) explanations in philosophy. My present aim is to show that: (i) Wittgenstein really rejected all 'philosophical explanations', and not just some of them; (ii) his position was not based upon 'antitheoretical', or empiricist, commitments, but on his awareness of the limits of explanation, which limits can easily be deduced fronl the very grammar of to explain as well as from (what Wittgenstein took as) the nature of philosophy. (AJ

The first point to be made, I think, is that Wittgenstein tended to see explaining less as a 'theoretical', and more as an 'empirical' enterprise; more precisely, he saw explaining as an activity that aims at discovering new factual connections and, therefore, that relies upon experience, hypotheses etc. It is thus not irrelevant, 1 think, that Wittgenstein systematically associates explaining with such things as hypothesis 3, conjecture4, discovering 'what is hidden'5, discovering me1 2 3 4

5

Philosophische Untersuchungen, hereafter PU, § 109 PU, §126 PU, §109 Zettel, hereafter Z, §447 PU, § 126

53

chanisms, formulating laws and predicting6 , looking for causes and making experiments7, confirmation by experience8• But perhaps we really begin to grasp his view when noting that, for Wittgenstein, seeking explanations means seeking connections: "What we call 'explanation' is a form of connection"9

Since, according to his metaphilosophical commitments, philosophical remarks should be 'grammatical', we should probably take this as a 'grammatical specification'. From a 'grammatical' point of view, thus, 'to explain' means to find an explanans for a certain explanandum; this, in tum, amounts to finding a connection between what we want to explain and a certain explanans. A historical explanation 10, a causal one I ], are, for instance, precisely ofthis type. Explaining means connecting. Now, a legitimate question would be: What kind of connection is Wittgenstein talking about? He does not say explicitly that the connections we seek, when explaining, are factual or empirical. But, in a often neglected remark, he specifies: "Jede Erklarung ist eine Hypothese"12

and, like any hypothesis, it is always "unsicher" (Ibidem). In other texts, Wittgenstein seems to suggest also that the standard meaning of 'explanation' inlplies the requirement of "agreeing with experience" 13. Although few, condensed and perhaps puzzling, these remarks do constitute, I think, important clues. One could dismiss the dictum "Jede Erklarung ist eine Hypothese" as mere overstatement or as a proof that Wittgenstein was committed to a narrow, empiricist view of explaining; but that would not be fair. Taking the dictum seriously, one can find two main reasons for which any explanation is bound to be (nothing more than) a hypothesis. One reason is that we always aim, when trying to explain, at finding 'what really happened', that is, at finding the actual connection or the real1ink between the ex-

6

7 8 9 10

11 12 13

54

Lectures & Conversations on Aesthetics, Psychology and Religious Belief, hereafter LCA, III, §8 LCA, III, §11 LCA, III, §27,28 LCA - in the Basil Blackwell edition (Oxford, 1970), p.16, footnote 3. See Bemerkungen tiber Frazers The Golden Bough, in Synthese, vol.17, 1967, p.241. (Hereafter, BFGB) Ibidem BFGB, op.cit., p.236 LeA, II, §39

planandum and the explanans, not at fmding a possible one. 14 But, if we try to find the actual connections, we are bound to engage in making hypotheses: that this or that is the real link between what we want to explain and what we take as 'the explanans' can never be more than a conjecture. Thus, explanation captures 'the real connection' only at a price: it boils down to a mere hypothesis. The other reason is that, when trying to explain, we always aim at fmding something new, which, added to the real facts, would made them understandable, i.e. would generate the explanation we need. When criticizing the "Tendenz, zu erklaren, start bloB zu beschreiben", Wittgenstein clearly suggests that explaining involves 'completing' the facts, adding something to them, whereas describing (that he recommends) has no similar implications. "Das bloBe Beschreibung ist so schwer weil man glaubt, zum Verstandnis der Tatsachen diese erganzen zu mlissen. Es ist, als sahe man eine Leinewand mit verstreuten Farbflecken, und sagte: so wie sie da sind, sind sie unverstandlich; sinvoll werden sie erst, wenn man sie sich zu einer Gestalt erganzt. - Wahrend ich sagen will: Hier ist das Ganze. (Wenn du es erganzt, verfalschst du es.)" 15

That is, 'to explain' means 'to understand by completing the facts', whereas describing may make us understand without adding anything to the facts. But if explaining implies 'adding something to the facts', it follows that (a) explaining relies upon experience, for only experience could provide new facts to be added; or (b) explaining relies upon hypothesis, for in any attempt of completing the facts by adding connections, patterns, 'missing links' etc., we are bound to engage in hypothesizing about how things are connected, what was the pattern of events, what made things happen and so on. Roughly speaking, the connections that we add to the facts are supposed to exist between them, which nleans that supposition and hypothesis are unavoidable. In both case (a) and case (b), explaining implies making a discovery (about facts or connections between facts), fmding something new, and thus relying upon expe-

14

15

This is shown by the fact that, while criticizing Frazer for his attempt to give explanations - see, especially, BFGB, op.cit., p.235 - , Wittgenstein remarks: "Das Richtige und Interessante ist nicht zu sagen: das ist aus dem hervorgegangen, soodem: es konnte so hervorgegangen sein" -BFGB, op.cit.,p.252. That is, explaining anlounts to saying 'how things really happened', to finding the actual connection, whereas one should only try to say 'how things could have happened'. Frazer's mistake was precisely to have tried to point out the actual link which supposedly 'explained' magic. Bemerkungen tiber die Philosophie der Psychologie, Basil Blackwell, Oxford, 1980, (hereafter BPP), Band I, §257

55

rience, confirmation by experience etc. Explaining is (in this particular sense) nlore or less empirical; it looks like an empirical enterprise. One is thus now able to understand why explaining is incompatible with doing philosophy. Philosophical problems are conceptual, not enlpirical (PU, §109), they are a matter of grammar not of experience; philosophical investigation is characterized precisely by the fact that it has nothing to do with discoveries, new results etc. (PU, §126). Seeking to explain, i.e. seeking new connections (between the explanandum and a explanans), is not a philosophical enterprise, because this enterprise is typically grammatical (PU, §89,90,120; Z, §458). Philosophy itself is interested in some connections, for it seeks 'die iibersichtliche Darstellung' and "Die libersiehtliehe Darstellung vermittelt das VersUindnis, welches eben darin besteht, daB wir die 'Zusammenhange sehen'" (PU, §122)

But the connections that philosophy is interested in are only possible connections (PU, §89-90), not actual or factual ones; philosophy concentrates on 'formal', not on 'real' connections: "Wie man eine interne Beziehung der Kreisfoml zur Ellipse dadurch illustrierte, daB TIlan eine Ellipse allrnahlieh in einen Kreis liberfuhrt; aber nieht urn zu behaupten, daB eine gewiBe Ellipse tatsaehlieh, historisch, aus einem Kreis entstanden ware (Entwieklungshypothese), sondern nur urn unser Auge flir einen formalen Zusammenhang zu scharfen" (BFGB, op.cit., p.241-242)

Now, someone could be strongly tempted to ask the following question: how on earth can Wittgenstein be so sure about what are philosophers interested in? does he possess any indisputable proofs for his claim, or is he simply deciding upon the matter, and if the latter is the case, isn't his position extremely dogmatic? Of course Wittgenstein cannot give a demonstration about that (as nobody else could, obviously). But his assumption on what philosophy is interested in is not purely dogmatic; he has an interesting argument to rely upon. The starting point of this argument is the principle that one should look at what philosophers actually do, in order to see what they are actually interested in: as he puts it in his Philosophische

Grammatik, "Sage mir wie Ou suchst, und ich werde Dir sagen was Du suchst" (PG, p. 370). Now, how do philosophers search into their topics? The obvious answer is that they mostly engage in conceptual analysis and mental experiments, neglecting activities such as collecting data, watching the flux of events, describing what actually happens at a particular point etc. This proves that they are interested in conceptual, not in empirical matters, in possible (logical, formal, conceptual) connections, not 56

in actual, enlpirical ones. If philosophers were really interested in empirical matters, in actual facts, they would have practiced observation, empirical experiments (of the kind scientists make), data gathering, measurements and so on; but they don't usually do that, and this shows what they are really searching for. They are not after 'what really happened' - they are rather investigating possibilities and they are fascinated by virtual alternatives. They don't focus upon the contingencies of reality, but rather upon (what they take as) logical necessity; that is why not every connection can be of philosophical interest. Now, although one could perhaps find some counterexamples to this description of philosophical activity, it is hard to deny that, as regards the Western tradition, there is nluch truth in it: Wittgenstein is not simply dictating dognlatically 'what philosophy should be interested in'; he has a point, and his interpretation cannot be simply dismissed, because it is based on acknowledging a major 'familyresemblance' among Western philosophical doctrines. Being interested exclusively in 'formal' or conceptual connections, in 'internal', i.e. grammatical, links, philosophy should never engage in explaining, for explaining points out only factual, actual links, which are, in a certain sense (see PU, Teil II, xii), philosophically irrelevant. Explaining is confmed to the field of factual items and actual connections, whereas philosophy should also deal with counterfactual items and possible connections. Now, some authors would perhaps say that, precisely because Wittgenstein took explaining as an empirical enterprise, he never rejected explanations in general, but only some particular forms of factual or 'empirical' explanation. Even Baker and Hacker seem to be attracted by such an idea, when they suggest that philosophy "explains by description whereas science explains by hypothesis"16. Couldn't Wittgenstein's remarks be interpreted as saying that philosophy is not interested in 'empirical' explanations, but only in conceptual explanations based on describing linguistic use? Such an interpretation is unacceptable first of all because there is no textual evidence to support it. Wittgenstein constantly invokes the opposition between describing and explaining17 and he never presents the former as a variant of the latter. But there is a deeper argument. Wittgenstein's dictum "Jede Erklarung ist eine Hypothese" shows that he considered explaining as a particular kind of language game, in which one finds 'new' elements (connections, causes, mechanisms, laws, processes, 'hidden' entities etc.) which supposedly account for what we want to ex-

16 17

Baker, G.P., Hacker, P.M.S. Wittgenstein Understanding and Meaning, Basil Blackwell, Oxford, 1980, p.490 This opposition is already present in The Blue Book, and then always evoked - see Z, §220, and especially PU, §109, 126,654; for a late mention, see BPP, Band I, §22

57

plain. Inside this game, we are bound to try to capture a reality that we have never perceived (BPP, Band I, § 256), to make hypotheses and to say more than we know (hoping that experience will confirm our conjectures). On the contrary, describing is a (different kind of) language game, in which we don't advance theses (PU, §128), and we stick to what we already know (PU, §126,127; BFGB, op.cit., p. 235; BPP, Band I, §257); when describing, we do not 'complete' the facts, we do not look for something 'new' - we only Judge the geography' as it now is. 18 Thus, describing cannot be seen as a particular kind of explaining; in a way, it is rather its opposite. There is no such thing as 'explaining by describing'. But someone could be tempted to raise the following objection: "Wittgenstein speaks as if there was only one kind of explanation, the explanation based upon factual connections; but there also are explanations based on conceptual connections and maybe 'die iibersichtliche Darstellung' is one of them. Couldn't philosophy give conceptual explanations, while ignoring the 'factual' ones as irrelevant?" Indeed, is there just one kind of explanation? The background of this argument includes the idea that, since there are different ways in which we come to understand - sometimes we understand by fmding new facts and empirical connections, some other times by simply examining conceptual connections - there must also be different kinds of explanations. Wittgenstein has explored mentally the difficulties created by this argument, for he remarked: "Not everything that causes understanding is an explanation" 19

This is connected with other Wittgensteinean argun1ents too, but it is, I hope, also connected with my present topic. It could be said that we are tempted to stress that there is not a single way of explaining and that there must be various kinds of explanations simply because we are aware of the diversity of our ways of coming to understand. But we should also be aware that not every way in which we reach understanding is a way of explaining. Wittgenstein seems to have been preoccupied to claritY that. In his famous Memoir, Norman Malcolm reports that he used arguments like the following: 'what if I hit you on the head and thereafter you were able to use the word red correctly /which, it is supposed, you were not able to do before - my note, A.P.L/; would thus be an explanation of the word red?' His intention was to reach the conclusion that 'An explanation is not anything that produces understanding, exactly as a ~ is not anything that opens the door'.

18 19

58

Bemerkungen tiber Die Grundlagen der Mathematik, (hereafter BGM), IV, §52 Wittgenstein's Lectures on Philosophical Psychology, 1946-47, P.T. Geach (ed), Harvester, Wheatsheaf,1988,p.22

Once established the fact that our inclination to universalize explanation (by supposing that behind any act of understanding there must be a successful explanation) is wrong, Wittgenstein stressed that we often come to understand by getting rid of confusions, by eliminating misunderstandings and ambiguities and also by seeing more (BOM, II, §85). For instance, die Ubersichtliche Darstellung is a way of coming to understand by seeing more. We can also reach understanding by clarifying meanings, the grammar of our expressions or conceptual connections; and clarifYing does not necessarily imply explaining. To be sure, there are many kinds of explaining; but it does not follow that philosophy should engage in some of them. Its task is to clarifY, not to explain, and the kind of understanding that philosophy provides is based on clarifications. But it is also true that Wittgenstein sometimes uses the expression "begriffliche ErkUirung" (see, for instance, BPP, vol.II, §381). Did he accept conceptual explanations, as opposed to the empirical ones? Not at all. As some authors have stressed20 , Wittgenstein often uses erkUiren in the sense of to clarify. And there is little doubt, I think, that "begriffliche ErkUirung" should be interpreted as 'conceptual clarification' and not as 'conceptual explanation', because only in this way can we account for the thesis that grammar "beschreibt nur, aber erkHirt in keiner Weise, den Oebrauch der Zeichen" (PU, §496). A "begriffliche ErkHirung" is a grammatical enterprise, and since grammar never explains, this enterprise must be one of clarifYing, not of explaining. An objection to this could be that PU, §496 does not reject explanations, but only the claim that philosophy should explain the use of words. Now, this does not take us very far, because, as long as meaning is use, if grammar does not explain linguistic use, then it does not explain meaning either; and if it does not explain meanings, then it simply can explain neither concepts, nor conceptual connections. For Wittgenstein, this is no problem, for there is nothing to explain: we already know what our concepts mean, even if we make confusions when philosophizing. We simply don't play here 'the game of explaining': we only play the game of describing and clarifYing.

20

See Eike von Savigny, We Must Do Away With All Explanation and Description Alone Must Take Its Place: Wittgenstein An Enemy of Science?, in Philosophie der Naturwissenschaften, Akten des 13. Internationalen Wittgenstein Symposiums, Paul Weingartner, Gerhard Schultz (Hrsg.), Wien, Holder-Pichler-Tempsky, 1989, p.386

59

(B) When dealing with aesthetic and religious matters, Wittgenstein associates again explaining with discovering mechanisms, formulating laws and predicting21 , looking for causes and making experiments (LeA, III, § 11) and confirmation by experience (LeA, III, §27,28). What it is said in these fragments illustrates the basic idea in PU, § 109, namely that "Die Erfahrung, 'daB sich das oder das denken lasse, entgegen unsell11 Vorurteil' - was immer das heiBen mag - konnte uns nicht interessieren" , for "Es darf nicht Hypothetisches in unsem Betrachtungen sein", and, consequently, that "Aile ErkHirung muB fort, und nur Beschreibung an ihre Stelle treten". It is quite clear that the reason for rejecting explanations is that "Die Probleme werden gelost, nicht durch Beibringen neuer Erfahrung, sondern durch Zusammenstellung des Hingst Bekannten" (Ibidem), which underlines the logic of this way of seeing: explaining is an activity based on experience and on finding (of new data), whereas philosophy should stick to description and to examination of the 'workings' of language; its problems "werden durch eine Einsicht in das Arbeiten unserer Sprache gelost" (Ibidem). There is a common element to philosophy and the Kulturwissenschaften, and this is the particularity that they are never interested in empirical matters. As Wittgenstein insisted, "in philosophy 'we know at the start all the facts we need to know'. "22 Philosophical difficulties are not generated by ignorance (concerning some facts); the main feature of a philosophical problem is "daB sich hier eine Verwirrung in Form einer Frage auBert, die diese Verwirrung nicht annerkennt"23. Thus, philosophical problems are confusions or misunderstandings (PU, §109,120), they have the general form "fch kenne mich nicht aus" (PU, § 123) and are the result of our being entangled in our own rules: "Dieses Verfangen in unsem Regeln ist, was wir verstehen, d.h. Ubersehen wollen" (PU, §125). Philosophical problems are conceptual, not empirical - "Diese sind freilich keine empirischen" (PU, § 109) - and this explains why philosophy is possible before any 'discoveries': "Philosophie konnte man auch nennen, was vor allen neuen Entdeckungen und Erfindungen moglich ist" (PU, § 126). Since empirical research is philosophically irrelevant, any empirical enterprise is doomed to failure in philosophy, and it is precisely the confusion between non-empirical (conceptual) and empirical (factual) enterprise that constituted the main root of old-fashioned, wrong-headed metaphysics: 21 22 23

60

Lectures and Conversations on Aesthetics, Psychology and Religious Belief, hereafter LeA, Basil Blackwell, Oxford, 1970, III, §8. Wittgenstein's Lectures in 1930-1933, in G.E. Moore, Philosophical Papers, George Allen and Unwin, London, Humanities Press, New York, 1959, p.323. Philosophische Grammatik, hereafter PG, Basil Blackwell, Oxford, 1969, I, § 141.

"Philosophische Untersuchungen: begriffliche Untersuchungen. Das Wesentliche der Metaphysik: daB sie den Unterschied zwischen sachlichen und begriftlichen Untersuchungen verwischt" (Z, §458). Since philosophical questions refer to concepts, philosophical investigations should only deal with concepts: "Deine Fragen beziehen sich auf Worter; so muB ich von Wortem reden" (PU, §120). This also explains the 'grammatical' character of these investigations, which aim only at conceptual clarification: "Unsere Betrachtung ist daher eine grammatische. Und diese Betrachtung bringt licht in unser Problem, indem sie MiBversUindnisse wegraumt" (PU, §90). Consequently, it is not a matter of contingencies that there are no explanations in this field, for grammatical elements are precisely what remains after all explaining has been abandoned (BPP, vol.l, §432); it is only after 'doing away with all explanation' that we really enter the field of philosophy. In a way, when we are trying to understand art we are also dealing with conceptual problen1s, because we want to clarify the meaning of 'beautiful', 'fme', 'right' (as in the case of 'right architectural proportions') etc. - LCA, I, §5-36. But the main thing is that, when interested in art and culture, we focus upon cultured taste, style, tradition, aesthetic appreciation (LCA, I, §18-31), and here again we don't deal with empirical matters. We are interested neither in facts (e.g., in aesthetic reactions - LCA, III, §11), nor in causes or n1echanisms (LeA, II, §1736) but only in reasons. Wittgenstein was convinced that the main question of Aesthetics "was not 'Do you like this?', but 'Why do you like this?'" and, of course, "To ask 'Why is this beautiful?' is not to ask for a causal explanation"24. As Moore reports, when speaking about Wittgenstein's lectures, "What Aesthetics tries to do, he said, is to give reasons, e.g. fOf having this word rather than that in a particular place in a poem, or for having this musical phrase rather than that in a particular place i~ music" (Moore, 1955, p.19). Understanding art is thus a matter of reasons and ideals; or, reasons and ideals are not empirical items, but normative ones, and therefore cannot be explained - they can only be described: "Reasons, he said, in Aesthetics, are of the nature of further descriptions" (Moore, 1955, p.19). Wittgenstein insists that the only thing one can do, in order to solve aesthetic puzzlements, is some sort of "grouping together of certain cases" (LCA, IV, §2) Of, in music, some sort of "arrangement of certain musical figures" (LCA, III, §9); but such operations like grouping, arranging, comparing are in fact similar to the conceptual investigations we do in philosophy, not to the empirical ones in science. We are not interested here in 'new' findings which would account for 'what happened'

24

G.E. Moore, Wittgenstein's Lectures in 1930-1933, in Mind, LXIV, No.253, 1955, p.1-27 (Kraus Reprint Lin1ited), pp.18-19. Hereafter Moore, 1955.

61

and we don't look for an explanans; we only try to get a better picture of what we already know, of the similarities and differences between the cases involved. On the other hand, Wittgenstein was aware that understanding art is not always a matter of normativity; for instance, HWhen we talk of a Symphony of Beethoven we don't talk about correctness" (LCA, I, §23). In many cases, understanding implies connecting what we are interested in (for instance, 'expressiveness') with a culture or a form of life. Is this an opportunity for explaining? Not at all. Features like 'expressiveness' are connected with the whole culture (UWas gehort also dazu? Eine Kultur, mochte man sagen" - Z, §164), and a whole culture is not the kind of particular item which could be singled out and used as explanans. Aesthetic appreciation is diverse and connected with the whole diversity of "arts and crafts u ; it is simply impossible to explain it, for "I would have - to say what appreciation is . . e.g. to explain such an enormous wart as arts and crafts" (LCA, I, §21). Since "What belongs to a language game is a whole culture" (LeA, I, §26), in order to explain 'cultured taste', 'tradition' or 'beauty' one has to describe whole different cultures (ibidem) and even whole "ways of living" (LCA, I, §35). Even the task of describing is here too difficult (LCA, I, §20), but the task of explaining, Le. of singling out particular items which play the role of an explanans, is impossible. Wittgenstein also claimed that religious matters are not empirical matters and that religious commitments are not opinions which could be explained by relating them to experience (LCA, p.54,57). Religious beliefs "are not treated as historical, empirical, propositions" and in this field "we don't talk about hypothesis" because empirical data are here "no evidence at all" (LCA, p.57). For instance, (it is stressed again and again in the Verolischte Bemerkungen) the question of empirical truth is irrelevant to Christianity. Religious commitments are deep disquietudes (Beunruhigungen) deeply connected with the whole of hunlan life; they cannot be explained, in the proper sense of the word, because explaining them would imply taking into account the whole life or a whole form of life. Once again, what can be done is only describing: "Nur beschreiben kann man hier und sagen: so ist das menschliche Leben" (BFGB, p.236). Exactly like in philosophy, where insatisfaction disappears by our seeing more, in order to understand religious matters we have to give up explaining and 'put together' what we already know (BFGB, p.235). Thus, as philosophy and the Kulturwissenschaftennever deal with empirical matters, explanation (as empirical enterprise) is here simply irrelevant; it is also impossible in principle, because one never can, in these fields, single out an explanans. Understanding should be then based here on description.

62_--------------------

(C) An objection which is often raised against this interpretation of the 'antiexplanation' metaphilosophical position is that Wittgenstein simply could not afford to reject all explaining in philosophy, because he so often engaged himself in explanatory enterprises. If we do not want to place him in the unpleasant (but not very unusual) position of a philosopher who, by his own activity, contradicts and invalidates the grand metaphilosophical theses he himself professes, then we simply have to find another interpretation; we should perhaps say that Wittgenstein used the term 'explanation' in a special sense, referring (for instance) only to some particular kinds of explanation (the scientific ones, say, or the causal ones) and that, consequently, despite his (too) general and categorical formulations, he did not reject all philosophical explaining. Now, it is of course true that Wittgenstein himself engaged in explaining and that he left us, scattered in his manuscripts, a lot of explanations (among which also a good number of causal ones - which, if we are to follow the line of thought promoted by this very objection, makes it doubtful that what he rejected is causal explanation; and, mind you, this was the most plausible alternative interpretation!). But I think that Wittgenstein's own 'sin' of having succumbed to the impulse of explaining proves completely irrelevant to his metaphilosophy, and this irrelevance becomes apparent as soon as one operates a (necessary) distinction between aiming at explanations (as final, or at least autononl0us, results of philosophical research) and simply using explanations (while aiming, philosophically, at something else). When Wittgenstein claims that there is nothing to explain in philosophy, he obviously rejects the idea that explaining should be a main task, a final aim, and an autonomous result of philosophical investigations. As long as (i) the philosophical questions refer to words (PU, § 120); (ii) the subject-matter of philosophy consists in grammatical conventions (or rules) in the web of which we are entangled (PU, § 125); (iii) conventions and rules, as prescriptive items, are not the kind of thing which needs explaining, but only clarification; what we aim at is clarity (PU, § 133); (iv) and the aim of philosophical activity is die iibersichtliche Darstellung (the perspicuous representation - PU, § 122), which, by itself: does not involve explanation, but only description; as long as (i)-(iv) are true then, explaining should not be the business of philosophy; rather, it is describing (Le., producing the perspicuous representation) which should occupy the central place, as the major task for philosophers, while clarity should be their final aim and the perspicuous representation the main result of their activity. But, of course, in the attempt to reach this fmal result, philosophers might have to eliminate misunderstandings, to change the prevailing ways of seeing and thus to 63

persuade; inside this therapeutic activity, they might come to use also explanations (exactly as they use comparisons, analogies, mental experiments, conceptual analysis, examples and even pictures or drawings). All these are nothing else than auxiliillY instruments: they are not constitutive, or sine qua non elements of philosophical activity. Some philosophers could use explanations, others could perhaps do without them: explaining is in no way a basic task for philosophy, and, as an instrument, it can be used or not, according to one's preferences. Even when it is used, explanation does not prove to be an autonomous result of philosophical activity, exactly as comparisons or analogies (even if useful) do not prove to be an autonomous result of research. Thus, the simple fact that some instruments (explanations or comparisons) are used does not contradict the idea that, in the end, philosophy is meant neither to explain, nor to compare. As therapeutic elements, explanations are facultative means the presence of which is irrelevant to 'what is to be done in philosophy'. Philosophers can make (mental) experiments; but this does not affect the truth that philosophy is, in its nature, non-experimental. They can also use pictures and drawings (and Wittgenstein uses a lot of drawings); but the presence of such auxiliary instrunlents as pictures or drawings does not disconfirm or invalidate the idea that 'there is nothing to depict in philosophy' - for, obviously enough, the topic of interest can never be depicted. Analogously, the fact that Wittgenstein uses explanations in his therapeutic or persuasive efforts does not invalidate the idea that 'there is nothing to explain'. That Wittgenstein's explanations are only auxiliary and therapeutic, but not 'constitutive' for his philosophy, is confirmed by the fact that what all of them show is 'simply how some philosophical mistakes, confusions and misunderstandings have appeared, and not 'how things are'. His explanations do not explain time, truth, meaning, thinking and so on; but rather how confusions about time, etc. were born. They do not elucidate the basic philosophical topics - they only 'clear up the ground', as he says in PU, § 118, in order to make the elucidation of these topics possible. Their role is purely destructive (they destroy 'the houses of cards') and therefore therapeutic; and, not being constructive or positive, i.e. not providing answers to the main questions of interest, such explanations do not disconfirm the idea that 'there is nothing to explain': indeed, the relevant answers are never arrived at by explaining. It is also important to note, I think, that the auxiliary role played by explanations in philosophy could be confirmed by many examples in classical thinking. To take just one of them: Hume does provide, indeed, an explanation for our confidence in causality - the explanation based on the existence of a 'habit of the mind' (to extend past experience to the future); but this explanation is neither important (any other

64

explanation would do as well, in that context), nor necessary for his main conclusion (that belief in causality is not rationally justified). The role of this explanation is therapeutic: it simply makes it easier for us to abandon the commitment to causality, by showing us why we are inclined towards such a commitment. I thus conclude that there is no contradiction between the fact that Wittgenstein uses some explanations and the fact that he says 'there is nothing to explain'. The point of this famous remark is not reductive (Le., what we have here is not an attempt to reduce all philosophical activity to one and only one kind of thing, excluding any other kind, and compelling all philosophers to do nothing else than describing); what Wittgenstein really tries to do here is riot to limit the philosophers' freedom of using different methods of inquiry and persuasion, but to remind them that the source of philosophical difficulties is not the lack of explanations and, consequently, that the results we look for do not consist in explanations. This is proved by the following fact: not only that Wittgenstein does not prohibit explaining, but he is perfectly aware that philosophy is neither reducible to, nor simply identical with, describing - for he says that "Philosophie ist nicht Beschreibung des Sprachgebrauchs, und doch kann man sie durch standiges Aufmerken auf aIle LebensauBerungen der Sprache lemel1"25. That is, when saying that philosophy should be descriptive, he does not mean to say that mere description of language use leads one to philosophical illumination; he was perfectly aware that, although the subject of philosophy "must say a great deal about language, it was only necessary for it to deal with those points about language which have led, or are likely to lead, to defmite philosophical puzzles or errors" .26 Wittgenstein, of course, never dreamt of claiming that doing philosophy is a matter of doing just one thing (a matter of simply describing) and nothing else. It is consequently wrong to separate his metaphilosophical ideas about description from their particular contexts, and to interpret them as restricting the range of theoretical instruments that a philosopher is allowed to use. What Wittgenstein always emphasizes upon, in such contexts, is the purpose of the activity, not simply its means. A certain description (of language use), with a certain purpose (eliminating confusions), can provide the desired result (a conspicuous representation); but this is not to say that any description of language is philosophically illuminating, nor that philosophers are not allowed to do something else. By his claims about 'what philosophy should do' and 'what is to be done in philosophy', Wittgenstein does not try to create a canon, or a methodological orthodoxy, by imposing a unique style of doing philosophy; he does not introduce restrictions upon the range of instruments which a 25 Letzte Schriften tiber die Philosophie der Psychologie, hereafter LSPP, Basil Blackwell, Oxford and Cambridge, 1992, Band I, § 121. 26 Wittgenstein's Lectures in 1930-1933, in G.E. Moore, op. cit., p. 323-324.

65

philosopher can use, but only reminds one that the true aim, task and result of philosophy (implied by its true difficulties) is not explanation, but rather an adequate description and a conspicuous representation. (D)

The above considerations might help us to answer also another objection that is sometimes raised in the same context. There is- an inclination of claiming that Wittgenstein's principle 'philosophy should not explain, but only describe' is nonsense, because any description implies the existence (in its background) of some explanations and even theories; there are no 'pure' descriptions, and therefore a requirement like the one implied in this principle simply doesn't make sense. Now, perhaps this objection does not deserve attention at all, because (as anyone familiar with Wittgenstein's thinking will immediately realize) the suggestion that he might have been committed to the idea of 'pure' description is simply ridiculous: it: in the first half of this century, there was a philosopher who could not be suspected of ignoring the 'theory-loaded' character of description, then certainly Wittgenstein was the one. But, nowadays, when the (intellectual) landscape is haunted by certain kinds of 'anarchism' and 'post-modernism' relying on the principle that 'if there are no absolute distinctions, then there are no distinctions at all', it is not, probably, useless to answer such an objection. The first point to make is that Wittgenstein's principle - 'philosophy should not explain, but just describe' - does not commit him to the idea of 'pure' description; this principle only commits him to the idea that there really is a distinction between describing and explaining, and to the consequence that one can engage in describing without also engaging in explaining. Both these commitments are harmless, and I know of no serious arguments against them. Even if behind any description there are to be found explanations and perhaps theories (a debatable clainl, because of its vagueness), this proves neither that describing and explaining are exactly the same thing, nor that one can never describe without simultaneously explaining. Descriptions are, of course, embedded in a web of cognitive elements (classifications, suppositions, theoretical commitments etc., and explanations too), which is the web of human language, thinking and knowledge, but they remain what they are, and distinct from other things they are connected with. It is perhaps the case that, when describing, one implicitly accepts some explanations and even some theories; but that in no way proves that, when describing, one is also explaining. Analogously, when writing a book, one implicitly accepts some semantic conventions, but this does not show at all that, when writing a book, one explains the semantic conventions on which the writing is based.

66

When advancing his 'anti-explanation'-principle, Wittgenstein does not claim, of course, that the philosophical enterprise should be reduced to describing and that describing should be done in a complete intellectual void (i.e. without any comn1itment to suppositions, categorizations, explanations etc.); such an enterprise is obviously impossible. He only claims that the task of philosophy is not to explain, but to describe; and that, even ifphilosophers (like any other people, in fact) are tacitly committed to many different theoretical elements (explanations included), which are always present 'in the background' (for, without them, no description is possible), they should still concentrate on describing (their proper task) and not on the (auxiliary) theoretical elements or efforts which are associated with describing. Perhaps what worried some people is that Wittgenstein once said "Philosophy really lli 'purely descriptive'." (BBB, p.18). It is 'purely' that can create suspicions; but, as I hope it is quite evident, 'purely descriptive' does not imply 'descriptive in such a way that no explanation could possibly be involved, not even implicitly or in the background'; it only implies 'descriptive in such a way that all intellectual efforts concentrate upon describing, not on something else'. Philosophy is 'purely descriptive' simply because it aims only at describing (at giving a synoptic view), not also at explaining or at building theories.

* Both objections I have just tried to answer are based on neglecting a very important distinction: the distinction between the characteristic preoccupation in a certain field and the background of this preoccupation, Of, in other words, the distinction between what is to be done and what is presupposed by what is to be done in a certain domain. The characteristic preoccupation in philosophy (according to Wittgenstein) is describing and thereby clarifying; but, of course, the necessary background for a successful clarification may involve many other things (explaining, as a means to eliminate misunderstandings, deducing, advancing arguments, giving examples etc.). The thing to be done is arriving at a good synoptic view, at the 'Ubersichtliche Darstellung'; but, of course, doing that implies and presupposes many other things, including criticizing, making mental experiments, analyzing, and also explaining, relying upon some explanations, theories etc. If one pays attention to this distinction, then the objections discussed above lose their force: they fail to prove that Wittgenstein's 'anti-explanation'-principle does not make sense. This, of course, does not compel one to accept this principle: but one has to admit that what Wittgenstein says in this respect is not confused.

67

3. 'THERE IS NOTHING TO DISCOVER'

"We seem to have made a discovery... "

(The Blue Book)

It is quite obvious that Wittgenstein's dictum "there is nothing to explain" is directly connected with his (less explicit but nevertheless hard to ignore) suggestion that "there is nothing to discover". Indeed, since explaining is seen as some sort of empirical enterprise (i.e., as an enterprise relying upon the discovery of something new, which accounts for what is to be explained), not being interested in explanations means not being in· search of new facts or connections and not looking forward to discoveries. It is thus no accident that in PU §126 Wittgenstein insists both upon the idea that "ist auch nichts zu erkHiren" and upon the idea that philosophical search is independent of any new discovery: "»Philosophie« konnte man auch das nennen, was vor allen neuen Entdekkungen und Erfindungen moglich ist". But the idea that philosophy doesn't care about new discoveries is bound to sound paradoxical - after all, philosophers claim to reach new results which improve our understanding of things; they are not just expressing metaphysical feelings (as RudolfCamap suggested), are they? And if philosophical activity is not merely expressive, but also intellectually productive, supplying (maybe not truths, but at least) abstract understanding, then how on earth could this activity achieve new results without making new discoveries? Or should we take this rejection of discoveries as part of a typical philosophical rhetoric, with not much conceptual substance behind it?

3.1 Are there philosophical discoveries? The answer to this question is that, as visible in PU § 119, Wittgenstein does not exclude the possibility of sonle discoveries: he admits that we can discover conceptual mistakes, misunderstandings, false analogies and so on, that new discoveries about our own way of thinking are possible and also vital. What he actually rejects is the possibility of discovering philosophically relevant new facts about time, mind, language etc.: any new element inside our subject-matter could be of

69

no possible philosophical interest - "was etwa verborgen ist, interessiert uns nicht". Thus, it is only the possibility of a discovery about the subject-matter that is rejected - the kind of discovery one can make in philosophy is a discovery about oneself, i.e. about one's own way of thinking, one's own errors - that is why, "the job to be done is... really a job on oneself' (as Wittgenstein says in MS 213, 407). Philosophical discoveries tell us something new, but only something about conceptual elements (i.e., about our own way of thinking), not about how real things are. They improve our understanding of things by improving our manner of thinking - not by completing directly our picture of reality. This distinction between two kinds of discovery might seem quite arbitrary, but, far from being so, it is in fact based upon strong arguments which are central to Wittgenstein's view of the nature of philosophical puzzlement. He insists that "there is puzzlement and mental discomfort, not only when our curiosity about certain facts is not satisfied or when we can't find a law of nature fitting in with all our experience" (BBB, p.59) - that is, philosophical disquietudes do not spring from lack of data or of theoretical constructs (laws, explanations, etc.). Were they to have such sources, the discovery of new facts or new connections between facts would have been philosophically relevant and important. But if the real source of philosophizing is misuse of language, conceptual confusion or entanglement in our own rules, no new informations or constructs, but only a deeper understanding of our own way of thinking and of its mistakes can be helpful.

As long as "Die Komplexitat der Philosophie ist nicht die ihrer Materie, sondem, die unseres verknotenen Verstandes"(PB, §2), the discovery we really need in order to handle this complexity is the discovery of those intellectual obstacles that hindered understanding and created Beunruhigungen. Thus, philosophy does lead to discoveries, but to discoveries of old conceptual errors and confusions, not of new facts or connections: "Die Ergebnisse der Philosophie sind die Entdeckung irgend eines schlichten Unsinns und BeuIen, die sich der Verstand beim Anrennen an die Grenze der Sprache geholt hat"(PU, § 119). The particular nature of philosophical discovery thus explains why a successful result should also put an end to philosophical activity: "Die eigentliche Entdeckung ist die, die mich fahig macht, das Philosophieren abzubrechen, wann ich will" (PU, § 133). Once thinking is cured of its illness, disquietudes are over, peace of mind is achieved and no more philosophizing is needed. 70

* But a slogan like "there is nothing to discover" still sounds unbearably dogmatic and (to some ears) even obscurantist. A quite natural reaction would be to ask: even supposing that Wittgenstein is right about the source of philosophical puzzlenlent, couldn't the discovery of new facts and explanations put the relevant things in a new light, making thus possible a new philosophical interpretation of the matter? Wittgenstein himself admits that, when doing philosophy,one is interested in the way concepts are used, in linguistic rules (for "Dieses Verfangen in unsem Regeln ist, was wir verstehen, d.h. iibersehen wollen" - PU, § 125) and in grammar.

He admits that "Es ist eine Hauptquelle unseres Unverstandnisses, daB wir den Gebrauch unser Worter nich iibersehen"(PU, § 122). Now, why couldn't a new factual discovery or a new explanation of speech acts shed light upon linguistic use and conceptual conventions, helping us to get a better picture of them? Wittgenstein keeps insisting upon what could be called 'the logical indeterminacy' of ordinary language use, which is not governed by laws (BBB, p.27), is not based on definitions or strict rules (BBB, p.25), nor on sharp senlantic boundaries (BBB,p.19). But is it impossible to fmd (some day) a 'hidden logic' behind this apparent indeterminacy? A scientifically-minded philosopher, impressed by the results of Chomsky, Fodor or Rawls, is inclined to answer negatively; it is certainly conceivable that one would discover some day a sort of 'hidden logic' behind our apparently arbitrary way of using words. For instance, one could fmd out that the grammatical conventions which we usually follow in ordinary speech (and which might seem arbitrary, vague, incomplete, sometimes even 'illogical') do in fact largely coincide with the consequences following from a rather small set of general principles. Of course, since we are not aware of these general principles, we don't actually apply thenl; but this coincidence can hardly be purely accidental. Wouldn't it be in such a case reasonable to suppose that there really is a 'hidden logic' behind our usual conventions, a logic based on these principles which somehow (although unawaringly) guide us in adopting and applying the usual linguistic rules? And if such a discovery is in principle possible, wouldn't it put language use in a new light, supplying a better understanding of it and solving a lot of philosophical difficulties? Why couldn't then such a discovery be philosophically relevant? This argument - based on the premiss that "what seems to be chaotic can always prove to have a 'deeper', invisible order", and meant to show that a relevant discove-

71

ry is always possible - appears as a quite natural and convincing one. Could Witt-

genstein reject it? My claim is that he could. (A) Coincidence is not enough

Suppose we have actually found out that the semantical conventions we ordinarily apply do coincide, to a great extent, with the consequences that follow from a small set of general principles. We are inclined to suppose that this coincidence is due to the fact that the 'hidden logic' implied by these principles is playing an active (although unnoticed) role, guiding our speech. That is, we speak as we do because the general principles exert their influence upon us; were they different, we would have spoken in a different way. Now, that much is clear: the coincidence we have discovered is not enough at all. In exactly the same way in which the nlere accord between a certain rule and a certain behavior cannot prove that what we have to do with is a case of behavior following that rule, the mere accord between the general principles and the usual linguistic conventions cannot prove that in adopting and applying the conventions one is guided by these principles. In both cases what must be shown is that the coincidence involved is not purely accidental, that there really exists a relevant connection which explains it. This is quite sure, at least since Quine has distinguished between 'behavior that fits a rule' and 'behavior that is guided by a rule', even if what exactly must be done in order to show that might not be completely clear. If in order to prove that the recently discovered general principles actually play a guiding role, we have to show that in case these principles were different, our own grammatical conventions would have been different, then our task is a hopelessly difficult one. The situation we are in is analogous in some important respects to another one, explicitly discussed by Wittgenstein: "Suppose I have invented a medicine and say: Every man who takes this medicine for a few months will have his life extended by one month. If he hadn't taken it, he would have died a month earlier." 1

Since it is impossible to test such a claim, Wittgenstein suggests that we might better think of it as meaningless: "Wouldn't it be better to say: »It is meaningless to say this medicine prolongs life, if testing the claim is ruled out in this way«. In other words, we are dealing with a correct

Ludwig Wittgenstein, Cause and Effect: Intuitive Awareness, in Philosophia, vo1.6, nos.3-4, sept-dec 1976,p.423.

72

English sentence constructed in analogy of sentences which are in common use, but you are not clear about the fundamental difference in the use of these sentences." (Ibidem)

This argument seems to show a certain commitment to some sort of verificationist theory of meaning; but if so, it was not a primitive kind of verificationism. Wittgenstein admits that claims of this kind 'tell us something', but he insists that the game in which we use them is not the authentical one; the words such claims are composed of "do perhaps evoke an idea (the prolongation of life, etc.), but the game with the sentence is so arranged that it doesn't have the essential point which makes useful the game with similarly constructed sentences. (As the »race between the hare and the hedgehog« looks like a race, but isn't one)" (Ibidem)

Analogously, one might claim that, since there is such a striking coincidence between our usual linguistic rules and what the general principles imply, the supposition that the latter play a guiding role is quite reasonable; and in so saying, one suggests that one might, some day, discover that, guided by principles of a different kind, we would have spoken in a different way. But if such a 'discovery' is not, in fact, possible, does such a claim make any sense? Now, even if there might be further controversy about that, what has to be accepted is the fact that the sin1ple coincidence just discovered is not the kind of reiation which could settle the matter, and Wittgenstein was aware of that. He suggested that "even if there is something common to all games, it doesn't follow that this is what we mean by calling a particular game a 'game'" (G.E.Moore, Wittgenstein's Lectures in 1930-33, in his Philosophical Papers, George Allen & Unwin, 1959, p.313). This is not only because we don't actually use the 'common element' as a criterion for being a game and for correct use of the word 'game', but also because we can always decide to apply the word 'game' to something which doesn't share this 'common element'. We are free to do that because we decide about meaning (Bemerkungen fiber die Philosophie der Psychologie, I, §547; hereafter, BPP) and this shows that the coincidence between the presence of the 'common element' and the use of the word does not commit us to an essential connection between them. Analogously, even if we discovered a coincidence between the 'hidden pattern' and our ordinary rules, it's not at all necessary that this pattern be 'behind' our rules and that our rules are 'guided' by it. At the same time, Wittgenstein insisted that a coincidence between patterns is irrelevant for meaning (BPP, I, §372). Now, if coincidence is not enough, what other kind of connection could be involved?

73

(B) Matters ofJustification Maybe the biggest difficulty - and the strongest objection - raised by the idea of a possible 'philosophical' discovery is that a 'hidden logic' of the kind mentioned above could easily prove to be, in some essential respects, irrelevant. In order to see that, one has first of all to realize the justification problems which inevitably arise here. The main source of these problems is the fact that, by simply claiming that "there is a hidden logic which guides our use of words, or the way we adopt and apply linguistic rules", one inevitably faces a quite new task: the task of Justifying' linguistic use and the ordinary rules on which it is based, or of 'checking the correctness' of these familiar rules. The essential point here is that, once this claim is made, it is implicitly but necessarily denied the common belief that our usual grammatical rules constitute the final authority in matters of semantic correctness - this authority is now inevitably transferred to the 'hidden logic' or to the general principles which are supposed to play the 'guiding role'. It: when speaking correctly, we are not simply following the usual and familiar linguistic rules, but we are in fact also unawaringly guided by the 'hidden logic' from which these rules spring, then, obviously, the final authority on linguistic correctness must belong to this logic or to the set of general principles which constitute its basis. And the authority of our ordinary linguistic intuitions must be somehow derived from this 'deeper' source of'grammaticality'. In other words: if I want to claim that there is a 'hidden logic' which, by guiding me unawaringly, generates the grammatical correctness of my speech, then I have to accept the idea that my (or our) usual intuitions about correctness (expressing, in general, the familiar linguistic rules) are not decisive any more; they are meresymptoms of correctness, but not genuine criteria. The criteria of semantic correctness should now be found in the 'hidden logic' and the general principles it relies upon. Now, somebody could be tempted to think that there is no real problem here: since in choosing, adopting and applying our ordinary semantic conventions we are actually guided by the recently discovered general principles and their logical consequences, there need be no difference and thus no conflict between the well-known logic of our familiar rules and the 'hidden logic' implied by the general principles. Consequently, we need not bother about the 'correctness' of our ordinary linguistic intuitions, for they are bound to coincide with what these principles imply. Unfortunately, this doesn't seem to be the case; the inclination to suppose that what we have here to do with must be complete coincidence or pure identity is wrong-headed. At least two elements speak against it. (i) the idea of guidance, which plays an essential role in this interpretation, does not, by itself, exclude the possibility of any error. Guidance is not infallible; despite it, some mistakes are sometimes possible, for the mecha'711

nisms involved in the process of guiding cannot be supposed to function perfectly. For instance, despite the guiding role played by different mechanisms involved in our capacity to estimate spatial dimensions (size, distance), we still make sometimes mistakes. This is an obvious point, but the argument might go further. Not only that the idea of guidance does not imply infallibility, but, in a way, it even requires that the logical possibility of error be acknowledged. For, in a case in which no mistake is possible, one doesn't speak about 'one being guided' to do something in a certain way, but rather about 'one being necessarily determined' to: a step which is never wrong, which cannot be wrong, is not a 'guided' step; its correctness is not a matter of guidance, but of logical or physical necessity. If there was no possibility for us to adopt in some cases 'the wrong' linguistic rule, then the very idea of guidance would have been misplaced here. (ii) Quite obviously, one sometimes is subject to error not only in applying rules, but also in wrongly considering something as 'the appropriate rule' for a certain word. Thus, children 'adopt' sometimes a 'wrong' rule for some unfamiliar expression and keep applying it, until corrected, in a wrong way which can still be very consistent with the rule they have mistakenly adopted. The same mistake is made, more or less often, by adults trying to learn a new language. And there are, of course, although not often, disputes even between natives, on which is the 'right' rule for the use of a certain word. Maybe some linguistic conventions never become subject of dispute between competent speakers of language, a fact which could create the impression that there is some sort of infallibility here; but, nevertheless, some others do, and this shows that such an infallibility doesn't really exist. It seems, thus, pretty clear that the supposed 'guidance' provided by the 'hidden logic' cannot be a perfect one. Rather, the situation could be interpreted as follows: if, in adopting and applying our gramnlatical conventions, we are guided by a 'hidden logic' of the mentioned kind, then it is likely that most, or at least a large part, of these conventions are 'right' (their adoption being the successful result of this invisible guidance); but it is and it must be possible that some of the conventions we actually use be 'wrong', and these are the ones the adoption of which is explained by failures ofthe 'guiding' mechanism. Now, all this leads to the conclusion that, once we have accepted the idea of a 'hidden logic' guiding us unawaringly, we really face a new task: due to the always present possibility of a failure in the guiding mechanism, we now have to check the correctness of our usual grammatical conventions, by comparing them to what the general principles (which have become the final authority) imply. 75

The defenders of the idea of a possible discovery relevant to ordinary linguistic use have thus two different duties: on one hand, to find the set of general principles which is supposed to provide the 'hidden logic' of normal speech, proving that it really functions 'behind' the usual linguistic rules; on the other hand, to check all these usual rules, in order to prove that most of them are 'correct' (in accord with the 'hidden logic') and to detect the few which could be 'wrong'. The question now is: are these two tasks epistemologically compatible? I shall tl)' to suggest that there are good reasons for a negative answer.

(C) Matters ofrelevance A set of general semantic principles does constitute a relevant discovery for us only in case that its implications are largely identical with the set of our ordinary grammatical conventions - an identity which could suggest that it is precisely this set of principles that 'functions behind' our way of speaking, shaping it and being thus susceptible to explain it. For, were these implications largely in conflict with our usual linguistic rules, how could one claim that it is precisely it and the associated 'hidden logic' that makes us speak as we do ? (In such a case, it would of course be much more reasonable to keep looking for another set of principles, more compatible with the nonns of ordinary speech.) Thus, coincidence (to a great extent) is here a sine qua non condition for accepting a certain set of principles as the real foundation of the 'hidden logic' of our way of speaking. On the other hand, such a set of principles is relevant for our topic only in case that it manages to confirm, and thus to justify, most of our usual linguistic rules, proving thereby that their medium-level authority comes, in fact, from it, which is the 'highest' authority (or that their value as symptoms of semantic correctness comes from their agreement with the criteria of correctness implied by it). Otherwise, how could one claim that, in using the rules we actually use, we are being guided by this particular set of principles? A successful justification of most of the usual rules is a necessary condition for accepting the set of principles as the 'final' authority, the one which can check even the correctness of linguistic norms. The trouble we are in now is quite evident: the set of general principles supposed to create the 'hidden logic' of our speech is bound to be in agreement with most of our usual grammatical conventions (otherwise, it would not be accepted as such); but ifit is necessarily in agreement with most of them, how could it also check their 'correctness'? Once the accord between general principles and usual rules is presupposed (and it really is presupposed as soon as a set of principles is accepted as the 'guiding' factor), how could this accord also function as a proof that most of our fU76

les are Justified'? Once accepted as a pre-condition, this accord, or agreement, cannot be relevant any more in matters ofjustification. Why is that so? One cannot justify the usual grammatical conventions by appeal to the implications of a set of general principles supposed to 'guide' speech, for the simple reason that any such set of principles will trivially coincide with these conventions; and this, in tum, is simply because we are bound to use this coincidence as a necessary condition for the acceptance of any such set as the set we are looking for. To nlake this clear, let us take another case in which an analogous coincidence is irrelevant. Wittgenstein himself evokes such a case, when saying: "Die Konventionen der Grammatik lassen sich nicht durch eine Beschreibung des Dargestellten rechtfertigen. Jede solche Beschreibung setzt schon die Regeln der Grammatik voraus" (PB, §7).

In other words, one cannot justify conventions like "no red object is green" by invoking the reality (Le., the fact that no red object is in fact green) which such conventions might be considered as aiming to describe, because in order to identify the relevant reality (the red objects), one nlust already use these conventions in a way which sinlply excludes any possibility of conflict between them and the relevant facts: in case that some object is green, one already applies the convention and consequently refuses to accept it as a red one. In this way, any possibility of'discovering' a red object that is green is obviously ruled out; reality could never 'confirm' or 'disconfirm' our convention - it is simply bound to agree with it. But since the coincidence between grammatical convention and real fact is thus always (necessarily) present, its presence becomes irrelevant - i.e., it cannot be taken as a proof that the convention is justified or 'correct'. My claim is that the case of a set of principles supposed to guide normallinguistic behavior is, in essential respects, quite similar. Suppose that one accepts the idea of such a set of general principles creating a 'hidden logic' of speech, and that one starts looking for it. Any such set of principles must fulfil the condition that its implications coincide largely with the usual gramnlatical conventions; one cannot 'discover' such a set which did not fulfil this condition, exactly as one cannot discover a red object that is green. Were a set of principles to disagree with the usual conventions, we would simply refuse to call it 'the set of principles which guides speech', as we refuse to call a green object 'red'. But what all that boils down to is that any set of principles of the kind we want is bound to agree with (most of) the usual linguistic rules, exactly as red objects are bound to fulfil the condition of not being green: there can be no failure here, and

77

precisely for this reason we don't have in fact to do with real confirmation or justification. But if the set of general principles we are looking for cannot be used (even if discovered) in order to justify the usual rules, then how can one claim that it plays a guiding role and that it is the fmal authority on linguistic correctness? How could a guide fail to be the higher authority by appeal to which one checks the correctness of one's own judgments? And if it actually fails, can it be considered a guide any more? Somebody n1ight feel that there is something wrong in the whole discussion, because a confusion is being systematically made between two completely different things: a matter of fact ('does really exist a set of principles which plays a guiding role?') and a matter ofjustification ('can such a set justify usual rules?') This feeling is not - as far as I can see - misplaced: such a confusion, or at least a mixture between two very different matters, is really implied here. But this objection does not run against my argument or against Wittgenstein's view: it runs directly against the idea of a 'hidden logic' guiding ordinary speech, for it is in this idea that the confusion is implied. It is precisely through the idea of some 'objective', hidden factor (a kind of logic, of abstract pattern or a set of general principles) playing an essential guiding role that the confusion is introduced: for this idea inevitably implies that the authority of our usual linguistic rules is not simply based upon the 'subjective' agreement which exists among competent speakers of a natural language, but also upon some sort of 'objective' element which supposedly supports these rules and their application. And this, in tum, is because those who defend a possible discovery of such 'objective' patterns do not simply say: "look, there might be a sort of pattern based on general principles which fits pretty well the multitude of different grammatical conventions that we actually apply"; this, of course, would not be an absurd claim, but it would not be very relevant either: it could only fascinate an als ob-theorist, Le. a theorist who would find it interesting that, although in fact such an objective pattern based on general principles plays absolutely no role, we use linguistic rules as if we were following the pattern. But, of course, the defenders of such a discovery say more than that; they don't try to follow Vaihinger, nor they suggest that the coincidence between the 'objective' pattern to be discovered and the ordinary rules is purely forn1al and accidental. No, what they say in fact is: "although our usual rules seem arbitrary, chaotic etc., it could be the case that behind them a deeper order manifests itself and this order, this 'hidden logic' based on general principles, plays an active role; it is it that makes us speak as we do". But if the objective pattern nlakes us speak as we do, if it shapes or guides the way we speak, it should be possible to justify linguistic use by appeal to it; its justificational power is presupposed in its guiding role.

78

It seems that what really happens here is that, after acknowledging the existence of the ordinary linguistic rules, one supposes that it could be the case that they don't constitute the final authority on correct speech, because they are themselves based on something else - one thinks that they are not (in the relevant sense) purely arbitrary, but 'founded' on something else, more objective; as Wittgenstein himself remarked in Tile Blue Book, one tends to think that they are imposed upon us, not simply created by us, and, consequently, that there could be some sort of scientific research on them (discoveries of new data, explanations, and hidden entities, reconstructions of hidden patterns and mechanisms). But, of course, the supposed hidden, objective element is to be identified by means of its accord with our ordinary rules. This creates a situation the strangeness of which could be brought to light by using the following analogy: it is as if we had a big book containing the usual rules of behavior, but we supposed that the correctness of all that is said in this book must be checked by comparison to what is said in another, smaller book, wh~ch contains the general principles which guided the elaboration ofthe usual rules (but which has been lost); due to this guidance, it is to be expected that most ofthe rules in the big book are confmned by the principles in the snlall one, but, as we never know where a mistake can be, everything must be checked. Now, the fITst task would be to find the small book containing the principles which (we think) played the guiding role. The only clue we have is the supposed agreement between the two books. We thus have to rely upon what is said in the big book in order to identify the snlall one (and consequently the general principles), and, of course, to rely then upon what is said in the small book in order to check the rules in the big one. But, since the correspondence between what is said in these books is both our only means of identifying the small one and our only means of checking the rules in the big one, we are now in trouble. The validity of the rules in the big book cannot be confirmed or discontinned by the principles in the small one for the simple reason that the latter are bound to agree with the former; if they didn't, we would have simply refused to accept the small book as the one we were looking for. In other words, we would accept a small book as being authoritative only if it confirmed most ofthe rules in the big book; but the validity of these rules is itself in question, because it is only after being confmned by what is said in the snlaH book that the statements in the big one become authoritative. We thus identify our 'fmal' authority (the small book) by its endorsement of what is said in the big book, although what is said in the big book is to be accepted only after being validated by the small book. Something (the small book) is recognized as the highest authority on behavior only if it confirms most of what comes from a source (the big book) which in itself has no authority, because it only becomes authoritative after being validated by the highest authority.

79

This, of course, leads us nowhere. We simply discover nothing about the validity of our rules if the authority meant to confirm it simply must validate most of them in order to be recognized as such.

* But, of course, this is not the end of the matter. An important objection to the above argument is simply the following: if the 'hidden pattern' to be discovered is bound to agree with most of our usual grammatical rules, but not with all of them, then it still can be helpful because it can indicate which of our ordinary rules do constitute a special case (not being in agreement with it). The fact that these (admittedly few) rules constitute a case apart could, for instance, mean that they are not the 'right' ones or at least that they are not coherent with the maj ority of our grammatical conventions. Wouldn't such a finding prove that there really is something to discover? A more sophisticated formulation ofthis objection could run as follows 2: although the set of general principles is bound to agree with most of our usual rules (and thus to 'validate' them), any of the rules we actually use could happen to be among the few ones which are not 'validated' in this way; we do not know beforehand which of the usual conventions are OK and which not, and that means that the discovery of a 'hidden logic' which could separate the sound ones from the remaining unsound ones would be very helpful. Due to it, some sort of coherentist justification would be possible; for instance, we could start by provisionally accepting the ordinary linguistic rules as authoritative; we could find the set of general principles which agree with most ofthenl, and then we could go back and use the authority of these principles in order to find which of the initial, usual, rules are OK and which not. That is (if we take again the analogy of the two books), we rely provisionally on the big book in order to find a small one (containing principles that confirm most of what is said in the former), and then we can perfectly well rely upon the latter in order to check every claim in the big one. We cannot check simultaneously the totality of claims (rules) in the big book, but we can check piecemeal all of them, without exception, by using coherence criteria. And the principles would thus prove to be very useful; it seems to show that their discovery is not only possible, but also extremely important.

* 2

I am indebted to Prof Richard Jay Wallace for an elaborate fonnulation ofthis objection.

80

The above objection seems to present convincingly a sort of two-ways procedu.. re which would make the discovery of a 'hidden logic' philosophically relevant. Unfortunately, I think, it is very doubtful that such a procedure can really be applied here. The main difficulty is connected with the fact that this procedure is based on a hidden assumption: the assunlption that the ordinary linguistic rules are or should be characterized by some sort of overall coherence and systematical character, which could be grasped by means of the recently discovered 'hidden logic' and then used in the attempt to detect a number of conventions which are not 'the right ones', i.e are not coherent with most of thenl. If we have a closer look at the kind of coherentist justification implied in the above objection, we notice that most of the usual rules are 'validated' because they can be 'embedded' in the coherent system founded on the general principles discovered, whereas the remaining few are rejected because they cannot - the latter appear to be 'wrong' because they look like 'inadmissible' departures from the coherent, systematic pattern based on our general principles. The model one has in mind here is an axiomatic system (in our case, the general principles play the role of the axioms): the statements which are 'validated' are those which can be 'embedded' in the system, whereas the remaining ones are rejected. Only against such a systematic, axiomatic-like background, does the procedure described work, and only in such a context would the discovery of a 'hidden logic' be helpful. Now, where is the problem?, someone could wonder. The problem is not the hypothesis that the axiomatic-like coherence implied by the hidden logic agrees with the axiomatic-like coherence supposed to exist among our usual rules, but rather that these rules do not have an axiomatic-like coherence at all. Wittgenstein, of course, constantly rejects the widespread, scientificallyinspired, inclination to see language as a calculus or as a logically-systematic pattern. If the ordinary linguistic rules are arbitrary (in the sense that they cannot be deduced fron1, and justified by, a set of universal principles or 'axioms'), if there is no 'law' in the way a word is used, if natural language is like an old town etc., then, of course, it is hardly to be expected that such a systematic pattern could be found 'behind' the visible surface of our usual way of speaking. There is (it seems) no reason to suppose that the set of our ordinary linguistic rules has an axiomatic-like coherence (nor is there any reason to think that it should have). If one supposes that there must be such a coherence here, then what really happens in this case is that: one nlakes the hypothesis that there must be a 'hidden logic' behind our usual way of adopting and applying linguistic rules; 81

one constructs such a logic based on some general principles meant to fit the existing linguistic rules; in order to prove that the logic just constructed fits and 'validates' the usual rules, one adopts the two-ways procedure; but as this procedure applies only to a coherent set of rules, one makes the supposition that the usual linguistic rules have such an axiomatic-like coherence. This strategy is obviously flawed, because it amounts to making two hypotheses which support each other: a hypothesis about a 'hidden' entity (logic) which is supposed to fit the facts (Le., our linguistic rules); and a hypothesis about the facts (Le., that the rules are coherent, that they have a 'hidden' logic), without the help of which there can be no proof that the 'hidden' entity actually fits the facts. That is, one not only tries to make the theoretical construct ('hidden logic') fit the facts (usual rules), but also tries to make the facts (the rules) fit the theoretical model, by claiming that they have a sort of coherence which is in fact clearly present only in the model. A normal, non-circular strategy would, of course, be to show (first) that the usual linguistic rules have some sort of 'natural', pre-theoretical, coherence, and (second) to prove that the coherence implied by the theoretical model (by the 'hidden' logic) coincides with the 'natural' one. But it is very doubtful that this can be done, because, as Wittgenstein insisted, the usual linguistic rules do not have a coherence similar to that of a formal calculus or of an axiomatic system.

(D) Matters ofauthority I think that the way of seeing which Wittgenstein rejected, and which I am trying here to criticize in a similar vein, ignores or at least obscures a very important aspect of our topic, namely linguistic authority. As many authors have already stressed - for instance, Hare: "people's linguistic 'intuitions' are indeed, in the end, authoritative for what is correct in their Ianguage"3, and Nagel: "the intuitions of native speakers are decisive as regards grammar. Whatever native speakers agree on is English... "4 - the usual linguistic rules that we follow when speaking are the final authority on what is to be taken as correct speech, and this is bound to have important consequences for the problem of a possible discovery of a 'hidden logic'. Suppose we accept the strategy of looking for a 'hidden logic' which 'guides' ordinary speech, and start the search for a set of general principles which would 'fit' 3 4

82

R.M. Hare Rawls' Theory of Justice, in Nonnan Daniels (ed), Reading Rawls, Basil Blackwell, Oxford, 1975, p.86 Thomas Nagel Rawls on Justice,in Daniels (ed), op.cit.,p.2

the usual linguistic rules that we use. There seems to be no reason for claiming that we shall find one, and only one, such set: if we consider our linguistic rules, or our linguistic judgements based on these rules, as a set of facts to be accounted for, then, like in any other case in which one tries to account for a set of facts, there might be different, and even incompatible, constructs which can do the job. Different 'discoverers' could, for instance, fmd different sets of general principles which are equally successful (i.e., account for most of our usual rules and thus 'validate' the same number of ordinal)' linguistic norms) but which are incompatible with each other, because (as it might happen) single out different, but equally small, groups of rules which are not 'validated' and should thus be rejected. What should one do in such a case? How could one determine which is the 'right' set of general principles and, consequently, which group of usual rules is to be rejected as 'wrong'? The only reasonable thing to do seems to be asking the native speakers to choose - i.e., to adopt a certain set of principles and, accordingly, to reject the small group of ordinary rules that is banned by the principles they have just chosen. It might be the case that different big groups of speakers make different choices, and, if so, then it is hard to see what else could be done in order to settle the matter. But even in case that all the native speakers agree on the same set of general principles as being the 'right' one, there still remains an essential question, I think: their choice is inevitably based upon their usual linguistic intuitions, and if, in the end, it is the set of ordinary linguistic intuitions that settles the matter, how could one claim that these intuitions are not enough and that (in order to understand natural language) we should look for a ''hidden logic' that guides us unawaringly? To see that this is a real problem, and not just a piece of philosophical casuistry, let us in1agine that we have found one and only one set of general principles which account for most of our usual rules and that, therefore, there is no need to ask the native speakers to choose. But we have to ask them son1ething else: since our set of rules does not 'validate' all the usual linguistic conventions, we have to ask the speakers to abandon those (few) rules which are not in agreement with the 'hidden logic' just discovered and to modify their way of speaking accordingly. (We could, for instance, tell them that these particular rules are the result of errors which appeared in the functioning of the 'guiding' mechanism, that they are not coherent with all the other rules we use, and should thus be eliminated). Now, they might, of course, agree and gradually change their way of speaking accordingly. But what if they do not? What if they refuse to make any change, continue to speak as before and claim that their (traditional) way of speaking is the (only) correct one? They could, for instance, claim that the changes which are recommended would transform their native language into a different one, one which they have no desire to adopt. What are we to say in such a case? Should we say (it: for instance, we have to do with speakers of English), that all the native speakers of 83

English actually use a sort of broken English, and that it is only the kind of English which the 'discoverers' recommend that is true English? Should we claim that only the 'discoverers' know what 'real English' is, whereas the native speakers are wrong about the language they speak? These claims seem, of course, ridiculous. But the fact that they seem to be absurd points out an important aspect of our problem: the peculiarities of a natural language can belong to its characteristic identity, and if so, no coherence requirements can justify their elimination. A natural language is not obliged to be coherent and uniform, to be 'logical' in the sense in which a calculus or an axiomatic system is. And if the 'hidden logic' one thinks one has discovered does not agree with some of the normal linguistic rules, and, moreover, the native speakers stick to these rules arbitrarily (i.e., despite any considerations of logical coherence, despite what one could think about the 'logic of language'), one still cannot say that the 'real' language is that which is in accord with the 'hidden logic', whereas the language spoken by the natives is incorrect. In such a case, whatever the native speakers choose, even arbitrarily, to speak is their real, genuine, native language; what the 'discoverers' claim to be 'the real, correct, language' is just a very coherent abstract invention. In short: one cannot discover what real English is, nor how real English should look like; real English can only be described. As Wittgenstein insisted, one cannot guess how words are used; one has to look and learn (PU, §340). This is, in a way, trivial, but also, in some ways, relevant to our topic. It shows that, in case one discovered a 'hidden logic' of some natural language, this 'hidden logic' would have to imply all the characteristic peculiarities of that language, even those which seem (from a logical or rational point of view) quite arbitrary, as long as the native speakers stick to them. Thus, such a logic would have to be quite illogical; and it is very unlikely that a logic, illogical enough as to fit the peculiarities of a natural language, could be some day discovered. And if the discovery of a 'hidden logic' is not to be expected, what else is to be discovered?

84

4. 'THERE IS NOTHING TO RECONSTRUCT'

"was etwa verborgen ist, interessiert uns nicht"

(Philosophische Untersuchungen)

Wittgenstein's whole attitude towards philosophy is unusual, and there are several amazing things about his metaphilosophical views, but perhaps the most surprising one is his insistence that the role of philosophy should be confined to describing what 'lies open to view' and to 'assembling reminders'. Many people would probably accept, in the end, the idea that philosophy should not try to explain and discover things, perhaps mainly because these tasks can be considered as typically scientific; but the claim that philosophers should only describe and 'assemble reminders' seems indeed wild, and, in any case, unbearably restrictive. Most classical attempts made in this field went nluch further, offering quite articulate intellectual constructions which we even today consider as representative and valuable; why, then, would philosophy be not allowed to advance (if not explanations, at least) theoretical constructions which provide a deeper understanding of 'the way things are'? Why couldn't a philosopher give 'rational reconstructions' of the things he is interested in, instead of giving 'descriptions' and of 'assembling reminders'? Shortly, why should philosophy be exclusively descriptive, and never constructive? These are (for us) legitimate questions, and if Wittgenstein has no convincing answers to them, a collapse of his whole metaphilosophical conception is to be expected. But, as I shall now try to show, there are strong reasons for clainling (as Wittgenstein implicitly did) that in philosophy there is actually nothing to reconstruct. Reconstruction aims, of course, at revealing a 'hidden, deeper, order' of things; but if truly 'nothing is hidden', then reconstruction becomes pointless. Now, the big problem is to prove that nothing is hidden; are there really strong arguments to that effect?

(A)

The best starting point for a discussion on this topic seems to be Wittgenstein's remark that philosophy is nothing else than a particular, sufficiently welldetermined, set of 'problems' (a quite reasonable claim, susceptible to be agreed upon by many people): "Denn die Philosophie, das sind die philosophischen Probleme, d.h. die bestimmten individuellenBeunruhigungen, die wir »philosophische Probleme« nennen"l. The philosophical 'problems', or, as Wittgenstein not accidentally insists, disquietudes have, as a rule, the form of 'What is X'-questions: "What is Time" (the Augustinean interrogation)2, "What is Knowledge?" (the Socratic interrogation)3, "What is Language" (PU, § 92), "What is a proposition?" (Ibidem), "What is thinking?" (PU, § 327) etc. One does not go so far as saying that there cannot be other kinds of philosophical problems, but one does suggest that questions like the ones mentioned above are very characteristic - the Augustinean interrogation, for instance, is presented as typically philosophical, in contrast to the questions asked by natural science (PU, § 89). Although an attempt to present philosophical questions as being, essentially, 'What is X'-questions might seem restrictive, such an attempt has the advantage of being in perfect agreement with the notorious philosophical interest for 'essences'. And, as many of his remarks show4, Wittgenstein closely associated traditional philosophy with a search for essences. Since it can be supposed that the other philosophical problems are derived from these questions about essence, the idea that philosophy is basically a set of 'What is XI-questions does not seem to be wholly inadequate. (B)

But if philosophers ask questions like "What is Time?", then surely their aim must be either (i) to elucidate a concept (time'), or (ii) to 'grasp the essence' of something (of time). (Under a certain reading of PU, § 116,120, the alternative disappears, because thetwo are simply identical; but we do not have to bother about that).

1 2 3 4

86

Philosophische Gramnlatik, hereafter PG, I, § 141. Philosophische Untersuchungen, hereafter PU, § 89. The Blue and Brown Books, hereafter BBB, Basil Blackwell, Oxford, 1969, p.20. PU, § 92, 116 but see also, for an explicit claim, Philosophische Bernerkungen, hereafter PB, Basil Blackwell, Oxford, 1964, IV, § 54.

The really important thing is that both (i) and (ii) imply, in a Wittgensteinean framework of thought, a requirement which appears as fundamental for any kind of philosophical work: the requirement of elucidating the use, or the grammar, of words. (i) obviously implies the elucidation of meaning, and, since meaning is use, also the elucidation of use; a philosopher preoccupied by "What is Time?" should thus try to concentrate upon the way 'time' is actually used, upon the grammar of this word (for, as PU, § 496 indicates, grammar too describes the use of words). But (ij) implies the same requirement too, not only because "Das Wesen ist in der Grammatik ausgesprochen" (PU, § 371), but also because "»Wesentlich« ist oie die Eigenschaft des Gegenstandes, sondem das Merkmal des Begriffes" 5, and this shows that "Wer tiber das Wesen spricht, konstatiert bloB eine Obereinkunft" (BGM, I, § 74). Philosophy, which is always interested in essence, should thus in any case concentrate upon linguistic use and gramnlar. But how should a philosopher approach this topic? Should he try to explain use? No, because, according to Wittgenstein, 'there is nothing to explain' (see above, the chapter with the same title). Should he try to justify grammatical conventions, by proving their adequacy to reality? No, because language is autonomous (PG, IV, § 55) and its conventions simply cannot be justified in this way: "Die Konventionen der Grammatik lassen sich nicht durch eine Beschreibung des Dargestellten rechtfertigen. Jede solche Beschreibung setzt schon die Regeln der Gramnlatik voraus" (PB, § 7). It seems that the only thing philosophers can do is to describe, and thus to claritY, the use of concepts and the grammatical conventions associated with it.

(C) Unfortunately, this is precisely what philosophers never do. They never manifest real interest in the actual way in which concepts are used; they always neglect the examination of real cases in which one normally applies concepts like 'time', and they completely ignore the task of (merely) describing such cases. Their typical attitude (Wittgenstein believes) is well illustrated by Socrates, who, although interested in the essence of knowledge, never pays attention to the real cases of knowledge, i.e. to the cases in which the term 'knowledge' is correctly applied: "When Socrates asks the question 'What is knowledge?' he does not even regard it as a preliminary answer to enumerate cases of knowledge" (BBB, p.20). Socrates tries to deduce or to 'discover' the essence of knowledge, but not to de5

Bemerkungen tiber die Grundlagen der Mathematik, hereafter BGM, I, § 73.

87

scribe what we actually call 'knowledge' or the real cases in which we use the word. And this is what philosophers generally do: they try to 'grasp the essence' without ever trying to describe the use of the relevant word or the grammatical conventions which actually express the essence. But why have philosophers been so reluctant to describe, while being so eager to deduce or to discover (essences)? The answer is: "Denn sie sieht in dem Wesen nicht etwas, was schon offen zutage liegt und was durch Ordnen libersichtlich wird. Sondem etwas, was unter der Oberflache liegt" (PU, § 92). It is because philosophers have always been committed to the idea of 'hidden essence', to the idea that the essence must be somewhere behind of what we can see, that they have always considered describing as irrelevant; they felt it to be their proper task not to describe, but to discover and to explain. What could be described always seemed insufficient to them: "Das bloBe Beschreibung ist so schwer, wei] man glaubt, zum Verstandnis der Tatsachen diese erganzen zu mUssen. Es ist, als sahe man eine Leinwand mit verstreuten Farbflecken, und sagte: so wie sie da sind, sine sie unverstandlich; sinnvoll werden sie erst, wenn man sie sich zu einer Gestalt erganzt. Wahrend ich sagen will: Hier ist das Ganze. (Wenn du es erganzt, verfalschst du es.)"6 Thus, philosophers have always been don1inated by the feeling that they still needed some extra facts, which would illuminate the general structure of things, and consequently indicate what was really essential. But this meant that they should not waste time in describing what is already known, but rather concentrate upon fmding new details and discovering general patterns. Seeking the essence and describing the facts (concerning the use of words, the contexts of use etc.) have always been considered by philosophers as incompatible enterprises.

(D) Instead of describing, philosophers have constantly tried to construct concepts. As they considered ordinary concepts uninteresting and the task of describing them trivial, they thought it better to create new concepts, meant to lead us 'faliher' or 'deeper' than any ordinary concept can do. The newly created concepts were supposed to grasp 'the real (hidden) essences' of things, and therefore classical philosophers thought highly of them. But Wittgenstein does not share their conviction; he thinks that such new concepts are arrived at through a process of

6

88

Bemerkungen tiber die Philosophie der Psychologie, hereafter BPP, Hrsg. von G.E.M. Anscombe und G.H. von Wright, Basil Blackwell, Oxford, Band I, § 257.

purely philosophical sublimation or idealization which involves all kinds of misunderstandings: "Mancher wird sagen, daB mein Reden tiber den Begriff des Wissens irrelevant sei, da zwar dieser Begrirt: wie die Philosophen ihn auffassen, allerdings nicht mit clem der alltaglichen Rede tibereinstimmt, aber eben ein wichtiger, interessanter Begriff sei, der durch eine Art Sublimierung aus dem landUiufigen und nicht sehr interessanten gebildet ist. Aber der philosophische Begriff ist aus dem landlaufigen durch allerlei MiBverstandnisse gewonnen worden und er befestigt diese MiBverstandnisse. Er ist durchaus nieht interessant; es sei denn als Warnung" (BPP, Band II, § 289). Did that happen only in the case of the concept of knowledge? Of course not. Many other philosophical concepts are created by the same (mistaken) strategy. The Platonic Ideas are the result of the same fallacious strategy: "fch will also sagen: Der 'reine' Farbbegriff, den man sich aus unsem gewohnlichen Farbbegriffen machen mochte, ist eine Chimare. (...) Start 'Chimare' hatte ich sagen konnen 'falsche Idealisierung'. Falsche Idealisierungen sind vielleicht die platonischen Ideen"? But the biggest mistake involved in 'constructivist philosophizing' is not that new concepts are invented; rather, it is that these newly constructed concepts are meant to explain or even to replace the old, usual, ones. Rudolf Carnap, for instance, explicitly defended a constructivism of this kind (which he preferred to call 'rational reconstruction'), pleading for 'the replacement of a pre-scientific, inexact concept' by (what he considered to be) 'an exact concept,g. Other philosophers did not emphasize the 'scientific' character of 'exact' concepts, but they nevertheless insisted to advance 'new' concepts which were more precise, or more 'logical' or better 'determined' (rationally) than the ordinary ones. But (one could wonder) what is wrong with that? Wittgenstein suggests that the mistake involved in this old philosophical strategy is that, after rejecting the 'old' concepts naturally adapted to our usual language games, the 'constructivists' try to impose their new concepts (their 'idealizations') in these games and contexts which they don't really belong to: "Wenn es so etwas gibt, dann muB, wer falsch idealisiert, Unsinn reden, - weil er eine Redeweise, die in einem Sprachspiel gilt, in einem andern, wo sie nicht hingehort, verwendet" (LSPP, Band II, p.48).

7 8

Letzte Schriften tiber die Philosophie der Psychologie, hereafter LSPP, Basil Blackwell, Oxford and Cambridge, 1992, Band II, p.48. See his answer to Strawson, in P.A. Schilpp, ed., The Philosophy of Rudolf Camap, La Sa)]e, Illinois, Open Court, 1963, pp.933-940.

89

In other words, one creates a concept (like 'pure colour', or 'Beauty' - understood as 'pure' or 'perfect' beauty, or defined by some 'essential' feature supposed to characterize any beautiful object) by idealization, a concept which, therefore, is not in agreement with ordinary use, and then one tries to introduce it in the usual language games, where it does not belong and consequently does not make any sense. The new concept (of beauty, or knowledge, etc.) actually belongs only to the philosophical jargon invented by that particular philosopher (and quite often accepted exclusively by him and his followers); it is only there that it functions adequately. But why, then, does one suppose that it could also be functional in other contexts, in ordinary language, in previous philosophical conceptions etc.? The answer is that the philosopher sees a 'law' (i.e., a perfect rule) in the way a word is normally used (BBS, p.2?), and he self-indulgently supposes that his new concept illustrates successfully this 'law'; that is, he instinctively claims that by his own linguistic invention he has managed to capture the law governing the use of the corresponding ordinary concept: his concept of beauty captures the (up to that moment ignored, or at least partially ignored) regularity characterizing the use of the ordinary concept of beauty, and so on. But since, in general, such laws do not exist, Le. the use of ordinary concepts is not characterized by such a 'geometrical' regularity as the one postulated by a philosopher, his new concept does not function in the 'nomlal' language games - its 'transplantation' in them leads to nonsense. (E)

But even if we accept the idea that philosophical constructivism has (often, or even always) failed, there still remains a question which, in a way, is the main question: is the failure of constructivist philosophy simply the result of mistakes accidentally made in the past by philosophers (of misunderstandings, or false idealizations), or is it unavoidable that constructivism would fail? In other words: is it just that classical constructivism happened to fail, or is it the case that any constructivist philosophy is bound to fail? Couldn't, in the future, (better equipped) constructivist philosophers build concepts which really grasped 'the essence' of things? This is an extremely important question; it forces one to choose between a position that can easily become trivial and another one, which is original, interesting and provocative. It is not so hard to admit that the conceptual constructions proposed by philosophers in the past were mere figments of imagination, which failed to account for the 'reality' they were meant to grasp; but if this is all Wittgenstein has to say about constructivisnl, his merit is not tremendously great, because many other thinkers criticized classical philosophy, and especially metaphysics, in a similar way. In such a case, Wittgenstein should only be recognized as belonging to a large group

90

of modem authors who denounced it. Moreover, if we accept this alternative, we may find it hard to account for the very general and categorical manner in which he dismisses philosophy: why does he refer to philosophy in general, and not to specific past contributions to it? On the other hand, if we choose the other alternative, which implies that constructivist philosophy not only happened to fail (in the past), but it is somehow bound to err, Wittgenstein's position becomes much more radical but also much more interesting, although, of course, we still have to discover what grounds could there be for such skepticism towards intellectual constructions. I shall argue that this last alternative is the right one, and that there are solid grounds (inside a Wittgensteinean framework of thought) for being very skeptical towards conceptual construction. (F)

Let us first remember that, according to Wittgenstein, (a) philosophers seek 'the essences of things'; (b) essence is expressed by grammar; and, (c) grammar captures the use of words. A constructivist philosopher proposes new concepts which are supposed to 'grasp' the essences, and this implies that his conceptual constructions should encapsulate grammatical conventions - according to (b) - and should correctly illustrate the use of words - according to (c). That is, our future (better equipped) constructivist must grasp correctly (by his new concepts) the characteristic grammatical conventions and patterns of use, in order to be entitled to claim that his newly constructed devices have successfully captured 'the essences'. For instance, a philosopher who proposes a 'concept of goodness' or 'of justice' must be able to show that his new concept grasps the characteristic granimar (and pattern of use) of the word 'good' (or justice') in a way which is acceptable to, and recognizable by (not only philosophers, but also) any competent speaker; if he is not, we should conclude that his 'concept of goodness' (justice') fails to account for what competent speakers call 'good' (just'), i.e. that our constructivist philosopher does not actually talk about goodness and justice, but about something else that he, and only he, wants to call 'good' or 'just'. The supposition involved here is that a philosopher simply cannot seriously say to the competent speakers of a natural language things like the following: "I am now going to tell you what should be called 'good', and no matter whether you like it or not, this is what you should call 'good'!"; i.e., if the philosopher aims at elucidating goodness, and not simply at developing his own, personal, idea of what goodness should be, then what he must do is to propose a concept which grasps the TIonnal grammar of 'good' in a way that is recognizable by, and acceptable fOf, any competent speaker. In the light of the huge amount of speculative thin91

king that has been produced by philosophers, this supposition is not unproblematic; but it is unavoidable for anyone who understands philosophy in the way which is typical to the main Western tradition - after all, when asking 'What is Goodness?', Socrates and Plato wanted to elucidate goodness (i.e., that particular quality that people used to refer to by this word) and not just to express their own, personal, feelings about what goodness should be. Now, the question is how can our philosopher grasp grammar and patterns of use by his new conceptual devices? Being a constructivist, he cannot do that by simply describing linguistic use; he is committed to constructing, not to describing, and, as a rule, he loathes describing as trivial and unphilosophical. The thing to do, then, for such a philosopher, is to deduce or to guess the patterns of use and the granlmatical conventions which 'capture the essence'; the concepts he constructs must somehow manage to coincide with the basic conceptual conventions and patterns, an achievement that must be arrived at by intellectual insight, not by description. It is at this point of the discussion that a very important (and often neglected) argument of Wittgenstein should be inserted - Le., the (quite simple and straightforward) argument that the basic conventions of grammar simply cannot be 'guessed', 'deduced' or 'predicted'. The way in which a word 'functions' is not the kind of thing that could be anticipated: "Wie ein Wort funktioniert, kann man nicht erraten. Man muB seine Anwendung ansehen and daraus lemen" (PU, § 340). The limit that Wittgenstein talks about here is not due to an epistemological insufficiency, and it is not purely contingent; the reason for which one cannot guess or deduce patterns of use (or the corresponding conventions) is deeper: these patterns simply are not the kind of well-ordered structures which could be 'grasped' by an insight into 'how things are'. They do not have the 'perfect' geometry which makes guessing or prediction or deduction possible; this semantic geonletry, which might be present in ideal or perfect languages, but which cannot exist in natural languages, is the main obstacle - and, obviously, it is a non-epistemological obstacle. The reason for which we are bound to fail in any attempt of deducing the patterns of use is that these patterns are of an empirical nature, i.e. they represent conlplex concrete structures to be found in linguistic experience, not abstract, geometrical, ones to be deduced or guessed. In order to reconstitute the way a word is used, one has to have access to empirical facts, because tt daB ein Wort das und das bedeutet, so und so gebraucht wird, ist wieder eine Erfahrungstatsache wie die, daB jener Gegenstand ein Buch ist tt9 •

9

92

Ober Gewissheit, hereafter 00, § 519.

There is, of course, the obvious objection that sometimes one manages to guess an empirical fact, but then there is also the obvious response that the use of a word is not just a simple, individual, fact, but rather a complex (or complicated) factual pattern that nobody can guess or deduce. You can guess the colour of a certain fruit, but not the whole pattern of the paths and roads crossing a certain forrest, and patterns of linguistic use are something similar to the latter, not to the former. Wittgenstein compares language to a labyrinth (PU, § 203) or to an old town (PU, § 18), and both comparisons suggest the 'empirical' nature of the linguistic institution. Trying to deduce or guess the patterns of use would be as hopeless a task as trying to deduce or guess the shapes of the streets in an old town, or the shapes of the paths in a (natural) labyrinth. Even if the streets had a geometrically standard structure, it would be absurd to try to guess it, instead of simply looking at it and describing it. But an old town has generally no well-ordered, deducible or predictable, structure and the shape of its streets cannot be anticipated or 'grasped' by pure intellectual insight. Ifwe keep in mind the fact that linguistic use is inherently imprevisible, Le. cannot be guessed or deduced a priori, then we are able to see what the 'predicament' of the constructivist philosopher consists in: as a philosopher, the constructivist is bound to seek 'the essences', which are 'marks' of our concepts, and he is thus compelled to grasp the basic grammatical conventions associated with the patterns of use; and, of course, these patterns are not to be deduced, but looked at and described; on the other hand, as a constructivist, the philosopher is bound to construct concepts meant to grasp 'the essences', and not to look at, or to describe, empirical facts. The constructivist philosopher faces thus two contradictory requirements, and, unable to find a way out, he usually comes to (simply) invent new concepts, which are supposed to grasp the patterns of use (and thereby 'the essences') but which actually fail to do that: as mere abstract inventions, they belong more to an ideal language than to the real one, and they never nlanage to account for, and to coincide with, the actual patterns of use. Therefore, what constructivist philosophers in fact do is just to "devise an ideal use, which turns out to be worthless" (LSPP, I, § 830). (G)

But, at this point, a scientifically-minded reader would be, tempted to raise the following question: why couldn't a constructivist philosopher do what is (in Chomsky's words) "just standard scientific practice", i.e. start with a description of the relevant facts (connected with linguistic use) and then propose a coherent construction, or a 'rational reconstruction', by means of new concepts (and perhaps explanations), accounting thus for what has been initially described? In other words, why couldn't a philosopher secure the adequacy of his conceptual 93

constructions to (linguistic) experience? Why would all philosophical reconstruction be bound to go wrong? The answer to this objection is twofold. First, it is even doubtful that patterns of use can be satisfactorily described; Wittgenstein came pretty close to saying that they cannot. First of all, since linguistic use is not determined strictly by categorical rules, necessary and sufficient conditions or exact definitions, the patterns of use have no 'perfect geometry'; they are not characterized by the simmetry and order one can find in a set of abstract, ideal shapes - rather, they are similar to the irregular forms one becomes familiar with through experience. But such forms simply cannot be properly described; they can only be 'presented': "Eine Form kannnicht beschrieben sondern nur dargestellt werden" (PB, XV, § 171). And it is precisely because the patterns of use cannot be described that, when trying to clarify the nleaning of a word, we have to evoke grammar rules: the complete foml of a particular pattern cannot be described, but the way that word should be used can be to some extent clarified by presenting the relevant rules. Grasping the main rules or conventions governing the use of a word is not the same thing with having an overview of the whole 'map' representing its pattern of use, but since we usually cannot provide the map, we teach someone by evoking the rules. A second argument is the following: in order to describe patterns of use, one would have to describe language games, and consequently the whole logic of language, because "zur Logik gehort alles, was ein Sprachspiel beschreibt" (00, § 56). But is the logic of language describable? Wittgenstein came to have doubts on that: "Komme ich nicht mehr und mehr dahin zu sagen, daB die Logik sich am SchluB nieht besehrieben lasse? Ou muBt die Praxis der Sprache ansehen, dann siehst du sie" (00, § 501). As to the reasons for which the logic of language remains undescribable, several arguments could be invoked: the games we play with words are never completely circumscribed by rules (PU, § 68), they are somehow 'unregulated' and it is hardly possible to encapsulate in a description the logic of an 'unregulated' game; moreover, the logic of language games is always context-dependent, so that a satisfactol)' description of it would imply a satisfactory description of all the relevant contexts, which is clearly an impossible task (a natural language is so strongly connected with a form of life, that a description of its logic would only be possible when combined with a description of the entire corresponding form of life); finally, since understanding a language means mastering a technique (PU, § 199), describing the logic of language implies describing a technique - and a technique which is not purely algorithmic, which is in fact a know-how, can never be satisfactorily described because

94

it consists in a set of practices one can only learn by exercising, not by studying a description of 'what one should do'. Now, these arguments may sound unconvincing to some people, especially to those who feel that one is never in the position to decide about what can and what cannot be described or explained, simply because future theoretical achievements can never be foreseen or confined within predetermined limits. The kind of epistemic precaution that prompts this objection is not unreasonable, but it is very doubtful, I think, that the 'one never knows'- objection is a powerful one. For the main question, as far as I see, is not whether future will bring us pleasant surprises (proving, for instance, that the logic of language can be described), but rather the following: what makes one think (and hope) that such an achievement is possible? Part of the answer could be that one is inclined to see semantic performance as analogous to some syntactic ones, and thus to expect an illuminating discovery concerning some algorithmic mechanisms which would explain meaning, understanding etc. But, if Wittgenstein was right in saying that language games are not everywhere circumscribed by rules, that they are based upon agreement in judgements and participation in forms of life, then their 'unregulated' (or, at most, partially regulated) character is a strong reason for rejecting this expectation as unrealistic. Some algorithmic elements (a certain logical nlechanism which is describable) might very well be involved in semantic performance, but, taken as a whole, the language games which make meaning and understanding possible are not algorithmic and their 'logic' (based on agreement in judgements and involvrnent in common forms of life) is much too complex to be captured in a theoretical description. The feeling that it should be possible to capture it comes, probably, from a commitment to the logicist, oversimplified, view according to which a limited set of strict, clearcut and therefore describable, rails-like rules determines entirely the way we speak. Unfortunately, such 'semantic rails' do not actually exist, and this is one of the few things connected with the present topic on which there is a large agreement. But the really important, the decisive (in a certain sense), part of Wittgenstein's answer to the above objection is still to come; this reply is based upon the 'arbitrary' character of linguistic patterns. In order that a philosopher be able to provide 'rational reconstructions' of linguistic use, there must be some non-arbitrary patterns of use (to be reconstructed), for it is only such patterns that can be effectively 'captured' by theoretical models. 'Arbitrary' patterns, characterizing evolutions which are in essential ways unpredictable, cannot be grasped and represented by a philosophical theory (and if science can give abstract reconstructions of factual patterns, this is precisely because these patterns are not arbitrary and the evolution of the events involved can be, at least to some reasonable extent, predicted).

95

Now, the main question is: are patterns of linguistic use non-arbitrary and predictable? Wittgenstein's answer is negative. "Du muBt bedenken, daB das Sprachspiel sOZllsagen etwas Unvorhersehbares ist. Ich meine: Es ist nicht begrUndet. Nicht vernlinftig (oder unvemilnftig). Es steht da - wie unser Leben" (DO, § 559). In order to grasp the full meaning of the philosophical view expressed by this aphorism, we have to remember that the language games we usually play have not been 'constructed' consciously and deliberately, answering the requirements of Reason and Logic. They are seldom, if ever, based on rational reasons concerning the best way 'in which language games should be played'. There are some rules saying how words should be used, but there are none saying how patterns of use should be introduced and developed. The pattern of use, for a word like 'game' or 'nunlber' (to take Wittgenstein's own examples), develops and becomes more and more complicated simply in virtue of our extending of a name from some particular cases to some other, new, ones, exactly as 'in spinning' ("wie wir beim Spinnen eines Fadens Faser an Faser drehen" - PU, § 67). This process of extending the use of a name is not governed by strict rules, i.e. we are not strictly compelled, by our own previous way of applying a word, to extend its use in a certain particular manner. But aren't there some reasons for extending the use of a word in a certain way, rather than in another? Of course there are such reasons, but they are never compelling ones (i.e., reasons which abolish our freedom to choose between alternative manners of developing a pattern of use). Although we generally have some reasons for extending the use of a name in a certain way (not in another), and these reasons influence our choice, they never determine it completely: despite their existence, we are still free, in any new particular case, to adopt an alternative or not - our right to choose remains unaffected: "Obrigens behalte ich mir vor, in jedem neuen Fall zu entscheiden, ob ich etwas zu den Spielen rechnen will oder nicht" (PG, VI, § 73). This freedom of choice comes from the fact that the use of a word like 'game' is not governed by necessary and sufficient logical conditions which would compel us to apply the nanle in any new case in which these conditions were fulfilled, while prohibiting its application when they were not. There are thus no categorical (strict) and complete 'logical' or 'rational' requirements, implied by some sort of coherence involved in the language-game (as we have played it before), which would compel us to apply words in certain pre-determined ways; the coherence of our language games is always partial, 'uncircumscribed' by strict rules and open ended: it always leaves open several alternatives, because the rules which generate it are not 'rails'. That is why it is up to the speakers to decide how the use of words should be extended in new situations:

96

"Ou entscheidest, ob nlm der und der Fall in diese Familie aufgenommen werden solI, oder nicht" (BPP,Band I,§ 547). It is the speakers who make decisions about how a pattern of use should be developed, because, as Wittgenstein insists, "a word hasn't got meaning given to it, as it were, by a power independent to us" (BBB, p.28) - we give meaning to words and we change their meanings (patterns of use) without being constrained by some 'objective' power. The pattern of use, for every particular word, is simply the result of a series of free decisions (which may involve the existence of some reasons, including 'coherence of use'-reasons, but which are not pre-determined by them). Therefore, the present ways of using words are, in an important respect, contingent - as any non-necessary result of 'subjective', free choices. From this point of view, it can be said that the patterns of use are, in an important sense, arbitrary. 10 (H)

In order to make cle~r the sense in which the use of words is arbitrary, let us first remember that "Unsere Sprache kann man ansehen als ein alte Stadt" (PU, § 18). The structure of an old city is not the result of previous planning, but rather the contingent result of many 'subjective' decisions of building a new house, a church etc. The citizens have surely had their particular reasons for erecting their houses in some places, rather than in others; their decisions have never been totally random. But they were not compelled to build their houses in the places in which they did, because there had been no general plan to conform to; there were no 'objective' constraints to detennine the way they should choose locations, no overall 'logic' to guide them in their choices (even if some loose coherence requirements could exist, they were not compelling since they did not have the character of strict regulations). That is why the resulting structure has no overall, 'objective', logic. And this, in tum, explains why there can be no 'theoretical reconstruction' of such an old town: there is no 'hidden' logic to be grasped, there is no predictable element in the way streets develop and change their shapes, etc. One can, of course, draw a map of such a city, but this implies looking at it and seeing what its real structure is like; one cannot guess or deduce its structure, one has to see it on the ground. Therefore, 'elucidating the structure' inlplies, in such a case, an effort of describing, not of guessing, deducing or predicting. The case of (natural) language is quite similar. One cannot deduce the pattern of use for a certain (ordinary) word, as one cannot deduce the shape (the meanders) of 10

Wittgenstein insists, in several places, that the rules of grammar are arbitrary, but since what I am mainly.interested in here are patterns of use, I shall leave aside his arguments about the arbitrariness of grammar.

97

an old street; and one cannot guess how the use of a name will be extended in a new case, as one cannot guess where a new house will be built and how its building will affect the shape of the street (in an old town where no regulations exist). The evolution of a pattern of linguistic use is not pre-determined by the existing way of using words, exactly as the shape that a street is going to take is not pre-determined by planning (or by compelling requirements generated by the positions of the already existing houses). In both cases, there is no non-arbitrary pattern to be followed and observed (by the speakers or by the builders), no 'logic' to be maintained; for any new case, the freedom of choice still exists, and the result of several new choices is arbitrary in the sense that it is not pre-determined or pre-contained in an overall, previsible, logic. A second clarification of the arbitrariness of patterns of linguistic use is prompted by another analogy: the foundational role played in Wittgenstein's thinking by the analogy between linguistic use and games makes it reasonable to expect that language use is arbitrary in the same sense in which games are arbitrary. Now, in what sense is a game arbitrary? A game has a logic of its own, a logic that follows from· its basic conventions and rules. As governed by rules, and as having 'a point' (a fact that Wittgenstein does not omit), the game is not arbitrary, but 'logical'. Its logic is obvious for anyone who understands its rules and its aims (or its 'point'), and it is one which can be made visible to newcomers by mere description. But, at the same time, a game 'is not everywhere circumscribed by rules' (PU, § 68) - for instance, there are no rules for how high one throws the ball in tennis. Some aspects of the game are regulated, but others are not, so that the different 'moves' nlade by the players and their final results are not predictable. Here we have an ineliminable element of 'arbitrariness', and it could thus be said that a game is both arbitrary and non-arbitrary. Analogously, there is a sort of 'logic' in the way a word is used, because there are some rules for applying it; in this sense, the patterns of use are not arbitrary. But the linguistic rules involved do not cover and do not regulate all the elements and all the aspects of use: there are no clear-cut boundaries (PU, § 68), etc. That is, although there are some rules involved in the patterns of use, there are no necessary and sufficient conditions to determine strictly the application of words. Linguistic use is also both arbitrary and non-arbitrary, exactly as any 'grammatical' system which "ist mit WillkUrlichen verwandt, und mit NichtwillkUrlichen"ll. But the analogy between the arbitrariness of games and the arbitrariness of patterns of linguistic use goes deeper than that. The really interesting aspects of this analogy become visible when one concentrates upon the justification of rules.

11

98

Zettel, hereafter Z, § 358.

It is immensely important that the logic of a game does not confoml to a preexisting pattern, which required that its rules were what they actually are; the rules do not follow 'necessarily' from some 'objective' constraints - they are not 'determined' or 'dictated' by 'hidden' principles, 'deeper' structures or previous choices. The rules of a game are simply the result of some free options previously made by people; they are nothing else than the rules which happened to be accepted once by players. In this sense, the rules and the 'logic' of the game are, once more, arbitrary; it is a simple matter of fact, and a matter of contingencies, that these rules, and not other ones, have been accepted. Nothing compelled people to accept them, nothing made it necessary that these rules be accepted; it was simply a series of free, 'arbitrary', options that led to their adoption. Accordingly, the resulting logic of the game is itself 'arbitrary' - Le., it does not follow necessarily from another, 'deeper', logic, it does not reflect a system of 'hidden' constraints and it is not shaped by inevitable restrictions. (This is, naturally, proved by the fact that the players can always change the rules according to their most recent wishes, it: of course, agreement can be reached as to what these wishes are). A main consequence of this kind of arbitrariness is the fact that nobody could have deduced or guessed the 'logic' of a game, or the rules which generate it; nobody could have predicted that this set of rules was to be accepted for this particular game, and this is because nobody could have anticipated the whole series of free options by which the rules have been adopted. At the same time, if the apparent logic of the game is not determined by another, 'deeper' or 'hidden' logic which we are not aware about, there is no question of 'reconstructing' the visible logic or the rules of the game. There is nothing to reconstruct, simply because these rules are arbitrarily chosen (in the sense explained above) and the only thing one can do is to describe them. Now, if Wittgenstein's analogy is correct, then much the same considerations apply to language games. From one point of view, the patterns of linguistic use are not arbitrary: neither in the sense that they have been randomly adopted, nor in the sense that they can be changed by anybody any time. People had their reasons to speak in some ways, and not in others, and it is not easy at all to change the linguistic rules: one does not simply change these rules as one pleases, when one pleases. But the connections between patterns of use and the practical interests which influenced their adoption is a matter of natural history, completely irrelevant philosophically, as Wittgenstein remarks 12 • The important point, for us, is that it would be wrong to clainl that people adopted the concepts they adopted because they had to, these concepts being the 'correct' ones (Ibidem); obviously, this is not the case, and alternative patterns of use 12

PU, Tei! II, xii, op.cit., p.230.

99

could be equally correct. The present patterns have been accepted freely, and were not 'determined' by inevitable constraints; in this sense, they are arbitrary, and the resulting logic is arbitrary too. The logic of linguistic use is the visible one, generated by the familiar rules who are arbitrary in exactly the same sense in which the rules of a game are arbitrary: nothing compelled us to accept these rules, and the logic generated by them does not follow necessarily from a 'deeper', hidden one. From this, second, point of view, the way in which we use words is inherently arbitrary, and therefore 'there is nothing to reconstruct'. The authors who advocate 'rational reconstruction' in the philosophy of language seem to be committed to the idea that 'there must be something behind' linguistic use: some principles which 'guide' or 'shape' the way we speak, a 'deeper' logic fronl which the apparent one follows, or more general patterns that the well-known patterns of use must conform to. The task of philosophy, they believe, is to discover and 'reconstruct' these 'hidden', determinant, elements that, in general, we are not aware about. But Wittgenstein's main premiss is that 'nothing is hidden': there simply are no such 'deeper' determinant elements, exactly as there are no such elements in a game. As arbitrary results of many contingencies, then, the patterns of use admit of no 'rational reconstruction'. It should be remarked that emphasizing this particular kind of arbitrariness does not amount to saying that words are simply used in a random, chaotic manner. We choose freely the ways in which words should be applied and their use should be extended, but of course that our choices are influenced by our interests - concepts" sind der Ausdruck unseres Interesses", as acknowledged in PU, § 570 - and by forms of life. But what this actually shows is only that concepts are instruments (PU, § 569) and, as it is always the case with instruments, their adequacy is important for us; we have always specific reasons to use them in some particular ways. But the existence of such reasons does not prove the existence of a 'hidden' logic which compels us to adopt and use concepts in the way we do, exactly as the existence of some reasons for using hammers in the way we do does not prove the existence of a 'hidden' logic governing the production and the use of hammers. Here, again, there is nothing 'behind', to be captured in a theoretical 'reconstruction'.

* There is a forgotten remark, in Bemerkungen fiber die Pbilosophie der Psychologie (the first volume), which could also be very helpful in clarifying the sense in which patterns of use are arbitrary: "Wir verleihen Worter, wie wir, bereits vorhandene, Titel verleihen" (§ 116). My reading of this remark is along the following lines: we do not confer titles randomly; we always have some reasons for conferring a certain title - but there is

]00

no algorithm to govern this social activity of conferring titles, Le. there are no necessary and sufficient conditions which determine exactly who will be given a certain title. Therefore, although we know perfectly well to whom a certain title has already been conferred, and (in general) for what reasons, we are never able to foresee with certainty who else is going to be given that title later on: for the decision to confer a title (although influenced by reasons) is not mechanically determined (by reasons concerning the coherence with previous similar decisions or by strict conditions), it is a free one, and, consequently, largely imprevisible and hardly 'deducible'. The 'pattern of use' of a certain title is, in this respect, largely arbitrary: there is no hidden, deeper, logic shaping this pattern in a strict, previsible, manner. Analogously, despite the fact that there are reasons for applying a word in a certain way, and not in others, in some cases, but not in others, there is no strict algorithm governing use; therefore, although we know quite well in what cases a word has been previously applied, we' can never tell (with absolute certainty) in which new cases it will be again applied and in which ones it will be not. The fonn that its pattern of use is going to take remains largely imprevisible; the pattern is open ended, susceptible of unexpected changes (due to new free decisions), and , therefore, in essential respects, 'indeterminate'. It is in this sense that patterns of use can be called 'arbitrary', and this feature is important. This arbitrariness also explains why there cannot be "a kind of scientific investigation into what a word really means" (BBB, p.28). Claiming that there could be such an investigation implies supposing that the meanings of words are not fully contained in the rules and the patterns of use we are all familiar with, that there must be something 'behind' them which science should discover and 'reconstruct', helping us thereby to get a 'deeper' understanding of meaning. Wittgenstein's answer to this supposition is that, since meaning is use and nothing nl0re than use, no attempt to understand meaning can go further than describing the patterns of use; we simply have to be content with noticing that "So werden diese Worte gebraucht" (PU, § 180), which actually amounts to acknowledging that "So spielen wir eben das Spiel" (PU, § 71). Somebody could suspect that this reply proves Wittgenstein's commitment to some sort of linguistic empiricism, according to which any investigation of language and meaning should be programnlatically confined within the narrow limits of empirical description. But what really makes hinl say that we have to be content with a description of language games is his presupposition that. there can be no 'hidden' logic behind these games; for if there was such a logic, then there would have been no arbitrariness of the kind we recognize in language. What makes people expect to find such a 'deeper', invisible, logic in our language games is the old conviction that language use is based on definitions, strict rules or even a 'law' concerning the application of words, a conviction that Wittgenstein constantly criticizes (see, for instance, BBB, pp. 67-70) - precise defmitions, strict rules or laws,

101

clear-cut boundaries etc. seem to be symptoms indicating the presence of a hidden logic. But once one realizes that this conviction is wrong, and that such symptoms are not, in fact, characteristic for natural language, one can see that there are no reasons for supposing that 'there must be a deeper logic behind our language games'; and once one abandons the idea of a 'hidden' logic, one can easily accept that 'there is nothing to reconstruct'. (I)

Now, somebody who is committed to philosophical constructivism would probably want to ask the following question: supposing that one accepts the idea that 'there is nothing to reconstruct', Le. that a reconstruction of our present concepts and patterns of use is pointless, why couldn't a philosopher construct new concepts, or even a whole new language which, from some points of view, might be superior to the ordinary ones? Even if the reconstruction of old concepts should be rejected for the reasons evoked above, there still remains the possibility of advancing new conceptual devices, the usefulness of which can arguably be proved either right now (in our forms of life) or later on (when new forms of life would appear); why would such an enterprise be a priori rejected? Let us answer this objection step by step. The fIrst step consists in pointing out the obvious fact that, if such a new set of concepts would be constructed, it would never account for our (present) way of using 'time', 'knowledge', 'goodness', 'beauty' etc. A new set of concepts would thus be irrelevant to the traditional philosophical questions, of the 'What is Time'-kind, because it would have nothing to do with our way of using words like 'time' and, consequently, nor with what we call tinle. It would only have something to do with what other (possible) speakers, in another (possible) world, inside another form of life, would call 'time'. By constructing such a set, one can devise an ideal use of words, but this is not the kind of linguistic use which would be philosophically illuminating: for what we are interested in simply is not what 'time' or 'goodness' ought or should mean (in an ideal case), but what they actually mean (in our case, for us, inside our own form of life). The philosophical puzzlements do not come from the absence of an ideal linguistic use, but rather from the (presence and the present difficulties of the) ordinary usage; it is consequently usual concepts that we have to clarify, not some ideal ones. You can always, of course, construct new concepts, exactly as you can always give (philosophical) definitions; the problem is that the new concepts, like the definitions proposed by philosophers, do not coincide with the actual use of words which is our real topic of interest. As Wittgenstein says, "If on the other hand you wish to give a defmition of wishing, i.e. to draw a sharp boundary, then you are free to draw it as you like; and this boundary will ne102

ver entirely coincide with the actual usage, as this usage has no sharp boundary" (BBB, p.19). The key words in this fragment are (I think) "as you like": you can construct concepts (as you can propose defmitions), but your new concepts will be built 'as you like' (exactly as your new defmitions), i.e. will have a 'personal' or 'private' character; your new concept of wishing (like the new definition) will not coincide with our (usual) concept of wishing, and, by using it, you will not speak about wishing (Le., about what we normally call 'wishing') but about something else (which is called 'wishing' ~, but perhaps not by others too). Newly constructed concepts prove thus to be philosophically irrelevant, even if a certain group of philosophers pays much attention to it. Now, for the second step of the answer: it is vel)' doubtful that a philosopher could construct a whole new language, which is an immensely (infinitely, indeed) complicated thing. What he could do, instead, is to propose some new concepts, or a sub-language, meant to deal with some particular topics which are of special (philosophical) interest. This is, in a way, similar to what one can do in an old city: when traffic becomes much too difficult, since streets are so narrow and 'illogically' interconnected, one can demolish some houses and create new (also better, from a logical point of view) streets. But - and this is the important philosophical point that one can expect to be made here by Wittgenstein - such a rationalization is only useful for some special purposes, like the one of improving traffic conditions; it is not useful from other points of view, and it does not help at all someone who was interested in the old town and its pattern of streets. Analogously, a rationalization of our concepts (by eliminating old, and constructing new, ones) can be very useful for some special purposes - as Wittgenstein insists, one can always 'improve' conceptual boundaries, but only for special purposes: "wir konnen - fur einen besondern Zweck - eine Grenze ziehen" (PU, § 69). But such a conceptual rationalization would tell nothing about how our usual concepts were actually used and about what they n1eant; consequently, it would do nothing to improve our understanding of 'time', 'wishing', 'goodness', nor to grasp the essence of the corresponding things. When analysing our (old) concept of time, in order to answer the question 'What is Time?', we are interested in the concept because we think that, by understanding what we n1ean by 'time', we could also understand what time actually is; exactly as, by analysing the structure of an old town, we can learn a lot about the form of life characteristic to the people who used to live there. But an analysis of some urban rationalization (of the shape of some newly opened streets) does not tell us anything about the old form of life; it only tells us something about the present wishes of its inhabitants (to have broader roads, direct access from some point to another, etc.). Analogously, a new (deliberately constructed) concept of time does not answer the 103

question 'what is time?' (which implies another one: 'what people mean by »time<
104

4.1 Naturalism and reconstructivism against descriptivism or Chomsky against Wittgenstein Wittgenstein says in The Blue Book that there can be no scientific investigation in the meaning of a word. This implies that no scientific reconstruction of language use is possible. Here is a very strong statement, that runs directly not only against the widespread conviction that anything can be a topic for scientific investigation, but also against contemporary naturalism. Theoretical reconstruction is intensely encouraged by naturalisnl. As pupils of the 20 th century naturalism, we are prone to think in the following way: language is not a supernatural, miraculous phenomenon; it is rather a normal kind of rational behavior, a 'natural' conduct, for humans. Therefore, there must be 'natural' laws and regularities governing it; why couldn't such laws be discovered by scientists? The speech capacity is not a divine gift, miraculously offered to us by God (as many learned people believed some centuries ago); it is something we do possess as natural beings. So there must be a 'natural' explanation for it... there must be causes determining it, which such an explanation could uncover. From this point of view, nothing appears more reasonable than trying to give a theoretical reconstruction of language use, in which the 'laws' governing it would be displayed. Such an attempt would be nothing more than normal scientific practice for naturalists. One of the advocates of both naturalism and reconstructivism is Noam Chomsky. Chomsky claims that his approach "considers language and similar phenonlena to be elements of the natural world" and takes linguistic elements as "things in the world, alongside of complex molecules, electrical fields" and so on. 13 And since linguistic phenomena are simply natural ones, we should expect them Hto be studied by ordinary methods of empirical inquiry" (Ibidem). In the best tradition of what Chomsky calls "Galilean style" of inquiry, linguistics should try to account for human speech in the same manner in which natural sciences account for physical, chemical or biological phenomena, that is in a naturalistic manner. More specifically, "a naturalistic approach to linguistic and mental aspects of the world seeks to construct intelligible explanatory theories, taking as 'real' what we are led to posit in this quest" (Ibidem). Chomsky claims that his own investigations illustrate precisely this kind of approach: he is not just the author of a Naturalistic Manifesto, but apparently also the Exemplary Hero of the Naturalistic Crusade. Now, what does his theory actually consist in? 13

Noam Chomsky Language and Nature, in Mind, vol. 104, January 1995, p.l.

105

Chomsky starts by stressing the obvious fact that every natural language is characterized by a particular grammar (i.e., a system of rules pairing sound and meaning). Under the stimulating influence of linguistic experience inside a certain conlmunity of speakers, a man (be it a child or a foreigner) gradually reaches the point where he/she masters this grammar. But Chomsky thinks that language learning cannot be explained solely by accumulated experience inside a conlmunity; he points out two main arguments against such an explanation: first, the actual amount of linguistic experience accumulated by a speaker during a certain period of time is always quite limited, and certainly extremely poor, as compared to the huge complexity of the gramnlar he/she manages to master at the end of that period; secondly, although such a speaker will make a lot of mistakes at different stages of language learning, there are nevertheless some possible mistakes he/she never nlakes. This could be taken as indicating that, experience apart, there must be also something else which helps one to learn a language. Chomsky's guess is that there must be some sort of innate universal grammar (hereafter UG) which makes fast language learning possible. According to him, the UG is a part of our genotype, i.e. a genetically determined system (of general rules or principles) or a schematism which "is specified, sharpened, articulated and refined under the conditions set by experience, to yield the particular grammars that are represented in the steady states attained" .14 If anybody manages to master quite quickly the grammar of a language, this should be explained by the influence of (by the unconscious guiding provided by) this 'hidden' UG. Now, what a scientist should do (according to Chomsky) is to build up theories about this VG. His own theory tries to provide a model or a reconstruction of the UG, and thus to account for language competence. It is a naturalistic theory, because it presents the UG as a natural structure (similar, in this respect, to the structure of benzene); it is also similar to theories in natural science because it aims at discovery (of the UG, a 'hidden' entity), explanation (speech capacity is explained by the unconscious guiding done by UG) and reconstruction (of the general rules in the UG). So, exactly as in natural science, if the truth of Chomsky's theory (for the time being only a hypothesis, as he himself admits) was proved, then we would have to admit the reality of the UG. Now, if such a theory about language can be developed, isn't that a proof that Wittgenstein was wrong, i.e. that there is something to discover, something to explain and especially something to reconstruct? 14

106

Noam Chomsky Rules and Representations, Columbia University Press, New York, 1980, p. 234.

My answer is negative, but before presenting my main arguments, let me make an important specification. Chomsky is reputed as a great specialist in syntax, and if he actually refers to syntactic aspects (when speaking about explaining and reconstructing language), then it is not much to argue about. In his later philosophy, Wittgenstein has never been interested in syntax, and his main claims (which are of interest here) do not refer to syntactic aspects. So, once again, if Chomsky is only talking about syntax, I have no comment. The trouble is that in fact Chomsky often talks indiscriminately about language in general, and his statements tend to be general, even philosophical, assertions about language. It is thus pretty clear that he has rlluch bigger ambitions than that of clarifying syntactic aspects, and he explicitly tries to combat Wittgenstein, which shows that his claims should be taken as referring to semantics too. In that case, I think that a confrontation with Wittgenstein's views is not only possible, but almost inevitable because what we have here is, after all, an open conflict between the two radically opposed conceptions about language, the pro-scientific (reconstructivist) one and the anti-scientific (antireconstructivist) one. Let us proceed then, keeping in mind that all the following Wittgenstein-type arguments should be taken as referring exclusively to semantic (and not to syntactic) aspects of language use.

* Chomsky claims that his UG (universal grammar) is meant to account for the familiar grammar (0) of a natural language, in the same way in which theoretical constructs in natural sciences account for empirical regularities in the natural world. But, when saying that 'UG accounts for 0', what does 'accounts' mean? It is undeniable that there can always be several different UO which 'account' for the same G, or for the same set of well-formed English sentences. (Quine, for instance, has evoked the case of two different groups of foreigners who could learn to speak English correctly by using two different sets of general principles: the speakers formulate exactly the same set of well-formed sentences, although they 'generate' them differently, because they are 'guided' by different systems of general principles). This is quite obvious. But then it is easily imaginable that different scientists could postulate different universal grammars which all account (equally well) for the familiar G of some natural language. Which one, if any, of all these universal grammars, is the 'real' one, the one active 'behind the curtain', which actually guides the speakers? We face here a problem called 'the problem of the psychological reality' of an VO: it is not enough to show that a certain model (of UG) is descriptively adequate (describes correctly, or can generate, the set of well-formed English sentences), because there obviously can 107

be several, distinct, such models; what is needed here is the proof that some model has 'psychological reality', i.e. 'is actually inside there' and guides actually the linguistic behavior. That is, coincidence is not enough: even if one shows that the set of sentences generated by a certain model coincides with the set of correct English sentences, this does not prove yet the 'psychological reality' of the model, as there can be many other models with the same capacity. It still remains to be shown that the rules belonging to the supposed ua are real, i.e. are 'actually there' and also 'causally active'. Objections of this kind, or similar to it, have been raised by many authors, Quine, Donald Davidson and John Searle among them. But, if the model devised by scientists is descriptively adequate, what are these skeptics actually asking for? Here is Chomsky's complaint: "I cannot see that anything is involved in attributing causal efficacy to rules beyond the claim that these rules are constituent elements of the states postulated in an explanatory theory of behavior and enter into our best account of this behavior. /.. ./ Clearly, we cannot obtain more evidence than all the evidence, or find better theories than the best theory. Nor is there any hope of identifying a kind of magical evidence that plays some unique role in determining that the rules attributed to Jones and invoked to explain his behavior, in the best theory of all the evidence, in fact guide Jones's behavior".15

In order to show that his hypothesis is not treated fairly, Chomsky compares the requirements inlposed by his critics to similar ones which could have been th imposed upon theories in natural science: would have been reasonable to ask 19 century chemists to prove that the models they had provided for the structure of benzene are not only "descriptively adequate" (i.e., in agreement with the available evidence) but also "actually out there", "behind the curtain" and "causally active"? If not, then one should admit that what Chomsky does is (as he claims) "just standard scientific practice", a practice we are familiar with from natural science: exactly as the best chemical theory attributes benzene a certain structure and nobody doubts its 'reality', our best linguistic theory attributes Jones a state characterized by the UG, and the ua should be accepted as no less 'real' than the structure of benzene. 16 If the UG has the same epistemological status like the structure of benzene, then Chomsky's naturalism seems vindicated. But there are strong doubts about this sameness of status. If the relationship between the 'deep' structure of benzene and the relevant empirical evidence (concerning benzene) is pretty clear (at least, this is what chemists tell us and we have no reasons for doubt), the relationship between the 'deep' UG and the 15 16

108

Noam Chomsky Knowledge of Language, Praeger, 1986, (hereafter KL), p.253. KL, p. 250.

relevant empirical evidence (linguistic behavior or the familiar G characterizing it) is very obscure. Chomsky keeps talking about 'guidance': the rules in the UO 'guide' linguistic behavior. 17 But this creates a big problem, because 'guiding' seems to suggest conscious use of some rules: 'we are guided by some rules' implies that we know them and use them accordingly. "Behavior fits the rule whenever it conforms to it; whenever the rule truly describes the behavior. But the behavior is not guided by the rule unless the behaver knows the rule and can state it". 18

A similar objection conles from the field of Wittgensteinean Orthodoxy. Authors like P.M.S. Hacker, Gordon Baker, Steven Davies insist upon the distinction between behavior "in accord with a rule" and behavior based upon ''following a rule". Normal linguistic behavior may well be in accord with the rules of UG; but this tells us nothing more than that VG correctly describes this behavior (and this behavior fits the rules in UG). But, in order to prove that ordinary speakers follow the rules in UG, one has to show that they know these rules and intend to obey them, because: "For a creature to follow a rule he must act in a way that accords with a rule and intend his action to accord with the rule".19 Chomsky, of course, cannot claim that ordinary speakers follow the 'deep' rules in VG (which they don't even know) in the sense evoked above. Consequently, he has suggested an idea of guidance based upon 'tacit knowledge', claiming that speakers follow the rules in VG because they know them tacitly or 'cognize' them. But, as Quine, Searle, Hacker and others have promptly remarked, this is hardly illuminating; in fact, it is a clear case of explaining the obscure by something even more obscure, for what does 'cognizing' mean and how are we to confirm its presence? The mystery deepens when we realize that the relationship between the 'hidden' rules in UG and the visible linguistic behavior (or its characteristic G) is not a causal one. That much Chomsky himself admits, for although he says that "Behavior is guided by the rules and principles of a system of knowledge", he stresses that "Our behavior is not 'caused' by our knowledge, or by the rules and 17 18

19

KL, p. 253, 260 and so on. W.V.O. Quine Methodological Reflections on Current Linguistic Theory, in D. Davidson, G. Harman (eds.), The Semantics of Natural Language, D. Reidel, 1972, p. 442. P.M.S. Hacker Language, Rules and Pseudo-rules, in Language and Communication, vol. 8, No.2, -1988, p. 163.

109

principles that constitute it".20 Thus guidance should not be understood as a causal relation. This, I claim, is very bad news for Chomsky's naturalism: it undermines his analogy between natural science and linguistics and it points out a big epistemological difference between the two fields. If the best chemical theory (in excellent accord with the empirical evidence) postulates a certain structure of benzene, then this can be the end of the matter, because this structure, seen as something real, causally explains the 'behavior' of benzene, Le. its properties (confimled by the relevant empirical evidence). Whereas in Chomsky's case, we face the unpleasant situation in which although the best theory postulates a 'deep' UG, this UG, even accepted as 'something real', does not explain causally the relevant empirical evidence (concerning linguistic behavior or its characteristic a). This raises justified doubts about the causal efficacy and indeed about the very reality of an UG, as many authors have claimed. But Chomsky was wise to accept that the relation between VG and G was not a causal one. Otherwise, he would have been in big trouble when asked about the normative status of G. The elements of G are the ordinary rules we know and follow in natural language. But a rule is nothing else than a convention made by people: it is not detemlined by some 'hidden causes'; it is 'arbitrary', both in the sense that it has not been dictated by some 'super-human' law or by some 'inevitable' cause and in the sense that the rule could have been different (take, for instance, the rule about driving on the right side of the road: it was not determined by 'hidden causes', it was chosen freely in Continental Europe and the choice could have been different, as it actually was in Britain). Thus, there can be no proper rules in G which are imposed upon us by a hidden UG. The connection between a particular G and the UG, or between rules and 'super-rules' (universal rules, general principles) simply cannot be a causal one: a rule can be derived by us from another, more general, rule, but it cannot be determined causally by the latter; the causal relation applies exclusively to factual events, never to normative items. If the ua caused us to speak in some specific ways, then those ways of speaking would have been natural, unchangeble, regularities of conduct, not the result of following rules that we can change at will. So, either UG causally determines a, but then the elements of G are not rules, but effects (patterns of linguistic conduct, similar to other biological patterns, and thus materialized in empirical regularities), or G are in fact rules, but then they cannot be detemlined by a hidden cause like the ua. Chomsky cannot have it both ways. We have thus reached what I consider to be the crux of the matter. Chomsky claims that the ua guides linguistic behavior and its characteristic G. I think that 20

110

KL, p. 260.

the vital question is the following: is the 'guiding' done by an UO compelling or not? That is, if an UG guides us to adopt and use a certain linguistic rule or way of speaking, does that mean that we are compelled to do that, or that we are still free to accept or not the UG-guidance in this respect? If Chomsky chooses to say that the UG-guiding is compelling, then we are much in the same case as if he said that UG causes us to speak in certain ways. Weare not free to speak in the way we want; we are deterministic speaking machines, compelled' to use certain pre-determined patterns of speech. We do not make our own linguistic conventions; we cannot change them; patterns of speech are imposed upon us by the UO. Natural language does not seem to have a nornlative, conventional, character any more; it looks more like breathing or walking, in that its regularities (patterns of speech) are natural, exactly like many patterns of physical' conduct. This could be excellent news for naturalism, but it is so implausible that very few people (I think) would support such an interpretation. After all, we all have the feeling that the ordinary speakers, not an invisible UO, are the final authority upon language: they can adopt or change ways of speaking, linguistic rules etc., without being 'compelled' by a 'deep, hidden' VO. But presumably Chomsky would not say that the guidance provided by UG is compelling. He is prepared to accept that, despite guidance from an VO, children can make mistakes and use 'incorrect' rules (for instance, they overgeneralize the Past Tense rule for regular verbs, saying 'sleeped' instead of 'slept') and that it is the authority of adults which compels them to correct such mistakes. 21 In the same context, he even goes that far as saying that "If all adults were to die from some sudden disease, the language would change and these irregularities would be erased" (the irregularities in the Past Tense form of irregular verbs). Now, such statements suggest that Chomsky accepts that the final authority on linguistic matters belongs to the community of native speakers, who are free to accept the regularities reconlmended by VO (such as the regular Past Tense form) or to reject them. And this, I submit, is the ruin of Chomsky's naturalism, for it can be turned against his analogy between natural science and linguistics. In natural science, if the theoretical model is correct (e.g., the structure of benzene has actually been discovered), there is no way in which the empirical evidence (concerning benzene) can differ from what the model implies: the structure we have just discovered explains the properties of benzene, and benzene can have no property which is different from those it should manifest (given the structure it has). In this sense, one can safely say that the theoretical model accounts for the relevant facts or properties of benzene.

21

KL, p.227.

III

On the contrary, in linguistics, as Chomsky himself seems to admit, it is easily conceivable that although the UG has actually been discovered, some rules are considered 'right' and some corresponding ways of speaking are considered correct, although they go against what is implied by the VG. This happens because native speakers of some natural language can simply refuse to accept as correct what a certain UG seems to imply, and because they have the final authority on what is to be considered correct. In this sense, the theoretical model (UG) does not account for (all) the relevant facts (linguistic rules), and some kinds of linguistic behavior can be correct no matter what the va requires. There is thus an important disanalogy between the way in which the 'hidden' structure (that we keep searching for in natural science) 'accounts' for the relevant natural facts and the way in which a 'hidden logic' like theUG is supposed to 'account' for linguistic behavior. The relation between the 'hidden structure' (the chemical structure of benzene) and the facts (concerning the observable properties of benzene) is such, that no contradiction between them is rationally acceptable. If such a contradiction comes up, then either the hypothesis about the 'hidden' structure is rejected as inadequate (incorrect or incomplete), or the facts are rejected as irrelevant (inaccurate, incomplete, wrongly interpreted, etc.). In natural science, permanent accomodation between the hypothesis and the facts is sought, and there is no reasonable way in which a hypothesis can coexist with facts that run against what it implies. And precisely because such contradictions are not accepted, when the hypothesis is confirmed, one is entitled to infer not only its truth, but also the 'reality' of the entities or the structures postulated by it; i.e., in Chomsky's terms, there is no gap between the truthfulness of a theory and the reality of what the theory claims to exist. But, contrary to what Chomsky maintains, things are different in the field of linguistic theory. In this field, it is perfectly possible that an VG be discovered which implied that one should say 'sleeped' instead of 'slept', or that one should not use the same expression ('bank') for both a financial institution and the borders of a river - and, nevertheless, the speakers of English keep using 'slept' and 'bank' as they have always did. The kind of usage banned by the (recently discovered) ua cannot be simply dismissed as incorrect, because (as already mentioned) the final authority on English language belongs to the ··English speakers, not to the scientists. (One generally says that 'facts are stubborn', but it appears that in the linguistic sphere they are even more stubborn than in other areas: the fact that a cel1ain linguistic use is taken by speakers as correct cannot be ignored, no matter what the cleverest theory claims). On the other hand, the (recently discovered) va should not be dismissed either, for there could be overwhelming empirical evidence in its suport. Thus, what in natural science seems to be out of the question appears as possible in the 112

linguistic field: a 'good' theory coexists with facts that contradict it. One has thus to agree that the analogy between the two fields is very problematic. But Chomsky could still raise some objections. On one hand, he could accept the existence of some accidental irregularities (which are not to be explained by his UG), but still claim that the main part of language is based upon regularities which can and should be explained by evoking the guidance of an UG. Moreover, we have discussed the case in which the VG failed to 'account' for all the irregularities of natural language; but what about the particularly fortunate case in which a newly discovered ua managed to 'account' for all the irregularities recorded in a certain natural language up to a certain moment of time? Let us then suppose that a very powerful UG has just been discovered, the implications of which coincide exactly with the linguistic bahavior, as it has been observed until that moment, including the normal speech irregularities (the syntactic ones, like the form 'slept', but also the semantic ones, like those manifested in the way we apply words to things, we group different things under the same name etc.). Is this absurd? Perhaps not, although it is of course extremely implausible. To suggest that all the irregularities in language use are not purely accidental and impossible to infer, but they belong to a certain 'hidden logic' and can thus be predicted by a strong UG, is to adopt a sort of providentialism which is very similar to the one advocated by old philosophers and theologians who claimed that no wrong act, sin or mistake was accidental, that every particular event that 'should not have happened' was in fact bound to have happened, having a special place and role to play inside the wise world drama written by God. But let us ignore this. The important thing is that it still seems in principle possible that an ua be discovered, which would perfectly well or reasonably well fit the facts of present Iinguistic use. Does that proves Chomsky right? I do not think so. The analogy that he proposes between natural science and linguistics entitles us to expect very much from an UG; i.e., to expect not mere agreement with the facts, but also accurate predictions. Not only that the UG should explain exactly the present way in which we speak (and would thus fit the available facts); like its counterparts in natural science, it should also predict and provide (up to that moment unrecorded) facts about changes in language, about the ways in which language use is going to be extended, modified, developed etc. After all, if an UG guides our present forms of speech, then it would also guide its future forms, and predictions about that should be possible for any expert in the UG. Is this a plausible thing to do? In fact, predictions in this field are so unlikely that even Chomsky stepped back: he candidly admitted that speech and language use cannot be predicted. But this is a very strong reason for rejecting his analogy between linguistics and natural science; the latter are so heavily dependent upon predicting, that its 113

absence in the former ruins the analogy. 'Standard scientific practice' (that Chomsky is so fond of) includes predicting. There are, of course, natural sciences (or branches of them) which provide only descriptions, and never predictions, but these have no aspirations at discovering 'hidden structures' with explanatory power; whereas Chomsky's DO is precisely the kind of theoretical entity the existence of which can only be proved by its success in explaining and predicting phenomena (like all the other similar entities in physics or biology). The fact that the UG is admittedly unusable for predictions is another big blow to Chomsky; anyone who takes his analogy seriously can tell him: no predictions, no 'standard scientific practice' (of the kind you claim to illustrate by your theoretical construction of an VO). The so-called 'psychological reality' of the UO cannot be accepted in the absence of some successful predictions, exactly as the 'physical reality' of electrons would not have been accepted if no such predictions were available. While Chomsky is at pains to explain why his approach cannot provide predictions, although it is 'standard scientific practice', Wittgenstein could easily explain the unpredictability of language use and of its changes. Chomsky takes language as a natural phenomenon; on the contrary, Wittgenstein insists that language is a social and historic phenomenon. The way 'we speak (i.e., the way in which we apply words to things and situations) is not determined by some 'hidden' logic, by laws that we still have to discover, but by rules which we have adopted in response to practical needs and concrete situations that we have faced in our forms of life; as Wittgenstein always reminds, words are instrunlents, and their specific use is shaped by needs they are to satisfy and by tasks they have to fulfil. This might seem very banal, but it actually has major consequences. If the way we use words is shaped by practical needs, springing from particular Lebensformen, then in order to predict future language use one would have to predict future practical needs and future Lebensformen. And because language use is so deeply context-dependent, in order to foresee future ways of applying words to things and situations, one has to know not only what are the relevant things and situations, but also what the future contexts of use would be like. Obviously, this is a hopeless task. One can describe a fanliliar form of life and can even imagine a possible fornl of life, but it is clearly impossible to predict a future form of life with all its details (relevant for people sharing it and for the way they speak). Consequently, it is impossible to foresee their future problenls and needs, and therefore impossible to predict their future language use. Linguistic behavior should not be compared to the (predictable) movement of celestial bodies, but rather to the totally unpredictable social practices and rituals: as Wittgenstein penetratingly renlarked, we apply words as we confer titles could someone reasonably hope to be able to predict the titles different

114

institutions would confer in the next thousand years, the way in which those titles would be conferred, and to whom? Specifying the classes of things which would in the future be assembled under the same names is no less difficult than specifying the categories of persons who would receive certain titles. Predicting the application of words (what words would be preferred? in what way would they be used? what they would be applied to?) is as hopelessly difficult as predicting the social rituals by which titles are conferred. Or, to use another Wittgensteinean analogy, could people in the 17th century predict the games we play today? The answer is obviously negative, although some games we play today (lawn tennis, for instance) are rooted in games that people used to play three centuries ago. But it was clearly impossible for them to foresee the evolution of their preferred games, exactly as it is impossible for us to foresee the evolution of our present language games. The unpredictability of language use is due to the adaptive character of language, which accommodates itself to particular life-situations and speechcontexts. This might sound trivial, but its implications are not so obvious and widely recognized, as they should be: the main implication of the adaptive character of language is that, if there is some sort of 'logic' which 'guides' linguistic use, then that is the visible 'logic of human practice' or the 'logic of life', not a 'hidden', innate, logic. That is, the secret of language use is to be found in specific human activities, in particular institutions, contexts of speech and practices, not in a 'deep', invisible, structure; language adapts itself to practical needs, not to 'hidden' requirements, because it is connected to public history, not to a genetically-rooted logic. It seems thus pretty clear that Chomsky is wrong when adopting a naturalistreconstructionist approach to language. There can be no naturalistreconstructionist approach to language as there can be no such approach to history. The reason for that is, of course, not that history and language are supernatural, but simply that evolutions in both history and language are so heavily dependent upon imprevisible changes in human practices and their context that no reconstruction (of the kind we do in natural science) is possible. Thus, one main source of Chonlsky's failure is his underestimation of the historical character of language. The second source of his failure is his underestimation of the normative character of the linguistic behavior. It is no accident that almost all his critics, authors as different as Quine and Dummett or Searle and Hacker, have insisted so much upon the normative aspects of language (rules, following rules, guidance provided by rules etc.). If language was exclusively based upon linguistic habits (i.e., upon natural regularities, devoid of any normative element), there would have been nluch easier to accept the idea that the (purely natural) linguistic 115

behavior could be explained (causally) by a 'deep', 'hidden" structure like the UG; anything implied by the UG would have been 'correct', and the native speakers would have had no authority to decide, as they please, about correctness. But, as Thomas Nagel once put it, "whatever native speakers agree on is English": in this sense, language use is arbitrary (i.e., it depends upon free agreement among speakers, not upon laws or imperative requirements) and that is why the UG has no role to play in determining what is English. It thus appears that naturalism is not the right strategy for such fields as language, law, or games, where the normative and conventional element is fundamental. There can be no naturalistic approach to language (of the kind advocated by Chomsky) as there can be no such approach to traffic regulations: in both cases, we have to deal with human-made conventions, heavily dependent upon practical needs and contingent situations, and in no way derivable from a hidden logic. Now, there still seems to be an important reply that naturalists could make. Even if one agrees that language, as a whole, has such historic and contextual characteristics that global reconstructions of the kind Chomsky aims at are not possible, one can still claim that inside language there can be or there actually are components or sub-structures which can be grasped by a naturalistreconstructionist approach. That is, even if language is not entirely algorithmic, this does not prove that there are no algorithmic elements in it; and if there are, they should of course be studied in a naturalist-reconstructionist manner. My answer to that is along the following lines. We have first of all to distinguish between syntax and semantics. In the field of syntax, which is Chomsky's true area of competence but which is of no interest for Wittgenstein and for our present topic, it may well be true that there are sub-structures which allow for a Chomsky-type reconstruction. I have doubts upon that too, but I will not make any judgement on that topic. But as to semantics (or, for that matter, to pragmatics), the only field of interest for Wittgenstein, the above argunlent is unacceptable. According to Wittgenstein's view, language use is developed by extending the application of words from some initial, paradigmatic cases, to new cases, in which some family-resemblances (with the original examples) are present. This process of extending linguistic use is entirely shaped by pragmatic requirements, springing from the relevant contexts of use, from the relevant forms of life and fronl the human interests connected to thenl. No abstract logic seem to be at work here; if there was, it would not have been possible for us to play the well-known games of denying what we want to affirm (or affinning what we want to deny) and of answering (quite seriously) some particular questions with a categorical 'yes and no'. The way we develop language use seems to be entirely a matter of 116

practical needs and local contingencies, and never a matter of theoretical reasons, logic requirements or algorithmic operations. The pattern of use is arbitrary (in the sense indicated earlier in this chapter), entirely contingent and purely adaptive: it can be elucidated by describing a form of life, a context of use, some practical needs or requirements, a sequence of relevant events, facts which have influenced the linguistic behavior etc.; it cannot be explained by some logical mechanism, some algorithm or some abstract, 'geometrical', constraints, simply because language use is historic (i.e., it belongs to the history of some local human community), and not 'geometric' or 'computational'. Take, for instance, the way in which the use of words like 'game' (Wittgenstein's example), 'family', or 'freedom' has changed and has been extended up to this moment; is there any boundary to be drawn between the influence of a historic element (to be simply described) and the influence of a 'computational' or 'logic' one (to be reconstructed in a Chomskyan manner)? If there is such a boundary, nobody has managed to indicate it. It seems that everything that is relevant for the evolution of such concepts and for language use in general can be captured in a 'historic' description of linguistic behavior and its pragmatic reasons; and if so, Occam's razor acts against any naturalist-reconstructionist approach. Why should one try to explain by sophisticated algorithms like the UG what can be perfectly well understood by describing what happened, what made people want to use some words in some ways and how the application of words has historically changed accordingly? Chomsky's approach could be justified by showing that in the way we apply words to things or use them in certain situations there are certain elements or aspects which simply cannot be accounted for by description: i.e., aspects which the history of speech (of concepts-use) cannot elucidate. It would be, for instance, very useful to prove that in the way we apply words as different as 'red', 'animal', 'category', 'influence' and so on, a boundary can and should be drawn between what can be explained 'historically' (by simply 'assembling reminders' about the relevant form of life, contexts and situations) and what cannot be explained in the same manner; between aspects of use which can be understood in the light of usual, visible, human behavior, and aspects of linguistic use which cannot be similarly understood because they seem to be connected to some invisible factors that influence speakers. It is only after the discovery of such a boundary, that the effort to discover, explain and reconstruct a 'hidden' logic of speech should start. Thus, the advocates of naturalist-reconstructionist explanations should first of all demonstrate that such explanations are actually needed, that major elements of language use cannot be elucidated by describing contexts of speech and language games; this, for the time being, they didn't. The burden of proof is upon them. 117

Appendix WITTGENSTEIN AND THE CLASSICAL GERMAN TRADITION OF LANGUAGE CRITICISM (A Case Study: The Critique of Assimilation from Nietzsche to Wittgenstein)

There is an obvious theoretical link between the very old philosophical dilemma "Uniformity versus Diversity" and the way in which the task of concepts is seen. If there actually is ('deep') ontological uniformity behind the (apparent) diversity of things, then concepts could legitimately ignore individual ('nlarginal') differences among real things and capture (through their meanings) only the 'deep', common elements which express the basic uniformity characterizing thenl. If uniformity and unity are real, then ('minor') differences and variations can easily and harmlessly be forgotten. For instance, if, despite their particularities, real things actually fall into categories, kinds or species, then concepts could safely encapsulate in their meanings only the 'essential' features of such a category, kind or species, leaving out 'inessential' individual particularities. Or, if particular things are 'mixtures' of general qualities (i.e., qualities which are common to all things of a certain kind) and of individual qualities, then of course concepts should only retain these general features, while simply ignoring the individual ones as irrelevant peculiarities. It is precisely this connection which justified the conceptual strategies used by Socrates in his philosophical conversations and by Plato in his dialogues. Being convinced that there is some sort of uniformity behind ontological diversity, they have all the time tried to grasp it by pointing out the general, essential, characteristic features of things as they were encapsulated in the meanings of just, true, good etc. But we don't have to go back so far in the past in order to fmd philosophers who accepted this theoretical link and worked in its light. A very good example, in this respect, is offered by Arthur Schopenhauer, maybe the last great philosopher who continued this tradition inaugurated by the ancient Greeks. Schopenhauer, who was interested in the philosophical problems raised by language and conceptual generalization, is clearly still committed to the classical view, according to which any individual has some general features (which must be captured by concepts) and a lot of particular and inessential ones which are to be ignored. He thinks that abstract concepts do precisely this, for

119

"Man HU~t namlich dabei das viele Unwesentliche, daher nur Verwirrende der realen Dinge weg und operiert mit wenigen, aber wesentlichen, in abstracto gedachten Bestimmungen".l

The description given by Schopenhauer to the process of concept-formation - a process in which many individual elenlents are being ignored - is similar to the description given later on by Nietzsche: "Die Bildung eines Begriffs geschieht iiberhaupt dadurch, daB von dem anschaulich Gegebenen vielen fallengelassen wird, urn dann das iibrige fur sich allein denken zu konnen".2

But while Nietzsche will see this process as a kind of 'falsification through conceptualization', Schopenhauer praises concepts (abstractions) for their achievement of throwing off the "useless baggage" ("Abwerfen unniitzen Gepackes" -Ibidem), i.e. individual, 'inessential' details. Thus, Schopenhauer iIIustrates, in this respect, an old philosophical tradition which (as I shaH try to show) is rejected by both Nietzsche and Wittgenstein.

* The aim of this paper is to point out a striking theoretical similarity between some philosophical efforts made by Nietzsche and some made by Wittgenstein in an attempt (that, I think, they both nlade) to explode a big metaphysical error largely due to the nature of language or to the ways in which the nature of concepts is seen. This could be called the error of assimilation, for in fact it consists in assimilating things which are sensibly different into a general, uniform type. Roughly speaking, nlY claim is that: (a) both Nietzsche and Wittgenstein criticize and aim at rejecting a deep-rooted intellectualist tendency to neglect diversity and to create uniformity, by ignoring or playing down individual differences and by postulating ontological elements which supposedly account for 'hidden' uniformities behind diversity; (b) they both reject the result of this tendency, i.e. the philosophical picture according to which diversity is always only the diversification of a basic, 'deeper' uniformity, a result that is deeply mistaken for it implies that things which 'seem' different are in fact the same;

2

120

Die Welt als Wille und Vorstellung, Band II, Erstes Buch, Kap.6, in A. Schopenhauer, Samtliche Werke, Suhrkarnp, 1986, band II, p.8? Ober die Vierfache Wurzel des Satzes vom Zureichenden Grunde, FUnftes Kapitel, §26, in A. Schopenhauer, Samtliche Werke, Suhrkamp, 1986, Band III, p.121.

(c) they both explain this unfortunate intellectual tendency by connecting it to the 'traps' set by language or by our inclinations to see and use language in certain ways. There are important differences between them at this point (as well as at all the others points), but in any case language appears to be the medium in which this mistake is made; (d) they both argue for an ontological picture in which things are characterized by apparent similarities (Ahnlichkeiten) as a better picture than the one in which things appear to have a 'hidden' unity or sameness (Gleichkeit). The 'unity' and 'sameness' of things are" presented by both Nietzsche and Wittgenstein as springing from human concepts-creating activity, rather than from a 'deep' ontological uniformity.

Assimilation and Its Sources "Jeder Begriff entsteht durch Gleichsetzen des Nichtgleichen" Nietzsche "die Kleider unserer Sprache alles gleichmachen" Wittgenstein

When trying to point out similarities, one may easily seem to be ignoring or neglecting dissimilarities. Moreover, in the present case there are so many differences between the aims, the methods and the intellectual backgrounds of Nietzsche and respectively Wittgenstein, that an attempt to compare their philosophical enterprises might seem hopeless or even wrong-headed. Many apparent sinlilarities may seem to be only superficial ones, and some of them really are inessential. There is no doubt that the danger of focussing on marginal and irrelevant 'comnl0n points' is great here. Nevertheless, I think that a comparison of their respective views on the present subject is not at all irrelevant. Of course there are very many differences in the way each of them tackles the problems connected to this topic; a long list of such differences could be given, but this is too boring a thing to do. But there is one major element which must be mentioned here. Nietzsche simply (and directly) criticizes language as an abstract schematism which provides a false picture of reality. Wittgenstein's view is, on one hand, much more subtle, and, on the other, it is situating its focus at another level. There are some fragments in which he blames, like Nietzsche, ordinary language as something which seduces and pushes us towards error (take, for example, the 121

well-known ones in Das B/aue Buch or in PU, Teil II, xi, p. 224); there are also some fragments in which Wittgenstein complains, not unlike Nietzsche, that language Keeps using the same word for things which are in fact very different, suggesting that this is a misfortune: "Das Ungliick ist, daB unsere Sprache so grundverschiedene Dinge mit jedem der Worte »Frage«, »Problenl«, 1.../ bezeichnet" (PG, p. 359). But in most cases what he actually blames is not language per se, not ordinary language (which, in his view, is OK) but rather its apparent uniformity or its appearance (of being uniform). That is the case in most parts of the Philosophische Untersuchungen, where Wittgenstein evokes the uniformity of how words appear to us ("die Gleichformigkeit ihrer Erscheinung" - PU, § 11). In general, what he points out is not a defect of human language, but rather hunlan errors in using language. The focus of his critique is placed at the level of language use: what he mostly criticizes is the way we understand concepts, we perceive their role and characteristics, we think about language and use it in solving some (philosophical) problem. It could be said that, while Nietzsche's critique is placed at the linguistic level, Wittgenstein's critique is placed at a metalinguistic level: the former blames the failure of language in itself (the gap between language and reality), the latter blames the way in which we use language and think about it, i.e. the gap between how language actually works and how we think that it works. To put it shortly, Nietzsche believes that language errs; Wittgenstein believes that we err, misinterpreting language and its true relation with reality. This, I think, does not create a complete opposition between the two kinds of language criticism. The error of assimilation diagnosed by Wittgenstein, even if not identical, is not unconnected with the same error described by Nietzsche. Despite many disanalogies, Nietzsche and Wittgenstein seem to have aimed at the same intellectual 'illness' and to have similarly identified several of its sources. In this respect, their positions seem to be quite close, although, of course, they are never identical. Both Nietzsche's critique and Wittgenstein's critique of assimilation have a common philosophical target: they aim at rejecting the old metaphysical trend, illustrated by Socrates and Plato, and much later by Hegel and Schopenhauer, which subordinates individuals to ontological types and postulates a 'deeper' uniformity behind the visible diversity of things. It is well-known (and consequently no textual evidence is needed any .more) that Nietzsche continuously attacked Socrates and Plato in this respect, trying to explode the old metaphysical myth of Ontological Unity (supposedly characterizing the 'real' world, in contradistinction to the 'apparent' one). His affirmation of multiplicity and diversity has been acknowledged by

122

many authors. 3 The existence of a similar attempt in Wittgenstein's work has also been noticed long ago. Pitcher, in his early book, wrote: "Whereas philosophers have traditionally looked for sameness and unity, Wittgenstein looks for difference and multiplicity".4 Pitcher says that the error criticized by Wittgenstein is illustrated by Plato's arguments in Menon and Parmenides. But, as far as I know, there is no analysis of the similarities which exist between Nietzsche's and Wittgenstein's critiques of ontological uniformisation and assimilation. These similarities can be highlighted by a comparison between the ways in which they have identified and explained the main sources of ontological uniformisation and conceptual assimilation. Source 1: The Neglect ofIndividual Differences "So gewiss nie ein Blatt einem andem ganz gleich ist, so gewiss ist der Begriff Blatt durch beliebiges Fallenlassen diese individuellen Verschiedenheiten, durch ein Vergessen des Unterscheidenden gebildet" Nietzsche

The first source of unifom1isation is the neglect of individual differences, which leads to assimilation, to the presentation of things which are actually different as being the same. One finds in Nietzsche's work the outline of a whole theory on the origins ,the n1echanisms and the functions of assimilation. Assimilation, "ein groBer Wahn "5, a "Grundfiktion "6 or even a "Grundirrthum "7, is nevertheless one of those falsifications without which life is not possible8 that Nietzsche insisted so much upon. Its origins are not in human thinking, but in the vital functions of even the most primi-

3 4 5

6 7 8

Recently, for example, by Alan D. Schrift, in his Nietzsche and The Question of Metaphysics, Routledge, New York, London, 1990, p.82. George Pitcher, The Philosophy of Wittgenstein, Prentice Hall, Englewood Cliffs, 1964, p.217. Almost all my quotations from Nietzsche are from Nietzsche Werke, Kritische Gesamtausgabe, Hrsg. von Giorgio Colli und Mazzino Montinari, Waiter de Gruyter, Berlin, New York (hereafter I use the abbreviation KGW). The present quotation is from KGW, V2, 11 (132), p.388. KGW, VII3, 35(37), p.259. KGW, IV3, §12, p.185. Die froliche Wissenschaft, § 111 - KGW, V2, p.149.

123

tive organisms;9 but assimilation occurs also in all mental activities, through sensations 10, memory11, imagination12 and especially through logical thinking13 and conceptual knowledge l4 . But what we are mostly interested in is conceptual assimilation. Like Schopenhauer before, Nietzsche thinks that concept-fonnation is a process of abstractisation by which many particularities are neglected or forgotten: "Das Uebersehen des Individuellen und Wirklichen giebt uns den Begrifr'15; but, contrary to his predecessor, he considers this process as nothing else than a falsification of reality: "Jede uns fordernde ErkenntniB ist ein Identificiren des Nichtgleichen, des Ahnlichen, .d.h. ist wesentlich unlogisch. Wir gewinnen einen Begriff nur auf diesem Wege und thun nachher, als ob der begriff Mensch etwas thatsachlich ware, wahrend er doch nur durch FalIenlassen alIer individuellen Ziige von uns gebildet ist. Wir setzen voraus, daB die Natur nach einenl solchen Begriff verfahre: hier ist aber einmal die Natur und sodann der Begriff anthropomorfisch. Das Obersehn des Individuellen giebt uns den Begriff und damit beginnt unsre ErkenntniB: im Rubriziren, in Aufstellungen von Gattungen. Dem entspricht aber das Wesen der Dinge nicht: es ist ein ErkenntniBprozeB, der das Wesen der Dinge nicht trifft."16

Conceptualization thus leads to a categorization of things, but - as Nietzsche insists - the resulting categories have no real counterpart: by postulating Gattungen, the intellect creates only a fictional sameness. The mistake involved here is that, by creating concepts which ignore individual differences, we do present things which are different as being the same - for "Jeder Begriff entsteht durch Gleichsetzen des Nichtgleichen"17. Although this view of conceptualization is already present in the early essay Uher Wahrheit und Luge..., it is not a mere folie de jeunesse; in 1881 Nietzsche is still preoccupied to show that there is no such thing as 'ontological sameness' and also that this is not sinlply a matter of fact, but also something necessary: "Ob es in Einer Gesammtlage etwas Gleiches geben kann, z.B. zwei Blatter? Ich zweifle: es wUrde voraussetzen, daB sie eine absolut gleiche Entstehung hatten, und damit

9 10 11 12 13 14 15 16 17

124

KGW, V2, 11(268), p.442. KGW, VII3, 40(15), p.367. KGW, V2, 11(138), p.391 - but there are many other fragments in which he says that memory is conditioned by assimilation. Ibidem. KGW, VII3, 40(15),p.367. KGW, III4, 19(236),p. 81. Ober Wahrheit und LUge im aussermoralischen Sinne, KGW, 1112, p.374. KG\V, 1114,19(236), p.81. KGW, 1112, p.373.

hatten wir anzunehmen, daB his in aBe Ewigkeit zurUck etwas Gleiches bestanden habe"18.

But if sameness is in principle impossible, the appearance of sameness or the uniformisation created by conceptual assimilation is (Nietzsche believes) a falsification of reality. Through conceptualization, things which are really different appear as being the same (kind of) thing, and consequently diversity is substituted by uniformity.

* "Hegel seems to me to be always wanting to say that things which look different are really the saIne. Whereas my interest is in shewing that things which look the same are really different" Wittgenstein

As the above-mentioned declaration 19 shows, Wittgenstein too was preoccupied to combat the old metaphysical conviction that diversity is more a matter of appearances than a matter of reality, whereas uniformity or sameness do constitute the 'deeper' reality. But, as analytic philosopher, Wittgenstein speaks less about 'how rea] things are' and more about language and concepts. When speaking about words and meanings, he is though eager to point out our neglect of diversity and also the roIe played by language in this respect: "Die unsagliche Verschiedenheit aller der tagtaglichen Sprachspiele kommt uns nicht zum BewuBtsein, weil die Kleider unserer Sprache alles gleichmachen. ,,20

Wittgenstein develops this idea in various contexts, by stressing that "in der Sprache haben wir verschiedene Wortarten" (PU, §17); that names are not one and the same thing, but many different things: "Name nennen wir sehr Verschiedenes" (PU,§38); that there are very many different kinds of sentences (PU,§23) and that we do most various things with them (PU, §27); that we apply words in many diffe-

18 19

20

KGW, V2, 11(202), p.421. It was a declaration made to M.O. Drury - see Rush Rhees (ed.), Ludwig Wittgenstein Personal Recollections, Basil Blackwell, Oxford, 1981, p.171. Philosophische Untersuchungen, (hereafter PU), Basil Blackwell, Oxford, 1968, Teil II, xi, p.224.

l25__

rent ways for the use of words is not gleich, but "ganz und gar ungleichartig" (PU, §IO). Unfortunately, we are generally reluctant to accept this linguistic multiplicity ("Mannigfaltigkeit"- PU, §23); although the functions of words are as diverse as the functions of tools (PU, § II) or of handles (PU, § 12), we tend to ignore this diversity (especially when we are doing philosophy) and "was uns verwirrt ist die Gleichformigkeit ihrer Erscheinung, wenn die Worter uns gesprochen, oder in Schrift und im Druck entgegentreten. ,,21

Although Wittgenstein speaks here explicitly only about the linguistic uniformisation we are guilty of: there is no doubt that he connected it with a sort of ontological uniformisation which is the result of our tendency to see similar things as 'having something in common': we tend to see words like hammer, pliers, saw as gleich precisely because we think that the corresponding objects are gleich. When told that these objects are not 'the same thing', because their functions are various (and "So verschieden die Funktionen dieser GegensUinde, so verschieden sind die Funktionen der Worter"-PU, §11), we object that they must be 'the same' because, despite their particularities, they all serve to modify something (PU, § 14). Wittgenstein's answer to this objection is: "Ware mit dieser Assimilation des Ausdrucks etwas gewonnen?" (Ibidem)

Now, this is very important, I think, because it shows how Wittgenstein thought about the matter. He clearly suggests that, by postulating an uniformity behind the visible diversity of things (in the present case, a functional uniformity behind the diversity oftools), we only get an assimilation of expressions. Linguistic or conceptual assimilation and ontological uniformisation are thus directly connected. Wittgenstein criticizes both of them. He doesn't accept the received opinion that there must be something common, and consequently a 'deeper' sameness, in the case of games (PU, §66), of leaves (PU, §73), of expectations (PU, §77, but see also Das Blaue Buch22 ) or of wishes (BB, op.cit., p.40 but see also Philosophische Grammatik23 ). This opinion does not reflect reality, but only our Streben nach Allgemeinheit:

21 22 23

126

PU,§l1. Das Blaue Buch, (hereafter BB), in Ludwig Wittgenstein, Schriften, Suhrkamp, Frankfurt am Main, 1970, Band 5, p.41. Philosophische Grammatik, (hereafter PG), in Schriften, op.cit., Band 4, teil IV, §76, p.120.

!tWir sind z.B. geneigt zu denken, daB es etwaB geben muB, das allen Spielen gemeinsam ist, und daB diese gemeinsame Eigenschaft die Anwendung der allgemeinen Bezeichnung Spiel auf die verschiedenen Spiele rechtfertigt, wahrend Spiele doch eine Familie bilden, deren Mitglieder Familieahnlichkeiten haben"24.

Wittgenstein compares this tendency to focus upon generality and to postulate something uniform (common to all similar things) to the old philosophical idea that features are ingredients of things (n daB Eigenschaften Bestandteile von den Dingen sind" - BB, op.cit., p.38), an idea that he attributes to Plat025 • The connection between the way we tend to see concepts (meanings) and the way we tend to see real objects is once again visible in Wittgenstein's discussion about leaves (BB, op.cit., p.38). We are tempted to explain the understanding of a concept like leaf by evoking a general representation which supposedly includes only what all things (leaves) have in common - "das nur das enthalt, das allenBHittern gemeinsam ist" (Ibidem). Our view of concepts (as capturing in their meanings sonle essential, common features) is obviously connected with our ontological conviction that 'there must be something common' to all things to which we apply the concept. The assimilation of linguistic uses (which takes place by our postulating of a uniform meaning behind all the diverse uses of leat) corresponds to an uniformisation of things (which takes place by our postulating of some common elements shared by all the relevant things). We forget the huge variety of uses which a word like leaf has and we focus on the uniform meaning (the general representation: "die allgemeine Vorstellung"); analogously, we forget the huge variety of things (or the individual differences, as Nietzsche said in his remark about the same concept, leaf) and we focus upon what 'they have in common'. It is in this way that the uniformity behind diversity comes into existence for us. But this uniformity doesn't in fact exist26 and we commit ourselves to it only because we have a wrong intellectual attitude towards individual differences and cases. For Nietzsche, this attitude consists in a tendency to forget individual differences and to treat as gleich what is really only similar (ahnlich). For Wittgenstein, it consists mainly in a tendency to find common features where, in fact, there are only Familieahnlichkeiten. But in both cases, the result is an underestimation of individuality - where Nietzsche speaks about omitting individual differences, Wittgenstein speaks about "die verachtliche Haltung gegenilber dem Einzelfall" (BB, op.cit., p.39). This attitude, deeply rooted in the Western thinking, encouraged philosophers to ignore particular cases as irrelevant: 24

25 26

BB, op.cit., p.3? Bemerkungen tiber Die Grundlagen der Mathematik, (hereafter BGM), Basil Blackwell, Oxford, 1967, Teil I, §71. BB, op.cit., p.40-41; PU, §66; PG, §35, op.cit., p.75.

L21

"diese Vorstellung hat nicht nur zu keinem Ergebnis gefl.1hrt, sondern dariiber hinaus den Philosophen veranlaBt, fiber konkrete FaHe als irrelevant hinwegzugehen"27.

Socrates appears to be a typical representative of this "verachtliche Haltung gegenliber dem Einzelfall": "Wenn Sokrates die Frage stellt: Was ist Erkenntnis?, denn halt er die Aufzahlung von tatsachlichen FaIle von Erkenntnis nicht einmal fUr eine vorHiufige Antwort"28.

This traditional inclination to underestimate individuality is of course the same with the venerable intellectual tendency to overestimate what is general.

* The best way to point out the similarities between Nietzsche's and Wittgenstein's positions, in this respect, is offered by a comparison of their views with Schopenhauer's conception, since Schopenhauer strongly influenced both of them. Schopenhauer, after Hegel, is the most remarkable representative of the classical, metaphysical trend favouring generality and dismissing individuality. He believes in a fundamental unity of the world - tIes eigentlich auch nur ein Weseo gibt"29 - and talks despisingly about the Illusion der Vielhait (Ibidenl). His attempt of 'naturalizing' Platonism results in an identification of Ideas/FomlS with natural species: "Was nun als bloB objektives BiId, bloBe Gestalt betrachtet und dadurch aus der Zeit wie aus allen Relationen herausgehoben, die Platonische Idee ist, das ist, empirische genommen und in der Zeit, die Spezies oder Art: diese ist also das empirische Korrelat der Idee u30 .

Fronl such a point of view, individuals are ontologically close to nothing, whereas species are eternal and grand: "die Lowen, welche geboren werden und sterben, sind wie Tropfen des WasserfalIs, aber die leonitas, die Idee oder Gestalt der Lowen, gleicht dem unerschfitterten Regenbogen darauf'. 31

27 28 29 30 31

128

BB, op.cit., p.40. BB, op.cit., p.41; see also PG, VI, §76. Die Welt als Wille und Vorstellung, op.cit., Chapter Transzendente Betrachtungen, op.cit., Band II, p.416. Idem, p.471. Idem, chapter Ober den rod, op.cit., p.617.

Schopenhauer illustrates perfectly the classical view according to which ontological diversity is nothing else than the diversification of a fundamental unity: "Die urspriingliche und wesentliche Einheit einer Idee wird durch die sinnlich und zerebral bedingte Anschauung des erkennenden lndividuums in die Vielheit der einzelnen Dinge zersplittert".32

And not only that individuals are 'rooted' in their Gattung, but they are individuals of some kind only in so far as they hide in themselves the essential, common features of their species: "Das Individuum wurzelt in der Gattung und die Zeit in der Ewigkeit: und wie jegliches Individuum dies nur dadurch ist, daB es das Wesen seiner Gattung an sich hat".33

Nietzsche explicitly rejects this way of seeing the relationship between generality (Gattung) and individuals, as being mistakenly anthropomorfic: "Es ist ein falsches Gesichtspunkt: unl die Gattung zu erhalten, werden unzahlige Exemplare geopfert. Ein solches Um giebt es nicht!" , because "Die Natur will nicht die Gattung erhalten!".34

Individuals are far from being mere means by which Gattungen are preserved; in fact, they are "similar beings" which naturally preserve thenlselves: "Thatsachlich erhalten sich viele ahnliche Wesen" (Ibidem). As it is well-known, Nietzsche does not believe in the absolute reality of individuals, for what we take as an individual is in fact a plurality, but, in comparison to Gattungen, they are closer to being real - "Die Gattung ist der grobere Irrthum, das Individuum der feinere Irrthum"35 - or have some sort of reality which Gattungen lack: "Ebenso giebt es keine Gattung, sondem lauter verschiedene Einzelwesen!"36. We need not deal here with all the subtleties of this view on the reality of individuals, because it is very clear that Nietzsche is against any subordination of individuals to types, and he strongly opposes the classical philosophical tendency to underestimate individuals. He deplores Schopenhauer's influence precisely because it led to a denial of plurality and of individuals, to "die Leugnung des Individuums ('alle Lowen sind im Grunde nur Ein Lowe', 'die Vielheit der Individuen ist ein Schein') - as he says in Die frohliche Wissenschaft (Zweites Buch, §99). 32 33 34 35

36

Idem, chapter Transzendente Betrachtungen, op.cit., p.472-3. Ibidem. KGW, V2, 11 (178), p.406. KGW, V2, 11(156), p.399. KG W, V2, 11 (178), p.406.

129

Nietzsche obviously refuses to see the diversity of individuals as a set of illustrations of the same basic thing, of a Type or of a Gattung. He implicitly rejects the classical view according to which things are composed of both general, connnon features and individual, specific qualities: not only that diversity and differentiation are universal - nVerschiedenheit herrscht in den kleinsten Dingen, Samenthierchen Eiern - die Gleichheit ist ein groBer Wahn"37, but there is no such thing as a 'general component' in an individual entity because "In einem Individuum ist bis in die kleinste Zelle hinein alles individuell".38 Precisely because any individual is individual through and through, conceptformation as a Fallenlassen of individual elements is a sort of falsification to be regretted, not - as Schopenhauer thought - a successful process of capturing 'what is general and essential" for a set of things. Concept-formation leads to Forms, Types and Gattungen, and all these are nothing else than Fiktionen. 39 Thus, in Nietzsche's thinking the classical picture is reversed. It is not the subordination of individuals to general types that makes these individuals what they are, i.e. particular examples of certain kinds of things (as Schopenhauer believed); on the contrary, by such a subordination things are 'assimilated' and thus reality is falsified. It is not the case that general and common elements are relevant, whereas individual ones (inessential) are not. It is in fact the other way round: 'what is general and common in things' is in fact fictional and irrelevant; only what is individual is real and relevant. The identity of things is not the result of their belonging to certain general categones. Moreover, instead of postulating Gattungen, and claiming that they capture the general elements hidden behind diversity, we should accept that in fact there are only similarities40 among individuals. But similarities do not make things gleich; thus, there is no uniformity behind diversity.

* This reversal of the classical view, according to which 'what is common and general' creates a sort of 'hidden' uniformity among things, is in a way completed by Wittgenstein. A particular thing is what it is, e.g. a game, not because it contains a 'core' of general features common to all similar things (to all games) - the common conviction that "Es muB ihnen etwas gemeinsan1 sein, sonst hieBen sie nicht

37 38 39 40

130

KGW, V2, 11(132), p.388. KGW, 1114, 19( 187), p.65. KGW, VIII2, 9(144), p.8t. See, for instance, KGW, V2, 11(237), p.429, where Nietzsche explicitly asks us to speak about 'ahnliche Qualitaten' instead of'gleich'.

'Spiele'" (PU, §66) is mistaken. There are no such 'general', identity-creating features; what we have in fact are only family-resemblances. Wittgenstein goes farther than Nietzsche, being more explicit and precise: not only that the various particular cases are not gleich, but in fact there is nothing common to all of them; for instance, in the case of language activities, "es ist diesen Erscheinungen gamicht Eines gemeinsam" (PU, §65; see also BB, op.cit., p.40-41). Analogously, there is nothing common to all wishes, expectations etc. Contrary to what one might have expected, Wittgenstein rejects not only the Assimilation des Ausdrucks, but also the prejudice that there must be something common to all things we give the same name or (as one could be tempted to say) the myth of the 'hidden' unity. This myth originates in Plato's inclination 'to find 'hidden' ingredients, i.e. common features behind the apparent diversity, but it is also supported by Hegel and Schopenhauer and it creates the metaphysical feeling that diversity is only a diversification of a basic unity (a common ingredient) exactly as all alcoholic drinks are variants of the same basic component (BB, op.cit.,p.38), which is 'hidden' in all of them. Wittgenstein insists that it would be wrong to suppose that all kinds of things have a similar uniformity springing from common ingredients. This widespread supposition is in fact the result of our Streben nach Allgemeinheit and of our tendency to generalize some cases (Zettel, §444) or to nourish one's thinking exclusively with one kind of examples (PU, §593). Wittgenstein does not plainly reject the existence of Gattungen or of types of things, as Nietzsche did, but he too rejects the idea that such categories had a unity of the kind described by classical philosophers. If things we give the same name ('gaole', 'wish', 'leaf) do have some sort of unity, this comes not from 'hidden' common ingredients, from a hidden uniformity, but rather frOTIl us, i.e. fronl our way of grouping together things which have family-resemblances. They are 'of the same kind' not because they illustrate the same common 'essence', but because we group them together in virtue of their similarities.

* A major common element, I think, is that both Nietzsche and Wittgenstein abandon the classical logico-metaphysical scheme of interpreting things as having 'common, general' features but also 'individual, specific' ones. Instead of that, they both plead for an interpretation in terms of similarities. Wittgenstein's view about family resemblances is so familiar to everybody that there is no need to present it here. The only thing to be mentioned, perhaps, is that precisely the same term has also been used by Nietzsche when speaking about indoeuropean languages (Jenseits von Gut und Bose, §20). But Nietzsche also seems to have been convinced that a description of things in terms of similarities is much

131

more realistic that the usual, classical one based on the supposition that things 'of the same kind' are essentially gleich. He keeps insisting that leaves, characters etc. can never be gleich, and that even in the field of Chemistry we should talk about similarities, not about qualities which are gleich: "Ahnliche QualiUiten, sollten wir sagen, statt 'gleich' - auch in der Chemie"41. His proposal that similarity should substitute sameness could be superficially interpreted as trying to combat an exaggeration, Le. as saying that, in fact, things are not completely identical, but only similar in certain degrees. But this is not what Nietzsche meant. According to his way of seeing, "Das Ahnliche ist kein Grad des Gleichen: sondem etwas vom Gleichen vollig Verschiedenes"42. This shows that Nietzsche did not think the difference between the classical view and his own to be a difference in degree; what he points out here is not that classical thinking rightly took things (that we call the same) as being alike, while making the mistake to consider them exactly alike; and his point is not that things are alike, but only to a certain extent or in some degree. It is not a matter of degrees; it is not the case that Nietzsche criticized an exaggeration or an overstatement, i.e. the idea that things are absolutely identical, whereas, in fact, they are only approximately identical. H is remark shows that, according to him, things are not to be seen as 'almost identical' or 'almost of the same kind'; on the contrary, they are to be seen as being in fact different, although they have certain similarities. While the classical view implied that things (which we give the same name) are basically of the same kind, although they have some (minor) particularities, Nietzsche insists that these things are basically different, although they have also some similarities (for us!43). Apart from reminding us of Wittgenstein's casual remark that "things which look the same are really different", this is important, I think, because it seems to indicate an implicit rejection of the old dichotomy between "the core of general, common features" and "the individual, marginal particularities". Nietzsche never says, of course, that there is no common feature shared by all things that we give the same name. In order to be entitled to say that, he would have had to analyse some particular examples (like games) in the careful and detailed manner used by Wittgenstein. As Nietzsche was far fron1 being an analytic thinker, such an analysis was not to be expected. But his above-mentioned remarks amount, I think, to an implicit rejection of the classical, logico-metaphysical scheme of in41 42 43

132

KGW, V2, 11 (237), p.429. KGW, V2, 11(166), p.403. For this subjectivistic nuance, see KGW, V2, 11(237), p.429. A similar nuance is not, of course, to be found in Wittgenstein's remarks.

terpreting things as 'mixtures' of general and respectively individual qualities; and, at least, they prove that Nietzsche abandoned the classical scheme still defended by Schopenhauer.

Source 2: Science and Formal Science Both Nietzsche and Wittgenstein connect ontological uniformisation with the Western tradition of systematic knowledge and stress the role played by formal science in this respect. Nietzsche's first point is that systematic knowledge should be seen as mere schematization: "nicht 'erkennen', sondem schematisiren"44 Systematic knowledge ~ .is only the expression ofa human need Uzu subsumiren, zu schematisiren" (Ibidem), and the categories which create uniformity - die Gattungen, in the fIrst place - are based on some sort of Rubriziren 45 Rubriziren implies, of course, a neglect of individual features - generalities prevail, things placed under the same heading appear as 'being the same'. Knowledge is always based on metaphors: 0

0

"Das Erkennen ist nur ein Arbeiten in den beliebsten Metaphem" 46 0

Consequently, truth is only "Ein bewegliches Heer von Metaphern, Metonymien, Anthropomorphismen" , as he says in his famous fragment of Uber Wahrheit und Luge. Metaphors, Nietzsche continues, are based on the permanent treating as identical, as gleich of what in fact is only similar: "Metapher heiBt etwas als gleich behandeln, was man in einem Punkte als ahnlich erkannt hat"4 7

Thus, to know means only "Die ganze Welt unter die richtigen Begriffe subsumiren".48 We aim at truth, but looking for truth is nothing more than "einem vorhandenen Begriffrichtig die einzelnen FaIle unterordnen" (Ibidem). What this whole view implies in the end is that knowledge, with its conceptual patterns, eliminates plurality and diversity:

44 45 46 47 48

KGW, KGW, KGW, KGW, KGW,

VIII3, 14(152), po125. 1II4, 19(236), p.8l. III4, 19(228), p.79. 1114, 19(249), p.86. 1114, 19(258), p.88.

133

"Der Intellekt als das Mittel der Tauschung mit seinem Zwangsformen 'Substanz', 'Gleichheit', 'Dauer' - er erst hat die Vielheit sich aus dem Sinne geschlagen"49

Intellectual judgement "arbeitet unter den Voraussetzung daB es iiberhaupt identische FaBe giebt"50 and therefore knowledge in general is nothing else than an "Identificiren des Nichtgleichen, des Ahnlichen"51. But although knowledge generally means simplification - for "Der ganze Erkenntnis-Apparat ist ein Abstraktions - und SimplifIkations - Apparat"52 - it is in fact formal thinking that plays the major role in this respect. Algorithmic thinking compells us to create unities 53 and to ignore (particular) qualities, to postulate identical cases: "tiberall wo etwas rein arithmetisch gedacht werden solI, wird die Qualitat weggerechnet. Ebenso in aHem Logischen, wo die Identitat der Faile die Voraussetzung ist, also der eigentliche spezielle Charakter jedes Vorgangs einmal weggedacht ist"54

Assimilation does not begin with logic, but it is logical thinking that amplifies and perpetuates it: "Vor der Logik, welche tiberan mit Gleichungen arbeitet, muG das Gleichrnachen, das Assimiliren gewaltet haben: lind es waItet noch fort, und das logische Denken ist ein fortwahrendes Mittel seIber fur die Assimilation, fUr das Sehen-wollen identischer Falle"55

The very existence of logic is conditioned by the supposition that there are identical cases 56 and Nietzsche insists that in all logical matters "die Identitit der Faile die Voraussetzung ist".57 Logical thinking is thus one of the main sources of conceptual assimilation.

*

49 50 51 52 53 54 55 56 57

134

KGW, V2, 12(35), p.480. KGW, VII3, 40( 15), p.367. KGW, III4, 19(236), p.8t. Der Wille zur Macht, §503. KGW, VIII3, 14(79), p.50. KGW, VII3, 40(37), p.379. KGW, VII3, 40(33), p.377. See also 42(7), p.433, in the same volume. KGW, VII3, 40(13), p.365. KGW, VII3, 40(37), p.379. See also Die frohliche Wissenschaft, §111. There are several quasi-identical remarks on this topic.

Wittgenstein too connects our tendency to ignore individual differences - our Streben nach Allgemeinheit - with the influence of science in general: "Unser Streben nach Allgemeinheit hat eine weitere Hauptquelle: unsere Voreingenommenheit fur die naturwissenschaftliche Methode. Ich meine die Methode, die Erklarung von Naturerscheinungen auf die kleinstnl0gliche Anzahl primitiver Naturgesetze zurlickzuflihren; in der Mathematik ist es die Methode, die Behandlung verschiedener Probleme mittels einer Verallgemeinerung zu vereinheitlichen"S8

But he also evokes the case of logic: "Die Raltung gegeniiber dem Allgemeineren und die Speziel1eren in der Logik ist mit dem Gebrauch des Wortes 'Art' verbunden, das uns leicht in Verwirrung bringen kann" (Ibidem), and he suggests that the wrong attitude towards particular cases is here based on the typically logical supposition that such cases are 'incomplete'. In Zettel, Wittgenstein explicitly attributes the tendency we have to generalize some (particUlarly clear) cases, and thus to ignore (other) particular cases, to logic; it is only in logic that such generalizations are justified: "Die Tendenz, den klaren Fall zu verallgemeinem, scheint in der Logik ihre strenge Berechtigung zu haben ltS9

Of course, we are often inclined to extend this tendency (which is justified in logic) to other fields (where it is not justified any more) and this is one way in which our metaphysical mistake appears. We think that some cases must be the exemplars of all cases (Z, §444) and that they show 'how things are in every case'. We thus assimilate cases which are actually different into general types, the resulting picture of things being characterized by uniformisation. Our seeing things 'from a logical point of view' seems to be a main cause of ontological unifonnisation.

Source 3: Commitment to Prototypes and Exemplars A nlain reason for which diversity is ignored and uniformity is postulated can be found in our commitment to prototypes and exemplars. Conceptual thinking overgeneralizes because it tends to focus on what is considered typical and exemplary. Individual differences are neglected as if they were insignificant, and this is because they do not belong to the acknowledged prototype, to the Urform or the Urbild. Nietzsche is very explicit upon this point:

58 59

BB, op.cit., p.38-39. Zettel, (hereafter Z), §444.

135

"So gewiss nie ein Blatt einem andern ganz gleich ist, so gewiss ist der Begriff Blatt durch beliebiges Fallenlassen dieser individuellen Verschiedenheiten, durch ein Vergessen des Unterscheidenden gebildet und erweckt nun die Vorstellung, als ob es in der Natur ausser den BHittern etwas gabe, das 'Blatt' ware, etwa eine Urform, nach der aile Blatter gewebt, gezeichnet, abgezirkelt, gefarbt, gekrauselt, bemalt waren aber von ungeschickten Handen, so dass kein Exemplar correct und zuverlassig als treues Abbild der Urform ausgefallen ware"60

Because our concepts ignore particularities and differences, focussing on general patterns, conceptual thinking is bound to interpret the variety of individuals as a mere diversification of the basic, unitary pattern. Unifonnisation appears to be the inevitable result of conceptualization: we are led to neglect diversity, and to take it as an insignificant by-product, because concepts come into existence by ignoring individual features. While Nietzsche thinks that overgeneralization is the result of conceptformation, in general ( Diese Verallgemeinerung geschiecht durch den dazwischentretenden Begriff'61), Wittgenstein's view is slightly different. He seems to suggest that overgeneralization is connected not with the concepts themselves, but rather with the way we see concepts and their meanings. But, once again, the roots of our error are to be found in the field of language: lt

"Es gibt eine Tendenz, die ihre Wurzeln in unsern gewohnlichen Ausdruckfonnen hat, namlich die Tendenz zu denken, daB der, der gelemt hat, eine allgemeine Bezeichnung zu verstehen, etwa die Bezeichnung 'Blatt', dadurch so etwas wie ein allgemeines Bild von einem Blatt gewonnen hat, im Gegensatz zu Bildern von bestimmten Blattern"62

But this is not only a matter of semantics, or of language-learning; it is also a philosophical matter, because this way of seeing meaning (that Wittgenstein criticizes) is clearly connected with the way we see the ontological relationship between 'general' and 'individual'. As long as one believes that there must be a general core of features, conlmon to all individuals that we give the same name, one is naturally led to think that the task of concepts is to capture in its meaning these conlmon features. Meanings can even be seen as "general representations" of 'what is common' to a set of things: "wir sind geneigt zu denken, daB die allgemeine Vorstellung von einem Blatt so etwas wie ein visuelles Vorstellungsbild ist, jedoch eines, das nur das enthalt, das allen Blattern gemeinsam ist. (Galtons zusammengesetzte Photographie.) Das ist wiederum mit der Idee verbunden, daB die Bedeutung eines Wortes ein Vorstellungsbild ist" (Ibidem). 60 61 62

136

KGW, 1112, p.373-374. KGW, 1114,19(177), p.62. BB, op.cit., p.38.

The temptation to see meaning as a 'general representation' is thus connected with the temptation to see a set of things as having a 'common core' of typical features. But Wittgenstein also points out a second source of overgeneralization: theoryconstruction. While Nietzsche insisted that the Verallgemeinerung achieved by concepts is reminiscent of our old commitment to the idea of an Urform (concepts ignore the individual features of leaves as though these features were departures from the Urform Leave), Wittgenstein remarks that the Verallgemeinerung achieved by theories suggests a commitment of ours to the idea ofUrbiid or of exemplar case. One reason for which Wittgenstein dismisses theorizing (in philosophy, at least) is precisely that theories tend to generalize a particular case: "Es ist ja das Charakteristische einer solcher Theorie, daB sie einen besonderen, klar anschaulichen Fall ansieht und sagt: 'Das zeigt, wie es sich iiberhaupt verhiilt; dieser Fall ist das Urbild aller Falle"'63

Theoretical overgeneralization is also encouraged by our essentialist presuppositions: "Denn wir sindja in der Tauschung, das Sublime, Wesentliche unserer Untersuchung bestehe darin, daB sie ein allumfassende Wesen erfasse" (Ibidem). Since we do believe in 'hidden' essences, once we have identified a particularly clear and relevant example, we tend to think that it is this example which typically captures the very essence we were looking for. The generalization of such an example does not appear to us as being harmful and the resulting uniforrnisation does not strike us as unjustified: on the contrary, it seems to be only natural that 'the essential element' is always present 'behind' the particularities of any individual case. Essentialisnl makes us accept as quite normal the fact that all the individual cases appear to 'illustrate' the same typical, general pattern: ontological diversity, in any particular field, is only the diversification of some 'hidden' essence, which, of course, is the source ofa 'deeper' tmiformity. Like Nietzsche before (in a sensibly different context, though), Wittgenstein rejects this kind of essentialist thinking. He always insists that his remarks do not amount to a theory and do not try to capture an essence: no 'deep' uniformity is being discovered here (for instance, in the case of wishes): "Denn ich wollte ja nicht sagen, diese Beispiele seien die Darstellung des Wesens dessen was man 'wiinschen' nennt. Hochstens Darstellungen verschiedener Wesenheiten die aIle man wegen gewisser Verwandtschaften nlit diesem Wort bezeichnet. Der Imum ist, daB angenommen wird, wir wollten durch Beispiele das Wesen, des Wiinschens etwa,

63

Z, §444.

137

illustrieren, und die Gegenbeispiele zeigten nun, daB dieses Wesen noch nicht richtig erfaBt sei. Das ist, als ware unser Ziel, eine Theorie des Wiinschens zu geben, die dann eben alle FaIle des Wiinschens erklaren muBte,,64

Wittgenstein denies, of course, that his efforts are meant to capture an unique essence of wishing, and also that he is trying to give a theory of his own. But it is not modesty, as to his aims, that prompts such a statement. What he really does here is to reject the whole way of thinking connected with theorizing and overgeneralizing 'typical' cases. The justification for his reluctance to give a theory and to capture 'the essence' by typical examples is simply that there is no unique essence that all particular examples (of wishing, for instance) illustrate, there is no uniformity 'hidden' behind the individual cases - for "schlieBlich gibt es nicht eine bestimmte Klasse von Merkmalen, die aIle FaIle des Wiinschens charakterisiert" (BB, op.cit.,p.40). The classical, metaphysical way of thinking ignored this, being, on the contrary, committed to the idea that typical examples indicate us 'what is essential, general and common'; it is for this reason that this old way of thinking tended to generalize some exanlples, ignoring many others, which only led to a 'philosophical disease': "Eine Hauptsache philosophischen Krankheiten - einseitige Oiat: man nahrt sein Denken mit nur einer Art von Beispielen" (PU, §593). The wrong-headed inclination of seeing diversity as mere diversification of a basic unity or of a prototype manifests itself not only in respect to essences, but also in respect to ideals: thus, "Wir miBverstehen die Rolle, die das Ideal in unsrer Ausdrucksweise spielt" (PU, § 100). For instance, Wittgenstein argues, we are reluctant to accept the idea of a ganle the rules of which are rather vague, because we are committed to a certain ideal of game (with clear-cut and complete rules) and we think that "das Ideal 'nliisse' sich in der Realitat fmden" (PU, §101). That is, we think that all games should illustrate the ideal of ganle, exactly as all leaves, illustrate (more or less successfully) the Urform Leave. Once again, we are 'captives' in a way of thinking based on the supposition that diversity is nothing more than diversification of a basic unifomlity.

Source 4: Abstractions Another source of ontological unifonnisation, identified by both Nietzsche and Wittgenstein, is the reification of abstractions. Diversity and multiplicity are eliminated by postulating "general qualities", which are nothing else than reified abstractions: abstractions, Le. the products of a logical process of schematization,

64 PG, VI, §76, op.cit., p.120.

138

simplification etc., are being treated as 'real things' and 'hidden' causes of the individual acts and events we are familiar with. For Nietzsche, this process starts, once again, with the creation of Gattungen, because "Die einzige Art, die Vielheit zu bezwingen, ist, daB wir Gattungen rnachen"65.But, according to Nietzsche, by identifying Gattungen we do not discover real, although 'hidden', common and essential qualities of things; rather, we simply apply rhetorical means ('Metonymien') in order to create an (artificial) unity: "die Vielheit der Dinge wird unter einen Hut gebracht, wenn wir sie gleichsam als unzahlige Handlungen einer QualiHit betrachten" (Ibidem)

By creating Gattungen, we postulate the existence of some 'unifying elements' which characterize all things belonging to the same Gattung; but such an 'unifying element' is not seen by Nietzsche as a genuine, 'deep' conlmon feature of things, but rather as an abstraction illegitimately considered to be an Ursache: "eine Abstraktion faBt zahllose Handlungen zusammen und gilt als Ursache" (Ibidem). For instance, we apply the same label to different human acts; after thus bringing several individual items under the same heading, we postulate an 'inner' (invisible) psychological trait which we treat as the cause of the particular acts we have already given the same name. "Wir nennen einen Mensch 'ehrlich'; warurn hat er heute so ehrlich gehandelt? fragen wir. Unsere Antwort pflegt zu lauten: seiner Ehrlichkeit wegen. Die Ehrlichkeit! Das heisst wieder: das Blatt ist die Ursache der Blatter. Wir wissen ja gar nichts von einer wesenhaften Qualitat, die 'Ehrlichkeit' hiesse, wohl aber von zahlreichen individualisirten, somit ungleichen Handlungen, die wir durch Weglassen des Ungleichen gleichsetzen und jetzt als ehrliche Handlungen bezeichnen; zuletzt formuliren wir aus Ihnen eine qualitas occulta mit den Namen: 'die Ehrlichkeit"'66

I'his is nothing else than an error of assimilation: one postulates a 'hidden' source of uniformity (a psychological quality) so that individual acts which are not at all gleich should now be seen as mere particular examples of the same general type, springing from the same 'inner' source: the general quality, 'die Ehrlichkeit', is only an abstraction which we postulate as 'real' in order to be able to present different acts as being all alike, i.e. 'ehrlich'. The 'common' element of all these acts is but a qualitas occulta, although we generally tend to consider it 'hidden', but quite real. It is particularly interesting that Wittgenstein seems to see this epistemological and metaphysical error in a sin1ilar light.

65 66

KGW, 1114, 19(215), p.74. KGW, III2, p.373-374.

139

"1st das nicht ahnlich, wie wenn man sagt: 'Wenn dieser Mensch anders handelte, dann mUBte auch sein Charakter anders sein'. Nun das kann in manchen Fallen etwas heiBen und in manchen nicht. Wir sagen: 'aus denl Charakter flieBt die Handlungsweise', und so flieBt aus der Bedeutung der Gebrauch"67

Thus, we postulate one's character as a hidden source for one's behaviour in the same way in which we postulate meaning as the hidden source of linguistic use: it is because the uniformity of conduct springs from one's (unchanged) character, that a change in one's conduct must be explained through a change in one's character; analogously, the uniformity of lingustic use springs from a stable 'hidden' entity, meaning, and so on. In both cases, we do postulate sources of uniformity which, we think, explain the facts as we know them. But what we actually do in such cases, is to obscure the real diversity of things and acts by assimilating them into types. It deserves renlenlbering that, according to the classical views, the meaning of a word reflects nothing else than the basic features common to all things which are designated by that word: the meaning of 'leaf reflects what is common to all leaves. Consequently, Wittgenstein's analogy between the epistemological roles played by character and meaning is pretty close to Nietzsche's analogy between the roles played by Blatt, as an Urform, and Ehrlichkeit. In all these cases, we postulate 'hidden' entities in order to create a unique, general source for what experience presents us as diverse, multiple, particular; by thus creating 'deeper' uniformities (springing from the 'hidden' entities), we treat different things as 'gleich' or as being 'the same" that is we commit the error of assimilation. Against this received view, both Nietzsche and Wittgenstein insist that 'sameness' is not to be found in reality neither as identity of features ('Gleichheit'), as Nietzsche stresses, nor as 'shared essential features' as Wittgenstein argues. Unifornlity and unity (in this respect) are created by our conceptualization techniques. In fact, there are only the individual things which are sometimes similar, for things "sind mit einander in vielen verschiedenen Weisen verwandt" (PU, §65). Source 5: Naming Uniformisation can also be seen as a product of the linguistic process by which similar things receive a name and thereby a common identity. This is, of course, a process by which "die Vielheit der Dinge wird unter einen Hut gebracht", that is a process by which some sort of unity is linguistically imposed upon things. Nietzsche is convinced that the unity thus created is purely linguistic and conceptual, having no real or 'natural' counterpart; he considers that unity is merely ap-

67

140

BGM, I, §13.

parent ("wir haben inlmer nur einen Anschein von Einheit"68) and, when referring to different particular cases, he insists that "auch hier, wie so oft, verbUrgt die Einheit des Wortes Nichts fUr die Einheit der Sache"69. Mentioning the case of mental activities, for instance, Nietzsche remarks that "diese sind eine Vielhait, hinter der es nicht nothig ist eine Einheit auszusetzen't7o. His main point is that unity comes into existence by our using of the sanle word for several, different things, as, for instance, in the case of states of mind which receive the same name (Morgenrothe, §502). He describes the process of naming (the creation of a common linguistic identity for things) as a process by which we extend the application of a name from some particular cases to many others, obscuring thus many individual differences: "Oer Begriff entsteht aus einem Gleichsetzen des Nichtgleichen: d.h. durch die Tauschung, es gabe ein Gleiches, durch die Voraussetzung von IdentiUiten: also durch falsche Anschauungen. Man sieht einen Mensch gehen: nennt es 'gehen' Jetzt einen Affen, Hund: sagtauch 'gehen"171

This way of extending the use of a name is not absolutely arbitrary; Nietzsche was prepared to admit that there are some similarities among the different cases/things which are covered by the same name. But he always insists that similarity is not sameness, and that what we do, in fact, is "das Aehnliche als gleich zu behandeln"72 or "Ahnliches mit Ahnlichenl identificiren"73. The fact that Nietzsche sees the process of naming as a process by which we extend the use, from some particular cases to others, is, I think, extremely relevant, proving his departure from the classical view. Classical thinkers (not only speculative philosophers like Schopenhauer, but also contenlporary authors like Rudolf Carnap), being convinced that the use of the same name for different things is justified by the existence of some basic uniformity ('common features', 'general elements' etc.), have never seen the same process in this light: for them, it is never a matter of extending a name from some particular things/acts to others, but rather a matter of applying the same name to all the particular things/acts which meet some condition, i.e. which have the relevant 'common features' which are 'captured' in the meaning of that name. (One is thus reminded of Camap's famous definition of meaning, 68

69 70 71 72 73

KGW, VIII 1, 5(56), p.209. See also the same volun1e, 2(87), p.l02, and Der Wille zur Macht, §536 where he says: "Was aber wirklich, was wahr ist, ist weder Eins, noch auch reduzierbar auf Eins". KGW, IV2, §14, p.13. KGW, VII3, 40(38), p.379. KGW, 1114, 23( 11), p.138. KGW,V2,§111,p.149. KGW, 1114, 19(217), p.75. See also the same volume, 19(236), p.8\.

141

which he tried to apply to natural language too: "the intension of a predicate 'Q' for a speaker X is the general condition which an object Y must fulfil in order for X to be willing to ascribe the predicate 'Q' to y tt 74) From such a classical point of view, the use of a name is not extended, since all things 'of a certain kind' share some characteristic features and they all receive simultaneously the same name in virtue of their shared features. One naturally interprets meaning in terms of 'extending' the use only when one has abandoned the scheme based on 'comnlon features' and 'individual particularities': if there are no 'common features' to guide us in applying names, then the thing to do is to give names to some particular things and, if possible and useful, to extend the use of these nanles for things which are sufficiently similar. Nietzsche's interpretation seems to be closer to this one, than to the classical one. But of course that the philosopher who explicitly defended this non-standard interpretation is Wittgenstein and his description of the way in which the use of names is extended sounds now familiar: "Warum nennen wir etwas 'Zahl'? Nun etwa, weil es eine - direkte - Verwandtschaft mit manchem hat, was man bisher 'Zahl' gennant hat; /.. ./ Und wir dehnen unseren Begriff der Zahl aus, wie wir beim Spinnen eines Fadens Faser an Faser drehen" (PU, §67).

There is no insistence here upon 'forgetting differences' or upon 'treating as gleich', but the interpretation of the process by which we create a common identity for things is quite the same. Both Nietzsche and Wittgenstein claim that we give the same name to similar things by extending the use of a word from some particular cases to others, not by grouping together things which are essentially 'of the same kind', Le. which are gleich or which share some common features. And for both of them the only justification of this extending-procedure is the existence of similarities; all others are dismissed. Nietzsche rejected the idea that individuals (that we give the same name) 'naturally' belong to the sanle 'Gattung' exactly as, later on, Wittgenstein rejected the idea that such individuals 'naturally' share the same core of 'common features'.

* There is also a common explanation for this extending-procedure. It is the obvious need to use the same words in several, different circumstances that makes us proceed in this way:

74

142

Rudolf Carnap, Meaning and Synonimy in Natural Languages, in Meaning and Necessity, The University of Chicago Press, 1970, p.242.

"Jedes Wort wird sofort dadurch Begriff, dass es eben nicht fur das einmalige ganz und gar individualisirte Urerlebnis, dem es sein Entstehen verdankt, etwa als Erinnerung dienen soIl, sondern zugleich flir zahllose, mehr oder weniger ahnliche, das heisst streng genommen niemals gleiche, also auf lauter ungleiche Falle passen muss"75

The general character of a concept is thus justified not by the existence of a 'core' of general features (captured jn the meaning of this concept) shared by all things to which the concept correctly applies, but by purely practical needs, viz. by the need to use the same word in many different contexts. ' Wittgenstein had essentially the same opinion. The fact that concepts are not 'one use-instruments' and that we apply them in different occasions makes us think that this uniformity of use is determined by a uniformity of facts, namely by the constant occurrence of the 'same' situations, although with some variations: "Aber wir, in unserer Begriffswelt, sehen immer wieder das Gleiche mit Variationen wiederkehren. So fassen es unsere Begriffe auf Die Begriffe sind ja nicht flir einmaligen Gebrauch"76

It is thus language which makes us inlpose (conceptual) uniformity upon real (ontological) diversity. But, since we do extend characteristics of language to the real world itself ("Man pradiziert von der Sache, was in der Darstellunsweise liegt', - PU, §104) and since we cannot get rid of the resulting ontological image ("Ein Bild hielt uns gefangen. Ond heraus konnten wir nicht, denn es lag in unsrer Sprache, und sie schien es uns nur unerbittlich zu wiederholen"- PU, § 115), or, as Nietzsche remarked, we are "in den Netzen der Sprache eingefangen"77, we are bound to believe in an ontological uniformity 'behind' the acknowledged diversity.

* Finally, a peculiarity of Nietzsche's conception about conceptual assimilation is that the inclusion of a particular thing in a Gattung (as being gleich to other things in the same category) appears to be the result of a voluntary decision, or, more exactly (and more speculatively), of a decision taken by Der Wille zur Macht. According to Nietzsche, it is the Will to Power which decides that something is soand-so, and therefore is 'the same thing as', and thus that it should be included in a certain category:

75 76 77

KGW, 1112, p.373. Z, § 568. KGW, 1114, 19(135), p.5l.

143

"Der Wille zur Gleichheit ist der Wille zur Macht - der Glaube, daB etwas so und so sei, das Wesen des Urtheils, ist die Folge eines Willens, es soli so viel als moglich gleich sein"78

One should not expect, of course, to find in Wittgenstein's work an analogue to this speculative (but not uninteresting) thesis of Nietzsche. But it is very remarkable, I think, that one does find there the idea that we want to decide freely (and we do) upon things being or not 'of the same kind'. As we build concepts by extending the application of names, we are not compelled by 'the nature of things' to include a certain thing in a certain category (among other things already designated by that particular name) - rather we 'reserve' the right to decide upon the matter in every particular, new case. For instance, as there are no 'common, essential features' characterizing all games, there is no general condition for 'being a game' to be imposed upon us, and thus "Obrigens behalte ich mir vor, in jedem neuen Fall zu entscheiden, ob ich etwas zu den Spielen rechnen will oder nicht"79

There is, of course, no question here of the role played by Der Wille zur Macht. But the idea that it is free hunlan will which decides what a thing is, whether it is or not sufficiently similar to certain other things, and consequently whether a certain name will be or not extended to it, is implied (at least tacitly) in both Nietzsche's and Wittgenstein's views. According to the classical philosophical views, it is the 'natural', 'objective' uniformity in the world (i.e., the fact that all things 'of the same kind' share some general features) that compells us to give certain names to certain things; it is not up to us to decide in such matters, all we can do is to check whether a pal1icular thing meets the condition required, the condition associated with a particular concept (this was the view behind RudolfCamap's definition mentioned above). Rejecting the idea that there must be such an ontological uniformity behind the diversity we perceive, both Nietzsche and Wittgenstein abandon this very special kind of (ontologico-semantic) detemlinism: there is nothing 'in the nature of things' to compel us to include a certain thing in a certain category and to give it a certain name. We are always free to decide on such matters, because we do not decide upon 'the true essence of things' or on the 'Ding an sich', but rather upon our own use of a concept. OUf decision is not arbitrary; we take into account similarities, and so on. But it is a free decision, not an act by which 'the general features' of a certain thing, its 'essence' makes us give it a certain name and include it among 78 79

144

KGW, VIllI, 2(90), p.104. PG, VI, §73.See also BPP, vol.I,§547.

other things 'of the same kind'. Once again, "Welche Art von Gegenstand etwas ist, sagt die Grammatik" (PU, §373).

145