Research and Perspectives on Processing Instruction

Research and Perspectives on Processing Instruction

≥

Studies on Language Acquisition 36

Editor Peter Jordens

Mouton de Gruyter Berlin · New York

Research and Perspectives on Processing Instruction by James F. Lee Alessandro G. Benati

Mouton de Gruyter Berlin · New York

Mouton de Gruyter (formerly Mouton, The Hague) is a Division of Walter de Gruyter GmbH & Co. KG, Berlin.

앝 Printed on acid-free paper which falls within the guidelines 앪 of the ANSI to ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data Lee, James F. Research and perspectives on processing instruction / by James F. Lee, Alessandro G. Benati. p. cm. ⫺ (Studies on language acquisition ; 36) Includes bibliographical references and index. ISBN 978-3-11-021532-8 (hardcover : alk. paper) 1. Second language acquisition. 2. Language and languages ⫺ Study and teaching ⫺ Psychological aspects. I. Benati, Alessandro G. II. Title. P118.2.L448 2009 418.00119⫺dc22 2009020707

Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de.

ISBN 978-3-11-021532-8 ISSN 1861-4248 쑔 Copyright 2009 by Walter de Gruyter GmbH & Co. KG, D-10785 Berlin. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Cover design: Sigurd Wendland, Berlin. Printed in Germany.

To Bernadette “She began her study of Italian wanting rules to memorize but then came to appreciate the power of Processing Instruction.”

Acknowledgements An anonymous reviewer once commented on our work that we cited ­VanPatten’s work to excess. The reviewer queried whether VanPatten was such a guru or if there were other theoreticians to cite. Without VanPatten there would be no Processing Instruction and indeed, he is the guru of this instructional intervention. We wish to acknowledge the impact that he and his work have had on our own. We are indebted to him for his work, his encouragement, and his friendship.

Table of contents Introduction Chapter 1: The theory of input processing underlying Processing Instruction Introduction: Working with input What is the current version of VanPatten’s theory of Input Processing? How have VanPatten’s processing principles evolved? How do the principles work? In what ways is the theory lacking? Conclusion Appendix: tables Chapter 2:  What makes Processing Instruction effective? Introduction What is Processing Instruction? What are the main components of Processing Instruction? How do you develop structured input activities keeping ­psycholinguistics principles in mind? What are the effects of structured input activities? Summary Appendix: tables Chapter 3: How does Processing Instruction compare to other types of instruction? Introduction What is the role of grammar instruction? What types of grammar teaching have been suggested? What are the main findings of studies comparing Processing ­Instruction and traditional instruction? What are the main findings of studies comparing Processing ­Instruction and meaning output-based instruction? What are the main findings of studies comparing Processing Instruction and other grammar-instruction approaches? Conclusion Appendix: tables

xiii 1 1 3 6 7 24 25 27 36 36 37 39 46 55 62 63 68 68 69 71 77 81 86 88 91

x 

Table of contents

Chapter 4: Can Processing Instruction be delivered effectively online as well as in classrooms? Is online delivery of Processing Instruction an effective means of ­delivering Processing Instruction? How does the online delivery of Processing Instruction compare to the online delivery of MOI? Summary Appendix: tables Chapter 5: Can you increase the positive effects of Structured Input on language development by enhancing it aurally and/or ­textually? Will enhanced and unenhanced structured input help learners process non-meaningful and redundant forms? Can enhanced Structured Input help push learners to process a ­meaningful form over a lexical item? Will enhanced and unenhanced structured input help direct learners’ attention to meaningful morphology in sentence-final position? Summary Appendix: tables Chapter 6: What are the transfer-of-training effects for Processing Instruction? Does training on Processing Instruction on a processing problem aid the acquisition of a different form affected by the same processing problem? What are the transfer-of-training effects and cumulative effects of processing instruction? Conclusion Appendix: tables Chapter 7: Are the effects of Processing Instruction durative (short‑term) and longitudinal (long-term)? Are the effects of Processing Instruction durative in the short-term? Are there longer term effects for Processing Instruction? Conclusion Appendix: tables

100 100 104 107 109

113 115 119 120 122 124 127 128 133 136 137 143 144 149 151 152

Table of contents 

Chapter 8: How have the effects of Processing Instruction been ­measured? What are the effects of Processing Instruction on discourse-level ­production tasks? What are the effects of processing instruction discourse-level ­interpretation tasks? Appendix: tables

xi

174 175 178 181

Chapter 9: Conclusion: What does the research on Processing Instruction tell us? Final comments

184 191

References Author index Subject index

195 211 213

Introduction This book will track the impact processing instruction has made since its conception. It will provide an overview of new research trends on measuring the relative effects of processing instruction. Firstly, we explain Processing Instruction, both its main theoretical underpinnings as well as the guidelines for developing structured input practices so that readers can critically evaluate this approach to grammar instruction. Secondly and more importantly we review the empirical research conducted, to date, on processing instruction so that readers will have an overview of new research carried our on the effects of processing instruction. Finally, we will reflect on the generalizability and limits of the research on processing instruction and we will offer future directions of processing instruction research. In chapter 1 we provide a synopsis of VanPatten’s theory of input processing, the one that most directly and greatly informs the practices of Processing Instruction (VanPatten 1996, 2003, 2004). We will then begin our explanation of the practice of Processing Instruction by defining and illustrating “structured input”, the concept most crucial to understanding PI. In chapter 2 we present and examine Processing Instruction. This approach to grammar instruction comprises of two elements, explicit information about the grammatical item including information about processing strategies and structured input activities. We refer to these elements as full processing instruction. The question is whether one or the other of these elements is the causal variable in processing instruction or whether both elements are necessary to achieve the effects of processing instruction. We will review research that has investigated what is the main variable responsible for the positive effects of processing instruction. We highlight the remarkable consistency of the findings. They are: (1) learners who receive no explicit information but only perform structured input activities perform equally to learners who receive full processing instruction; (2) receive full processing instruction outperform learners who receive only explicit information. We conclude that the element of full processing instruction that causes changes in language development is structured input. In chapter 3 we present studies which have compared processing instruction to other types of instruction. The effects of processing instruction have been compared to those of two other types of instruction, both of which

xiv 

Introduction

e­ mphasize language production, not language processing. They are traditional instruction with form-focused output practice and meaning-based output instruction with communicatively-focused output practices. In the review we will provide detailed analyses of the languages examined, the linguistic items tested, and the assessment tasks used. For this synopsis, we highlight the remarkable consistency of the findings. They are: 1. In all studies, learners who receive processing instruction outperform learners who receive traditional instruction on an interpretation task; 2. learners who receive processing instruction perform the same as learners who receive traditional instruction on a production task; and, a finding limited to just two studies, 3. learners who receive processing instruction perform the same as learners who receive meaning output-based instruction on both the interpretation and production tasks. As will be seen, we find that learners who receive processing instruction outperform learners who receive meaning output-based instruction on the interpretation task and are equally successful as meaning output-based instruction learners on the production test. All in all, we will present processing instruction as an “educational bargain” in which learners receive two abilities (interpretation and production) as a result of practicing how to process input appropriately. In chapter 4 we review classroom research that has been carried out to measure the positive effects of processing instruction in a different mode of delivery. Processing instruction seems to be equally effective in promoting second language development no matter the mode of delivering the instruction. Research has shown that classroom and computer delivery of processing instruction yielded identical results across three languages (Spanish, Italian, French) and three linguistic items/structures (preterit–imperfect distinction, negative informal commands, subjunctive of doubt). In chapter 5 we review research that has examined a variety of grammatical items that present processing problems to second language learners. Both structured input activities and enhanced structured input activities have been designed to address these processing problems. Empirical research has been conducted to examine different languages (Italian, Spanish and Japanese) and linguistic features (adjective agreement, future tense, past tense, and subjunctive). Learners improved equally by being exposed to structured input activities and enhanced structured input activities. In chapter 6 we review an unique line of research within the processing instruction model that has attempted to assess the secondary and cumulative effects of this approach on grammar instruction. Research on processing instruction has mainly focused on measuring its direct and primary effects by

Introduction 

xv

comparing this type of instruction with traditional and meaning-output based instruction. The results of the empirical research have shown that processing instruction is a better approach to output-based approaches to grammar instruction. Processing instruction is very effective approach towards altering inappropriate processing strategies and instil appropriate ones in L2 learners. The main aim of this chapter is to review classroom research that has investigated whether learners receiving processing instruction can transfer that training on the acquisition of other forms without further instruction. In chapter 7 we review and examine research which has demonstrated that processing instruction is an effective approach to grammar teaching not only at sentence level interpretation and production measures but also at discourse level production tasks. In chapter 8 we review and discuss research that has measured long-term effects of processing instruction. In the final chapter we critically review classroom research and main findings of processing instruction research. We will then provide an agenda for future research in this area.

Chapter 1 The theory of input processing underlying Processing Instruction Introduction: Working with input Different researchers and different theoreticians focus on various aspects of second language acquisition. Some focus on input while others on output. Some model the cognitive mechanisms that occur in the brain while others model the social and interactional dimensions of second language development. In essence, researchers can easily look at SLA from different perspectives.1 VanPatten and Williams (2007b: viii) assert that researchers understand that in order to understand the whole of SLA, they may need to concentrate on the smaller parts first. VanPatten (2004b: 27) likens SLA to building construction. “In a sense, understanding SLA is like understanding how a building works. There is the electrical system, the plumbing, the foundation, the frame, the heat and air system, and so on. All are necessary; one alone is insufficient. But like those who work in house construction and are electrical contractors or plumbing contractors, in SLA some of us are interested in matters dealing with input. Others are interested in output.” In the present work, we are concerned with input and VanPatten’s theory of input processing. We are, in particular, interested in a  pedagogical intervention called Processing Instruction that teaches learners target-language appropriate ways to work with input. 1.  VanPatten and Williams (2007) offering the following regarding looking at SLA from different perspectives. “To understand [why there isn’t just one theory to account for SLA], one might consider the parable about the four blind men and the elephant. These sightless men chance upon a pachyderm for the first time and one, holding its tail, says, “Ah! The elephant is very much like a rope.” The second one has wrapped his arms around a giant leg and says, “Ah! The elephant is like a tree.” The third has been feeling along side the elephant’s massive body and says, “Ah! The elephant is very much like a wall.” The fourth, having seized the trunk cries out, “Ah! The elephant is very much like a snake.” For us, SLA is a big elephant that researchers can easily look at from different perspectives …Thus, researchers have grabbed onto different parts of the elephant as a means of coming to grips with the complex phenomenon. (VanPatten and Williams 2007: vii–viii).

2 

The theory of input processing

What is input? The following four definitions suffice to demonstrate that input is language, presented orally or in written form. “The raw linguistic data (oral or written) to which learners are exposed.” (Farley 2005: 109). “Samples of language that learners are exposed to in a communicative context or setting.” (Wong 2005: 119). “Samples of second language that learners hear or see to which they attend for its propositional content (message).” (VanPatten 1996: 10). “Input is defined as language the learners hears (or reads) and attends to for its meaning.” (VanPatten and Williams 2007b: 9). Of additional importance is that input is language presented in a  communicative context, meaning that learners are attending to the meaning of the message(s) encoded through the language directed to them. What is input processing? Processing Instruction is rooted in what we know about what learners do with input. The process it for its meaning and that meaning is formally encoded. Input processing, then, refers to the processes by which learners make the initial connection between a  grammatical form and its meaning. That is, we are concerned with how learners make sense out of the language they hear or read (input) and how they get linguistic data or intake from the input (Wong 2005: 28) “Indeed, it is common ground among all theorists of language learning, of whatever description, that it is necessary to interpret and to process incoming language data in some form, for normal language development to take place. There is thus a  consensus that language input of some kind is essential for normal language learning.” (Mitchell and Myles 2004: 20). Ortega (2007: 236) reviews the role of input in the nine theories of second language acquisition included in VanPatten and Williams (2007a). The role of input in the theories varies. Input might be one ingredient only that is necessary for language acquisition but not sufficient to account for all language acquisition. Input may be a trigger. Or, input may be the driving factor in learning. VanPatten and Williams (2007b: 9) assert that any theory of second language acquisition will have to address in some way the observed phenomenon, or consensus as Mitchell and Myles call it, that exposure to input is necessary for SLA. As stated above, we are working with VanPatten’s theory of input processing as presented in its initial form in VanPatten 1996, its modified form in VanPatten 2004b and its most recent form in VanPatten 2007. The purpose of the present chapter is twofold. First, we seek to explain what VanPatten’s theory of input processing entails. This theory provides the background for understanding Processing Instruction. Second, because the present work is a retrospective on Processing Instruction, we will highlight

VanPatten’s theory of Input Processing 

3

the development and evolution of the theory from 1996 to 2007 (the most recent material ­available). What is the current version of VanPatten’s theory of Input Processing? VanPatten’s theory of input processing in adult second language acquisition frames the research questions, methods, and procedures used in all of the many investigations we review in this book. It is critical, then, that we begin with an explication of this theory. We draw from several sources to present our account. We draw extensively from the work of its principal theorizer, Bill VanPatten (VanPatten 1996, 2000, 2004b and 2007), as well as from our own work with and within this theoretical framework (Benati and Lee 2008; Lee and Benati 2007a, 2007b). As a theoretical framework, “Input Processing is concerned with three fundamental questions that involve the assumption that an integral part of language acquisition is making form–meaning connections: –– Under what conditions do learners make initial form–meaning connections? –– Why, at a given moment in time, do they make some and not other form- meaning connections? –– What internal strategies do learners use in comprehending sentences and how might this affect acquisition?” (VanPatten 2007: 116) We can add to this list of three umbrella questions more specific ones, the answers to which the research on input processing has attempted to illuminate. –– What linguistic data do learners attend to during comprehension? Why? –– What linguistic data do learners not attend to? Why? –– How does a formal feature’s position in the utterance influence whether it gets processed? –– What grammatical roles do learners assign to nouns based on their position in an utterance? In its current form, VanPatten’s theory consists of two overarching organizing principles of input processing, each of which is further explicated with (sub)principles. The two overarching principles address two different aspects of processing. The first, The Primacy of Meaning Principle, asserts that when learners are engaged in communicative, meaningful interchanges, they are

4 

The theory of input processing

primarily concerned with meaning. That is, “… learners are driven to look for the message or communicative intent in the input.” (VanPatten 2004b: 7) The second, The First Noun Principle, asserts that the order in which learners ­encounter sentence elements is a powerful factor in assigning grammatical ­relations amongst sentence elements. In relation to this principle, VanPatten has commented that, “… the human mind may be predisposed to placing agents and subjects in a  first noun position.” (VanPatten 2004: 15). These principles and their associated (sub)principles appear in Table 1.1. In this table we have traced the evolution of VanPatten’s theorizing. We have presented the principles at three points in time, specifically 1996, 2004 and 2007. In their current form, the two main principles are as follows. Principle 1. The Primacy of Meaning Principle. Learners process input for meaning before they process it for form. (VanPatten 2004: 11) Principle 2. The First Noun Principle. Learners tend to process the first noun or pronoun they encounter in a sentence as the subject. (VanPatten 2007: 122) The Primacy of Meaning Principle is further subdivided into six (sub)principles, labelled a through f. Some of these subprinciples had previously been referred to as the corollaries of the main principle (VanPatten 1996) whereas others are new developments to the framework. Each new development is meant to add to the explanatory adequacy and predictive capability of the theory. As we read through these subprinciples, we find that they are meant to capture the interplay of various linguistic and cognitive processes that take place during comprehension. The principles associated with the Primacy of Meaning Principle are, in their most current formulation, as follows. P 1a. The Primacy of Content Words Principle: Learners process content words in the input before anything else. (VanPatten 2007: 117) P 1b. The Lexical Preference Principle: If grammatical forms express a meaning that can also be encoded lexically (i.e., that grammatical marker is redundant), then learners will not initially process those grammatical forms until they have lexical forms to which they can match them. (VanPatten 2007: 118) P 1c. The Preference for Nonredundancy Principle: learners are more likely to process nonredundant meaningful grammatical markers before they process redundant meaningful markers. (VanPatten 2007: 119)

VanPatten’s theory of Input Processing 

5

P 1d. The Meaning-Before-Nonmeaning Principle: learners are more likely to process meaningful grammatical markers before nonmeaningful grammatical markers. P 1e. The Availability of Resources Principle: for learners to process either redundant meaningful grammatical forms or nonmeaningful forms, the processing of overall sentential meaning must not drain available processing resources. (VanPatten 2004b: 14) P 1f. The Sentence Location Principle: learners tend to process items in sentence initial position before those in final position and those in medial position. (VanPatten 2007: 125) When we listen to an utterance or read a sentence we are presented the linguistic elements of the sentence in a rigidly linear fashion. One sentence element precedes the next such that we must, there are no options, comprehend and interpret the sentence “as it comes” to us. While regression is possible in some reading contexts, it is rarely possible in listening contexts. Research in both first and second language acquisition has found that the order of the words plays a role in comprehension and hence in language acquisition (e.g., Slobin 1973 for first language acquisition and Lee 2003 for the second language acquisition of Spanish). VanPatten’s First Noun Principle captures one powerful and pervasive processing strategy, that is, assigning the grammatical role of subject or agent to the first noun encountered in an utterance. VanPatten has recently acknowledged that learners might transfer L1 parsing procedures to the L2 processing context (VanPatten 2004c: 330; 2007: 122). In doing so, he has proposed the possibility that his theory might yet incorporate “The L1 Transfer Principle. Learners begin acquisition with L1 parsing procedures.” (VanPatten 2004: 330). This principle has not been fully incorporated into the theory; L1 transfer remains a possibility, and yet, emerging data suggest that L1 transfer may not take place (VanPatten and Keating 2007). Between 1996 and 2004, researchers gathered more data on the conditions that favour or attenuate learners’ misassignment of the first noun as subject so that VanPatten developed a  set of (sub)principles that delineate various factors that attenuate learners’ misassignment of the first noun. The subprinciples are as follows. P 2a. The Lexical Semantics Principle: learners may rely on lexical semantics, where possible, instead of the First Noun Principle to interpret sentences. (VanPatten 2007: 124)

6 

The theory of input processing

P 2b. The Event Probabilities Principle: learners may rely on event probabilities, where possible, instead of the First Noun Principle to interpret sentences. (VanPatten 2007: 123) P 2c. The Contextual Constraint Principle: learners may rely less on the First Noun Principle (or L1 transfer) if preceding context constrains the ­possible interpretation of a clause or sentence. (VanPatten 2007: 124) These principles model “what guides learners’ processing of linguistic data in the input as they are engaged in comprehension” (VanPatten 2007: 116). In the following section, we will explicate each of these principles in turn and, in doing so, demonstrate some of the evidence that supports them. It is important to keep in mind that learners are doing two things with the language to which they are exposed and with which they are engaged. They are, firstly, making meaning and they are, secondly, making form–meaning connections (Lee and VanPatten 1995, 2003). Making meaning is comprehending, arriving at an idea of what the propositional content of the message is. Making form–meaning connections is input processing, attending to the grammatical forms/features in the input so as to connect the forms with their meanings or functions. While related, these are not the same processes. As we further explore VanPatten’s theory of input processing we will see both types of processes at work. How have VanPatten’s processing principles evolved? In Table 1.2 we have placed VanPatten’s original formulation of the theory along with the research on which it was based; VanPatten (1996) is the source for Table 1.2. In this table we see that the theory first contained three principles with associated subprinciples. It’s current formulation is quite different. We also see in this table that VanPatten relied on a wide range of research and theorizing in order to develop his principles. The basis of the theory includes studies of both first and second language acquisition as well as work on both processing and production. The 1996 work remains the most detailed account of the previous research and theorizing that led to the formulation of VanPatten’s processing principles. VanPatten’s subsequent work focused on explaining the workings of his principles, clarifying them, accounting for criticisms, and revising the principles as needed (VanPatten 2004b, 2004c, 2007). Lee and Benati (2007a, 2007b) have brought to the discussion of input processing research other than that cited by VanPatten (1996) that support his formulation of the principles.

How do the principles work? 

7

How do the principles work? The push to make meaning To assert the primacy of meaning in input processing is to take as the point of departure that learners are primarily motivated to understand messages, be they delivered orally during an interaction or visually while reading print. If someone is talking to us, we assume they have something to say that we are meant to understand. Our task as listeners is to put forward at least an effort, if not our best effort, to understand the speaker. When we see an advertisement, for example, and read what it says, we assume that someone has something to communicate to us about a product, event, or service. There is a message that we are meant to grasp and we put forth the effort to do so. Second language learners assume the same thing; there are messages in what they hear and read and they are meant to put forward an effort to understand them. “Simply put, P1 states that learners are driven [emphasis added] to look for the message in the input (“What is this person saying to me?”) before looking for how that message is encoded” (VanPatten 1996: 17). Meaning first. VanPatten (1996: 17) supported Principle 1 The Primacy of Meaning Principle with work in first and second language acquisition. For first language acquisition, he cited Peters’ (1985) operating principle that guides children during input processing. The principle states that children pay attention to utterances that have a readily identifiable meaning. For second language acquisition, he cited the work of Sharwood Smith (1986), who posited the difference between processing for communication, i.e., meaning, and processing for acquisition, i.e., form. Research has repeatedly uncovered the varying conditions under which learners successfully make meaning from the input. Lee (1987), for example, showed that L2 learners of Spanish can extract the lexical meaning of verbs that are morphologically marked as subjunctive even though they had never been exposed to subjunctive forms in the classroom setting. They extracted meaning as successfully as a  group of learners who had already been taught subjunctive forms. Lee and Rodríguez (1997) compared the effects of morphosyntactic modifications on passage comprehension. Keeping content constant, they manipulated subordination and whether that subordination required subjunctive mood or not. They found that L2 learners of Spanish comprehended the three versions of the passage equally well. Additionally, they substituted the target verbs (those that were subordinated and made into subjunctive mood forms) with nonsense words that conformed to the

8 

The theory of input processing

o­ rthographic structure of Spanish. This substitution had no effect on passage comprehension. Manipulating both verbal and lexical forms did not affect passage comprehension because the readers’ task was to get the meaning of the text and they did. More evidence for how learners process input for meaning before they process it for form comes from the recall data reported in Lee (2002, 2003). The learners read a passage about the future of telecommunications technologies in which the last few sentences in the passage warned of the dangers of a  society increasingly dependent upon technology. The last sentence they encountered was as follows. Text: El hombre, Homo sapiens, se convertirá en Homo electrónicus. Translation: Man, Homo sapiens, will become Homo electrónicus. Most learners understood the meaning of the sentence and understood its meaning in the context of the passage. Few, however, wrote the exact form of what they had seen in the text. That is, few learners wrote Homo electrónicus. Some learners produced graphemically-based renderings such as Homo electricity and Homo erectus. Others made a semantic substitution such as Homo technology and Homo technologicalus. Clearly, the semantic substitutions show us that learners processed the input for meaning before they processed it for form. Lee and Rossomondo (2007) analysed other elements of the input passage reported on in Lee (2002, 2003) and Rossomondo (2006). The passage they used in their research targeted learners’ processing of future tense verb forms in Spanish, which are morphologically marked word finally for person/ number and tense. The first verb in the passage was dependerá will depend. Their analyses revealed that learners recalled this verb in a variety of forms. The forms varied but the meaning always centred on the idea of dependence. They noted both verbal and nominal renderings of the target verb. Among the verbal forms they found will depend on, depend or depends on, rely upon, relies, and will use. Among the nominal forms they found dependency, dependence upon, are dependent, and will become dependent. These forms show us that learners were primarily working to get the meaning not the form. Content words. VanPatten’s theory differentiates between the value of content words and function words for their contribution to meaning from the learners’ perspective and from the perspective of the push to get or make meaning. Which words are the most helpful for getting the meaning out of the input? The answer is content words, those words that represent major lexic-

How do the principles work? 

9

al categories as opposed to functional or minor lexical categories. VanPatten (1996: 19–20) supports the subprinciple of content words with research by Klein (1986) and Mangubhai (1991). Klein (1986) showed that when asked to repeat utterances, early stage learners tended to repeat only the content words. Only advanced level learners could repeat the utterances correctly with content words plus functors. Mangubhai (1991) showed that learners who were being taught via Total Physical Response methodology routinely extracted the content words from the stimuli commands in order to physically respond. Carroll (2004: 299) hypothesized that “a number of distinct phonetic cues might lead learners to segment and phonologically encode words from the major lexical classes of English precisely when they are realized as prosodic words, and that would lead them not to segment and encode clitics (determiners, auxiliaries, complementizers, tense morphemes, number morphemes, etc.) This is true despite the fact that many of the functional categories express important semantic distinctions.” Her hypothesis is useful for considering why content words are attended to over function words. In layman’s terms we might refer to content words as the “big” words and functional words as the “little” words. Learners must bring some metalinguistic knowledge with them to the task of second language acquisition such that they can differentiate content and function words in the L2. In other words, “… second language learners in particular know there are “big words” that can help them get the meaning of what is being said to them and their internal processors attempt to isolate these aspects of the speech stream during comprehension.” (VanPatten 2004: 8). We now present several empirical works that have demonstrated the greater value of content words to second language learners. Bernhardt (1992) discussed the different text processing strategies employed by native and inexperienced nonnative readers of German. In tracking their eye movements across the lines of a text, she showed that native readers of German fixated (i.e., placed their central focal point) far more frequently than inexperienced nonnative readers did. In other words, they read more densely and intensely than the nonnatives did. Moreover, she found that among the native readers’ more numerous fixations were those they placed on the ends of words, that is, on word final morphology. The nonnative readers tended to fixate on the centers of words leaving word final morphology in peripheral vision. And, with their fewer in number fixations, nonnative readers tended to process content words over function words. This eye movement data is very interesting because it contrasts the approaches natives and nonnative readers take to processing. Nonnative readers, the language learners,

10 

The theory of input processing

valued content words highly and valued word final morphology to a lesser degree. The eye movement evidence very directly supports the value of content words to learners. In VanPatten (1990) we see what happens when we focus learners on word final morphology. VanPatten (1990) conducted an experiment in which he demonstrated the interplay of content words, function words, and verb morphology with comprehension. He asked learners of Spanish to listen to a short passage on inflation in Latin America. He assigned learners to one of four conditions. One group simply listened to the passage and had no secondary processing task to perform; he termed this the content only group. Another group listened to the passage and indicated each time they heard the word inflacíon, which was also the title of the passage. He termed this the content + lexical item group. A third group listened to the passage and indicated each time they heard the word la, the feminine singular form of the definite article. It occurred prior to each occurrence of the word inflación. He termed this the content + functor group. Finally, a fourth group listened to the passage and indicated each time they heard an n at the end of a word. This verbal inflection is the morpheme that marks third person plural in Spanish. He termed this group the content + inflection group. As they listened to the passage, the three groups with simultaneous listening tasks placed a checkmark on a page for each occurrence of the target item. After listening to the passage, the learners recalled as much as they could of what they had heard. VanPatten found that listening for content alone and listening for content + lexical item were complimentary activities insofar that learners in both conditions comprehended equal amounts of the passage. Listening for the functor and for the verbal inflection were equally detrimental activities in that comprehension fell off significantly in these conditions. From Bernhardt (1992) we know that learners tend not to process word final morphology and from VanPatten (1990) we know that if they are directed to process the word final morphology, they loose some of the meaning. These data support the thesis that content words are the building blocks of comprehension. Drawing learners’ attention to non-content elements of a passage, be they verb morphemes or definite articles, causes learners to loose some of the content. We have more evidence of how learners process content words over other sentence elements from examining the products of comprehension. VanPatten and Wong (2004) demonstrated that learners misinterpret French causative constructions using an inappropriate word order processing strategy. They give the following example (VanPatten and Wong 2004: 98–99).

How do the principles work? 

11

Jean fait promener le chien à Marie. John makes to walk the dog to Mary ‘John makes Mary walk the dog.’ The target sentence contains two verbs each with its own subject/agent. Learners, however, tend to take the first subject, Jean, and make it the agent of the second verb, promener. The second subject, Marie, tends to be interpreted as the dog’s owner. In the end the learners misinterpret the sentence (by misassigning grammatical and semantic roles) to mean the following. John walks the dog for Mary. Whereas VanPatten and Wong address this processing problem from the perspective of word order and P2 The First Noun Principle, we can also see the effects of processing content words over other sentence elements. We underline the content words to demonstrate that they are the words learners focused in on. Jean fait promener le chien à Marie. Two important grammatical elements are not processed, fait and à. They are important because they signal the underlying semantic relationships between Jean and Marie. Another example of how learners collect content words to make meaning comes from Lee (1990). After early stage learners read a short passage about feudalism, they were asked to recall in writing as much as they could remember. The first two sentences of the passage follow.  ext: Entre los años 900–1000, Europa Occidental estaba en gran desT orden. El imperio de Carlomagno, que había logrado unir esa parte del mundo, se estaba dividiendo en pequeños estados. Translation: Between the years 900 and 1000, Western Europe was in great disorder. The empire of Charlemagne, that had managed to unite that part of the world, was being divided into small states. One subject recalled these two sentences as follows (misspellings in the original are preserved here).

12 

The theory of input processing

I n the years around 900–1000, there was an emporer Carlomagna (sp?) that ruled an area in the world. (Lee 1990: 147) If we go back to the original text and underline the words that appear in the recall, we literally see the way the learner has collected words from the input and put them together. We also see the way he left out some words most likely due to the effortfulness of the task (Lee 1999; VanPatten 2007).  ext: Entre los años 900–1000, Europa Occidental estaba en gran desT orden. El imperio de Carlomagno, que había logrado unir esa parte del mundo, se estaba dividiendo en pequeños estados. This learner did not process an important function word, the preposition de in the noun phrase imperio de Carlomagno (empire of Charlemagne). The preposition is important because it establishes the relationship between the two “big” words. The learner put the two content words together and misassigned the meaning of one of them, making it emperor not empire. What do emperors do? They rule, which is something the learner inserted into the recall. Another learner recalled these two lines as follows. Charlemagne was the ruler. He divided the country into small states … (Lee 1990: 146). The content words that this learner collected to make meaning attribute agency to Charlemagne. Attributing agency to Charlemagne occurred quite frequently in the recalls. The learners who did so did not process the entire verb phrase, se estaba dividiendo, but rather only used the content word, dividiendo, to make meaning. The verb phrase in the original text is marked for past (estaba) as well as for imperfective aspect (-aba) and progressive aspect (-ndo). The learner uses a perfective past form in his recall. We can, therefore, infer that he did not process these formal elements in the input. He collected content words but did not make use of their formal features in order to make meaning. We find additional evidence of the value of content words to language learners from their own accounts of what they do to understand. Lee (1999) analysed think aloud protocols for the interplay between input processing strategies and comprehension strategies. He asked learners to perform a retrospective think aloud of a  passage in which eight past tense verb forms were the targeted linguistic items. The following excerpt is one learner’s think

How do the principles work? 

13

aloud protocol that was not included in Lee (1999) but comes from the data set which he analysed. The tasks of the second language learner were to read the passage sentence by sentence and then think aloud his comprehension process. With regard to two of the target sentences he stated the following about his use of key words. From the context, we can understand “key words” as content words. Um, you can see from like different words kike, um, contacto and consecuencia, like the, um, different like structure of the sentence, like if you can get the key words like that then you can kinda see what should follow it… Um, a word like disco is kind of a humorous tip-off ’cause you can tell that’s something related to music. This learner collects content words to build up his comprehension. In all he refers to his use of words to comprehend sentence meaning seventeen times. At a later point in the think aloud he offers another example of using key/ content words and other words, too. Um, “filme”, um, “ofer-”, “oferta”, “aceptó”, “Hollywood” all give you like, you know, key words, like “offer”, “accept”, “Hollywood”, uh, would tell you what the sentence is going to be about. And those are all cognates, and then other words that you already know or can figure out would tell you pretty much the meaning of that sentence. Finally, the learners also makes a distinction between key words and another type, the small words. He stated the following. Um, múltiples is a good cognate. And then there’s a lot of pretty small easy words like “with” and “to see”, “is”. Another type of evidence supporting the primary role of content words in comprehension comes from Carroll (2004), who points to their role in the negotiation of meaning. She specifies content words as those in major lexical categories and refers to them as prosodic words. In a footnote to her commentary on VanPatten’s model of input processing she notes that content words have the linguistic properties that make them repeatable “as single utterances in situations where a speaker has failed to make herself understood and believes that the learner has limited language abilities” (Carroll 2004: 298). She provided the following example to underscore the point.

14 

The theory of input processing

NS: The exercises are all on my homepage. NNS: (…) NS: EXERCISES… HOMEPAGE NNS: oh… yes… EXERCISES   (Carroll 2004: 298) The native speaker has isolated and repeated the two content words from her initial utterance. Content words are not only important to learners but to native speakers also who, apparently, assign them value for insuring comprehension. Lexical preference. Languages have ways of encoding and th`us signalling the same information in multiple ways; we refer to this as the natural redundancy of languages. Given the availability of limited resources to second language learners VanPatten’s theory tries to account for where learners direct their processing resources. The background research VanPatten uses to support the Lexical Preference Principle has all been focused on tense assignment. Likewise, the research on the effects of Processing Instruction framed by the Lexical Preference Principle has focussed on tense assignment. The background research on tense assignment has manipulated the input to include or exclude lexical and grammatical cues to tense. Preferring lexical cues to tense is connected to learners’ use of content words to make meaning. “Learners will thus seek out lexical forms of semantic notions in the input before they seek grammatical forms that encode the same semantic notions.” (VanPatten 2007: 118). Lee, Cadierno, Glass and VanPatten (1997) gave two groups of learners of Spanish different versions of the same passage to listen to. One version contained lexical temporal adverbs referring to the past whereas the other version contained no such lexical temporal markers. In this second version, only the verb final morpheme indicated tense, for example, admitió ‘he admitted’. After listening, the learners performed a tense identification task. The results showed that the learners who listened to the passage with adverbials identified correctly more of the temporal references than did the learners who listened to passages with only verb morphology to mark tense. Lee (1999) examined the comprehension and input processing strategies of a small number of learners of Spanish as they performed a retrospective think aloud. Half the learners read a passage that contained lexical temporal adverbs whereas the other half read a version of the passage that did not contain the adverbs. As Lee (1999: 53) put it, “when subjects have adverbs they use them [to comprehend temporal reference]. Those in the +adverb condition only sporadically refer to verb forms.”

How do the principles work? 

15

Rossomondo (2007) showed a very dramatic difference in tense assignment due to the presence or absence of lexical temporal markers. She conducted introspective think aloud protocols on two groups of learners. One group was reading and introspecting on a  passage that contained Spanish future tense verb forms along with lexical temporal markers. The other group read and introspected on a version of the passage that contained only verb forms but no lexical temporal markers. She found that in the presence of a lexical temporal marker, learners comprehended (i.e., rendered it in the introspection) the future meaning of the verb form on average 52%; the range of scores per individual verb was from 0% comprehension of the future meaning to 92%. In the absence of the temporal marker learners comprehended the future meaning of the target verbs only .8%. The learners did not comprehend the future meaning of 12 of the 13 target verbs. They managed a future meaning for one verb only 11% of the time. Lee (1990) examined the recall protocols of first year learners of Spanish in order to reveal the processes through which they constructed the meaning of the passage. The passage referred to the social-political system known as feudalism. As noted above, the first line of the passage referred to the years 900–1000. The passage also contained the number 400, which was the number of years the feudal system helped maintain order, and a reference to the year 1200, the year feudalism was at its zenith. All learners read the same version of the passage, i.e., there was no version of the passage in which temporal markers had been removed. Relevant to the current discussion of the Lexical Preference Principle is that 11 of the 13 learners Lee (1990) examined used the years 900–1000 to construct the temporal framework for their recalls. One learner, who had a particularly difficult time comprehending used the years to structure the second part of his recall. In the beginning there was nothing. (STOP!) There was a society change in the era of Charlemagne a change from honor to a society of justice. I remember many years 900–start of Charlemagne 1100– ? 1200– justice and equality   (Lee 1990: 148) Musumeci (1989) conducted a cross-linguistic study (Italian, French and Spanish) in which she examined how successfully learners assigned tense at sentence level under different exposure conditions. She manipulated both

16 

The theory of input processing

linguistic and nonlinguistic cues to temporality. In one condition, the baseline condition, she presented learners with sentences in which the only cue to temporal reference was the verbal inflection. In each of the subsequent conditions the learners received the verbal inflections as well as some other cue or cues. In the next condition she added a lexical temporal adverbial as an additional cue to temporal reference. In the third condition, she supplied learners the additional cue of a typical teacher gesture performed as the learners heard the sentences. For example, to indicate past, the speaker pointed her thumb over her shoulder, pointing backwards. To indicate future, the speaker gestured with her hand moving it away from her body and outward toward the side. In the fourth condition Musumeci supplied learners with all the cues: verbal inflections, adverbials, and gestures. Overall, the results confirmed that the main factor influencing correct tense assignment was the presence or absence of temporal adverbials in the input sentences (Musumeci 1989: 127). Specifically, she found that learners assigned tense more correctly in the two conditions that included a lexical temporal adverbial than in the other two conditions, i.e., verbal inflections only and verbal inflections plus gestures. Lexical items were more useful to the learners for assigning tense than were the verb forms or a teacher’s ­gesturing. Redundancy and meaningfulness. VanPatten (1985b) introduced the idea that formal features of a second language, from a learner’s perspective, would be either of high or low communicative value. “Communicative value refers to the relative contribution a  form makes to the referential meaning of an utterance and is based on the presence or absence of two features: inherent semantic value and redundancy within the sentence-utterance.” (VanPatten 1996: 24). VanPatten extrapolated his original principle regarding processing meaningful versus nonmeaningful morphology (1996: 24) from the morpheme acquisition studies. Why, for example, in the progressive aspect marker -ing acquired before third person singular ‑s in English? VanPatten’s answer is that because ‑ing contributes unique information to the sentential meaning (an event in progress) learners direct attention to it during processing. They do not direct attention to the third person ‑s because it does not offer them unique information. He provides two examples (VanPatten 2007: 119). The cat is sleeping. The cat sleeps ten hours everyday. “Thus, if learners are confronted with something like ‑ing on verb forms, they will be forced to make this form–meaning connection sooner than, say, third

How do the principles work? 

17

person ‑s because the latter is redundant and the former is not” (VanPatten 2007: 119). As we saw with the Lexical Preference Principle, a grammatical marker might well have semantic value but other sentence elements might make it redundant. Removing the lexical item, as is done in Processing Instruction, makes the grammatical marker nonredundant. And so, the Preference for Nonredundancy Principle is highly related to the Lexical Preference Principle. Some grammatical markers do not carry any meaning; they express no real world semantic information. We can offer several examples (Lee and Benati 2007b). Grammatical gender marking in Romance languages is an example of non-meaningful morphology. The surface–level agreement features are such that adjectives agree in number and in grammatical gender with the nouns they modify. In other words, the characteristics of the noun determine the form of the adjective. That an adjective is marked as masculine or feminine does not change its meaning. In Italian, both bassa and basso mean short. Which form to use would be determined by the noun they would be describing. Additionally, gender markings on adjectives are often a redundant marking. Consider the noun phrase mio fratello basso (‘my short brother’) and the gender markings on the three words. The gender marking on basso is the third masculine marker in the string. Being nonmeaningful as well as redundant contribute to the processing problems second language learners face in processing these forms. Subjunctive mood verbal morphology is another grammatical form that is non-meaningful and redundant in sentences that express doubt and opinion in Italian and other Romance languages. We use the subjunctive mood markings on the verb in a subordinate dependent clause when the verb of the main clause expresses doubt or opinion. The following two sentences demonstrate the processing problems learners encounter with subjunctive mood morphology. Dubito che George sia intelligente. I doubt that George is intelligent. So che George e intelligente. I know that George is intelligent. The meanings of the verbs dubito and so trigger the forms sia and e, respectively. Sia and e mean exactly the same thing, is. In other words, “… such [nonmeaningful] formal features of language will be processed in the input later than those for which true form–meaning connections can be made.” (VanPatten 2007: 120).

18 

The theory of input processing

Resources. It is not impossible for learners to direct their attention to meaningful but redundant grammatical forms or to nonmeaningful grammatical forms. As VanPatten (1990) and Bransdorfer (1991) showed with their simultaneous processing tasks, learners can be directed to attend to nonmeaningful forms but at a loss to comprehension. “Comprehension for learners is initially quite effortful in terms of cognitive processing and working memory. This has consequences for what the input processing mechanisms will pay attention to. At the same time, learners are limited capacity processors and cannot process and store the same amount of information as native speakers can during moment-by-moment processing.” VanPatten (2007: 116). And so, VanPatten proposed the Availability of Resources Principle. Getting overall sentential meaning can not be overly effortful if learners are to process redundant meaningful grammatical forms or nonmeaningful forms. Lee (1999) analysed the comprehension and input processing strategies of second language learners. He states, “the comprehension strategies of low comprehenders may circumvent processing text for form. It is an interesting paradox to consider that learners’ attempts to manage their comprehension has the less than desirable effect of dislocating from their attention key aspects of the input.” (Lee 1999: 57). Comprehension difficulties can impede processing forms in the input. In the following example, the target form is the verb decidió (decided) or more specifically, the ‑ó morpheme indicating past tense. Learners performed a retrospective think aloud of a passage that contained eight target items. In the following think aloud, comprehension was so effortful that the learner miscomprehended the temporal and lexical meanings of the target form. The learner ultimately abandoned the attempt to make meaning. Text: Hace siete años decidió volver a la universidad para hacer un Master en leyes. Translation: Seven years ago, [he, Rubén Blades] decided to return to university to do a Masters in Law. Think aloud: Um, for seven years he, seven years old he’s coming to the university… but I’m not sure and seven sounds awfully young. I would probably just disregard that sentence. (Lee 1999: 50) Another example from this research also demonstrates the interplay between comprehension and input processing. In this example, the learner initially miscomprehends the verb actuó ‘he acted’ as the adverb actually. What follows is the learner’s efforts to make the meaning of the sentence make sense.

How do the principles work? 

19

Text: En la década pasada , actuó en varias películas, como por ejemplo, Crossover Dream y The Milagro Beanfield War. Translation: In the past decade, [he, Rubén Blades] acted in various films, for example, Crossover Dream and The Milagro Beanfield War. Think aloud: So, in the decade past, actually in various films. I don’t think that’s actually but it looks like actually so that’s why I guessed that. Um, maybe it’s act. Act in various films, like the example Crossover Dream and The Milagro Beanfield War. So that just speaks of like if you know the context of the paragraph you can follow along and probably guess that, that, uh, it’s talking about how he worked on the different [films]. Interestingly, the learners’ second rendering of actuó was ‘act’, not ‘acts’ nor ‘acted’. He uses an uninflected form. Moreover, he assigned no subject even though this is the sixth of eight statements about Rubén Blades. In this think aloud we can see both the push to get and make meaning as well as the effortful nature of comprehension. In his own words the learner tells us he used context to follow along the meaning and he used guess work. His last statement about the meaning of sentence demonstrates that he is aware of the past temporal framework of the passage he read; he states that Blades worked on films. As part of the think aloud procedure, the researcher made inquiries about the target forms, in this sentence actuó. This learner does make a  form–­ meaning connection with the target, but only after he is certain about the meaning, with which he works intensively. The rest of the think aloud protocol now ­follows. Researcher: Why did you decide that this word “actuó” does or doesn’t mean “actually”? Learner: Um, well, actually wouldn’t be like the word to fit in there. Well as least to me it wouldn’t because it wouldn’t make sense to say actually in various films. Researcher: So what makes more sense to you then? Learner: Um, he, he acted in various films. And also, we just learned the past tense last night. And it looks like the past tense, actuó. I don’t know. Good guess at least. Location. Again referring to the availability of processing resources, ­VanPatten (2004) proposed the Sentence Location Principle (a principle not included in the 1996 version). He states, “… elements that appear in certain positions of an utterance are more salient to learners than others, namely,

20 

The theory of input processing

s­ entence initial position is more salient than sentence final position that in turn is more salient than sentence internal or medial position.” (VanPatten 2004: 13). That sentence initial position is the most favourable processing position is logical. Those elements that are the first encountered and are the first on which processing resources get aligned. Through the medial portion of a sentence, the processing resources may likely still be processing the initial elements but then get redirected when the end of the sentence comes into focus. The evidence for the Sentence Location Principle strongly affirms that initial position is the most favoured processing position. Barcroft and VanPatten (1997) and Rosa and O’Neill (1998) varied the location of target elements in sentences; the locations being initial, medial and final position in the sentence. They also used both acoustically stressed and unstressed forms. They asked learners to repeat the sentences they heard and then determined how successfully the learners repeated the target items in each position. Barcroft and VanPatten (1997) found that learners repeated items most successfully in initial position, more so than in medial position and more so than in final position. They did not find a difference between medial and final position. They also found that learners more successfully repeated the stressed targets over the unstressed ones. Rosa and O’Neill (1998) found interactions between location and acoustic stress. Both factors affect processing. By and large their results confirm that initial position is the most favourable processing position and that final position is more favourable than medial position. We use the phrase ‘by and large’ here to indicate that in three out of four processing contexts the results demonstrated that learners more successfully repeat the targets that occur in sentence initial position than in medial position. Likewise, in three out of four processing contexts learners more successfully repeated target items that occur in sentence final position than in sentence medial position. Processing grammatical and semantic relationships As noted in our introduction, VanPatten’s theory of input processing encompasses two main principles (VanPatten 2004b). In this section we address the second of these, the First Noun Principle that asserts that learners tend to misassign the grammatical role of subject or semantic role of agent to the first noun or pronoun they encounter in a sentence. This processing strategy has been documented in child first language acquisition (­Slobin 1966; Bever 1970), child second language acquisition (Ervin-Tripp 1974; Nam 1975), and adult second language acquisition (VanPatten 1985a; Lee 1987; LoCoco 1987). In

How do the principles work? 

21

a sense, initial position works against a grammatical object being correctly interpreted. The problem with learners’ use of this processing strategy goes beyond miscomprehension but to the heart of acquisition. VanPatten states, this particular principle may have a variety of consequences in a variety of languages. It is not just that learners may get word order wrong, it is also that they may not process case markings for some time, will have difficulties with the pronoun system in some languages, and so on.  (VanPatten 2004: 16)

In other words, misprocessing the form leads to incorrect information being supplied to the developing system. Word order. Even though languages have a typologically canonical word order, such as SVO for English and SOV for Japanese, other word order permutations are permissible. González (1997) documented the acquisition of different word orders for learners of Spanish with SVO being the first acquired word order pattern and OSV and OVS being the last acquired. Strings in which the object precedes the subject present learners with difficulties. Children acquiring Spanish as a first language also acquire these word order patterns last (Echevarría 1978 cited in González 1997). LoCoco (1987) presented learners of Spanish and German with three different sentence types to process. In each of these sentence patterns, the first noun was an object of some kind: direct object, indirect object or the object of a preposition. She found that learners assigned the first noun the grammatical role of subject from 7% to 72% of the time. VanPatten (1985a) and Lee (1987) presented learners of Spanish with OVS and OV sentences, respectively, in which the objects were pronominalised. VanPatten documented that learners assigned the grammatical role of subject to the object pronoun from 35% to 70% of the time. Lee (1987) documented learners’ use of the first noun strategy between 27% and 73% of the time. VanPatten and Wong (2004) and Allen (2000) demonstrated that learners of French use the first noun strategy to assign the semantic role of agent to the first noun in faire-causatif sentences. As mentioned above in the section on the Primacy of Content Words Principle, learners incorrectly interpret that the agent performing the action of the second verb is the first noun. Instead of indicating that Henri reads the newspaper in the sentence Jean-Paul fait lire le journal à Henri (Jean-Paul makes to read the newspaper to Henri/jeanPaul makes Henri read the newspaper) they indicate that Jean-Paul reads it (VanPatten and Wong 2004: 104). Lexical semantics. Learners do not absolutely and categorically use only the first noun strategy to assign grammatical and semantic roles. They are sen-

22 

The theory of input processing

sitive to other factors, one of them being lexical semantics, that attenuate their use of the first noun strategy. The lexical semantics of the verb ‘kick’, for example, requires an animate agent. A sentence such as The ball was kicked by the child is unlikely to be misinterpreted because a ball cannot perform the action. Among her target sentences LoCoco (1987: 124) included the ­following. La cerveza le trae el muchacho a la muchacha./ Das bier bringt der Junge dem Mädchen. the beer to her brings the boy to the girl ‘The boy brings the beer to the girl.’ Las flores le da el muchacho a la niña./ Die blumen gibt der Junge dem Mädchen. the flowers to her gives the boy to the girl ‘The boys gives the flowers to the girl.’ The learners of Spanish and German never identified the first noun in these sentences as the grammatical subject/agent. The lexical semantics of the verbs bring and give do not allow inanimate subjects. In short, when “… only one noun is capable of the action …” learners correctly identify that noun as the agent (VanPatten 2007: 124). Event probabilities. We also use what we know about the world to interpret sentences. In the following sentences, both nouns are capable of performing the action but one interpretation is more likely than another. The event probabilities are low for the first noun being the agent and are higher for the second noun being the agent. In these scenarios, children and adult second language learners correctly identify that the second noun performed the action of the verb. The farmer was kicked by the horse. The child was bitten by the dog. Likewise, learners of French are unlikely to interpret that the professor is doing the studying in the following sentence (VanPatten and Wong 2004: 101). Le professeur fait étudier le verbe “être” à l’élève. the professor makes to study the verb “être” to the student. ‘The professor makes the student study the verb “être”.’ Learners may well use event probabilities to attenuate their use of the first

How do the principles work? 

23

noun strategy. That is, “… it is possible (though not necessary) that real-life scenarios might override the First Noun Principle …” (VanPatten 2007: 123). We might also consider learners’ background knowledge of a set of characters as a type of real world knowledge. Houston (1997) showed that learners use their knowledge of a set of characters to overcome their use of the first noun strategy. He had learners interpret two sets of sentences in both of which he used OVS word order. One set of sentences used the names of characters from a video series the learners were seeing in class. The other used random names. As can be seen in the following example, the sentences are structurally and semantically identical. Characters Random names A Raquel la contrata don Pedro. A Silvia la contrata Ricardo. Learners used the first noun strategy only 28% of the time to misassign the grammatical role of subject to the character-based sentences. They misassigned the grammatical role of the first noun in the random-name sentences 48% of the time. Malovrh (2006) presents a findings consistent with Houston (1997). He found that learners processed OVS strings that referred to characters from the Simpsons more accurately than they processed OVS strings about made up characters. The accuracy rate was 65% for the Simpson-based sentences versus 50% for the other type. Contextual constraints. Background knowledge (or topic familiarity as Malovrh labelled it) is an extralinguistic type of context. VanPatten (2004b, 2007) has added sentence-internal linguistic context as a possible constraint on learners’ use of the first noun strategy. VanPatten and Houston (1998) demonstrated the effects of context on sentence interpretation. They created ten target sentences containing OVS word order in which a clause preceding the object pronoun provided contextual information. The target sentences were paired with ten sentences that contained a preceding clause that did not provide a contextual cue. The target sentences were constructed with the verbs attacked, insulted, rejected, greeted, and kissed. Note the following examples from VanPatten and Houston (1998). The OVS constructions are italicised. Context Ricardo está enojado porque lo insultó Susana en la reunión. Ricardo is angry because him insulted Susana in the meeting ‘Ricardo is angry because Susana insulted him in the meeting.’

24 

The theory of input processing

Roberto está en el hospital porque lo atacó María con un cuchillo. Robert is in the hospital because him attacked María with a knife ‘Robert is in the hospital because María attacked him with a knife.’ No context Ricardo me dice que lo insultó Susana en la reunión. Ricardo me tells that him insulted Susana in the meeting ‘Ricardo tells me that Susana insulted him in the meeting.’ Gloria contó a sus amigas que la atacó Ramón en su casa. Gloria told to her friends that her attacked Ramón at home ‘Gloria told her friends that Ramón attacked her at home.’ VanPatten and Houston found that sentence-internal context attenuated learners’ use of the first noun strategy for assigning grammatical roles. In the context condition, learners assigned the grammatical role of subject to the object pronoun 59% of the time. In the no context condition, learners assigned the grammatical role of subject to the object pronoun 84% of the time. Learners use of the first noun strategy to assign grammatical roles is quite strong in both the context and no context sentence types but context does provide learners an additional clue for processing the formal elements of the sentence. Malovrh (2006) investigated whether the placement of contextual information prior to or after the targeted object pronoun would differentially affect learners’ accurately assigning agent/subject. He found no difference in learners’ performance based on the placement of the contextual information. It is the presence of the contextual information that helps learners not its placement. In other words, “… contextual information…would push [learners] away from interpreting the targeted clause the wrong way” (VanPatten 2004: 17). In what ways is the theory lacking? There is no single comprehensive theory of second language acquisition and so the theories available to researchers and practitioners focus on different aspects of SLA. VanPatten’s theory focuses on input processing. In ­VanPatten (2004) he invited commentary on his model. He addressed some of the issues in the concluding chapter (VanPatten 2004c) and in subsequent work (­VanPatten 2007). Harrington (2004: 80) comments on VanPatten’s model by raising issues of internal validity. That is, he addresses areas in which he asserts that ­VanPatten’s

Conclusion 

25

model does not account for the observed difficulties that beginning learners have with learning certain grammatical morphemes. First, Harrington expresses concern over the use of the term “meaningful” in some of the principles. Indeed, the principles have referred to more or less meaningful forms and, at one time, to their relative communicative value. Harrington asserts that “meaningfulness” is a  notion difficult to operationalise and then test. Harrington also takes exception to the locus of input processing. VanPatten’s model claims that the difficulty of making a form–meaning connection arises at the initial encoding of that form–meaning connection. Harrington asserts that the claim is untestable. He points to other loci: at the earlier stage of perceptual processing or at the later stages of storage and retrieval. The third issue Harrington refers to is the model’s reliance on processing capacity constraints to explain processing/learning outcomes. He asserts that “a capacity account of input processing has to show that the outcome was due to capacity limitations and not just to lack of L2 knowledge” (Harrington 2004: 89). As the model is currently formulated, Harrington states that “… the notions ‘processing resources’ and ‘capacity limitations’ will be interchangeable with ‘the relative amount of L2 knowledge available.’” (Harrington 2004: 90). These issues will have to be addressed. Carroll (2004) points to two issues with VanPatten’s theory: scope and processing assumptions. In terms of scope, she asserts that the theory needs a theory of perception in reference to processing forms. She also asserts that the theory needs a theory of parsing as the First Noun Principle and the Sentence Location Principle involve word order. Additionally, she questions the processing assumptions. She affirms that there can be no processing of input for meaning before there are forms onto which those meanings can be mapped. In general, she encourages VanPatten to continue to clarify terminology and definition. VanPatten addresses some of these concerns in his formulation of the revised Lexical Preference Principle, the L1 Transfer Principle, and the Early Constraint Principle. As he notes, the model continues to evolve (VanPatten 2007: 117). Conclusion Input processing is one part, albeit a rather important one, of what we could refer to as the entirety of second language acquisition. VanPatten and Williams (2007) offering the following regarding looking at SLA from different perspectives. “To understand [why there isn’t just one theory to account

26 

The theory of input processing

for SLA], one might consider the parable about the four blind men and the elephant. These sightless men chance upon a  pachyderm for the first time and one, holding its tail, says, “Ah! The elephant is very much like a rope.” The second one has wrapped his arms around a giant leg and says, “Ah! The elephant is like a tree.” The third has been feeling along side the elephant’s massive body and says, “Ah! The elephant is very much like a  wall.” The fourth, having seized the trunk cries out, “Ah! The elephant is very much like a snake.” For us, SLA is a big elephant that researchers can easily look at from different perspectives … Thus, researchers have grabbed onto different parts of the elephant as a means of coming to grips with the complex phenomenon. (VanPatten and Williams 2007: vii–viii) We do not know what part of the elephant input would be, but that is the part we are working with. Exposure to input is necessary for second language acquisition to take place. In this chapter we have presented VanPatten’s theory of input processing and traced its development over time. The theory offers a set of principles and subprinciples designed and formulated to explain how learners work with input, that is, how they make a connection between a form in the input and its meaning. Mitchell and Myles (2004: 187–188) see VanPatten’s theory as explaining “the apparent failure of second language learners to process completely the linguistic forms encountered in second language input, and hence to explain their impoverished intake which in turn restricts the development of grammatical form.” As long as the statements that learners fail to process completely and that their intake is impoverished are not read in a judgmental fashion, we concur with Mitchell and Myles’s assessment. Learners do not process completely the forms in the input. Why not? Their intake restricts the development of grammatical form. Of course. Learners have a limited capacity to process as well as possess an incomplete second language linguistic system. The idea behind exposing learners to input is increase their capacity to process and to make the developing system a bit more complete. Both events happen slowly over time. In this chapter we not only presented and explained each principle and subprinciple, we also supplied empirical evidence to support them. These are the foundations on which Processing Instruction has been built (VanPatten 1993, 1996; Lee and VanPatten 1995, 2003, and elsewhere). When we know what learners do with input, how they work with it, we can then derive instructional techniques and write instructional materials that intervene at the time learners are working with input to make form–meaning connections and not at the time when they are practicing making output.

revision 1: clarification Principle 1b. The Lexical Preference Principle. Learners will tend to rely on lexical items as opposed to grammatical form to get meaning when both encode the same semantic information. (VanPatten 2004b: 14)

P1(b). Learners prefer processing lexical items to grammatical items (e.g., morphology) for semantic information. (VanPatten 1996: 21)

revision 2 (Revised) Lexical Preference Principle. If grammatical forms express a meaning that can also be encoded lexically, then learners will not initially process those grammatical forms until they have lexical forms to which they can match them. (VanPatten 2004c: 331)

unchanged The Primacy of Content Words Principle. Learners process content words in the input before anything else. (VanPatten 2007: 117)

unchanged Principle 1a. The Primacy of Content Words Principle. Learners process content words in the input before anything else. (VanPatten 2004b: 14)

P1(a). Learners process content words in the input before anything else. (VanPatten 1996: 18)

revised: clarification (Revised) Lexical Preference Principle. If grammatical forms express a meaning that can also be encoded lexically (i.e., that grammatical marker is redundant), then learners will not initially process those grammatical forms until they have lexical forms to which they can match them.

unmentioned as a principle, but used to frame the discussion “Learners are driven to get meaning while comprehending.” (VanPatten 2007: 116)

unchanged Principle 1. The Primacy of Meaning Principle. Learners process input for meaning before they process it for form. (VanPatten 2004b: 14)

P1. Learners process input for meaning before they process it for form. (VanPatten 1996: 17)

VanPatten (2007)

VanPatten (2004b, 2004c)

VanPatten (1996)

Table 1.1.  The evolution of VanPatten’s principles of input processing

Appendix Appendix  27

revision: wording The Preference for Nonredundancy Principle P1c. The Preference for Nonredundancy Principle. Learners are more likely Principle. Learners are more likely to process to process nonredundant meaningful nonredundant meaningful grammatical form before grammatical markers before they process they process redundant meaningful forms. redundant meaningful markers. (VanPatten 2007: 119)

VanPatten (1996)

P1(c). Learners prefer processing “more meaning­ ful” morphology before “less or nonmeaningful morphology.” (VanPatten 1996: 24)

P2. For learners to process form that is not meaningful, they must be able to pro­cess informational or communi­ cative content at no or little cost to attentional resources. (VanPatten 1996: 27)

VanPatten (2004b, 2004c)

Table 1.1. (cont.)

revised: clarification Principle P1e. The Availability of Resources Principle. For learners to process either redundant meaningful forms or nonmeaningful forms, the processing of overall sentential meaning must not drain available processing resources. (VanPatten 2004b: 14)

Principle 1d. The Meaning-Before-Nonmeaning Principle. Learners are more likely to process meaningful grammatical forms before nonmeaningful forms irrespective of redundancy. (VanPatten 2004b: 14)

revision: 2 new principles

The theory of input processing

unmentioned as a principle “Comprehension for learners is initially quite effortful in terms of cognitive processing and working memory. This has consequences for what the input processing mechanisms pay attention to.” (VanPatten 2007: 116)

revision: wording The Meaning-Before-Nonmeaning Principle. Learners are more likely to process meaningful grammatical markers before nonmeaningful markers. (VanPatten 2007: 120)

VanPatten (2007)

28 

P3. Learners possess a default strategy that assigns the role of agent to the first noun (phrase) they encounter in a sentence. We call this the “first noun strategy.” VanPatten (1996: 33)

unchanged The First Noun Principle. Learners tend to process the first noun or pronoun they encounter in a sentence as the subject [or agent]. (VanPatten 2007: 122) unchanged The L1 Transfer Principle. Learners begin acquisition with L1 parsing procedures. (VanPatten 2007: 122)

alternative: The L1 Transfer Principle. Learners begin acquisition with L1 parsing procedures. (VanPatten 2004c: 330)

unmentioned

possible alternative The Early Constraint Principle. The processing of the initial element(s) constrains the processing of the rest of the sentence. (VanPatten 2004c: 328)

revised: clarification P2. The First Noun Principle. Learners tend to process the first noun or pronoun they encounter in a sentence as the subject/agent. (VanPatten 2004b: 18)

unchanged The Sentence Location Principle. Learners tend to process items in sentence initial position before those in final position and those in medial position. (VanPatten 2007: 125)

new principle Principle 1f. The Sentence Location Principle. Learners tend to process items in sentence initial position before those in final position and those in medial position. (VanPatten 2004b: 14)

Appendix  29

The theory of input processing

revision: wording The Lexical Semantics Principle. Learners may rely on lexical semantics, where possible, instead of the First Noun Principle to interpret sentences. (VanPatten 2007: 124) revision: clarification The Event Probabilities Principle. Learners may rely on event probabilities, where possible, instead of the First Noun Principle (or an L1 parsing procedure) to interpret sentences. (VanPatten 2007: 124) revised: clarification The Contextual Constraint Principle. Learners may rely less on the First Noun Principle (or L1 transfer) if preceding context constrains the possible interpretation of a clause or sentence. (VanPatten 2007: 124)

VanPatten (2004b, 2004c)

revision: 2 principles Principle P2a. The Lexical Semantics Principle. Learners may rely on lexical semantics, where possible, instead of word order to interpret sentences. (VanPatten 2004b: 18)

Principle P2b. The Event Probabilities Principle. Learners may rely on event probabilities, where possible, instead of word order to interpret sentences. (VanPatten 2004b: 18)

new principle P2c. The Contextual Constraint Principle. Learners may rely less on the First Noun Principle if preceding context constrains the possible interpretation of a clause or sentence. (VanPatten 2004b: 18)

Table 1.1. (cont.)

VanPatten (1996)

P3(a). The first noun strategy may be overridden by lexical semantics and event probabilities. (VanPatten 1996: 36)

VanPatten (2007)

30 

In other words

“Simply put, P1 states that learners are driven to look for the message in the input (“What is this person saying to me?”) before looking for how that message is encoded.” (VanPatten 1996: 17)

“… what in the input would learners attend to get meaning. A logical place to begin would be with content words.” (emphasis original, VanPatten 1996: 18)

Principle

P1. Learners process input for meaning before the process it for form. (VanPatten 1996: 17)

P1(a). Learners process content words in the input before anything else. (VanPatten 1996: 18)

L1 production English children L1 processing and production English children L1 production English children L2 production German adults L2 processing Hindi adults L2 processing Spanish adults

Peters (1983, 1985)

Radford (1990) Klein (1986) Mangubhai (1991) VanPatten (1990)

L2 processing English adults

Sharwood Smith (1986) Lightbown and Spada (1993)

L1 processing English children

Focus

Peters (1985)

Supporting work

Table 1.2.  The original formulation of VanPatten’s theory of input processing and a summary of the works from which it was derived

Appendix  31

In other words

“… learners first attend to and detect lexical markers of plurality before grammatical markers of plurality, that they attend to and detect lexical markers of temporal reference before verbal inflections of tense, and so on.” (VanPatten 1996: 22)

Table 1.2.  (cont.)

Principle

P1(b). Learners prefer processing lexical items to grammatical items (e.g., morphology) for semantic information. (VanPatten 1996: 21)

L2 production English adults Pidgins and creoles

Pica (1985)

L2 processing Italian, English adults L2 processing Italian, Spanish, French adults

Glass (1994)

Musumeci (1989)

Todd (1974) L2 processing Cadierno, Glass, VanPatten Spanish adults and Lee (1991) published as Lee, Cadierno, Glass and VanPatten (1997)

L2 production English adults

Bardovi-Harlig (1992)

Focus

The theory of input processing

Supporting work

32 

“… it is the relative communicative value [to the learner] of a grammatical form that plays a major role in determining the learner’s attention to it during input processing and the likelihood of its being detected and thus part of intake …” (VanPatten 1996: 24)

“… it seems logical that the detection of grammatical form of little communicative value is inhibited by the allocation of processing resources to getting meaning from the input… Detection of these grammatical forms will occur only if processing for meaning is relatively easy for the learner.” (VanPatten 1996: 29)

“… regardless of the actual syntactic configuration of the sentence, learners interpret the first noun as the agent.” (VanPatten 1996: 33)

P1(c). Learners prefer processing “more meaningful” morphology before “less or nonmeaningful morphology.” (VanPatten 1996: 24)

P2. For learners to process form that is not meaningful, they must be able to process informational or commu­ nicative content at no or little cost to attentional resources. (VanPatten 1996: 27)

P3. Learners possess a default strategy that assigns the role of agent to the first noun (phrase) they encounter in a sentence. We call this the “first noun strategy.” VanPatten (1996: 33)

L2 production Spanish adults L2 processing various languages adults L2 processing Spanish adults

Van Naerssen (1981) VanPatten (1985b)

Bransdorfer (1989, 1991)

L1 processing English children L1 processing English children

Bever (1970) Slobin (1966)

theoretically motivated but not empirically supported

L2 production English adults

VanPatten (1984a)

Appendix  33

L2 processing Spanish adults L2 processing Spanish adults L2 processing Spanish adults L2 processing Spanish, German adults L2 processing Spanish adults

Binkowski (1992) Glisan (1985) Lee (1987) LoCoco (1987)

VanPatten (1984b)

L2 processing Italian, English adults MacDonald and Heilenman L2 processing (1992) French adults

L2 processing French children

Ervin-Tripp (1974)

Gass (1989)

L2 processing English children and adults

Nam (1975)

Principle

Focus

Table 1.2.  (cont.) Supporting work

The theory of input processing

In other words

34 

“Lexical semantics refers to the constraints on a situation imposed by the semantics of the verb involved. The verb kick, for example, requires an animate being with legs to do the kicking … An event probability refers to the likelihood that a given situation would exist in the real world, even though lexical semantics allows it.” (VanPatten 1996: 36)

“Concerning the assignment of semantic roles, what guides learners’ processing of input also changes over time.” (VanPatten 1996: 37)

P3(a). The first noun strategy may be overridden by lexical semantics and event probabilities. (VanPatten 1996: 36)

P3(b). Learners will adopt other processing strategies for grammatical-semantic role assignment only after their developing system has incorporated other cues (e.g., case marking, acoustic stress, subject–verb agreement).

L1 processing children English L1 processing Walpiri children L2 processing Italian, English adults L2 processing Dutch adults L1 processing French children L1 processing Turkish children L1 processing Hungarian children L2 processing French, English adults L2 processing Italian, English adults

Bavin and Shopen (1989) Gass (1989)

Issidorides and Hulstijn (1991) Bates and MacWhinney (1989) Slobin and Bever (1982) Plèh (1989) McDonald and Heleinman (1992) Gass (1989)

L1 processing Hungarian children

Bever (1970)

Plèh (1989)

Appendix  35

Chapter 2 What makes Processing Instruction effective? Introduction In the last ten years, like many of our colleagues in the United States, we have conducted research to measure the possible effects of Processing Instruction in different languages while focusing on the different processing principle of VanPatten’s (1996, 2002, 2004) input processing model (Lee and Benati 2007a, 2007b; Benati and Lee 2008). We have also presented our research findings at different conferences and research/annual meetings. We have however, developed the clear impression that there is still an obvious misunderstanding about what VanPatten’s approach to grammar instruction actually is, among scholars and practitioners. Some practitioners believe that Processing Instruction is just another non-communicative and very explicit type of grammar instruction. Others think that it is another comprehension-based approach to grammar teaching, while another group failing to replicate previous Processing Instruction studies correctly, have developed structured input activities for their experiment that are not in effect structured input activities (DeKeyser and Sokalsky 1996; Salaberry 1997; Allen 2000). These studies were criticized (Sanz and VanPatten 1998; Wong 2004) because they did not accurately replicate previous studies comparing Processing Instruction and traditional instruction. In these studies the structured input activities did not push learners to make appropriate form–meaning connections. Meaning was not in focus and learners were not pushed to alter their default strategies. All these studies have been criticized because they do not follow Processing Instruction guidelines to produce the materials. The materials and instruction develop in Dekeyser and Sokalsky (1996), Salaberry (1997) and Allen (2000) studies was not Processing Instruction. Norris and Ortega (2000) have classified the instruction used in the above studies as ‘focus on forms’ (see chapter three) and a type of instruction that does not integrate form and meaning and includes mechanical drills. We believe that it is time to clarify Processing Instruction once and for all and to show what makes this approach to grammar instruction an extremely effective approach. In this chapter we describe, in simple terms this new pedagogical approach to grammar teaching and examine its three main

What is Processing Instruction? 

37

c­ omponents. We will also examine how Processing Instruction and structured input activities have been used in altering some of the processing problems outlined in the input processing model. What is Processing Instruction? As an approach to grammar instruction, Processing Instruction has many characteristics: –– it is based on the input processing model; –– it is a psycholinguistics motivated approach; –– it is intended to make learners make better form meaning connections in the input learners receive; –– it is a type of instruction that keeps meaning in focus; –– it is intended to bring learner’s attention to incorrect processing strategies; –– it is input-based as opposed to output-based; –– it is ‘‘focus on form’’ as supposed to ‘‘focus on forms’’; –– it is communicative as opposed to traditional; –– it is a three-component approach to grammar instruction. In 1996, VanPatten (1996: 60) argued that the main goal in this approach ‘‘is to alter the processing strategies that learners take to the task of comprehension and to encourage them to make better form–meaning connections than they would if left to their own devices’’. Processing Instruction is therefore a deliberate instructional intervention predicated on VanPatten’s model of input processing (see chapter one in this book). When learners received or are exposed to input they tend to rely on internal strategies (called principles in VanPatten’s input processing model, see 2004a, 2007) to process the input. As a result of their internal processing they might not be able to make correct form–meaning connections. As underscored by Wong (2004: 33) ‘‘the goal of Processing Instruction is to help L2 learners derive richer intake from input by having them engage in structured input activities that push them away from the strategies they normally use to make form–meaning connections’’. Processing Instruction is an input-based approach to grammar instruction. It is input based as supposed to output-based because the main focus of this approach to grammar instruction is to intervene when L2 learners process the language at input level. In this approach the main focus is to help learners to develop their internal linguistics system and in this way this approach is

38 

What makes Processing Instruction effective?

input-based as opposed to output-based instruction (we discuss this in relation to output-based approaches in chapter 3). However, Processing Instruction is not simply a comprehension/interpretation-based approach to grammar instruction as maintained by VanPatten (1996: 82), because its main purpose is to ensure that L2 learners process correctly and efficiently forms/structures (one at a time) in the input they receive. It is more than a comprehension-based approach to grammar teaching because in Processing Instruction learners are asked to focus on small parts/ features of the targeted language when they process the input. Psycholinguistics processing (learners‘ strategies, see processing principles in chapter 1) is kept in mind as the main goal of Processing Instruction is for learners to intake a ‘better’ input. As emphasised by VanPatten (1996: 83-84) ‘‘ Processing Instruction is a specific approach to explicit grammar instruction and thus falls more clearly within the category of instructional treatments called ‘‘focus on form’’. Long (1991) and more recently Long and Robinson (1998) have distinguished two types of focus on form: ‘‘focus on form’’ and ‘‘focus on forms’’. Doughty and Williams (1998) have defined “focus on forms” as any type of instruction that isolates specific linguistic forms in order to teach them one at a time. Focus on forms refers to synthetic approaches to language teaching where the L2 is analyzed in different parts such as grammar and vocabulary and these elements are taught in isolation from context. This model of focus on forms has been criticized by scholars (Long and Robinson 1998; Wong and VanPatten 2003) particularly on the basis of the fact that L2 learners, rather than learning discrete lexical or grammatical items one at the time, follow predictable sequences in certain L2 features. Focus on forms refers to synthetic/traditional approaches to language teaching where the L2 is analyzed in different parts such as grammar and vocabulary and these elements are taught in isolation from context. Focus of form approaches, like Processing Instruction, are more congruent with a different kind of language syllabus, one that focuses on meaning. As outlined by Lee and VanPatten (1995) Processing Instruction because of its nature and characteristics is certainly an appropriate and effective approach to grammar instruction. It is one possible way to incorporate explicit grammar instruction in a communicative framework. It is a communicative approach to grammar teaching as one of its goals is to increase learners’ opportunities to receive good comprehensible and meaning bearing input. At the same time it provides L2 learners with opportunities to focus on linguistic properties of the language. Lee and VanPatten (1995: 94) maintain that it ‘‘is

What are the main components of Processing Instruction? 

39

a way to ­incorporate explicit grammar instruction into classes without sacrificing either communication or learner-centered activities’’. The characteristics of Processing Instruction have been described in detail in previous literature (Lee and VanPatten 1995; VanPatten 1996; Wong 2004, 2005; Farley 2005; Lee and Benati 2007a, 2007b; Benati and Lee 2008, 2009) and in the next section we will examine the three components of this ­approach in detail. What are the main components of Processing Instruction? As emphasized in the previous paragraph, Processing Instruction is a new type of grammar instruction which is concerned with learners’ awareness of how grammatical forms and structures are acquired. First and foremost it is a type of focus on form which draws on the principles of the input processing model (VanPatten 1996, 2002, 2004). In the model of L2 acquisition proposed by VanPatten ‘input provides the data, input processing makes (certain) data available for acquisition. Processing Instruction seeks to intervene in the processes learners use to get data from the input. Research on input processing has attempted to describe what linguistic data learners attend to during comprehension and which ones they do not attend to, for example how learners assign particular roles to nouns or how position in an utterance influences what gets processed. These processing principles seem to provide an explanation of what learners are doing with input when they are asked to comprehend it. As a result of the way learners attend to input data VanPatten (1996) has developed a new kind of grammar instruction which guides and focuses learners’ attention when they process input. This new type of grammar instruction is diametrically opposed to traditional instruction which consists of drills in which learner output is manipulated and instruction is divorced from meaning or communication. Processing Instruction is a more effective method for enhancing language acquisition as it is used to ensure that learners’ focal attention during processing is directed toward the relevant grammatical items and not elsewhere in the sentence. Its main objective is to help learners to circumvent the strategies used by them to derive intake data by making them to rely exclusively to on form and structure to derive meaning from input (in the next paragraph we will provide practical examples as to how structured input activities are designed keeping in mind VanPatten’s psycholinguistics principles).

40 



What makes Processing Instruction effective?

Processing Instruction consists of three main components:

1. Learners are given explicit information about a linguistic structure or form. Forms or structures are presented one at a time (e.g. regular past forms, regular future forms). In the example below learners of Japanese are presented with explicit information about the passive forms (Hikima, forthcoming). Attention! The first noun is not always an agent in the SOV (subject object verb) sentence. Pay attention! Who did what to whom?

The passive sentence When a passive sentence is constructed, a verb must be changed to a passive verb form. R1V + れる たたく (tataku) たたかれる

→ drop u (tatak) → put a (tataka) たたか → connect with れる →

make stem of ない form → change to a past form たたかれた → change to a past masu form たたかれました (yobu) → drop u (yob) → put a (yoba) よば → connect with れる → よばれる make stem of ない form → change to a past form よばれた → change to a past masu form よばれました こわす →　こわされる →　こわされた →　こわされました

よぶ

R2 V + られる たべる → drop る form たべられた

→

たべ

→ connect with られる →

たべられる

→ change to past

→ change to past masu form たべられました みる → みられる → みられた → みられました IRR

くる → こられる → こられた → こられました する →される → された → されました 1. active sentence トムは　クリスを　たたきました。Tom



hit Chris agent patient 2. passive sentence クリスは　トムに　たたかれました。Chris was hit by Tom patient agent

What are the main components of Processing Instruction? 

41

2. Within the linguistics information provided to learners about a targeted form or structure, learners are also given information on a particular processing principle that may negatively affect their picking up of the form or structure during comprehension (see the previous example for Japanese passive forms. Learners are warned that the first noun is not always an agent in a SOV(subject object verb) sentence: ‘Pay attention! Who did what to whom?’). The explicit information provided should help L2 learners to be careful when they process input. In a different example (see below) the case of the presence of temporal adverbs in a sentence (Lexical Preference Principle) might cause other processing problems. Learners are therefore warned to pay attention to tense endings to understand when the action takes place rather than relying on temporal indicators. Spanish simple past tense is called pretérito indefinido. This past tense has different forms from the present tense. The pretérito indefinido serves to report actions, events, and states that are viewed as having been completed in the past.   Notice that: Past form is usually accompanied by temporal adverbs that will indicate that the action has already happened in the past. Here are some of the most common ones: ayer (yesterday), la semana pasada (last week), anteayer (the day before yesterday), anoche (last night), etc. However, although these adverbs are a good clue to know that an action has occurred in the past, they are not always present in the sentences. That is why it is very important for you to recognise past tense forms.

  (Benati and Romero 2004)

In our final example the location (Sentence Location principle) of the form or structure (in this case the Subjunctive) in a sentence would determine whether or not that particular form or structure will be processed more or less efficiently (see the information provided below for the subjunctive forms in Italian.). Learners are given this information to help them to focus on the middle part of the sentence when processing input containing this linguistic feature. The subjunctive verb is located generally in the middle of the sentence (e.g. Io non credo che parli bene il francese). ATTENZIONE Pay attention to the middle of the sentence!! When the verb or expression in the independent clause denotes certainty, the present indicative is used in the dependent clause. When the verb or expression in the independent clause expresses opinion, doubt or uncertainty, the present subjunctive is used

  (Lee and Benati 2007b)

42 

What makes Processing Instruction effective?

3. Learners are pushed to process the form or structure during activities in which the input is manipulated in particular ways to push learners to become dependent on form to get meaning. After receiving the explicit information about the targeted linguistic feature and the information about the processing principle affecting that feature, learners are pushed to process the form or structure through structured input activities. In structured input activities the input is manipulated in particular ways to push learners to become dependent on form and structure to get meaning. As outlined by Wong (2004) Processing Instruction ‘pushes learners to abandon their inefficient processing strategies for more optimal ones so that better form–meaning connections are made’ (p. 35). As the main component of this input approach to grammar teaching, structured input activities help learners to make those form–­ meaning ­connections. VanPatten and Sanz (1995) have produced the following guidelines for structured input activities: 1. Present one thing at a time. 2. Keep meaning in focus. 3. Move from sentences to connected discourse. 4. Use both oral and written input. 5. Have the learner do something with the input. 6. Keep the learner’s processing strategies in mind. As Wong (2004) has pointed out we need to follow these guidelines to the letter in order to develop effective structured input activities. She rightly argues (Wong 2004) that structured input activities are not any other inputbased practice. According to Wong (Wong 2004: 37) ‘for an activity to be an SI activity, that activity must somehow push learners to circumvent an inefficient processing strategy’. Identifying the processing problem in a target language is the most important step in developing structured input activities. One form or structure in a given L2 can be affected by one single processing problem or a combination of processing problem which makes learners process input inefficiently and sometime misinterpret sentences. One processing problem is the tendency for L2 learners to rely temporal adverbs to establish the time frame (Preference Lexical Principle). In a sentence such as: ‘Yesterday I walked with John in the park’ L2 learners can establish the temporal framework with a lexeme (the temporal marker ‘Yesterday’), and they do not need the verb form -ed to do that. According to the Lexical Preference Principle: learners will tend to rely on lexical items as opposed to grammatical

What are the main components of Processing Instruction? 

43

form to get meaning when both encode the same semantic information. In addition to that, the lexeme (temporal adverb in this case) makes the verb form redundant. Learners do not have to process the verb form for its meaning because it is redundant (according to the Preference for Non-redundancy Principle: learners are more likely to process non-redundant meaningful grammatical form before they process redundant meaningful forms). Now that we have identified some processing problems for L2 learners, we can create structured input activities to resolve them. Let’s turn on the guidelines developed by VanPatten and Sanz (1995). 1. Rules should be broken down into smaller parts and taught one at the time during the course of the lesson. Students are presented with the linguistic feature before being exposed to structured input activities. In fact the type of input L2 learners receive in PI is meaningful as it should help them to make correct meaning–form connections. We should avoid providing L2 learners with lots of information and grammatical rules as learners possess a limited capacity for processing information. Presenting L2 learners with a smaller and more focused amount of information will clearly enhance the opportunity for learners to pay more focused attention (Lee and VanPatten 1995, 2003). Therefore, we must develop structured input activities that focuses on one form at the time so that it will be easier for learners to map one form to one meaning. In the example in Italian below, the instructor reads some sentences (all containing the present tense third person singular form ‑a-) the following sentences and ask students whether these statements are true or not true for a typical university college student. Vero Falso Lavora a tempo pieno (works full time) ☐ ☐ Studia tutto il giorno (studies all day) ☐ ☐ Lavora mezza giornata (works part-time) ☐ ☐

Students are exposed to the forms (third person present tense) and at the same time meaning is in focus. 2. Learners should be encouraged to make form–meaning connections through structured input activities. As pointed out by VanPatten (1996: 68) ‘if meaning is absent or if learners do not have to pay attention to meaning to complete the activity, then there is not enhancement of input processing’. Keep meaning in focus is crucial when we develop structured input activities. The task in structured input activities must be completed with focused ­attention to the referential meaning of the input L2 learners are exposed to.

44 

What makes Processing Instruction effective?

A good structured input activity is the one where students must attend to the meaning of the sentence to complete the task. In the example below the gender agreement in Italian is expressed morphologically by the endings –a- for feminine nouns and ‑o- for masculine nouns. Learners will have to pay attention to the ending of the adjectives describing either Sofia Loren or Luciano Pavarotti (see sentences herd by learners) to establish whether the sentence is referring to one or the other. However, after learners have established which person is being described, they need to express their opinions on whether or not they agree or disagree. In order to complete the task and express their opinions they must understand the meaning of each utterance. 1. 2.

Luciano Pavarotti ☐ ☐ agree Luciano Pavarotti ☐ ☐ agree

Sofia Loren ☐ ☐ disagree Sofia Loren ☐ ☐ disagree

Sentences heard by learners

1. È grasso 2. È brutta

3. L2 learners are first exposed to sentences. As previously said, L2 learners have limited capacity for processing and they need to direct their attention and efforts to process input for its meaning first. However, at a later stage we should be able to move from providing L2 learners with short sentences to expose them to connected discourse. This should happen only when learners have already had opportunities to process the new form or structure. In the example below learners must listen to a story that a student told about his first day in London and decide which statements accurately describe what happened. TRUE OR FALSE   THE STUDENT… ­­_______________   1 … wanted to see Buckingham Palace _______________   2 … showed pictures of Hyde Park to his parents _______________   3 … stayed in bed most of the morning _______________   4 … liked Covent Garden _______________   5 … phoned his parents _______________   6 … liked English Trains _______________   7 … visited Madame Tussaud’s _______________   8 … traveled on the Bus _______________   9 … really liked the clowns in Oxford Circus _______________ 10 … arrived home early

What are the main components of Processing Instruction? 

45

(Instructor’s script) I really wanted to see Madame Tussaud’s, so early in the morning I took a quick shower and got washed, and traveled by train to Baker Street station. I had my picture taken with David Beckham, well the wax model that is. I also visited Trafalgar square and saw Nelson’s Column. There were so many pigeons there I think they owned the place! In the afternoon I waited for ages for a train to take me to Covent Garden. I really hated London trains; they are awful. Covent Garden was fantastic, I watched the different street performers, and I really enjoyed listening to the musicians. I finished my exploration and sunbathed for a bit in Hyde Park – London’s most famous and largest open space. I arrived home quite late and called my parents back in China and talked about my day. All in all, I had a wonderful time today.

4. Structured input activities which combine oral and written input should be used as some L2 learners respond better to one mode of input than to the other. This is in order to account for individual differences. In addition to that, as noticed by Farley (2005: 15) “hearing the forms allow only for sound– meaning connections, whereas written form–meaning connections are made via reading.” 5. Structured input activities should be designed to make learners do something with the input they receive (i.e. agreeing or disagreeing; false or true; likely or unlikely). During structured input activities L2 learners should be encouraged to make form–meaning connections. Learners must be engaged in processing the input (having a specific reason for processing input) sentences and must respond to the input sentence in some way. 6. Learners’ attention should be guided so as not to rely on natural processing strategies. Activities in which the input is structured to alter learners reliance on one particular processing principle should be created. This is the main goal for structured input activities: correcting inefficient processing strategies and instilling in L2 learners more efficient ones. Structured input activities are of two types: referential and affective. Referential activities are those for which there is a right or wrong answer and for which the learner must rely on the targeted grammatical form to get meaning. In the following sentence ‘I walked with John in the park’ learners must decide whether this even refers to the past or the present. Affective structured input activities are those in which learners express an opinion, belief, or some other affective response and are engaged in processing information about the real world. In the following sentence ‘I visited Rome with my friends’ learners might be asked what they think about this even and express an opinion on whether they think it is interesting or boring for example. According to VanPatten (1996, 2002, 2004) affective activities

46 

What makes Processing Instruction effective?

reinforce form–meaning connections established during referential structured input activities. Learners must be engaged in processing the input sentences and must respond to the input sentence in some way through referential and affective types of structured input activities. How do you develop structured input activities keeping psycholinguistics principles in mind? In this section we will provide some practical examples as to how we can make structured input activities very effective in altering processing strategies. The main goal of structured input activities is to make sure that L2 learners will process input more effectively and efficiently and therefore the input must be restructured to allow this. Over the last ten years, VanPatten (VanPatten 2004: 15) has developed his input processing model which has two main principles and their corollaries as a crux: Principle 1. Learners process input for meaning before they process it for form. P 1a. The Primacy of Content Words Principle: learners process content words in the input before anything else. P 1b. The Lexical Preference Principle: learners will tend to rely on lexical items as opposed to grammatical form to get meaning when both encode the same semantic information. P 1c. The Preference for Non-redundancy Principle: learners are more likely to process non-redundant meaningful grammatical form before they process redundant meaningful forms. P 1d. The Meaning-Before-Non Meaning Principle: learners are more likely to process meaningful grammatical forms before non-meaningful forms irrespective of redundancy. P 1e. The Availability of Resources Principle: for learners to process either redundant meaningful grammatical forms or non-meaningful forms, the processing of overall sentential meaning must not drain available processing resources. P 1f. The Sentence Location Principle: learners tend to process items in sentence initial position before those in final position and those in medial position.

Structured input activities and psycholinguistics principles Structured input activities and psycholinguistics principles 

47

Principle 2. The First Noun Principle. Learners tend to process the first noun or pronoun they encounter in a sentence as the subject or agent. P 2a. The Lexical Semantics Principle: learners may rely on lexical semantics, where possible, instead of word order to interpret sentences. P 2b. The Event Probabilities Principle: learners may rely on event probabilities, where possible, instead of word order to interpret sentences. P 2c. The Contextual Constraint Principle: learners may rely less on the First Noun Principle if preceding context constraints the possible interpretation of a clause or sentence. The Lexical Preference Principle In the second corollary, (P1.b) the lexical preference principle, VanPatten (1996) asserts that learners prefer processing lexical items to grammatical items (e.g., morphology) for semantic information. This principle is a direct consequence of VanPatten’s first principle. A great number of grammatical features encodes some kind of semantic information (e.g. Italian verbal inflection -ato encodes past as in guardato; English verbal inflection ‑ed encodes past as in watched; Japanese verbal inflection -mashita encodes past Activity A Ascolta e determina se l’azione è passata (passato) o si riferisce ad una azione abituale nel presente. Presta attenzione al verbo in posizione finale!

1. 2. 3. 4. 5.

Kinō Mainichi □ □ □ □ □ □ □ □ □ □

Sentences heard by learner: 1. 2. 3. 4. 5.

Kyoto ni ikimashita. Kaisha ni ikimasu. Marason o shimashita. Totemo ii hon o yomimashita. Gek.ijyou ni ikimasu.

48 

What makes Processing Instruction effective?

Activity B Things people did now and last summer. Listen to the following statements and decide whether each statement refers to an activity that takes place now or took place last summer in London.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.



NOW □ □ □ □ □ □ □ □ □ □

LAST SUMMER □ □ □ □ □ □ □ □ □ □

Sentences heard by learners: 1. People worked overtime at work. 2. People visit London for the first time. 3. People celebrated different festivals. 4. People handed out bread to the pigeons at Trafalgar square. 5. People protest in London about the war. 6. People walked to work. 7. People danced at nightclubs. 8. People watch at lot of TV. 9. People started a new hobby. 10. People receive gifts from their friends and family.

as in ikimashita). The same semantic notion is, however, also expressed in a sentence by a lexical items such as temporal adverbs (e.g. in Italian ieri; in English yesterday; in Japanese kinoo). Given, as postulated in the first corollary (Principle 1a), that learners are driven to process content words before anything else, they would attend to lexical temporal references of ‘pastness’ before verbal inflections of the past tense. Learners will prefer to mark tense lexically before they mark it morphologically and therefore will skip the grammatical forms. SIA can then be designed to help learners paying attention and process these forms affected by the Lexical Preference Principle. In the referential activities A, B and C learners must process the input to determine whether the statement they hear is referring to a present or past action. Learners are obliged to attend to the grammatical markers (present vs.

What are the main components of Processing Instruction? 

49

Activity C Listen to the following statements made by a journalist about the life of Pele and decide whether each statement is referring to his past life as a football player or his present life as a sports minister.

Football player □ □ □ □ □ □ □ □ □ □

Sports minister □ □ □ □ □ □ □ □ □ □

Sentences heard 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

ha giocato per il Brasile fa una nuova legge per l’educazione sportiva ha vinto due coppe del mondo ha rappresentato il suo paese all’estero ha segnato molti goal incontra spesso il primo ministro aiuta i giovani ha giocato per il Santos lavora per il governo brasiliano va spesso nelle scuole in Brasile

past in Japanese, Italian and English). Considering that learners are affected by the Lexical Preference Principle (P1 d.) temporal adverbs for past and present were removed from the input sentences so that students were forced to attend to the past or present tense forms to encode the meaning. The Preference for Non-redundancy Principle VanPatten’s third corollary (P1.c) is currently formulated as the meaningfulbefore-non meaningful principle. Its roots lie in the construct of a grammatical item’s communicative value to the learner. VanPatten (1996: 24) stated that ‘it is the relative communicative value of a grammatical form that plays

The effects of structured input activities 50 

What makes Processing Instruction effective?

a major role in determining the learner’s attention to it during input processing and the likelihood of its becoming detected and thus part of intake’. Communicative value refers to the relative contribution a grammatical item makes to the meaning of an utterance from a second language learner’s perspective. In order to establish whether a linguistic form has low or high communicative value, we need to follow two criteria: (1) inherent referential meaning, and (2) semantic redundancy. Word-final inflections on adjectives (o and a) in Italian are, for example, low in communicative value because they are redundant and lack inherent semantic value. In the Italian phrase la penna rossa (the red pen), rossa ends in -a- because the noun it modifies is grammatically feminine. The -a does not carry any inherent referential meaning in this phrase. In addition, -a characterises how three words in the noun phrase end; the grammatical marker is, therefore, highly redundant. If learners do not need the element to process the meaning of the phrase, then they most likely will not process it. In affective activity D and referential activity E learners attention is to the gender agreement forms in order to process the sentence correctly.

Activity D Read the following statements about Uma Thurman and indicate whether they are true or false. Vero Falso 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

è brutta è acuta è una brava attrice è chiusa è limitata è dotata è aperta è stupida è fredda è una persona calda

□ □ □ □ □ □ □ □ □ □

□ □ □ □ □ □ □ □ □ □

Compare your answers with a classmate: Sei d’accordo? Or Non sei d’accor­do? And establish the following:

I have a positive view of Uma □

I have a negative view of Uma □

What are the main components of Processing Instruction? 

51

The Sentence Location Principle The sixth corollary (P1.f) lays out a hierarchy of difficulty, or rather a hierarchy of processing saliency, with regard to L2 features. In a sentence like Penso che Paolo sia un buon giocatore (I think that Paul is a good player) the easiest forms to process are those located in salient positions, specifically, initial position (Penso) within an utterance. The second easiest forms to Activity E Listen to each sentence in which a person is described and determine which person is described.

Claudia Schiffer



□ □ □ □ □ □ □ □ □ □

Brad Pitt

□ □ □ □ □ □ □ □ □ □

Sentences heard by learners: 1. È bella 2. È bello 3. È Americano 4. È Tedesca 5. È bruno 6. È bionda 7. È viva 8. È antipatico 9. È magra 10. È alto

52 

What makes Processing Instruction effective?

Activity F Ascolta e determina se la frase è affermativa o negativa. Presta attenzione al verbo in posizione finale!

1. 2. 3. 4. 5.



Koutei 肯定 □ □ □ □ □

Hitei 否定 □ □ □ □ □

Sentences heard by learner: 1. 2. 3. 4. 5.

Watashi wa italiago o benkyou shimasu Watashi wa italiago o benkyou shimasen Watashi wa London ni sumimasu Watashi wa London ni sumimasen Watashi wa gekijyou ni ikimasen

私はイタリア語を勉強します。 私はイタリア語を勉強しません。 私はロンドンに住みます。 私はロンドンに住みません。 私は劇場に行きません。

process occur in the other salient position, utterance-final position (un buon giocatore). Finally, the most difficult forms to process are those that occur in utterance-medial position (sia). We tend to pay attention to what comes first (primacy) and what comes last (recency). Initial and final positions are privileged in terms of processing. The problem for second language learners is that many grammatical features occur in sentence medial position. In the French utterance Je ne pense pas que Gérard Depardieu soit très elegant (‘I do not think that Gérard Depardieu is very elegant’) the subjunctive form is located in medial position and therefore the most difficult form to process in the sentence. Learners do not attend to sentence medial position. The consequence for second language acquisition is that no intake is derived around the form. In referential activity F learners’ attention is directed to the final part of the sentences which learners have heard. This is because they can establish whether the sentence in Japanese is affirmative or negative. In affective activity G learners’ attention is directed to the subjunctives forms in medial position as they have to decide if the speaker is expressing certainty about the statements or is expressing a value judgment about the statements.

What are the main components of Processing Instruction? 

53

Activity G Read each phrase about Ricki Martin and decide if the speaker is expressing certainty about the statement or is expressing a value judgement about the statement 1. … è celebre in tutto il mondo Non penso che… So che…

6. … è bello e affascinante Penso che… Si sa che…

2. … sappia cantare molto bene Sono sicuro che… Non credo che…

7. … sia popolarissimo tra le donne giovani So che… Dubito che…

3. … possa ballare Dubito che… Sono certo che… 4. … ha molti soldi Non credo… So che… 5. … ha una bella voce E certo che… Dubito che…

8. … abbia una bella fidanzata Si sa che… Non credo che… 9. … non è molto intelligente Sono certo che… Non penso che… 10. … non piaccia a molte persone Temo che… So che…

Which statements best describes your feeling about Ricki Martin?

The First Noun Principle In the second principle (P2) VanPatten (1996) argues that learners tend to process the first noun or pronoun they encounter in a sentence as the subject or agent. In a language such as Japanese an object is often placed before the subject (OSV) and the verb at the end of the sentence (word order). The First Noun Principle might affect language processing. In the sentence Chris hit Maria (see below), learners might process Maria as the subject of the sentence and this will lead to a misinterpretation of the sentence and delay in acquisition.

Maria o Chris wa nagutta Maria Chris hit

54 

What makes Processing Instruction effective?

Japanese for example allows L2 learners to express the same content by more than one word order like SOV, OSV, OV. Apart from the word order example provided, other linguistic features are affected by the First Noun principle in Japanese: a. case marker b. comparative c. passive a. Kumakun wa Yoshikocahn o sukidesu (SOV) Yoshikocahn o Kumakun wa sukidesu (OSV) Both sentences are possible and mean ‘Kuma likes Yoshiko’.

Activity H Listen to each sentence and select a sentence that matches what you hear. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

□ Yoshiko asked Tom. □ Yoshiko complimented Tom. □ Yoshiko kissed Tom. □ Yoshiko woke Tom up. □ Yoshiko broke Tom’s cup. □ Yoshiko called Tom. □ Yoshiko looked at Tom. □ Yoshiko invited Tom. □ Yoshiko scolded Tom. □ Yoshiko said to Tom.

□ Tom asked Yoshiko. □ Tom complimented Yoshiko. □ Tom kissed Yoshiko. □ Tom woke Yoshiko up. □ Tom broke Yoshiko’s cup □ Tom called Yoshiko. □ Tom looked at Yoshiko. □ Tom invited Yoshiko □ Tom scolded Yoshiko. □ Tom said to Yoshiko.

Sentences heard by learners: 1. よしこちゃんは　トムさんに　たのまれました。 2. よしこちゃんは　トムさんに　ほめられました。 3. トムさんは　よしこちゃんに　キスされました。 4. よしこちゃんは　トムさんを　おこしました。 5. トムさんは　よしこちゃんに　コップを　こわされました。 6. トムさんは　よしこちゃんに　よばれました。 7. よしこちゃんは　トムさんを　みました。 8. トムさんは　よしこちゃんに　さそわれました。 9. よしこちゃんは　トムさんを　しかりました。 10. トムさんは　よしこちゃんに　いわれました。

What are the effects of structured input activities? 

55

b. watashi no hooga anata yori utsukushii (‘I am more beautiful than you’) anata yori watashi no hooga utsukushii (‘I am more beautiful than you’) c. neko wa inu ni oikakerareta (‘A cat was chased by a dog’) cat dog was chased Sentence (c) must be interpreted by L2 learners as if it was the cat who chased the dog as L2 learners would process the first item in the sentence as the agent (subject) of the sentence. In referential activity H learners are asked to pay process sentences keeping in mind that the first noun they encounter is not always an agent in the SOV (subject object verb) sentence. Learners are asked to find out who did what to whom. In affective activity I learners are asked to pay attention to the word order of the sentence and then they must agree or disagree with the statements provided. What are the effects of structured input activities? One line of research in the Processing Instruction research agenda is the one which has investigated what are the effects of the three components of Processing Instruction (see Table 2.1 for overall summary of this line of research conducted to measure Processing Instruction individual components). The main objective purpose of this classroom-based research was to establish which factor is the most effective component in the Processing Instruction approach. Is it the explicit information or the structured input practice? Or perhaps a combination of the two? The first study which attempted to address these questions was conducted by VanPatten and Oikkenon (1996). Fifty-nine subjects participated in this classroom experiment. The students were studying Spanish in secondary school and were in their second year. The item investigated was the object pronouns in Spanish and the processing principle under investigation was the First noun principle. Three groups were compared . One receiving only explicit instruction, the other structured input activities and the third both components (full Processing Instruction). The groups received 4 hours of instruction. As shown in Table 2.2 the performance of the three groups at pre-test level on the interpretation and production tasks indicates a low level of knowledge on the use of object pronouns in Spanish. The mean scores on the pretests for the interpretation task ranged from 8% to 11%; and for the production task ranged

56 

What makes Processing Instruction effective?

Activity I Each sentence below indicates the possibility of happening in your daily life. Have you had same experiences? Read each sentence and tick “I have had a same experience” or “I have never had a same experience”. Same Different 1. せんせいに　しかられました。 □ □ 2. せんせいに　ほめられました。 □ □ 3. せんせいに　おこされました。 □ □ 4. せんせいに　ペンを　つかわれました。 □ □ 5. せんせいに　パブに　さそわれました。 □ □ 6. せんせいに　かいものを　たのまれました。 □ □ 7. せんせいに　ビールを　のまれました。 □ □ 8. せんせいに　クリスプを　たべられました。 □ □ 9. せんせいに　「こんにちは」といわれました。 □ □ 10. せんせいに　たすけられました。 □ □

from 0% to 9%. After instruction the mean scores on the posttests for both the Processing Instruction group and the structured input only ranged from 45% to 44% in the interpretation task and from 21% to 30% in the production task indicating an improvement of from 16% to 36% in both assessment measures (see Table 2.2). The statistical analyses showed that the Processing Instruction and the structured input only group made similar gains whereas the explicit information only group did not. The statistical analyses revealed that the gains made (on both the interpretation and production tasks) by both the Processing Instruction and the structured input activities group were greater than the group receiving only explicit instruction on the targeted form. The outcomes of this study was that structured input activities were found responsible for learners gains. A very significant finding of this study is that the structured input activities group performed as well as the Processing Instruction group. As indicated by VanPatten (1996: 126), these findings strongly suggest that it is the structured input activities itself and the form–meaning connections being made during input processing that are responsible for the relative effects observed in the present and previous studies. Benati (2004a) replicated this study comparing the effects of Processing Instruction, structured input activities and explicit information on the acquisition of Italian future tense. This study addressed the Lexical Preference Principle. Thirty eight subjects studying Italian at undergraduate level participated in this classroom experiment. The population was divided, as in the

What are the effects of structured input activities? 

57

case of the original study, into three groups receiving respectively: Processing Instruction, structured input only, explicit information only. The three groups received instruction over a period of two consecutive days (six hours in total). One interpretation and one production task were developed for this study. The interpretation task consisted of twenty aural sentences (ten were distracters) where subjects had to interpret whether the sentences heard referred to a present or future event. The test was developed to measure the learners ability to interpret sentences containing the target form. The written production task was developed to measure subject’s accuracy in producing correct future tense forms (subjects had to produce five sentences). The mean scores of the pretests (see Table 2.2) on both interpretation and production tasks showed that subjects in the two groups were comparable in terms of showing a very low level of knowledge about the future tense in Italian. The mean scores on the pretests for the interpretation task ranged from 21% to 27%; and for the production task ranged from 18% to 22%. After instruction the mean scores on the posttests for both the Processing Instruction group and the structured input only ranged from 78% to 83% in the interpretation task and from 65% to 73% in the production task indicating an improvement of from 46% to 55% in both assessment measures. The results of a second post-test revealed similar improvements (see Table 2.3). The statistical analyses showed that the Processing Instruction and the structured input only group made similar gains whereas the explicit information only group did not (its gains were minimal). The results of this study confirmed the findings obtained in the VanPatten and Oikkenon’s study (1996). It was the structured input activities component that is responsible for learners improved performance. A second replication’s study was conducted by Benati (2004b) on the acquisition of Italian of gender agreement. This study addressed the Preference for Non-Redundancy Principle. The structured input activities were developed with the intention of helping learners to process the target form efficiently and correctly. Thirty one students studying Italian at undergraduate level were the population in this study. Even in this case subjects were divided into three groups: the first received Processing Instruction; the second group structured input only; the third group explicit information only. The three groups were exposed to instruction over two consecutive days for a total of four hours instruction. As in previous studies measuring the effects of Processing Instruction components, interpretation (one measure) and production (two measures) were used in a pre and post-test design. The ­interpretation task consisted in twenty audiotaped sentences (ten targets and

58 

What makes Processing Instruction effective?

ten distracters).Learners were asked to make the right gender agreement connections. The written production task was develop to measure the accuracy in which learners can produce accurate gender agreement forms. In the oral production task learners were given two pictures illustrating life in a city and a small town and were asked to describe and make comments about life in these two environments. The task was designed to elicit the use of correct gender agreement in more spontaneous speech. The mean scores of the pretests (see Table 2.2) on both interpretation and production tasks (written and oral) indicate the groups had a very little knowledge of gender agreement in Italian. The mean scores on the pretests for all the three tests ranged from 16% to 26%. The mean scores on the interpretation posttest for both the Processing Instruction group and the structured input only ranged from 71% to 760% indication an improvement of from 48% to 50%. The improvement of the explicit information only group was minimal (10%). The mean scores for the Processing Instruction group and the structured input only group on both production tasks ranged from 55% to 63% indicating an improvement of from 35% to 42% in both assessment measures. Again the explicit information only group showed very little improvement (from 3% to 7%). Once more the results of the statistical analyses were similar to those of VanPatten and Oikkenon (1996). The Processing Instruction group and the structured input group made significant gains on a sentence-level interpretation test and a sentence-level production tests, while the explicit information group made very little gains. The structured input group also made identical gains to the Processing Instruction group in the oral production task, compared to the explicit information group. Wong (2004) found positive results for structure input practice alone in a study where she made the same comparisons as in the studies described previously in this paragraph. She compared the effects of Processing Instruction, structured input activities and explicit information in the acquisition of French negative + indefinite article (Lexical Preference principle). ­Ninety-four undergraduate, intermediate students of French participated in this study. Subjects received instruction which lasted one day. The materials were designed to alter the processing problem and an interpretation and a production task were developed. The interpretation task was a sentence level task consisting of twenty items in the target linguistic feature and ten distracters. The production task had twelve items with six distracters. The mean scores on the pretests for both the interpretation and the production task ranged from 6% to 16% indication a very little knowledge of the target feature under investigation. After the instructional period the mean scores

What are the effects of structured input activities? 

59

on the interpretation posttests for both the Processing Instruction group and the structured input only ranged from 85% to 55% showing an improvement of from 43% to 79%. In the production task the mean scores of the posttest for the Processing Instruction group and the structured input only agree were similar ranged from 17% to 36% with an improvement from 17% to 36% (see Table 2.5). The explicit information only group made some gains from the pretests to the posttests for both measures (26% for the interpretation task and 14% for the production task). The control group made no gains. The statistical analyses revealed that in both the interpretation and the production task both the Processing Instruction group and the structured input only group were not different and better than the explicit information group and the control group. The little improvement on the interpretation and production tasks (but not statistically significant) of the explicit information only group was attributed to the characteristic of the particular target structure under investigation. Wong (2004: 195) argued that ‘‘ learners might be able to rely on the EI to help them make gains on this type of task. The fact that the EI group was not better than the control group, however, also suggests that the impact of this information was minimal’’. The structured input component seemed to be the causative factor for the beneficial effects of Processing Instruction. These results are confirmed for Spanish, French and Italian and on a variety of linguistic items that represent various processing problems. The addition of explicit grammatical information in the Processing Instruction treatment did not cause greater improvement. Sanz (2004) provides further evidence in support of the great the effects of structured input activities in the acquisition of Spanish object pronouns (First Noun Principle was the processing strategy which was attempted to alter by providing structured input practice). Forty-two students studying Spanish participated in the study. Sanz used two groups: one receiving structured input activities only; and a second group receiving the same structured input treatment plus explicit feedback on their wrong answers. Three tests were used. The interpretation and production sentence completion tasks were the same task used by VanPatten and Oikkenon (1996). The videoretelling task was developed to elicit the ability for participants to produce a connected short composition. The mean scores (see Table 2.2) on the pretests for the interpretation task were around 26% for both groups. The mean scores of both the structured input group and the structured input plus feedback group ranged from 62% to 66% indication a similar improvement on the posttest of the interpretation task (ranged from 36% to 40%).

60 

What makes Processing Instruction effective?

The mean scores on the pretests for the production sentence level task were also very similar and around 42% for both groups. The mean scores of both groups in the posttest ranged from 66% to 83% indicating an improvement of 23.6% for the structured input only group and 41.5% for the structure input plus feedback group. The mean scores on the pretests for the production video-retelling task ranged from 22% and 31% for both groups. The mean scores of both groups in the posttest ranged from 55% to 58%. This was indication of an improvement of 24% for the structured input only group and 35% for the structure input plus feedback group. The statistical analyses carried out on the test results (interpretation or production tasks) showed no significant differences in the two groups. Farley (2004) also conducted a replication study in which he measured the effects of Processing Instruction and structured input activities only on the acquisition of Spanish subjunctive of doubt (Sentence Location Principle). Fifty-four students participated and were divided into two groups which received two hours instruction. One group received full Processing Instruction and the other structured input practice. Likewise most of the previous studies, one interpretation and one production sentence level tasks were used to measure learner’s performance. The mean scores on the pretests for the interpretation sentence level task were very similar for both groups and ranged from 31% and 39%. In the first posttest and the delayed posttest the Processing Instruction group (mean scores for the first posttest = 64%, mean scores for the delayed posttest = 67%) outperformed the structured input group (mean scores for the first posttest = 43%, mean scores for the delayed posttest = 50%). The improvement of the Processing Instruction group was higher than the structured input group (see Table 2.2). The mean scores on the pretests for the production sentence level task were very similar for both groups and ranged from 2 and 5%. In the first posttest and the delayed posttest the Processing Instruction group (mean scores for the first posttest= 44%, mean scores for the delayed posttest = 37%) outperformed the structured input group (mean scores for the first posttest= 23%, mean scores for the delayed posttest = 22%). The improvement of the Processing Instruction group was higher than the structured input group (see Table 2.2) The results of the statistical analyses conducted in this study were slightly different than the previous ones. Despite the fact that both groups made

What are the effects of structured input activities? 

61

significant improvements from pre to post-tests, the Processing Instruction group outperformed the structured input practice group both in the interpretation and the production task. Farley attributes this finding to the complexity of the linguistic item he investigated, the Spanish subjunctive. Lee and Benati (2007a) also investigate the effectiveness of structured input activities in the acquisition of Japanese past forms and affirmative vs. negative present forms. These forms are affected by a combination of different processing problems (The Lexical Preference Principle, and The Sentence Location Principle. Participants in this study were beginning level students of Japanese. All subjects were adult native speakers of Italian and were studying Japanese in a private school in Italy. Twenty seven participants were assigned to two groups. One group received structured input activities and one group received traditional instruction over a four hours instructional period. A pretest/posttest approach to examine the short term effects of the two instructional treatments was adopted. Two assessment measures were produced for each linguistic feature: one for the interpretation tasks and one for the production tasks. These tasks were both sentence level task and very similar to the ones previously described in this chapter. In the case of the interpretation and production task on the affirmative vs. negative forms in Japanese, the mean scores (see Table 2.2) on the pretests were fairly similar ranging from 4% to 6% for both groups. However, the mean scores of the structured input group (62%) was different than the mean scores of the traditional instruction group (13%)on the posttest interpretation task. The structured input group made greater and significant improvement (56%) compared to the traditional group (7%). The mean scores of both groups in the posttest production task ranged from 59% to 61% indicating a similar improvement ranged from 55% to 57%. The results of the Japanese past tense were very similar (see Table 2.2). The statistical analyses carried out on the test results (interpretation or production tasks) showed significant differences in the two groups. The groups who carried out structured input activities performed better than the traditional instruction groups on the interpretation tasks on both linguistic features of Japanese (affirmative and negative present tense and past tense). The results from the production tasks also confirmed previous findings. The structured input activities groups performed equally to the traditional instruction groups on the production tasks on both linguistic features of Japanese (affirmative and negative present tense and past tense).

62 

What makes Processing Instruction effective?

Summary In this chapter we have examined the characteristics and the nature of this new approach to grammar instruction which takes into consideration how learners first process the input they are exposed to. PI is very effective as it provides learners with opportunities through structured input activities, to process forms and structures in a target language more efficiently than any other approach to grammar teaching. In PI we first of all identify a particular linguistic feature in the language we want to teach. This form or structure might be affected by one of the processing principles we outlined in chapter one and therefore it could be very difficult for learners to process. In order to address possible processing problems we then provide L2 learners with information about the linguistic feature and the processing problem. Finally, and more importantly we engage L2 learners in structured input activities practice. The main finding of the line of research teasing out the main factors responsible for the effectiveness of Processing Instruction confirmed that it is the structured input component that is responsible for the changes in learners developing system and eventually in their output. As a result of the empirical evidence collected in the research which has compared Processing Instruction versus its components, we are able to conclude that the causative factor in the positive effects for Processing Instruction is due to the effects of the structured input activities. These has been proved and observed in different processing principles, languages, linguistic items and assessment task. Structured input activities represent the most significant variable within the Processing Instruction approach. As indicated by VanPatten (1996: 126), structured input activities and the form–meaning connections being made during input processing are responsible for the relative effects observed.

Linguistic feature/ language

Spanish object pronouns

Italian future tense

Spanish subjunctive

Spanish object pronouns

Study

VanPatten & Oikennon 1996

Benati 2004a

Farley 2004

Sanz 2004

First noun

Lexical preference

Lexical preference

First noun

Processing principle

Number

First and second year courses

53

Intermediate 54 English native

Beginners 38 English native

Intermediate 59 English native

Subjects/L1

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

Design

Int. = SI = SI + EFB Interpretation Production (sentence completion and video retelling

Prod.= SI = SI + EFB

Int. = PI > SI Prod. = PI > SI

Int. = (PI = SI) > EI Prod. = (PI = SI) > EI Interpretation (aural) Production (written) Interpretation (aural) Production (written)

Int. = (PI = SI) > EI Prod. = (PI = SI) > EI

Results

Interpretation (aural) Production (written)

Tests

Table 2.1.  Summary of primary effects for the components of Processing Instruction (PI – TI- SI – EI)

Appendix Appendix  63

Lexical preference Sentence location

Adjective agreement

Japanese past tense

Japanese affirma­tive vs. negative present tense

Benati 2004b

Lee and Benati 2007a

Number

Beginners Italian native

27

Beginners 31 English native

Intermediate 94 English native

Subjects/L1

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate

Design

Interpretation (aural) Production (written)

Interpretation (aural) Production (written)

Interpretation (aural) Production (written) Production (oral)

Interpretation (aural) Production (written)

Tests

Int. = SI> TI Prod. = SI = TI

Prod. = (PI = SI) > EI Prod. = (PI = SI) > EI Int. = SI> TI Prod. = SI = TI

Int. = (PI = SI) > (EI = C) Prod. = (PI = SI) >C PI > EI EI=SI EI = C Int. = (PI = SI) > EI

Resutlts

PI = Processing Instruction comprising EI and SI; SI = Structured Input activities only; EI = Explicit Information only; C = Control group; EFB = Explicit feedback; Int. = Interpretation test; Prod. = Production test/

Preference for nonredundancy

First noun

French negative + indefinite article

Wong 2004

Processing principle

Linguistic feature/ language

Study

Table 2.1.  (cont.)

64  What makes Processing Instruction effective?

Spanish object pronouns

Italian future tense

Spanish subjunctive

VanPatten & Oikennon 1996

Benati 2004a

Farley 2004

PI vs. SI PI > SI

PI vs. SI vs. EI (PI = SI) > EI

PI vs. SI vs. EI (PI = SI) > EI

Linguistic Treatment and feature/language results

Study

Int. PI = 3.1 SI = 3.9 Prod. PI = .2 SI = .5

Int. PI = 2.7 SI = 2.1 EI = 2.6 Prod. PI = 2.2 SI = 1.9 EI = 1.8

Int. PI = 0.8 SI = 0.8 EI = 1.1 Prod. PI = 0.9 SI = 0.5 EI = 0.0

Means scores Pretest

Posttest 1 Int. PI = 6.4 SI = 4.3 Prod. PI = 4.4 SI =2.3

Posttest 1 Int. PI = 8.3 SI = 7.8 EI = 4.6 Prod. PI = 7.3 SI = 6.5 EI = 3.1

Int. PI = 4.3 SI = 4.4 EI = 1.5 Prod. PI = 3.0 SI = 2.1 EI = 0.2

Means scores Posttest

Posttest 2 Int. PI = 50% SI = 54% EI = 12% Prod. PI = 43% SI = 29% EI = 7% Posttest 2 Int. PI = 36% SI = 11% Prod. PI = 36.8% SI =21.5%

Posttest 1 Int. PI = 55% SI = 57% EI = 20% Prod. PI = 51% SI = 46% EI = 13% Posttest 1 Int. PI = 33% SI = 4% Prod. PI = 43.8% SI = 22.5%

Posttest 2 Int. PI = 7.7 SI = 7.5 EI = 3.8 Prod. PI = 6.5 SI = 4.8 EI =2.5 Posttest 2 Int. PI = 6.7 SI = 5.0 Prod. PI = 3.7 SI =2.2

Int. PI = 35% SI = 36% EI = 2% Prod. PI = 21% SI = 16% EI = 2%

Improvement

Table 2.2.  Summary of the improvement from pretest to posttest(s) of studies investigating the effects for the components of ­Processing Instruction

Appendix  65

Spanish object pronouns

French negative + indefinite article

Sanz 2004

Wong 2004

PI vs. SI vs. EI vs. C Int. (PI = SI) > (EI = C) Prod. (PI = SI) > C PI > EI EI = SI EI = C

SI vs. SI+EFB SI = SI-EFB

Linguistic Treatment and feature/language results

Study

Table 2.2.  (cont.)

Int. PI = .6 SI = 1.2 EI = .6 C = .9 Prod. PI = 1.2 SI = 1.6 EI = 1.0 C = 1.2

Int. SI = 26.2 SI+EFB = 26.2 Prod. sent. SI = 42.4 SI+EFB = 42.4 Prod. video SI = 31.3 SI+EFB = 22.6

Means scores Pretest

Int. PI = 8.5 SI = 5.5 EI = 3.2 C = .9 Prod. PI = 4.8 SI = 3.3 EI = 2.4 C = 1.2

Int. SI = 62.5 SI+EFB = 66.6 Prod. sent. SI = 66.0 SI+EFB = 83.6 Prod. video SI = 55.7 SI+EFB = 58.1

Means scores Posttest

Int. PI = 79% SI = 43% EI = 26% C = 0% Prod. PI = 36% SI = 17% EI = 14% C = 0%

Prod. sent. SI = 23.6% SI+EFB = 41.5% Prod. video SI = 24.4% SI+EFB = 35.5%

Int. SI = 36.3% SI+EFB = 40.5%

Improvement

66  What makes Processing Instruction effective?

Japanese affirmative vs. Negative present tense

Japanese past tense marker

Lee & Benati 2007a

Lee & Benati 2007a

SI vs. TI Int. SI> TI Prod. SI = TI

SI vs. TI Int. SI> TI Prod. SI = TI

PI vs. SI vs. EI (PI = SI) > EI Int. PI = 2.6 SI = 2.3 EI = 2.5 Prod. written PI = 2.1 SI = 2.1 EI = 1.6 Prod. oral PI = 1.7 SI = 1.5 EI = 1.6 Int. SI = .6 TI = .6 Prod. SI = .4 TI = .4 Int. SI = .6 TI = .4 Prod. SI = .4 TI = .6

Int. PI = 7.6 SI = 7.1 EI = 3.5 Prod. written PI = 5.6 SI = 6.3 EI = 1.9 Prod. oral PI = 5.9 SI = 5.5 EI = 2.3 Int. SI = 6.2 TI = 1.3 Prod. SI = 5.9 TI = 6.1 Int. SI = 7.3 TI = 1.2 Prod. SI = 6.0 TI = 6.0

Int. PI = 50% SI = 48% EI = 10% Prod. written PI = 35% SI = 42% EI = 3% Prod. oral PI = 42% SI = 40% EI = 7% Int. SI = 56% TI = 7% Prod. SI = 55% TI = 57% Int. SI = 67% TI = 8% Prod. SI = 56% TI = 54%

PI = Processing Instruction comprising EI and SI; SI = Structured Input activities only; EI = Explicit Information only; C = Control group; EFB = Explicit feedback; Int. = Interpretation test; Prod. = Production test/

Adjective agreement

Benati 2004b

Appendix  67

Processing Instruction and other types of instruction

Chapter 3 How does Processing Instruction compare to other types of instruction? Introduction Over the last fifty years we have witnessed some dramatic changes in the way language is taught in the language classroom. In the grammar translation approach, teachers prioritized the explicit teaching of grammatical rules and they expected learners to be able to translate texts from L1 to L2. The Direct Method focused on developing learners’ ability to listen and make associations of words and phrases. According to this method, grammar should be taught inductively. Learners should learn grammar by interpreting contextual and situational cues rather than receiving explicit explanations. In the 50s the Audiolingual method, based on the habit formation theory, argued that good language habits are learned through the process of repetition, imitation and reinforcement. This method emphasized the use of memorization and pattern drills. During the 60s and 70s the Cognitive-Code Method, influenced by Chomsky (1965), in contrast with the Audio-lingual Method, sustained that rather than mimic language structures learners need to understand and to analyze grammar in order to build linguistic competence. Learning is not the result of habit formation but of an analytical process that moved from previously acquired knowledge to unknown knowledge. With the advent of the communicative language teaching approach the role of grammar diminished. Heikel and Fotos (2002) have described communicative approaches to language instruction as instruction that does not include formal grammar instruction and the correction of learner errors. The assumption is that grammar instruction does not help learners develop any kind of communicative ability in the L2. In communicative language teaching learners are asked to perform tasks with large quantities of meaning-focused input containing target forms and vocabulary. Classroom research in the early eighties on the effects of formal instruction on second language acquisition principally focused on one main issue: whether grammar instruction per se makes a positive impact on the acquisition of a second language (Long 1983). Formal instruction was relegated to a fragile and peripheral role in Krashen’s theory (1982), according to which formal grammar instruction permits the learners to monitor their

What is the role of grammar instruction? 

69

L2 production but does not have any effects on learners’ competence. In order to determine whether instruction might have a positive role, Long (1983) reviewed the results of studies examining the effect of formal instruction on the rate and success of second language acquisition. According to Long there is enough evidence to sustain the view that second language instruction makes a difference in terms of being beneficial for adults as well as for children, for intermediate and advanced students, despite the way it is measured (by means of integrative or discrete-point tests) and in acquisition-rich and acquisitionpoor environments (that is, in settings where learners have little opportunity to hear the language outside their language class). He concluded that a combination of instruction and exposure to the language was more beneficial than exposure alone, as instruction seems to speed up the acquisition processes. Savignon’s (1972) measured the communicative and linguistic abilities of three groups of beginners enrolled in a French audio-lingual programme. The findings of this study suggests that a combination of formal and informal instruction is very effective in the development of communicative language skills in foreign language learners. A second experimental study also provided support for this view. Spada (1987) carried out an investigation to measure the effects of different instructional programmes (with an highly communicative syllabus) on learners’ improvement in proficiency. The evidence provided by Spada suggests that learners make more rapid progress when they experience both form-focused and meaning-focused instruction. Spada’s findings echoed the results of a previous study carried out by Montgomery and Eisenstein (1985). Likewise Spada, Montgomery and Eisenstein concluded that a combination of focus on form and focus on meaning in language teaching is more beneficial than an exclusive focus on form. What is the role of grammar instruction? There is still a very open debate on the role and the effects of grammar instruction in second language acquisition. Despite the fact that we cannot draw any definite conclusion about its role, we are able to claim today that although the effects of grammar instruction appeared to be limited, grammar instruction might have a facilitate role. There is empirical evidence which has shown that grammar instruction seems to promote more rapid SLA and to contribute to higher levels of ultimate achievement (Long 1983, 1991; Ellis 1994, 2001). Schmidt (1990, 2001) has emphasized the role of consciousness. He has argued that one of the main conditions for learners to acquire

70 

Processing Instruction and other types of instruction

a target language is the development of conscious attention to form (‘noticing’). However, it seems that learners have an internal mechanism which organizes linguistic data independently of the order, explanation and practice in which the linguistic forms are presented. There seems to be psycholinguistic constraints which determine whether grammar instruction is successful or not. Instructors might be successful, as claimed by Pienemann (1984, 1998) if learners have reached a stage in the developmental sequence that enables them to process the target structure. As instructors, our task seems to be rather than teaching language we must create the conditions in which learners can acquire the language (Corder 1981). As suggested by Ellis (1994) there is some evidence to support the thesis that formal instruction helps L2 learners to develop greater L2 proficiency, particularly if opportunities for natural exposure are given. The important question seems to be, not whether grammar instruction ‘per se’ makes a difference but whether certain types of grammar instruction that are more effective than others in second language acquisition. One of the question raised by VanPatten (2004) is: how do we teach grammar so that instruction works with acquisition processes and not against them? Language classroom is becoming more and more communicative, however the way grammar is taught has hardly changed at all. As argued by Lee and VanPatten (1995, 2003) the challenge today is not whether or not we should teach grammar but to find a way to incorporate grammar in a communicative framework. The question is: which is the better approach or technique to grammar instruction that can be best incorporated in a communicative language teaching programme? These questions have been addressed by recent classroom-based research. This research has been conducted on one hand to ascertain the role of grammar instruction, and on the other hand to measure the effectiveness of specific approaches and techniques to grammar teaching (see full reviews in Spada 1997; Doughty and Williams 1998; Norris and Ortega 2000; Nassaji and Fotos 2004). According to Sharwood Smith, making certain features salient in the input might help drawing learner’s attention to that specific feature. Enhancing the input through different techniques might be sufficient in helping learners paying attention to the formal properties of a targeted language without the need of metalinguistic discussion. Rutherford and Sharwood Smith (1988) coined the term ‘consciousness raising’ to refer to external attempts to drawn learners’ attention to formal properties of a target language. The goal of this approach is to make learners conscious of the rules that governs the use of particular language forms while it provides the opportunity to engage in meaningful interaction.

What types of grammar teaching have been suggested? 

71

VanPatten (1996: 59) has proposed a positive role for grammar instruction in the acquisition process since instruction can make certain grammatical forms more salient in the input. He recognizes the essential role of input in second language acquisition and takes the view that grammar instruction provided through the input phase can be beneficial What types of grammar teaching have been suggested? The theoretical views briefly reviewed in the previous paragraph and many empirical findings have challenged the way grammar is taught and practiced. There has been a dramatic shift from traditional grammar-oriented methods to more communicative grammar approaches. This shift has meant a change in the way grammar is taught and practised in the language classroom. Traditional instruction In traditional methods, grammar was provided through long and elaborated explanations of the grammatical rules of the target language. Paradigms of those grammatical rules were provided and followed by output-based practice (written and oral exercises) where the main focus was to practice the grammatical rules to obtain accuracy. Paulston (1972) has argued that traditional instruction is usually provided following a particular sequence which goes from mechanical to communicative drills practice. Most books used to teach foreign languages, approach the teaching of grammar in a very traditional way. The Grammar section of these book is generally characterised by paradigmatic explanations of linguistic structures and grammatical principles in learners L1. The paradigmatic explanation is followed by pattern practice and substitution drills. Real life situations are completely ignored and practice is implemented in a completely decontextualised way. Inn this type of practice learners are provided with activities in which they have to transform or substitute grammatical forms. The main purpose of this type of practice is for learners to memorise the various forms before they are asked to practice them through mechanical drills. The question addressed in this paragraph is whether there are types of grammar teaching that could be incorporated successful in the teaching of a foreign language. The theoretical view briefly reviewed in the previous paragraph have on one hand indicated the limited role for instruction, and on the

72 

Processing Instruction and other types of instruction

other hand highlighted the importance of incorporating grammar in a more communicative framework of language teaching by devising grammar tasks that enhance the grammatical features in the input. The question is to determine what type of grammar is more successful in terms of helping learners internalise the grammatical features of a target language. Our challenge is to provide an alternative way to introduce the linguistic properties of this language in a communicative framework of language teaching. We should develop tasks that introduce learners to patterns of the language not through explanations and memorisation but rather through exposure and process of those linguistics features. Rather than using a deductive approach which involves the use of translation we advocate more inductive approaches. Meaning-based output instruction Lee and VanPatten (1995, 2003) have suggested that we should move from structured input activities to structure output activities when we aim at practicing grammar. Both activities follow the guidelines provided by Lee and VanPatten 2003: 154): 1. 2. 3. 4. 5.

present one thing at a time; keep meaning in focus; move from sentences to connected discourse; use both written and oral output; others must respond to the content of the output.

Both activities are based on Lee and VanPatten (1995: 121) guidelines for developing structured output activities. This is to say that learners of a foreign language should be involved in the exchange of previously unknown information while at the same time use and produce a particular target form or structure to express meaning. Input enhancement Comprehensible and meaning-bearing input is one of the main factor in SLA. Learners must be exposed to comprehensible and meaning-bearing input for acquisition to take place. Learners acquire a L2 by having access to two types

What types of grammar teaching have been suggested? 

73

of evidences: positive and negative. Positive evidence is the various utterances learners are exposed in the input. As Wong (2005) has highlighted, there are techniques that would help teachers to expose learners to comprehensible input and positive evidence while at the same time drawing learner’s attention to some linguistics properties of the target language. Wong (2005) has identified two main techniques that would help learners notice and possibly acquire a targeted feature: input flood and the textual enhancement techniques. The advantage of input flood is that it provides comprehensible meaning-bearing input. It is also effective as it does not disrupt the flow of communication. However, because this technique is so implicit it is difficult to know whether learners are learning something through the flood. Input flood As Wong (2005: 37) has highlighted in input flood ‘‘the input learners received is saturated with the form that we hope learners will notice and possibly acquire. We do not usually highlight the form in any way to drawn attention to it nor do we tell learners to pay attention to the form. We merely saturate the input with the form’’. When we design input flood activities we should follow the following guidelines: a. grammatical tasks using input flood should either be used in written or oral input; b. the input learners receive must be modified so that it contains many instances of the same form/structure; c. input flood must be meaningful and learners must be doing something with the input (i.e. reconstruct a story, draw a picture for instance). Textual enhancement Wong (2005: 48) has defined textual enhancement as the use of typographical cues such as bolding or italics to draw the reader’s attention to particular information in a text. Designing this type of grammar tasks will involve following these guidelines: a. grammatical tasks using textual enhancement should use written input; b. the target form is enhanced visually altering its appearance in the text (i.e. the form can be italicized, bolded or underlined). The font, character and

74 

Processing Instruction and other types of instruction

can be enlarged or the form could be altered by highlighting the form with different colors. Consciousness raising As defined by Ellis (1991: 236) a consciousness raising task ‘‘is pedagogic activity where the learners are provided with L2 data in some form and required to perform some operation on or with it, the purpose of which is to arrive to an explicit understanding of some linguistic property or properties of the target language’’. Consciousness raising tasks can be inductive or deductive. In the case of an inductive task learners are provided with some language data and are required to provide an explicit representation of the target linguistic feature. In the case of a deductive task learners are given a description of the target linguistic feature and are required to use that description to apply it to L2 data. Consciousness raising tasks have been designed with the following guidelines in mind: a. the task focuses on a source of difficulty for English speaking learners who are learning a targeted foreign language; b. the data provided is adequate to make learners discover the rule; c. the task requires minimal production on the part of the learner; d. there is an opportunity for applying the rule to construct a personal statement in order to promote its storage as explicit knowledge. Consciousness raising is an approach to grammar teaching in accordance with new views about education as a process of discovery through problem solving tasks. It does not conflict but provide a supplement to the teaching of grammar communicatively. Processing Instruction Unlike traditional instruction where the focus of instruction is in the manipulation of the learners’ output to effect changes in their developing system, Processing Instruction aims to change the way input is perceived and processed by language learners (see for a full description of Processing Instruc-

What types of grammar teaching have been suggested?  Intake

→

Developing system

→

Output

→

→ →

Input

75

Processing Instruction

Traditional instruction

Figure 3.1.

tion chapter 2). As pointed out earlier, this approach to grammar instruction is consistent with the input processing perspective in second language acquisition. It is therefore evident in Figure 3.1 that Processing Instruction in its attempt to alter the way L2 learners process input should have a greater impact on learners’ developing system than an output-based approach to grammar instruction (see Figure 3.1) whose aim is to alter how L2 learners produce the target language. Unlike output-based instruction which emphasized grammar rules and oral\written production practice, the purpose of Processing Instruction is to alter how learners process input and to encourage better form–meaning mapping which results in a grammatically richer intake. In the case of tense markers, Processing Instruction can make these redundant and non-salient grammatical meaning–form relationships more salient in the learner’s input. Given the emphasis on learners’ input rather than focusing in on the output, the type of practice provided by the Processing Instruction approach consists in activities which offer the opportunity to interpret the meaning–form relationship correctly without any practice in producing the targeted form or structure. This is accomplished (as also suggested by Terrell 1991) by providing learners with meaningful input that contains many instances of the same grammatical meaning–form relationship. This would appear to be a step forward to Sharwood Smith’s position (1993). He suggests a way to provide formal instruction which is based on making some forms more salient in the input so that they come to learners’ attention. Processing Instruction does this but it also provides opportunities for form–meaning mapping in activities. As outlined by VanPatten (1996: 84) ‘‘simply bringing a form to someone’s attention is not a guarantee that it gets processed … for acquisition to happen the intake must continually provide the developing system with examples of correct form–meaning connections that are the results of input processing’’. Processing Instruction, contrary to ‘negative enhancement’, does not address the role of output errors since it is solely concerned with the processing of input data. Processing Instruction might be considered, as mentioned earlier, as a type of

76 

Processing Instruction and other types of instruction

‘­consciousness-raising’; in the sense that, as indicated by VanPatten (1996: 85) it ‘‘does not seek to pour knowledge of any kind into learners’ heads; it assists certain processes that can aid the growth of the developing system over time’’. However, the ultimate scope of Processing Instruction is not about raising learners’ consciousness about a grammatical form but rather to enrich learners’ intake. Processing Instruction is as VanPatten (1997) claimed: ‘a psycholinguistic motivated approach to focus on form’ whose main aim is to teach grammar without sacrificing either communication or learning-centered activities and ultimately gets learners to practice grammar. Processing Instruction is a new type of grammar instruction (as described in chapter two) that is concerned with learners’ awareness of how grammatical forms and structures are acquired. It is a type of focus on form which draws on the principles of the input processing model (VanPatten 1996, 2002, 2004a). In the model of SLA proposed by VanPatten ‘‘input provides the data, input processing makes (certain) data available for acquisition, other internal mechanisms accommodate data into the system (often triggering some kind of restructuring or a change of internally generated hypotheses), and output helps learners to become communicators and, again, may help them become better processors of input’’(2002: 760). This new pedagogical approach, based on the input processing model (see VanPatten 1996, 2002, 2004), seeks to intervene in the processes learners use to get data from the input. Research on input processing has attempted to describe what linguistic data learners attend to during comprehension and which ones they do not attend to, for example what grammatical roles learners assign to nouns or how position in an utterance, influences what gets processed. These processing principles seem to provide an explanation of what learners are doing with input when they are asked to comprehend it. As a result of the way learners attend to input data VanPatten (1996) has developed a new kind of grammar instruction which guides and focuses learners’ attention when they process input. This new type of grammar instruction called Processing Instruction is diametrically opposed to traditional instruction which consists of drills in which learner output is manipulated and instruction is divorced from meaning or communication. Processing Instruction is a more effective method for enhancing language acquisition as it is used to ensure that learners’ focal attention during processing is directed toward the relevant grammatical items and not elsewhere in the sentence. Its main objective is to help learners to circumvent the strategies used by them to derive intake data by making them to rely exclusively to on form and structure to derive meaning from input.

Processing Instruction and traditional instruction 

77

What are the main findings of studies comparing Processing Instruction and traditional instruction? In this paragraph we briefly review five main empirical studies which have investigated the role and effects of Processing Instruction compared to traditional instruction (see Table 3.1). The so-called ‘original study’ was conducted VanPatten and Cadierno (1993) in order to investigated the relative effects of Processing Instruction in altering a processing problem known as The first noun principle. They investigated the impact of Processing Instruction on the acquisition of direct object pronouns as learners seem to misinterpret sentences containing direct object pronouns in Spanish. Three groups of students of Spanish at intermediate level (Eighty subjects participated to this study) who received different instructional treatments over two consecutive days of instruction were compared. One group received traditional instruction which emphasized grammar explanation and oral-written production; the second group received Processing Instruction as described in the previous chapter; and a third group was used as a control receiving no instruction. A pre-test/post-test design was used to measure the possible effects of instruction. Two different assessment tasks were developed by the researchers. (interpretation task and a sentence-level written production task). In Table 3.2 the performance of the three groups at pre-test level on the interpretation and production tasks indicates a low level of knowledge on the use of object pronouns in Spanish. The mean scores on the pretests for the interpretation task ranged from 11% to 17%; and for the production task ranged from 15% to 26%. Three posttests were administered and the mean scores of the Processing Instruction group on the interpretation tasks ranged from 71% to 80% showing an improvement which ranged from 54% to 63%. The mean scores of both the traditional instruction group and the control group (this group received no instruction) were similar and ranged from 14% to 38% in the three interpretation posttests. The traditional group improved (mean scores ranged from 17% to 25%) slightly more than the control group (mean scores ranged from 3 to 11%). In the production tasks the mean scores of both the Processing Instruction group and the traditional group ranged from 81% to 90% (see Table 2.3) indicating an improvement of from 55% to 67%. The means scores of the control group in the three posttests ranged from 34% to 45% indicating an improvement of from 19% to 30%. The results from the statistical analyses showed that Processing Instruction is superior to traditional instruction as learners receiving Processing

78 

Processing Instruction and other types of instruction

I­ nstruction improved in their ability at interpreting object pronouns in Spanish correctly and furthermore the study demonstrated that this approach was also effective in improving learners’ production. Cadierno (1995) carry out an experimental study measuring the relative effects of Processing Instruction on a different processing problem (Lexical Preference Principle). This study was a partial replication of the previous study in terms of the design used (pre-post tests) and the overall aims. In the case of this experiment, the linguistic items chosen was the Spanish preterite tense. Considering the effects of the processing problem affecting the targeted form, the aim of Processing Instruction in this study was to push learners to attend the grammatical element in the input that might be otherwise get missed. The participants of this study were sixty English native speakers studying intermediate Spanish at undergraduate level. Once again Processing Instruction was compared to traditional instruction on the acquisition of this grammatical feature. Likewise the previous study two assessment tests were used (interpretation and production written task). In this case, as in the previous study, the scores of the three groups on the pre-tests for both the interpretation (the mean scores on the pretests for the interpretation task ranged from 38% to 43%) and production tasks (ranged from 17% to 23%) indicated that the three groups were homogeneous in terms of knowledge on the use of Spanish preterite tense. The mean scores on the interpretation tasks of the Processing Instruction group showed clearly an improvement from pre to posttest. On the interpretation task the scores ranged from 74% to 78% indicating an improvement which ranged from 32% to 36%. Likewise the previous study, the mean scores of both the traditional instruction group and the control group were similar and ranged from 42% to 52% with a similar improvement ranging from 6% to 14%. In the production tasks the mean scores of both the Processing Instruction group and the traditional group ranged from 59% to 78% (see Table 3.2) indicating an improvement of from 53% to 55% for the Processing Instruction group and from 42% to 43% for the traditional instruction group. The means scores of the control group in the three posttests ranged from 32% to 36% indicating an improvement of from 9% to 13%. The statistical analyses carried on the scores of the three groups in the pre-test and posttests can be summarized as follow: the results of the interpretation task showed that the group who received Processing Instruction outperformed the group exposed to traditional instruction and the control group in the interpretation task; the results of the production task were the same as the results of the original study (VanPatten and Cadierno 1993). Both the Processing Instruction group and

Processing Instruction and traditional instruction  

79

the traditional group improved from pre to post-test on the production task. Benati (2001) investigated the effects of Processing Instruction and traditional instruction on the acquisition of Italian future tense. The processing principle was in this case the Lexical Preference Principle. Participants were beginner undergraduate students (39 subjects) learning Italian as part of their degree (all English native speakers). Three groups were formed: one received Processing Instruction; a second group received traditional instruction; the third group was used as a control group. Instruction lasted for six hours over two consecutive days. One interpretation test and two production tests were used for this study and consisted of an aural interpretation task, and a written com The mean scores of the three groups on the pre-tests for both the interpretation and production tasks were similar and ranged from 20% to 38%. On the interpretation tasks the mean scores of the Processing Instruction group (83%) were much higher than the traditional group (58%) showed clearly an improvement from pre to posttest (55% for the Processing Instruction group and 17% for the traditional group in the first posttest and 43% and 15% in the second posttest). However the mean scores of the traditional group were higher that the control group (see Table 3.2). On the production tasks the two instructional group performed (PI = 73% and TI = 77% on the written; PI=71% and TI = 76% on the oral and improved very much equally and better that the control group. Similar mean scores were obtained in a second posttest (see Table 3.2). Participants in the PI group outperformed the other two groups in the interpretation task and performed as well as the traditional group in both production tasks. Cheng (2004) also carried out an investigation to measure the effects of Processing Instruction on the acquisition of copular verbs in Spanish (ser and estar). The relevant processing principle was the The Preference for Nonredundancy Principle as copular verbs in Spanish are of low communicative value for L2 learners and redundant features of Spanish. Eighty-tree subjects studying Spanish (final year) at undergraduate level participated in this study. Cheng compared three groups (Processing Instruction, traditional instruction and control group which did not receive any explicit instruction on the targeted linguistic feature) using in a pre and post-test format including three type of tests (Interpretation, Sentence completion, Guided composition). The results of the interpretation task confirmed (this results are reported in Cheng 2002) previous results. Processing Instruction is an effective form of instruction at helping learners to interpret sentences correctly. The Processing Instruction group also performed as well as the traditional instruction group in the production tasks. Cheng has (2004) analyzed the results of the guided

80 

Processing Instruction and other types of instruction

composition data. As reviewed by Cheng the use of estar increased among the three groups. However the Processing Instruction group and the traditional group made greater gains (improvement in the range of 20% to 28%) compared to the control group (improvement in the range of 6% to 10%) in the first and second posttests (see Table 3.2). The statistical analyses showed that the Processing Instruction and the traditional instruction were better than the control group. Both instructional groups produced more estar tokens than the control group. The positive findings obtained in studies measuring the effects of Processing Instruction were further generalized through another experimental study (VanPatten and Wong (2004). VanPatten and Wong (2004) compared again the effects of Processing Instruction and traditional instruction on the acquisition of French faire causative. Like in the case of the original study, on relevance to this study was the First Noun Principle (P2). Seventy-six undergraduate students learning French were the subject of this study. They were enrolled in two different universities. Three groups were randomly formed (Processing Instruction, traditional instruction and the control group) in each university. As in previous studies to separate instructional packets were used. The assessment tasks consisted of an interpretation and a production test. The mean scores of the three groups on the pre-tests for both the interpretation and production tasks were similar and ranged from 0% to 23% in both universities. On the interpretation tasks the mean scores (see Table 3.2) of the Processing Instruction groups (49% to 50%) were higher than the traditional group (27% to 34%) and showed an improvement from pre to posttest (48% to 50% for the Processing Instruction groups and 26% to 32% for the traditional group). No improvement for the control group. On the production tasks the instructional groups made gains and improved. The Processing Instruction group performed better that the traditional group (PI groups = 82% to 83% and TI groups = 66% to 68%). The instructional group improved from pre to posttest (PI groups = ranged from 60% to 79%; TI groups = ranged from 49% to 64%). The control group made minimal improvement (4% to 7%). Participants in the Processing Instruction group outperformed the other two groups in the interpretation task and performed as well as the traditional group in the production task. Once again the results confirmed the findings of the studies reviewed in this paragraph. The results of the studies comparing Processing Instruction and traditional instruction and reported in this chapter (see Table 3.1 and Table 3.2) have reached the following conclusion:

Processing ­Instruction and meaning output-based instruction 

81

a. Processing Instruction is a more effective approach to grammar instruction than traditional instruction as it seems to have a direct effect on learners’ ability to process input (various processing problems, various linguistic forms, different languages and populations). L2 learners have gained the ability to interpret sentences efficiently and correctly. Processing Instruction is responsible for learners increase rate of processing; b. Processing Instruction provides L2 learners with the ability to produce the target linguistic features during output practice. Processing Instruction is responsible for the increase rate of accuracy in production. The Processing Instruction groups performed as well as the traditional groups on the various production tasks used in the experimental studies previously reviewed and this is a remarkable finding given that participants in the Processing Instruction group were never asked to produced the target features through output practice. As underscored by Collentine (2004) there is enough evidence to support the view that Processing Instruction is a better and more efficient approach to grammar teaching than traditional instruction. Although there is a large database indicating the superiority of Processing Instruction, one of the question which have been raised is whether the same results can be found when Processing Instruction is compared to a more meaning output based approach to grammar teaching. What are the main findings of studies comparing Processing ­ Instruction and meaning output-based instruction? In Table 3.3 we present the results of studies which have compared Processing Instruction with meaning output-based instruction. These studies have measured the effects of Processing Instruction with a different type of outputbased instruction; namely meaning output-based instruction. In the studies reviewed in the previous chapter Processing Instruction was compared to traditional instruction as traditional instruction is still the dominant approach to grammar teaching in foreign language classrooms in both the United States and Europe. The materials developed for these studies was very traditional and involved mechanical output practice. The question was whether Processing Instruction should be compared to a more meaningful oriented type of output practice. Processing Instruction was therefore compared to a different output based approach called ‘meaning output instruction’. This approach

82 

Processing Instruction and other types of instruction

i­nvolves the use of structured output activities which are meaningful activities in nature. These activities carry a meaningful context and the target forms are produced not with the sole intention of practicing the target item, but rather to communicate opinions, beliefs, or other information related to designated topic. The results in all the studies comparing Processing Instruction and meaning output instruction, reported in this chapter are displayed in Tables 3.3 and 3.4. Farley (2001a, 2001b, 2004) was the first who compared the effects of Processing Instruction versus meaning output- based instruction on the acquisition of the Spanish subjunctive in two consecutive studies. Processing Instruction was compared to an output based approach and the activities learners were exposed to did not contain any mechanical drills. The activities for the meaning output-based approach were based on the tenets of structuredoutput activities proposed in Lee and VanPatten (1995, 2003). The effects of the two instructional treatments were measured on the acquisition of Spanish subjunctive. This feature is affected by the Sentence Location Principle. In Spanish the subjunctive is located in medial position where it is least likely to be processed. In the sentence Non pienso que entienda el problema (I do not think he/she understands the problem) the subjunctive inflection (the ‑a‑ of intienda) is in the middle of the sentence and the Sentence Location Principle predicts that learners will overlook the subjunctive inflection because it is not located in a more salient positions. In his first study (2001a) twenty-nine subjects enrolled in a fourth-semester Spanish course were assigned to two different instructional treatments: Processing Instruction and meaning-based output instruction. The instructional period lasted for two days. Pre and posttests consisted of an interpretation and production tasks. The mean scores on the pretests for the interpretation and production tasks ranged from 21% to 31%. There was not statistical difference between the two groups prior the beginning of the instructional period on interpretation and production of subjunctive forms in Spanish. Two posttests were administered and the mean scores of the Processing Instruction group on the interpretation tasks were higher than the meaning-base output instruction (see Table 3.4). The results were even more evident in the second posttest (Processing Instruction group = 66%; meaning-based output instruction group = 39%) indicating an improvement for the Processing Instruction group of 35%. The mean scores of both the Processing Instruction and meaning-based output instruction group were similar and ranged from 81% to 90% in the two production posttests.(see Table 3.4) indicating an improvement of from 59% to 67%.

Processing ­Instruction and meaning output-based instruction  

83

The results showed that Processing Instruction had a greater effect than meaning-based output instruction on how L2 learners interpret subjunctive forms in Spanish. Both instructional treatments had a similar effects on how L2 learners produce sentences containing those forms. One hundred and twenty nine undergraduate students participated to Farley second experiment (2001b, 2004). They were assigned to two instructional groups (Processing Instruction vs. meaning output-based instruction). Both groups were assessed following a pre-test/post-test design which included interpretation and production tasks. Farley’s results differed from his previous study. The mean scores on the pretests ranged from 32% to 35% for the interpretation task and less than 1% for the production task. The mean scores of both instructional groups in the two posttests on the interpretation task ranged from 65% to 70% indicating similar improvements between pre and posttests (31% to 36%). The mean scores of both the Processing Instruction and meaning-based output instruction group were similar and ranged from 38% to 42% in the production posttests.(see Table 3.4) indicating an improvement of from 35% to 41%. Both groups made equal and significant improvements on both the interpretation and the production tests. Farley attributed the equal performance of the two treatments to one main factor. The meaning output-based treatment is different from traditional instruction practice as it does not contain mechanical drills practice and its communicative and interactive nature might have resulted in incidental input. Benati (2005) investigated the effects of Processing Instruction, traditional instruction and meaning output-based instruction on the acquisition of English past simple tense. The relevant processing principle in this case was the Lexical Preference Principle. The subjects involved in the present studies were Chinese (47 subjects) and Greek school-age learners (30 subjects) of English residing in their respective countries. The participants in both schools were divided into three groups. The first group received Processing Instruction; the second group was exposed to traditional instruction; the third group received meaning output-based instruction. One interpretation and one production measure were used in a pre and post-test design. The results are very consistent in both studies (study 1 = Chinese participants; study 2 = Greek participants). In the first study the mean scores on the pretests for the interpretation task ranged from 16% to 18%; and for second study from 43% to 45%. The posttest on the interpretation task in both studies showed that the Processing Instruction group obtained much higher scores than the traditional and meaning-based output instruction group (see Table 3.4). Particularly in the case of the first study the scores of the

84 

Processing Instruction and other types of instruction

­ rocessing Instruction group were 76% showing an improvement of 60%. P The mean scores of the traditional instruction group was 19% indicating only 1% improvement. The mean scores of the meaning-based output instruction group showed an improvement of about 9% from pre to post-test (see Table 3.4). In the production task of both studies the mean scores of the groups were very similar ranging from 30% to 56% (see Table 3.4) indicating a consistent improvement of from 25% to 29%. The results showed that Processing Instruction had positive effects on the processing and acquisition of the target feature. In both studies the Processing Instruction group performed better than the traditional and meaning outputbased groups in the interpretation task and the three groups made equal gains in the production task. The Processing Instruction treatme4nt was superior than the other two output treatments in terms of helping learners to interpret utterances containing the English past. The findings on the sentence-level task involving the interpretation of the English past simple tense support previous findings on Processing Instruction research which indicated that Processing Instruction is successful in altering learners processing default strategies. The results of the statistical analysis indicated that Processing Instruction, traditional instruction and meaning output-based instruction made an equal improvement (from pre-test to post-test) on the production task (sentence-level task). Even in this case, the findings from the present experiment support the main results of previous research on Processing Instruction which showed that the Processing Instruction group made significant similar gains from the pre-test to the post- test compared with the two output groups in production tasks. The evidence obtained in this study on the production task suggests that the effects of Processing Instruction not only have an impact on the way learners interpret sentences but also on the way learners produce sentences. Processing Instruction has clearly altered the way learners processed input and this had an effect on their developing system and subsequently on what the subjects could access for production. Gely (2005) investigated the effects of Processing Instruction and meaning output-based instruction on the acquisition of French imperfect tense (Lexical Preference Principle).Thirty-three college students learning French (second year) participated in this experimental study. One group received Processing Instruction, the second group meaning output-based instruction and the third group acted as a control group. Gely (2005) used three different tasks to measure the effects of instruction: an interpretation task and two production tasks (completion text and a written production task). The results of this study were very similar to Farley (2001a) and Benati (2005).

Processing ­Instruction and meaning output-based instruction  

85

The mean scores on the pretests for the interpretation task ranged from 18% to 21%; and for the production task from 24% to 27%. The posttest on the interpretation task indicated that the Processing Instruction group obtained much higher scores than the meaning-based output instruction group and the control group (see Table 3.4). This group made steady improvement (56%) compared to meaning-based output instruction (2%) and the control group which did not improve. In the production task the mean scores of both groups were similar ranging from 85% to 86% (see Table 3.4) indicating a consistent improvement of from 67% to 70%. These results were confirmed in the delayed posttests given three weeks later (see Table 3.4). The statistical analyses confirmed that the Processing Instruction group outperformed the other two groups on the interpretation task and that the Processing Instruction group and the meaning output-based group were equal on the production tasks. Morgan-Short and Bowden (2006) carried out a study to investigate the effects of Processing Instruction and meaningful output-based instruction. Forty-five first semester Spanish students were assigned to three groups: Processing Instruction, meaningful output-based instruction and a control group. One interpretation and one production task were developed and used in this study. The scores on the two tasks were converted in percentages (see Table 3.4). The mean scores of the two instructional groups were significantly higher than the control group. There was a significant loss for the meaning output-based group from posttest1 to posttest2 (from 63.1% improvement to 34.7%). However, the two groups made equal gains from pre to posttest 1 with mean scores ranging from 59% to 63%. In the production task the mean scores of the instructional group were higher than the one of the control group (see Table 3.4). In the first posttest the meaning output-based instruction outperformed the Processing Instruction group and the control group (MOI= 69%; PI = 30%; C=13%). However, this difference between groups was not maintained in the second posttest (MOI= 31%; PI = 26%; C=15%). As pointed out by Morgan-Short and Bowden (2006: 56) ‘‘the analysis of our results is straightforward: in PI, input based practice leads to form–meaning connections evidence in both interpretation and production test gains, whereas in the MOBI, learners might have been able to establish or strengthen form–meaning connections via both output-based practice and the input they receive in instruction and feedback’’. Lee and Benati (2007a) also provided further support to the effectiveness of Processing Instruction when compared to meaning output-based approach. In a parallel study they contrasted the effects of these two instructional treatments on the acquisition of Italian subjunctive of doubt and opinion and

86 

Processing Instruction and other types of instruction

French subjunctive of doubt (Sentence Location Principle and as in the case of Farley’s studies the subjunctive was chosen). This study has been reviewed in chapter four of this book (see Table 4.2). The overall findings from the present study support the results obtained by the majority of studies investigating the effects of Processing Instruction, which show that Processing Instruction is superior to meaning output-based instruction. The results of Lee and Benati study differ from Farley’s research (Farley 2001b, 2004) as it indicates that Processing Instruction is better than meaning output-based instruction in helping learners to process the subjunctive of doubt in French and subjunctive of doubt and opinion in Italian no matter the form of delivery. The results obtained in the two parallel studies confirm the effectiveness of Processing Instruction in improving learners’ performance in both interpretation and production tasks. The results obtained in Lee and Benati’s (2007a) study might explain the difference in the results between s the majority of studies comparing Processing Instruction and meaning output-based instruction and Farley and Morgan-Short and Bowden ‘s studies . As stated by Lee and Benati (2007a: 122) ‘‘could be the fact that both treatments were compared through different modes of delivery. In the case of the meaning output-based instruction and Processing Instruction in the treatments delivered via computer terminals, learners did not receive input from an instructor or from interacting with other learners’’. In the case of the two parallel studies conducted by Lee and Benati (2007a) the group receiving the meaning output-based instruction treatment delivered via computer terminals, received no incidental structured input, unlike in the case of Farley’s studies (Farley 2001a, 2001b) and Morgan-Short and Bowden’s study (2006). This would explain the difference their results and the results obtained in these parallel studies. What are the main findings of studies comparing Processing Instruction and other grammar-instruction approaches? Marsden (2005) carried out one experimental study comparing Processing Instruction to a different input-based approach to grammar teaching (enriched input). Twenty-seven school age learners of French were assigned randomly to two groups, one receiving Processing Instruction materials and the other which was exposed to enriched input material. Assessment tasks consisted

Processing Instruction and other grammar-instruction approaches 

87

of interpretation and production tasks in two modalities (oral and written) and including different performance modes (sentence level, narratives and guided conversations). The acquisition of French regular verb inflections in the perfect and present tense was the target feature under investigation (Lexical Preference Principle). Significant gains were made by the Processing Instruction group from pre-tests to posttests and the gains were maintained at delayed posttests. This was not the case for the enriched input group that made no gains (see Table 3.6 and 3.6). The mean scores on the two posttests for the Processing Instruction group showed an improvement of about 9% for the listening test, 11% for the reading test, from 7% to 11% for the writing task and 6% to 8% for the speaking task. Overall the findings from this study supports the view that Processing Instruction is a very effective approach to grammar teaching. The Processing Instruction group who received structured input activities (combination of referential and affective structured input activities) made gains in all the assessment measures and those gains were maintained in delayed posttests at 14–16 weeks after the posttests. Toth (2006) compared the effects of Processing Instruction with a different instructional approach called communicative output tasks. Participants were eighty English native speakers studying beginning Spanish at undergraduate level. The subjects were assigned to six different groups (two for each treatment): two for the Processing Instruction group; two for the communicative output tasks and two control groups. Instruction lasted for seven days and the target linguistic feature was the anticausative ‑se‑ in Spanish. This feature was chosen as learners fail to link an obligatory ‑se‑ to anticausative meaning and this is captured by the Primacy of meaning principle (VanPatten 2004). Data were collected (pre-tests-posttests and delayed posttests) on grammatical judgment and production tasks. This study is not a replication of previous Processing Instruction studies as no interpretation measure was used. The mean scores on the two posttests for the production task for the Processing Instruction group and the communicative output tasks group showed an improvement. Both groups were better than the control group (see Table 3.6). The statistical analyses showed that the communicative output tasks group was significantly better than the Processing Instruction group. In the grammar judgment tasks both the instructional groups made some gains from pre-tests and posttests. However, the statistical analyses showed that the communicative output tasks groups was statistically better that the other two groups. This study provides further evidence that Processing Instruction has a direct impact on learners’ developing system as the Processing Instruction

88 

Processing Instruction and other types of instruction

group in this study made gains form pre-test to post test in the production task. This study provides further evidence to Lee and VanPatten’s (2003) argument that the best conditions for L2 learners to acquire a target language is a combinations of structured input activities and meaningful output tasks. Qin (2008) carried out a study to compare Processing Instruction to dictogloss tasks which is an output-based type of instruction on the acquisition of English passive choice. The grammatical feature was chosen because it is affected by the First Noun Principle. One hundred and ten Chinese students were the participants of this experiment. They were intermediate school age learners. One comprehension and one production tasks were developed and a pre-test and posttests procedure was adopted. The mean scores on the pretests for the interpretation task for both groups ranged from 86% to 90%. The means of the two groups were almost equivalent at pre-test level. The posttest on the interpretation task showed that the Processing Instruction group obtained much higher scores than the dictogloss tasks group (see Table 3.6). The mean scores of the Processing Instruction group improved of 86% from pre-test to posttest, whereas the improvement of the other group was of 55%. The statistical analyses conducted on the raw scores of the interpretation task for both groups indicated that the Processing Instruction group was statistically better than the dictogloss group. The mean scores on the pretests for the production task for both groups ranged from 28% to 30%. In the posttests for the production task both groups made very similar gains, ranging from 8% to 11% (see Table 3.6) indicating a consistent improvement of from 71% to 84%. The statistical analyses conducted on the raw scores of the production task for both groups indicated that the Processing Instruction group was statistically similar to the dictogloss group. Overall the results concurred with the finding of previous studies comparing Processing Instruction and other types of output-based instruction. The Processing Instruction group performed better than the dictogloss group in the interpretation task and the two groups made equal gains in the production task. Conclusion The role of grammar in language teaching has dramatically changed in the last few decades. As previously argued, this is due to a number of theoretical factors and various empirical findings in instructed second language acquisition research. We know that learners seem to acquire grammatical features in

Conclusion 

89

a predictable order regardless of their L1 and formal instruction (including correction). There are predictable sequences in second language acquisition (Pieneman 1998) and certain features in L2 have to be acquired before others (stages of development). Many scholars would disagree with Anderson (1983) that all or most of second language grammar is initially learnt through the conscious study and application of explicit rules. There is a consensus that much grammar is learnt without conscious awareness. Learners have internal strategies to process and organise language data. Pienemann (1998) claims that we need to understand how learners process L2 in order to understand second language acquisition. It is fairly clear from current perspectives on instructed second language acquisition that the effects of instruction are limited (see Doughty and Williams 1998). In addition, acquisition orders are immutable (R. Ellis 1994; Lightbown and Spada 1999). Explicit grammar instruction does not affect ‘natural’ stages of development. VanPatten and Oikkenon (1996) and Benati (2004a, 2004b) have also called into question the role and the necessity of explicit information. Krashen (1982) has hypothesised that if learners are exposed to comprehensible input and provided with opportunities to focus on meaning and messages rather than grammatical forms and accuracy, they are able to acquire the L2 in the same way as L1. Processing Instruction is a very effective approach to grammar instruction in that it has never failed to yield significant improvement in learner performance on either interpretation or form production tasks. When presented with a processing problem in the L2, learners can be taught to alter their processing strategies thereby delivering better intake (i.e., grammatically richer) to their developing systems. This better intake yields improved performance on both interpretation and production tasks. The primary effects of Processing Instruction, investigated in a number of studies (see Tables 3.1 and 3.2), have shown that Processing Instruction is a better approach to grammar instruction than traditional instruction at improving learners’ rate of processing. Processing Instruction is also successful at increasing learners’ accuracy in production. Processing Instruction has also been compared to a different output based approach called meaning output-based instruction which consists of structured output activities which are meaningful activities in nature. They all carry a meaningful context and the target forms are produced not with the sole intention of practicing the target item, but rather to communicate opinions, beliefs, or other information related to designated topic. Overall, the results in the studies comparing Processing Instruction vs. meaning output-based instruction and reported in this chapter (see Table 3.3 and Table 3.4) showed that Processing Instruction (delivered in two different modes, ­classroom and com-

90 

Processing Instruction and other types of instruction

puter) was superior to meaning output-based instruction. The parallel studies carried out by Lee and Benati (2007a) clearly showed that the meaning output-based instruction treatment delivered through two different modes and containing meaning based activities was not successful at producing positive effects (altering the processing problem) on students’ performance. The other studies comparing Processing Instruction and other types of grammar instruction (reviewed in Tables 3.5 and 3.6) have confirmed previous findings on the role and effect of Processing Instruction and meaning output-based instruction. Based on the findings measuring primary effects for Processing Instruction we conclude the following: –– Processing Instruction is an effective approach to alter a variety of L2 learners default strategies in different languages and with native speakers of a variety of L1s; –– Processing Instruction is more effective than traditional instruction. Primary effects studies have clearly confirmed the superiority of Processing Instruction; –– Processing Instruction is overall more effective than meaning output-based instruction.

Spanish preterite tense

Italian Lexical preference future tense

Spanish copula

French causative

Cadierno 1995

Benati 2001

Cheng 2004

VanPatten & Wong 2004

Number

39

Intermediate 76 English native

Intermediate 83 English native

Beginners English native

Intermediate 60 English native

Intermediate 80 English native

Subjects/L1

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

Design

Int. = PI > (TI = C) Prod. = (PI = TI) > C

Int. = PI > TI > C Prod.= (PI = TI) > C

Interpretation (aural) Production (written)

Int. = PI > TI > C Prod.= (PI = TI) > C

Int. = PI > (TI = C) Prod.= (PI = TI) > C

Int. = PI > (TI = C) Prod. = (PI = TI) >C

Resutlts

Interpretation (aural) Production (written)

Interpretation (aural) Production (written and Oral)

Interpretation (aural) Production (written)

Interpretation (aural) Production (written)

Tests

PI = Processing Instruction; TI = Traditional Instruction; C = control group; Int. = Interpretation test; Prod. = Production test

First noun

Preference for nonredundancy

Lexical preference

First noun

Spanish object pronouns

VanPatten & Cadierno 1993

Processing principle

Linguistic feature/ language

Study

Table 3.1.  Primary effects of processing Instruction vs. traditional instruction

Appendix Appendix  91

Benati 2001

PI TI C

PI TI C

PI TI C

Production (written)

Production (oral)

PI TI C

Production

Interpretation

PI TI C

Interpretation

PI TI C

Production

Cadierno 1995

PI TI C

Interpretation

VanPatten and Cadierno 1993

  2.6   2.9   2.0

  2.5   2.3   2.4

  3.8   4.1   3.8

  2.3   1.7   2.3

  4.2   3.8   4.3

  2.1   2.6   1.5

  1.7   1.3   1.1

Treat­ Pre­ ment test

Task

Study

Table 3.2. 

  7.1   7.6   2.7

  7.3   7.7   3.2

  8.3   5.8   4.0

  7.8   6.0   3.6

  7.4   5.0   4.9

  8.8   8.5   3.4

  8.0   3.0   1.4

Post­ test 1

  6.5   7.2   2.3

  6.7   7.5   2.7

  8.1   5.6   3.4

  7.7   5.9   3.2

  7.8   5.2   4.2

  8.1   9.0   3.4

  7.1   3.4   2.2

Post­ test 2

  7.5   6.0   3.2

  7.8   5.1   4.5

  8.1   8.1   4.5

  7.4   3.8   2.2

Post­ test 3

42% 52%   3% 39% 43%   3%

45% 47%   7%

43% 15% −4%

54% 42%   9%

36% 14% −1%

60% 64% 19%

54% 21% 11%

48% 54%   8%

55% 17%   2%

55% 43% 13%

32% 12%   6%

67% 59% 19%

63% 17%   3%

53% 43%   9%

36% 13%   2%

60% 55% 30%

57% 25% 11%

Prod.= (PI = TI) > C

Prod.= (PI = TI) > C

Int. = PI > TI >C

Prod. = (PI = TI) > C

Int. = PI > (TI = C)

Prod. = (PI = TI) > C

Int. = PI > (TI = C)

Improvement Improvement Improvement Results from pretest to from pretest to from pretest to posttest 1 posttest 2 posttest 3

92  Processing Instruction and other types of instruction

Interpretation

VanPatten and Wong 2004

  8.2   6.6   2.7   8.3   6.8   1.9

UNI 1 PI   2.0 TI   1.7 C   2.3

UNI 2 PI   0.4 TI   0.4 C   1.2

  5.0   3.4   0.7

UNI 2 PI   0.0 TI   0.2 C   0.7

79% 64%   7%

60.2% 49.9%   4%

50% 32%   0%

48% 26% −3%

Improvement

Pretest Posttest   4.9   2.7   0.7

25.8% 26.8%   6.5%

50% 52.4% 24.2% 53.3% 46.9% 26.5% 28.8% 32.5% 22.3%

UNI 1 PI   0.1 TI   0.9 C   1.0

PI TI C 28.2% 20.4% 10.2%

Prod.= (PI = TI) > C

Int. = PI > TI >C

Results

(PI = TI) > C

PI = Processing Instruction; TI = Traditional Instruction; C = control group; Int. = Interpretation test; Prod. = Production test

Production

Production Guided composition (use of estar)

Cheng 2004

Appendix  93

Linguistic feature/ language

Spanish subjunctive

Spanish subjunctive

English simple past

French imperfect

Spanish object pronouns

Study

Farley 2001a

Farley 2001b 2004

Benati 2005

Gely 2005

Morgan-Short & Bowden 2006

First noun

Lexical preference

Lexical preference

Lexical preference Sentence location

Lexical preference Sentence location

Processing principle

Intermediate English native

Beginners English native

  45

  33

  47

  30

129

Intermediate English native

Beginners Greek and Chinese native

  29

Number

Intermediate English native

Subjects/l1

Interpretation (aural) Production (written)

Immediate & delayed

Interpretation (aural) Production (written)

Prod. = Posttest 1 (PI > (PI = C) Pottest 2 (PI = MOI) > C

Int. = (PI = MOI) > C

Int. = PI > TI = MOI Prod. = PI = TI = MOI

Int. = PI > TI = MOI Prod. = PI = TI = MOI

Int. = PI = MOI Prod.= PI = MOI

Interpretation (aural) Production (written) Interpretation (aural) Production (written)

Int. = PI > MOI Prod. = PI = MOI

Resutlts

Interpretation (aural) Production (written)

Tests

Pre-test Post-tests

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

Design

Table 3.3.  Primary effects of Processing Instruction vs. meaning output-bas instruction

94  Processing Instruction and other types of instruction

French subjunctive

Lee and Benati (2007a)

Sentence location

Sentence location

Beginners English native

Intermediate English native

  61

  47

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate

PI > MOI classroom PI > MOI computer PI = MOI classroom PI = MOI computer

Production (written)

PI = MOI classroom PI = MOI computer

Production (written) Interpretation (aural)

PI > MOI classroom PI > MOI computer

Interpretation (aural)

PI = Processing Instruction; TI = Traditional Instruction; MOI: Meaning-based Output Instruction; C = control group; Int. = Interpretation test; Prod. = Production test

Italian subjunctive

Lee and Benati (2007a)

Appendix  95

Benati 2005

  4.5   4.3   4.3

Study 2 PI TI MOI   7.5   4.3   4.8

  7.6   1.9   2.5

Posttest

Pretest

  1.6   1.8   1.6

  4.2   4.1

  6.5   7.0

  8.8   8.5

  6.8   5.5

Posttest 1

  0.1   0.3

  3.2   3.5

  2.1   2.6

  3.1   2.8

Pretest

PI TI MOI

Study 1

PI MOI

Production

Interpretation

PI MOI

Interpretation

PI MOI

Production

Farley 2001b, 2004

PI MOI

Interpretation

Farley 2001a

Treatment

Task

Study

Table 3.4. 

  3.6   3.8

  6.8   6.6

  8.1   9.0

  6.6   3.9

Posttest 2

30%   0%   5%

60%   1%   9%

35% 35%

36% 31%

60% 64%

35%   1.1%

Int. = PI > TI = MOI

Results

Prod. = PI = MOI

Int. = PI = MOI

Prod. = PI = MOI

Int. = PI > MOI

Improvement Results from pretest to posttest 2

Improvement from pretest to posttest

41% 38%

33% 35%

67% 59%

37% 27%

Improvement from pretest to posttest 1

96  Processing Instruction and other types of instruction

PI MOI C

PI MOI C

Production

PI MOI C

Production

Interpretation

PI MOI C

  1.9%   0.5%   2.4%

  7.6%   8.1%   9.9%

  1.8   1.6   2.1

  2.7   2.5   2.4

  2.4   2.7   2.6

Study 2 PI TI MOI

Interpretation

  0.8   0.9   0.9

PI TI MOI

Study 1

30.8% 69.9% 13.1%

67.3% 71.2% 13.6 %

  8.5   8.6   2.1

  8.5   3.2   2.4

  5.1   5.6   5.5

  3.0   3.0   3.4

26.3% 31.5% 15.8%

62.2% 46.4% 35.7%

  7.9   7.6   1.9

  8.3   2.7   2.2

28.9% 64.9% 10.7%

59.7% 63.1%   3.7%

67% 70%   0%

58%   7%   0%

27% 29% 29%

28% 29% 25%

24.4% 26% 13.4%

54.6% 34.7% 25.8%

61% 60% −2%

56%   2% −2%

Prod. = Posttest 1 (PI > (PI = C) Pottest 2 (PI = MOI) > C

Int. = (PI = MOI) > C

Prod. = PI = MOI

Int. = PI > MOI

Prod. = PI = TI = MOI

PI = Processing Instruction; TI = Traditional Instruction; MOI: Meaning-based Output Instruction; C = control group; Int. = Interpretation test; Prod. = Production test

Morgan-Short and Bowden 2006

Gely 2005

Production (written)

Appendix  97

110

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate & delayed

COT > (PI = C) PI > C COT > C COT > PI

GJ task Production

Interpretation Int. = PI >DT (aural) Production (written) Prod.= PI = DT

PI > EnI PI > EnI PI > EnI PI > EnI

Resutlts

Listening Speaking Writing Reading

Tests

EnI = Enriched input; DT = dictogloss tasks; COT = communicative output tasks; Int. = Interpretation test; Prod. = production test

First Noun Principle

School age Chinese students intermediate

English passive voice

Qi 2008

  80

Adults English native speakers

Primacy of Meaning Principle

Spanish ‑seanticausative

Pre-test Post-tests Immediate & delayed

Number Design

Toth 2006

Subjects/l1 English school age   27 native

Linguistic feature/ Processing principle language

Marsden 2006 French Lexical Preference regular verb inflections in the perfect and present tense

Study

Table 3.5.  Primary effects of Processing Instruction vs. other types of instruction

98  Processing Instruction and other types of instruction

  0.03   0.04   0.01

  0.30   0.86   0.71

  2.8   3.0

  0.47   0.63   0.00

  1.58   2.17   0.61

10.4 10.9

17.2 14.5

22.78 14.12

13.47 12.93

36.75 28.53

49.11 43.59

  0.24   0.50   0.03

  1.35   1.93   0.80

  8.9 11.4

12.7 12.5

20.54 13.58

17.22   9.74

34.16 27.83

48.96 41.51

 –

  1.88%   2.69%  –

76% 79%

86% 55%

  8%   0%

  7%   2%

11%   2%

  9%   2%

Posttest 1 Posttest 2 Improvement from pretest to posttest 1

 –

  1.65%   2.45%  –

71% 84%

41% 35%

  6% −1%

11% −1%

  9%   1%

  8%   0%

Improvement from pretest to posttest 2

PI > C COT > C COT > PI

COT > (PI = C)

Prod.= PI = DT

Int. = PI >DT

PI > EnI

PI > EnI

PI > EnI

PI > EnI

Results

PI = Processing Instruction; C= control; EnI = Enriched input; DT = dictogloss tasks; COT = communicative output tasks; Interpretation test; Prod. = Production test; GT = grammatical judgement task

Toth 2006

PI COT C

PI DT

Production

Production

PI DT

Interpretation

PI COT C

14.95 14.78

PI EnI

Speaking

GJ task

  6.80 10.14

PI EnI

Writing

  8.6   9.0

25.32 26.85

PI EnI

Reading

Qin 2008

40.33 41.51

PI EnI

Listening

Pretest

Marsden 2006

Treatment

Task

Study

Table 3.6. 

Appendix  99

Online delivery of Processing Instruction

Chapter 4 Can Processing Instruction be delivered effectively online as well as in classrooms? Prior to Lee and Benati (2007a) the two foci of Processing Instruction research were whether Processing Instruction was effective grammar instruction vis a vis other types of instruction and what element of Processing Instruction, be it Explicit Information or Structured Input practices, was the causative variable in bringing out the positive effects found for Processing Instruction (VanPatten 2004). In a series of empirical investigations, Lee and Benati and their colleagues took up a question VanPatten posed in 1996. He queried the following, “Because Processing Instruction is input-based, can computers deliver effective Processing Instruction?” (VanPatten 1996: 158). Pedagogical discussions of whether or not to accept technology as part of language teaching have long passed. More relevant to contemporary discussions of effective language teaching is how to best integrate technology so that language teaching and learning are enhanced (Liu, Moore, Graham and Lee 2003: 262). Processing Instruction is a type of grammar instruction in which learners are directed to focus on a form and connect the form to its meaning. Lee and Benati and their colleagues designed in-class and online versions of Processing Instruction materials for various linguistic items in Spanish, Italian, and French. Lee and Benati with Aguilar-Sánchez and McNulty (2007a) compared the same Processing Instruction materials for Spanish preterite–imperfect aspectual distinction and negative informal commands delivered in three modes: in-class, online, and a hybridised in-class/online mode. Additionally, lee and Benati (2007a) created in-class and online materials for meaning-based output instruction so as to compare it to Processing Instruction for Italian and French subjunctive of doubt and opinion. The results of these studies are presented below. Is online delivery of Processing Instruction an effective means of ­delivering Processing Instruction? In the past tenses, Spanish distinguishes between perfective and imperfective aspect through the use of preterite and imperfect verbal morphology, respect-

Efficacy of online delivery Efficacy of online delivery 

101

ively. (English also makes an aspectual distinction.) This verbal morphology can also co-occur with lexical temporal/aspectual indicators. Perfective aspect, hence the preterite tense, can be signalled by such phrases as ‘last summer’ and ‘last night.’ Imperfective aspect, hence the imperfect tense, can be signalled by phrases such as ‘every summer’ and ‘in the evenings.’ Note the following examples. Last summer we visited Melbourne. El verano pasado, visitamos Melbourne. Every summer we would visit Melbourne. Todos los veranos, vistábamos Melbourne. Last night we watched a movie. Anoche, miramos una película. In the evenings, we would watch a movie. En las noches, mirábamos una película. The verb morphology of Spanish is meaningful; it, in and of itself conveys aspect. But the presence of lexical temporal/aspectual indicators makes the verb morphology redundant. VanPatten (2004b) has captured the dual processing problems as follows. “P 1c. The Preference for Nonredundancy Principle: learners are more likely to process nonredundant meaningful grammatical form before they process redundant meaningful forms. P 1b. The Lexical Preference Principle: learners will tend to rely on lexical items as opposed to grammatical form to get meaning when both encode the same semantic information.” (VanPatten 2004b: 14) Previous Processing Instruction research has effectively directed learners to process verb morphology for temporal distinctions: the Spanish preterite (Cadierno 1995), Italian future (Benati 2001), Japanese past tense (Lee and Benati 2007a), and English simple past (Benati 2005). Lee and Benati with Aguilar-Sánchez and McNulty (2007a) examined, not temporal distinctions, but aspectual ones by comparing the effectiveness of the same Processing Instruction materials for Spanish preterite/imperfect aspectual distinction. They did not alter the materials but rather altered the mode of delivering them. The basis for the materials was the textbook, Vistazos: Un curso breve (VanPatten, Lee and Ballman 2006). These textbook/classroom materials were very carefully transposed for computer-delivery. Additionally, a hydridised mode of delivering the materials was created. The computer screens were

102 

Online delivery of Processing Instruction

d­ ownloaded and put into packets that individual learners received in a classroom setting. The instructor used overhead transparencies of the computer screens to deliver the explicit information, but then learners carried out the activities ­individually. Lee and Benati with Aguilar-Sánchez and McNulty (2007a) examined 25 learners of Spanish enrolled in a first-year, university-level Spanish language course. All subjects were native speakers of English who spoke English at home. They created a form recognition/interpretation task to measure changes in learners’ knowledge. The task contained two parts. The first part presented 5 sentences in which two verbs were deleted. Subjects selected the form of the verb they thought correctly completed the sentence. The second part presented a brief paragraph in which learners selected either the preterite or imperfect form of the verb, according to context, for 11 items. Across the three modes, the pretest mean scores ranged from 46.2% to 46.4%. Although no learners who scored above 60% on the pretest were included in the study, 46% is a relatively high pretest mean score in Processing Instruction studies. The high score is probably due to the fact that the learners had previously studied the forms and uses of the preterit as well as the forms and uses of the imperfect. The instruction they were receiving as part of the study was their introduction to the aspectual distinctions between the preterite and imperfect. Learners performed both affective and referential activities during a single 50 minute class period at the end of which they performed a posttest. The mean scores on the posttest ranged from 51% to 60% indicating an improvement of between 10.2% and 14.7%. One week later they performed another posttest. Their scores on the delayed posttest ranged from 54% to 64% indicating an improvement between 13.8% and 17.9%. The statistical analyses revealed that there was no significant effect for Treatment but there was a significant effect for Time. These results indicated that the three treatments were equally effective in bringing about improved performance and that improvement endured after one week of instruction. The same 25 learners also took part in another study that focussed on Spanish negative informal commands. Negative informal commands in Spanish are formed with a preverbal negative particle (no) that co-occurs with particular verbal morphology. Learners’ processing difficulties with these verb forms are similar to those Wong (2004b) examined for French indefinite articles in negative sentences. The negative particle is an obligatory lexical item in these sentences. The verb morphology in Spanish or the indefinite article in French are triggered by the negator; they are, then, non-meaningful and redundant. VanPatten (2004b) captured learners’ processing difficulties with

Efficacy of online delivery 

103

non-meaningful and redundant forms in two of his processing principles. They are stated as follows. “P 1c. The Preference for Nonredundancy Principle: learners are more likely to process nonredundant meaningful grammatical form before they process redundant meaningful forms. P 1d. The Meaning-Before-Nonmeaning Principle: learners are more likely to process meaningful grammatical forms before nonmeaningful forms irrespective of redundancy.” (VanPatten 2004b: 11) Lee and Benati with Aguilar-Sánchez and McNulty (2007) once again compared learners’ performance across three modes of delivering Processing Instruction: in-class, online, and in a hydridised mode. The assessment task used for the pretest and posttest was a contextualised preference task. It contained 5 scenarios with multiple items for each (a total of 14 items) for which learners selected the appropriate response for recommending to a friend what they should not do. The options contained the correct verb form as well as present tense, preterite and imperfect forms. The mean scores on the pretest ranged from 5.7% to 6.3% indicating a low level of knowledge about negative informal commands. Learners performed both affective and referential activities during a single 50 minute class period, at the end of which they performed a posttest. The mean scores on the posttest ranged from 67.9% to 84.7%, indicating a very large improvement between 62.2% to 78.6%. One week later the performed another posttest. Their scores on the delayed posttest ranged from 65.3% to 77.9% indicating a large sustained improvement (from the pretest) of between 59.2% and 71.4%. The statistical analyses revealed no significant effect for Treatment but there was a significant effect for Time. As was the case with preterite/imperfect aspectual distinction, these results indicated that all three treatments were equally effective in bringing about improved performance and that the improvement endured for one week following the ­instruction. As mentioned, the same 25 learners participated in both studies. Accordingly, Lee and Benati with Aguilar-Sánchez and McNulty (2007) included Linguistic Item in their statistical analyses. They found a significant interaction between Linguistic Item and Time. The interaction was due to there being significant differences between the scores on the two linguistic items on the pretests and then again on the posttests. The direction of the differences changed as a result of instruction. Prior to instruction, the pretest mean

104 

Online delivery of Processing Instruction

scores were significantly higher for preterite/imperfect but after instruction, the posttest mean scores were significantly higher for negative informal commands. These difference held across all three modes of delivering Processing Instruction. Lee and Benati with Aguilar-Sánchez and McNulty comment that Processing Instruction is effective for both linguistic items but that the degree to which Processing Instruction is effective may be due to the linguistic item and learners’ prior experience with elements of that linguistic item such as its forms or its other functions. How does the online delivery of Processing Instruction compare to the online delivery of MOI? Subjunctive mood verbal morphology presents learners several processing challenges. This verb morphology occurs in nominal, adjectival and adverbial dependent clauses. In the vast majority of its uses, the morphology is triggered by the semantics of the independent clause. Thus, the morphology is non-meaningful; its use is a reflex to another sentence element. Insofar as its use co-occurs with another sentence element, the morphology can also be seen as being redundant. Finally, the form tends to occur near the end of a multiclause sentence, a position that is not the most favoured processing position. Note the following examples of the use of subjunctive mood morphology in Italian following expressions of doubt and opinion versus the use of indicative mood morphology following expressions of certainty. Dubito que Giorgio canti. ‘I doubt that Giorgio sings.’ Sono certo que Giorgio canta. ‘I am sure that Giorgio sings.’ The two verb forms are identical in meaning; the both mean ‘sings’ and refer to the third person singular. The -i versus -a ending does not alter the meaning of the verb. The processing difficulties presented by subjunctive mood morphology are captured by several of VanPatten’s Processing Principles. “P 1c. The Preference for Nonredundancy Principle: learners are more likely to process nonredundant meaningful grammatical form before they process redundant meaningful forms. (VanPatten 2004b: 11)

Processing Instruction vs. MOI Processing Instruction vs. MOI 

105

P 1d. The Meaning-Before-Nonmeaning Principle: learners are more likely to process meaningful grammatical forms before nonmeaningful forms irrespective of redundancy. (VanPatten 2004b: 11) P 1f. The Sentence Location Principle: learners tend to process items in sentence initial position before those in final position and those in medial position.” (VanPatten 2004b: 14) Previous Processing Instruction research on subjunctive mood has consistently demonstrated that Processing Instruction is effective in bringing about significant improvement in learners’ performance on both interpretation and production tasks. These results have been found for Spanish (Farley 2001b, 2004a, 2004b), but the research presented below extends the work to Italian and French. Lee and Benati (2007b) examined 24 learners of Italian enrolled in a second-year, university-level Italian language course. All subjects were native speakers of English who spoke English in the home. They compared the effects of Processing Instruction versus Processing Instruction using textual enhancements taught in a classroom setting versus Processing Instruction using textual enhancements delivered online to individuals as computer work stations. They focussed their work on the Italian subjunctive mood in subordinate clauses after main clauses that express doubt and opinion. They used both interpretation and production tasks as pretests and posttests. Across the three modes of delivering Processing Instruction, the pretest means scores for interpretation ranged from 1.0 to 1.7 whereas the production mean scores ranged from .33 to .83. With maximum score being 10, these means indicate a low level of knowledge regarding Italian subjunctive forms and use. Learners performed both affective and referential activities during two consecutive class periods for a total of four hours of instruction. The mean scores on the posttest for interpretation ranged from 6.7 to 7.0, indicating an average improvement of between 50% and 67%. The mean scores on the production posttest ranged from 4.2 to 4.6, indicating an average improvement of between 37.7% and 38.7%. The statistical analyses for both types of tests revealed no significant effect for Treatment but a significant effect for Time. These results indicated that the three treatments were equally effective in bringing about improved performance on both interpretation and production. Computers can effectively deliver Processing Instruction on the Italian subjunctive of doubt and opinion. Lee and Benati (2007a) examined the online delivery of French and Italian subjunctive of doubt and opinion. For these studies, they compared not only

106 

Online delivery of Processing Instruction

classroom versus computer-delivery, but also instruction type. Specifically, they compared Processing Instruction with MOI (meaningful output instruction) for both modes of delivery. A final pool of 47 subjects participated in the examination of the Italian subjunctive. The subjects were enrolled in a second semester, intermediatelevel university Italian language course. They were all native speakers of English who spoke English as a home language and who had no contact with the Italian language outside of class. Across the four treatment conditions, the learners’ mean pretest scores for interpretation of subjunctive forms ranged from 2.3 to 3.1 whereas their mean pretest scores for production ranged from 1.6 to 2.2. These mean scores indicated a low level of knowledge concerning Italian subjunctive forms and use. They received a total of 4 hours of instruction across two consecutive class periods. They then performed interpretation and production posttests. Their posttest scores for interpretation ranged from 3.6 to 8.1, indicating an average improvement between 5% to 57%. The improvement for the MOI conditions was 5% and 33% whereas it was 54% and 57% for Processing Instruction. Their posttest scores for production ranged from 7.5 to 7.9 indicating an average improvement between 54% and 63%. The statistical analyses of the interpretation scores revealed no significant difference for mode of delivering Processing Instruction. The classroom and computer delivery modes were equally effective in terms of delivering Processing Instruction. There was a significant difference for mode of delivering MOI. The classroom delivery of MOI yielded significantly higher results than its corresponding computer delivery. Interestingly, learners scores for either mode of delivering Processing Instruction were significantly higher than the MOI scores in either mode of delivery; Processing Instruction was a more effective intervention for teaching the Italian subjunctive of doubt and opinion as measured by the interpretation task. The statistical analyses of the production scores revealed no significant differences between the scores for the four treatments. Each was equally effective for helping learners to produce subjunctive forms in the context of expressions of doubt and opinion. Learners who receive Processing Instruction, during which they do not produce the target form, are equally able to produce the target form as those who receive MOI, during which learners repeatedly produce the target form. Lee and Benati (2007a) also examined the French subjunctive of doubt. These 61 learners were all native speakers of English, enrolled in a beginning-level, university French language class. The learners were assigned to the same four treatment conditions as Lee and Benati used for examining the Italian subjunctive. Across the four treatments, the learners’ mean scores

Summary 

107

on the interpretation pretest ranged from 1.1 to 1.3. Their mean scores on the production pretest ranged from .6 to 1.4. These mean scores indicate a low level of knowledge about the forms and use of the French subjunctive. Learners received instruction across two consecutive class periods for a total of four hours. After receiving instruction in Processing Instruction or MOI, in class or online, learners then performed an interpretation and a production posttest. The mean scores on the interpretation posttest ranged from 1.7 to 7.8. The improvement for the Processing Instruction groups was higher (59% and 67%) than that for the MOI groups (6% and 18%). The mean scores on the production posttest ranged from 5.9 to 6.7, indicating improvement from 48% to 61%. The results of the statistical analyses on the French subjunctive mirror those Lee and Benati found for the Italian subjunctive with one exception. The statistical analyses of the interpretation scores revealed no significant difference for mode of delivering Processing Instruction or MOI. The classroom and computer delivery modes were equally effective in terms of delivering the two types of instruction. As was the case with the Italian subjunctive, learners’ scores for either mode of delivering Processing Instruction were significantly higher than the MOI scores in either mode of delivery; Processing Instruction was a more effective intervention for teaching the French subjunctive of doubt and opinion as measured by the interpretation task. The statistical analyses of the production scores revealed no significant differences between the scores for the four treatments. Each was equally effective for helping learners to produce subjunctive forms in the context of expressions of doubt and opinion. Once again, learners who receive Processing Instruction, during which they do not produce the target form, are equally able to produce the target form as those who receive MOI, during which learners repeatedly produce the target form. Summary In this chapter, we have reviewed the results of a the studies that examined possible differential effects delivering instruction either in a classroom setting to a group of students or on a computer to individuals. These studies are summarized in Table 4.1. As can be seen in this table, the research has examined Spanish, Italian, and French, three Romance languages. The Processing Principles, and hence the processing challenges to learners, are varied, although all five studies focus on verb morphology. Whereas the number of

108 

Online delivery of Processing Instruction

studies is somewhat limited, the findings for Processing Instruction are perfectly consistent. We can answer VanPatten’s 1996 question, Can computers deliver effective Processing Instruction?, in the affirmative. The five comparisons of classroom versus computer delivery of Processing Instruction show that both are equally effective modes of delivery. The rationale for exploring possible effects for classroom versus computer delivery of Processing Instruction was to assess the effectiveness of computer delivery with the idea that perhaps the computer might be a more effective mode of delivering Processing Instruction. In Table 4.2 we present a summary of the degree of improvement, expressed as a percentage, documented between pretest and posttest scores for the various treatments examined. This summary clearly shows that Processing Instruction can be delivered quite ­effectively by a computer to an individual learner. It also shows that the computer is not superior to an instructor when it comes to Processing Instruction.

Linguistic feature/ language

Spanish preterite/ imperfect aspectual distinction

Spanish negative informal commands

Italian subjunctive

Study

Lee & Benati with AguilarSánchez & McNulty 2007

Lee & Benati with AguilarSánchez & McNulty 2007

Lee & Benati 2007a

Beginners L1 English

Subjects/L1

Nonredundancy; Intermediate Nonmeaningful; L1 English Sentence location

Non-meaningful; Beginners Redundancy L1 English

Lexical Preference; Nonredundancy

Processing principle(s)

47

25

25

Number

Pretest Posttest Immediate

Interpretation (aural) Production (written)

Int.: a. PI-Cl = PI-Co b. PI-CL > MOI-Cl c. PI-Cl > MOI-Co d. PI-Co > MOI-CL e. PI-Co > MOI- Co f. MOI-Cl > MOI-Co Prod.: PI-Cl = PI-Co = MOI-Cl = MOI-Co

PI-Cl = PI-Co = PI-Hy

Contextualized Pretest preference Posttest Immediate & delayed

Results

PI-Cl = PI-Co = PI-Hy

Tests

Pretest Form recognition Posttest Immediate & delayed

Design

Table 4.1.  Summary of the studies investigating modes of delivering Processing Instruction

Appendix Appendix  109

French subjunctive

Italian subjunctive

Lee & Benati 2007a

Lee & Benati 2007b

Subjects/L1

61

Number

Nonredundancy; Intermediate L1 24 Nonmeaningful; English Sentence location

Nonredundancy; Beginners Nonmeaningful; L1 English Sentence location

Processing principle(s)

Pretest Posttest Immediate

Pretest Posttest Immediate

Design

Interpretation (aural) Production (written)

Interpretation (aural) Production (written)

Tests

PI-Cl = PI-E-Cl = PI-E-Co

Prod.: PI-Cl = PI-Co = MOI-Cl = MOI-Co

Int.: a. PI-Cl = PI-Co b. PI-CL > MOI-Cl c. PI-Cl > MOI-Co d. PI-Co > MOI-CL e. PI-Co > MOI- Co f. MOI-Cl > MOI-Co

Results

Int. = interpretation task results; Prod. = production task results. PI-Cl = Processing Instruction delivered in a classroom; PI-Co = Processing Instruction delivered online (computer); MOI-Cl = Meaningful Ouput Instruction delivered in a classroom; MOI-Co = Meaningful Ouput Instruction delivered online; PI-hyb = Processing Instruction delivered as a hybrid of classroom and computer formats and materials; PI-E Processing Instruction with textual and aural enhancements

Linguistic feature/ language

Study

Table 4.1.  (cont.)

110  Online delivery of Processing Instruction

PI-Cl vs. PI-Co vs. PI-Hyb PI-Cl = PI-Co = PI-Hyb

PI-Cl vs. PI-Co vs. PI-Hyb PI-Cl = PI-Co = PI-Hyb

Spanish Preterite/ imperfect aspectual distinction

Spanish Negative informal commands

Italian Subjunctive

Lee & Benati with AguilarSánchez & McNulty 2007

Lee & Benati with AguilarSánchez & McNulty 2007

Lee & Benati 2007a

PI-Cl vs. PI-Co vs. MOI-Cl vs. MOI-Co Int. PI-Cl = PIComp MOI-Cl > MOI-Co PI-Cl > (MOICl; MOI-Co) PI-Co > (MOICl; MOI-Co) Prod. PI-Cl = PI-Co = MOI-Cl = MOI-Co

Linguistic Treatment and feature/language results

Study

Int. PI-Cl = 8.0 PI-Co = 8.1 MOI-Cl = 5.6 MOI-Co = 3.6 Prod. PI-Cl = 7.9 PI-Co = 7.8 MOI-Cl = 7.5 MOI-Co = 7.5

Posttest 2 PI-Cl = 68.6% PI-Co = 77.7% PI-Hyb = 65.3%

Posttest 1 PI-Cl = 67.9% PI-Co = 81.3% PI-Hyb = 84.7%

PI-Cl = 5.7% PI-Co = 6.3% PI-Hyb = 6.1% Int. PI-Cl = 2.6 PI-Co = 2.4 MOI-Cl = 2.3 MOI-Co = 3.1 Prod. PI-Cl = 1.6 PI-Co = 2.2 MOI-Cl = 1.7 MOI-Co = 2.1

Posttest 2 PI-Cl = 60.0% PI-Co = 64.3% PI-Hyb = 53.7%

Posttest 1 PI-Cl = 60.0% PI-Co = 56.6% PI-Hyb = 51.0%

Means scores Posttest

PI-Cl = 46.2% PI-Co = 46.4% PI-Hyb = 46.3%

Means scores Pretest

Int. PI-Cl = 54% PI-Co = 57% MOI-Cl = 33% MOI-Co = 5% Prod. PI-Cl = 63% PI-Co = 56% MOI-Cl = 58% MOI-Co = 54%

Posttest 1 PI-Cl = 62.2% PI-Co = 75% PI-Hyb = 78.6%

Posttest 1 PI-Cl = 13.8% PI-Co =10.2% PI-Hyb = 14.7%

Improvement

Posttest 2 PI-Cl = 62.9% PI-Co = 71.4% PI-Hyb = 59.2%

Posttest 2 PI-Cl = 13.8% PI-Co = 17.9% PI-Hyb = 17.4%

Table 4.2.  Summary of the improvement from pretest to posttest of studies investigating modes of delivering Processing I­ nstruction

Appendix  111

French subjunctive

Italian subjunctive

Lee & Benati 2007a

Lee & Benati 2007b Int. PI-Cl = 1.0 PI-E-CL = 1.7 PI-E-CO = 1.3 Prod. PI-Cl = .33 PI-E-CL = .83 PI-E-CO = .51

Int. PI-Cl = 1.1 PI-Co = 1.3 MOI-Cl = 1.1 MOI-Co = 1.1 Prod. PI-Cl = .6 PI-Co = 1.0 MOI-Cl = .73 MOI-Co = 1.4

Means scores Pretest

Int. PI-Cl = 6.7 PI-E-CL = 6.7 PI-E-CO = 7.0 Prod. PI-Cl = 4.2 PI-E-CL = 4.6 PI-E-CO = 4.3

Int. PI-Cl = 7.8 PI-Co = 7.2 MOI-Cl = 2.9 MOI-Co = 1.7 Prod. PI-Cl = 6.7 PI-Co = 5.9 MOI-Cl = 6.7 MOI-Co = 6.5

Means scores Posttest

Int. PI-Cl = 57% PI-E-CL = 50% PI-E-CO = 67% Prod. PI-Cl = 38.7% PI-E-CL = 37.7% PI-E-CO = 37.9%

Int. PI-Cl = 67% PI-Co = 59% MOI-Cl = 18% MOI-Co = 6% Prod. PI-Cl = 61% PI-Co = 49% MOI-Cl = 60% MOI-Co = 51%

Improvement

Int. = interpretation task results; Prod. = production task results. PI-Cl = Processing Instruction delivered in a classroom; PI-Co = Processing Instruction delivered online (computer); MOI-Cl = Meaningful Ouput Instruction delivered in a classroom; MOI-Co = Meaningful Ouput Instruction delivered online; PI-hyb = Processing Instruction delivered as a hybrid of classroom and computer formats and materials

PI-Cl vs. PI-ECl vs. PI-E-Co PI-Cl = PI-ECl = PI-E-Co

PI-Cl vs. PI-Co vs. MOI-Cl vs. MOI-Co Int. PI-Cl = PI-Co MOI-Cl = MOI-Co PI-Cl > (MOICl; MOI-Co) PI-Co > (MOICl; MOI-Co) Prod. PI-Cl = PI-Co = MOI-Cl = MOI-Co

Linguistic Treatment and feature/language results

Study

Table 4.2.  (cont.)

112  Online delivery of Processing Instruction

Structured Input enhancement and language development

Chapter 5 Can you increase the positive effects of Structured Input on language development by enhancing it aurally and/or textually?

Previous Processing Instruction-based research, reviewed in Chapter 2, isolated the effects of Structured Input vis à vis full Processing Instruction and explicit information (EI). The findings overwhelming point to Structured Input as the causative variable, over EI, in bringing about the effects we find for Processing Instruction. Engaging learners in processing structured input is the element of Processing Instruction that brings about changes to learners’ developing systems. Structured Input is, arguably then, the heart of Processing Instruction. Lee and Benati (2007b) focused in on Structured Input and in a series of experiments compared the effects of enhancing Structured Input both aurally and textually, depending on the mode of the input, with those of unenhanced Structured Input which is Structured Input as we know it (Lee and VanPatten 1995, 2003; VanPatten 1996; Wong 2004a; Farley 2005). Structured input is a type of manipulated input, manipulated such that learners must use the targeted form in the input and connect it to its meaning. Learners, for example, would listen to a sentence such as the following that contained a verb form inflected for future tense. The verb was the only indicator of futurity in this sentence; there were no additional lexical indicators of time such as ‘next month’ or ‘after graduating’.

Lavorerò per l’UNICEF.

Based solely on the morphological ending on the verb, learners were asked to decide if the statement referred to Tony Blair’s current or future plans (Lee and Benati 2007a: 149). This item referred to the future but the input items mixed future and present tense forms so that the correct answer was “current plans” for some of the items. Sharwood Smith (1991, 1993; Rutherford and Sharwood Smith 1985) advocated that forms in the input could be made more salient to learners in a variety of ways including typographically as well as perceptually. The more salient the form the more likely it would be to be processed in the

114 

Structured Input enhancement and language development

input ­because the form’s salience would draw learners’ attention. Researchers have utilized textual enhancements under a variety of conditions with a variety of intentions. Textual enhancements have been combined with input floodings, a technique by which learners are exposed to an artificially large number of target items (White 1998). Others have compared the effects of enhancements with those of orienting learners to a formal rule for forming the targeted linguistic feature (Shook 1994; Alanen 1995). Several researchers examined whether enhancements promoted learners’ noticing the forms in the input (Alanen 1995; Jourdenais et al. 1995; Leow 2001). Much research has investigated the effects of enhancements on learners’ comprehension of passage content (Leow 1997, 2001; Leow et al. 2003; Overstreet 1998; Wong 2003). The primary focus has been to examine the intended effects of input enhancements on the intake of grammatical forms, that is, on whether learners not only notice the forms in the input, but make a form– meaning connection (Shook 1994; Alanen 1995; Leow 1997, 2000, 2001; Overstreet 1998; White 1998; Leow et al. 2003; Wong 2003). The research we review in the present chapter has the aim of investigating whether enhanced input promotes greater form–meaning connections than does unenhanced input. Typographical enhancements include underlining or capitalizing a target form as well as using bold typeface (Wong 2005). Perceptual enhancements include speaking the form more loudly than the surrounding input (i.e., raising your voice), producing the form with tenser muscle movements (think of children saying ‘please’ in that extra-nice way), and by pausing before and/or after the target in order to draw attention to it (Leeser 2005). Lee and Benati (2007b) undertook a series of empirical investigations of second language processing. In their research, they compared the effects of enhanced versus unenhanced structured input on learners intake of a variety of grammatical items in Italian and Japanese. The input activities included both written and aural input. They chose bold typeface with underlining as the textual enhancement of written input. They decided to use bold typeface and underling only on the target feature in three of the studies, but underlined the entire word in which the target occurred in one of the studies. The following examples illustrate the point. artistico dica troverà ikimashita

Italian noun–adjective agreement Italian subjunctive mood Italian future tense Japanese past tense marker

Processing non-meaningful and redundant forms Processing non-meaningful and redundant forms 

115

When the input activities utilised aural input, the target forms were enhanced using the vocal apparatus. They increased the muscular tension used to produce the target form and combined it with raised volume as the perceptual enhancement. The purpose in enhancing structured input was to draw learners’ attention to the target forms and thereby increase the effects that structured input would have on language development vis a vis the effects of ­unenhanced input. Will enhanced and unenhanced structured input help learners process non-meaningful and redundant forms? VanPatten (2004b: 11) captured learners’ processing difficulties with nonmeaningful and redundant forms in two of his processing principles. They are stated as follows. “P 1c. The Preference for Nonredundancy Principle: learners are more likely to process nonredundant meaningful grammatical form before they process redundant meaningful forms. P 1d. The Meaning-Before-Nonmeaning Principle: learners are more likely to process meaningful grammatical forms before nonmeaningful forms irrespective of redundancy.” (VanPatten 2004b: 11) Two forms in Italian present learners the processing problems of non-meaningfulness and redundancy. They are noun-adjective gender agreement and subjunctive mood verb morphology in nominal subordinate clauses. Some morphology is meaningful; that is, the morphology has semantic value. For example, the Spanish and English plural marker is word-final ‑s, as in sillas and chairs, respectively. Other morphology is non-meaningful in that it does not carry semantic value. Grammatical gender marking on nouns (masculine, feminine, and, sometimes, neuter) in Romance languages is an example of non-meaningful morphology. That the Spanish word silla ‘chair’ is marked as feminine (-a) and its near-synonym asiento ‘seat’ is masculine (-o) has no bearing on their meanings. The surface agreement features are such that adjectives agree in number and gender with the nouns they modify; the noun determines the form of adjective. Whether the adjective is marked as masculine or feminine does not change the meaning of the adjective. Both cómoda and cómodo mean ‘comfortable’. Gender marking on adjectives is usually a redundant marking. Consider the noun phrase una silla cómoda. The gender marking on cómoda is the third feminine marker in the string.

116 

Structured Input enhancement and language development

As is the case of noun–adjective agreement in Spanish, adjectives in Italian must agree in number and gender with the noun they modify (la casa famosa, il castello famoso). This agreement feature (a = singular feminine) has no inherent semantic value and, depending on the structure of the noun phrase, is highly redundant (la casa famosa and il castello famoso). Finally, there is no difference in meaning between famosa and famoso; both mean famous. the form of the adjective itself, its feminine/masculine marking, does not contribute to the meaning of the sentence. Benati (2004b) established that the most effective component of Processing Instruction for Italian noun–adjective agreement was the structured input activities. Specifically, he compared the performance of learners who received full Processing Instruction (explanation plus structured input practice) to that of learners who only received Structured Input (structured input practice with no explanation). The two groups made equal gains from pretest to posttest on both production and interpretation of the forms. Taking that research as a point of departure, Lee and Benati (2007b) compared the performance of learners who received enhanced Structured Input with that of learners who received unenhanced or typical Structured Input. They addressed the question of whether learners make greater grammatical gains on Italian noun–adjective agreement by performing structured input activities in which the targeted adjective form is enhanced textually/aurally as compared to performing structured input activities without enhanced forms as measured by both form interpretation and form production tests. They examined the performance of 20 native speakers of English as well as beginning level learners of Italian who were all enrolled in a first-semester university-level Italian language class. The mean scores on the two pretests ranged from 15% to 19% indicating a very low level of knowledge about noun adjective agreement prior to the treatments. The subjects were divided into two groups both of which performed the same structured input activities. One group’s activities were enhanced textually and aurally whereas the other group’s were not. These activities consisted of both referential and affective activities. For the referential activities learners selected the person to whom a sentence referred based solely on the form of the adjective. For the affective activities learners agreed or disagreed with a statement that referred to different albeit specific male or female persons. The instruction took place over two consecutive class periods lasting a total of four hours. After instruction learners performed a 20-item interpretation posttest on which 10 of the items targeted noun–adjective agreement and a 10-item production test for which five items referred to feminine gender and the other five to masculine gen-

Processing non-meaningful and redundant forms 

117

der. The mean scores on the posttests ranged from 54% to 65% which indicated an improvement of from 39% to 47%. The statistical analyses revealed that there was a significant difference between the pretest and posttest scores (i.e., the variable Time) on both types of posttests. There was no statistically significant difference between the scores on either test for enhanced versus unenhanced input (i.e., the variable Treatment). There was no significant interaction between the two variables. All in all, enhanced and unenhanced structured input activities proved equally effective in promoting grammatical gains on Italian noun–adjective agreement. What grammarians call complex sentences are those that contain independent and dependent clauses. The terminology illustrates certain grammatical interrelationships. For Italian and other Romance languages (French, Spanish) when the semantic intent of the independent clause expresses doubt, disbelief, uncertainty or, in some languages, opinion, then the verb in the dependent clause must be inflected for the subjunctive mood. In other words, the semantics of the independent clause triggers the form of the verb in the dependent clause. The subjunctive mood is encoded morphologically and is distinct in form from the indicative mood. The following two sentences exemplify the subjunctive–indicative distinction in Italian. Dubito che Paulo non parli inglese. ‘I doubt that Paolo does not speak English.’ So che Paolo non parla inglese. ‘I know that Paolo does not speak English.’ The meanings of the verbs dubito and so trigger the forms parli and parla, respectively. Parli and parla mean exactly the same thing, speaks. If the semantic intent of the independent clause expresses knowledge, affirms or asserts belief, or certainty, then the verb in the dependent clause must be inflected for the indicative mood (… Paolo non parla inglese). The processing problem for second language learners with the subjunctive/indicative alternation is that the subjunctive form in expressions of doubt, uncertainty, disbelief or opinion is nonmeaningful. The subjunctive form itself does not contribute to the semantics of the sentence. Another processing problem for second language learners is that the semantics of the utterance is rendered by the lexical meaning of the verb or verb phrase in the independent clause, not by the formal properties of the verb in the dependent clause. Due to the cross-clausal interrelationship and that Italian reads left to right, the subjunctive morphology is a redundant marker.

118 

Structured Input enhancement and language development

Subjunctive mood morphology also presents learners with an additional processing problem captured by another processing principle. “Principle 1f. The Sentence Location Principle. Learners tend to process items in sentence initial position before those in final position and those in medial position.” (VanPatten 2004b: 14) As can be seen in the example sentence above, the position of the subjunctive verb forms in the middle of the subordinate clause puts them in a lesspreferred processing location. Previous research on the effects of Processing Instruction on the subjunctive in Spanish (Farley 2001, 2004a, 2004b), Italian (Lee and Benati 2007a), and French (Lee and Benati 2007a) demonstrated that Processing Instruction is an effective form of instruction. Taking that research as a point of departure, Lee and Benati (2007b) compared the performance of learners who received unenhanced or typical Processing Instruction with that of learners who received enhanced Processing Instruction delivered in one of two modes, either in the classroom or via computer in a laboratory setting. They addressed the question of whether learners make greater grammatical gains on Italian subjunctive of doubt and opinion by receiving Processing Instruction, in a classroom or in a computer laboratory, in which the targeted subjunctive mood verb form is enhanced textually/aurally as compared to receiving Processing Instruction without enhanced forms as measured by both form interpretation and form production tests. They examined the performance of 24 native speakers of English as well as intermediate level learners of Italian who were all enrolled in a first-­ semester second-year Italian language class. The mean scores on the pretests ranged from 3% to 17% indicating low levels of knowledge about subjunctive mood use or forms prior to the treatments. The subjects were divided into three groups with each one performing the same activities. Two groups’ activities were enhanced textually and aurally (classroom versus online materials) whereas the third group’s were not (classroom materials). Learners carried out both referential and affective activities. The instruction took place over two consecutive class days for a total of four hours. After instruction, learners performed a 20-item interpretation posttest on which 10 of the items targeted subjunctive forms and a 10-item form production test. The mean scores on the posttests ranged from 42% to 70%. The statistical analyses revealed that there was a significant difference between the pretest and posttest scores (i.e., the variable Time) on both types of posttests. There was no statistically

Processing meaningful forms vs. lexical items Processing meaningful forms vs. lexical items 

119

significant difference between the scores on either test for enhanced versus unenhanced input delivered in a classroom or on computer (i.e., the variable Treatment). There was no significant interaction between the two variables. All in all, enhanced and unenhanced Processing Instruction, in a classroom or on a computer, proved equally effective in promoting grammatical gains on Italian subjunctive mood of doubt and opinion. Learners’ scores improved from 38% to 57%. Can enhanced Structured Input help push learners to process a meaningful form over a lexical item? VanPatten (2004b: 14) captured learners’ processing difficulties with co-occurring lexemes and morphemes in two of his processing principles. They are the Primacy of Content Words Principle and the Lexical Preference Principle. “P 1a. The Primacy of Content Words Principle: learners process content words in the input before anything else. P 1b. The Lexical Preference Principle: learners will tend to rely on lexical items as opposed to grammatical form to get meaning when both encode the same semantic information. (VanPatten 2004b: 14) Many studies have documented the positive effects that Processing Instruction has on directing learners attention to meaningful forms in the input. The forms investigated have ranged from the Spanish preterite tense (Cadierno 1995), Italian future tense (Benati 2001), English simple past tense (Benati 2005), Japanese past tense (Lee and Benati 2007a), among others. In these works, the morphology is meaningful. In the studies just mentioned, the morphology signals the time frame in which something took place: present, past, or future. Yet, when we speak or write about events and when they take place we often use lexical temporal indicators in addition to the verb morphology to refer to the time frame. Words and phrases such as ‘tomorrow,’ ‘yesterday,’ and ‘later today’ convey the same information as the verb morphology. Second language learners tend to use these lexical items to establish the time frame rather than the verb morphology. If they do not process the verb morphology, they can not begin to acquire it. Benati (2001) established that Processing Instruction for Italian future tense was an effective type of instruction to help learners process meaningful verb morphology. Specifically, he compared the effects of Processing Instruc-

Meaningful morphology in sentence-final position 120 

Structured Input enhancement and language development

tion to those of traditional instruction (TI). The Processing Instruction group made more gains than the TI group on the interpretation posttest while both groups made equal gains on the production posttest. Taking that research as a point of departure, Lee and Benati (2007b) compared the performance of learners who received enhanced Structured Input with that of learners who received unenhanced or typical Structured Input. They addressed the question of whether learners make greater grammatical gains on Italian future tense verb morphology by performing structured input activities in which the targeted verb form is enhanced textually/aurally as compared to performing structured input activities without enhanced forms as measured by both form interpretation and form production tests. They examined the performance of 20 native speakers of English who were also enrolled in a first-semester, beginning-level Italian language class. The mean scores on the two pretests ranged from 15% to 29% indicating a low level of knowledge about Italian future tense verb morphology prior to the treatments. The subjects were divided into two groups and given the same structured input activities to perform. One group’s activities were enhanced textually and aurally whereas the other group’s were not. The learners performed both affective and referential activities over two consecutive class periods, for a total of four hours of instruction. After carrying out the activities, the learners performed a 20 item interpretation test for which 10 items referred to the future tense and a 10 item production test. The mean scores on the posttests ranged from 56% to 74% indicating an improvement of between 41% and 49%. The statistical analyses revealed that there was a significant difference between the pretest and posttest scores (i.e., the variable Time) on both types of posttests. There was no statistically significant difference between the scores on either test for enhanced versus unenhanced input (i.e., the variable Treatment). There was no significant interaction between the two variables. All in all, enhanced and unenhanced structured input activities proved equally effective in promoting grammatical gains on Italian future tense morphology. Will enhanced and unenhanced structured input help direct learners’ attention to meaningful morphology in sentence-final position? Japanese word order places the verb in sentence-final position. Additionally, Japanese marks the verb for tense using a word final inflection, -mashita. Both features of Japanese conspire to make the past tense marker difficult for

Meaningful morphology in sentence-final position 

121

second language learners to process. VanPatten (2004b) captured these difficulties in several processing principles. “P 1f. The Sentence Location Principle: learners tend to process items in sentence initial position before those in final position and those in medial position. P 1b. The Lexical Preference Principle: learners will tend to rely on lexical items as opposed to grammatical form to get meaning when both encode the same semantic information.” (VanPatten 2004b: 14) These principles are exemplified in the following sentence. Kino kaisha ni ikimashita. ‘Yesterday, I went to the office.’ Both the lexical item kaisha and the verb marker -mashita communicate past temporal reference. Lee and Benati (2007a) established that performing structured input activities is an effective treatment for helping learners make grammatical gains with the Japanese past tense marker. Specifically, they compared the results of Structured Input (structured input practice with no explanation) with those of traditional output-based instruction (TI). The Structured Input group outperformed the TI group on the interpretation posttest whereas both groups made equal gains on the production posttest. Taking this research as a point of departure, Lee and Benati (2007b) compared the performance of learners who received enhanced Structured Input, enhanced Structured Input and a no instruction control group. The conducted immediate interpretation and production posttests and, one week later, delayed posttests. They examined the performance of 26 adult, native-speakers of Italian who were enrolled in a beginning-level Japanese language course at a private language school in Italy. The mean scores on the two pretests ranged from 3% to 9% indicating a very low level of knowledge of the targeted past tense marker. The subjects were divided into three groups, two of which carried out the same structured input activities (one set enhanced, the other not) while the third acted as a control group receiving no activities. The instruction took place over two class meetings for a total of four hours of instruction. Learners performed both referential and affective activities. After instruction learners performed a 20-item interpretation posttest on which 10 of the items targeted the past tense marker and a 10-item production test. The mean scores

122 

Structured Input enhancement and language development

on the immediate posttests for the two groups performing structured input ­activities ranged from 46% to 60% which indicated an improvement of from 43% to 51%. The mean scores on the delayed posttests for these two groups ranged from 41% to 52% indicating an improvement from the pretest scores of between 38% and 43%. The statistical analyses revealed that on both the interpretation and production tests a significant effect for Time, a significant effect for Treatment, and a significant interaction between the two. The effect for Time was due to the improved performance of the two groups that performed structured input activities compared to the control group that did not improve. Likewise, the effect for Treatment showed that the two Structured Input groups outperformed the control group and performed equally well to each other. The significant interaction is due to the control groups lack of improvement. All in all, enhanced and unenhanced structured input are equally effective in helping learners process word-final, sentence-final verb morphology. Summary In this chapter, we have reviewed the results of a handful of studies that examined possible differential effects of enhanced versus unenhanced structured input on learners’ grammatical development. These studies are summarized in Table 5.1. As can be seen in this table, the research has examined both Italian and Japanese, two languages that represent quite different typologies. The Processing Principles, and hence the processing challenges to learners, are quite varied. Whereas the number of studies is quite limited, the findings are perfectly consistent and can be phrased in at least two ways, depending on one’s perspective. First, we can say that enhanced structured input is not more effective than unenhanced structured input in promoting second language grammatical development. Textual and aural enhancements do not bring about greater improvement in learners’ performance; they do not make the input more salient to the learners. Second, we can say that enhanced and unenhanced structured input are equally effective in promoting second language grammatical development. Structured input, in its enhanced or unenhanced form is sufficient to direct learners attention to a targeted form and for getting them to use that form to make a form–meaning connection. The rationale for exploring possible effects for enhanced structured input was that of trying to improve upon the results we get using unenhanced struc-

Summary 

123

tured input. In Table 5.2 we present a summary of the percent of improvement documented between pretest and posttest scores for the enhanced and unenhanced conditions. Across the studies we see that the greatest degree of improvement is in the area of form interpretation vis a vis form production. The improvement in form interpretation ranges from 42% to 57%. The improvement in form production ranges from 37% to 48%. Future research will have to focus on something other than textual and aural enhancements as the means by which to increase learners’ rates of improvement from pretest to posttest.

Italian noun–adjective agreement

Italian subjunctive

Italian future tense

Japanese past tense marker

Lee and Benati 2007b

Lee and Benati 2007b

Lee and Benati 2007b

Lee and Benati 2007b

Sentence location’ Lexical Preference

Lexical Preference

Nonredundancy, Non-meaningful, Sentence location

Nonredundancy and non-meaningful

Beginners L1 Italian

26

20

Pretest Posttest Immediate & delayed

Pretest Posttest Immediate

Pretest Posttest Immediate

Intermediate 24 L1 English

Beginners L1 English

Pretest Posttest Immediate

Number Design

20

Beginners L1 English

Processing principle Subjects/l1

Int. = SI = SI-E Prod. = SI = SI-E

Int. = PI = PI-E = PIE-C Prod. = PI = PI-E = PI-E-C Int. = SI = SI-E Prod. = SI = SI-E

Posttest 1 Int. = (SI = SI-E) > C Prod. = (SI = SI-E) > C Posttest 2 Int. = (SI = SI-E) > C Prod. = (SI = SI-E) > C

Interpretation (aural) Production (written) Interpretation (aural) Production (written) Interpretation (aural) Production (written)

Results

Interpretation (aural) Production (written)

Tests

PI = Processing Instruction (unenhanced); PI-E = enhanced Processing Instruction delivered in a classroom; PI-E-C = enhanced Processing Instruction delivered via computer; SI = Structured Input (unenhanced); SI-E = enhanced Structured Input; C = control group receiving no instruction on the targeted linguistic feature; Int. = interpretation test; Prod. = production test

Linguistic feature/ language

Study

Table 5.1.  The effects of enhanced versus unenhanced Structured Input

124  Structured Input enhancement and language development

Italian noun–adjective agreement

Italian subjunctive

Italian Future tense

Lee & Benati 2007b

Lee & Benati 2007b

Lee & Benati 2007b

SI vs. SI-E SI = SI-E

PI vs. PI-E vs. PI-E-Co PI = PI-E = PIE-Co

SI vs. SI-E SI = SI-E

Linguistic Treatment feature/language and results

Study

Int. SI = 7.4 SI-E = 7.1 Prod. SI = 5.6 SI-E = 5.9

Prod. SI = 1.5 SI-E = 1.8

Prod. SI = 41% SI-E = 41%

Int. SI = 49% SI-E = 42%

Prod. PI = 38% PI-E = 38% PI-E-Co =38%

Prod. PI = 4.2 PI-E = 4.6 PI-E-Co =4.3

Prod. PI = .33 PI-E = .83 PI-E-Co = .51 Int. SI = 2.5 SI-E = 2.9

Int. PI = 57% PI-E = 50% PI-E-Co = 57%

Prod. SI = 37% SI-E = 39%

Int. PI = 6.7 PI-E = 6.7 PI-E-Co = 7.0

Prod. SI = 5.5 SI-E = 5.4

Prod. SI = 1.8 SI-E = 1.5

Int. SI = 44% SI-E = 47%

Improvement

Int. PI = 1.0 PI-E = 1.7 PI-E-Co = 1.3

Int. SI = 6.5 SI-E = 6.4

Means scores Posttest

Int. SI = 1.9 SI-E = 1.7

Means scores Pretest

Table 5.2.  Summary of the improvement from pretest to posttest(s) of studies investigating enhanced versus unenhanced Structured Input

Appendix Appendix  125

Prod. SI = 4.1 SI-E = 4.7 C = .29

Posttest 2 Int. SI = 5.0 SI-E = 5.2 C = .43 Prod. SI = 43% SI-E = 48% C = na

Posttest 1 Int. SI = 51% SI-E = 51% C = na

Improvement

Prod. SI = 38% SI-E = 42% C = na

Posttest 2 Int. SI = 43% SI-E = 43% C = na

PI = Processing Instruction (unenhanced); PI-E = enhanced Processing Instruction delivered in a classroom; PI-E-Co = enhanced Processing Instruction delivered via computer; SI = Structured Input (unenhanced); SI-E = enhanced Structured Input; C = control group receiving no instruction on the targeted linguistic feature; Int. = interpretation test; Prod. = production test; na = not applicable

Prod. SI = 4.6 SI-E = 5.3 C = .57

Prod. SI = .33 SI-E = .50 C = .43

Means scores Posttest Posttest 1 Int. SI = 5.8 SI-E = 6.0 C = .71

Japanese past tense marker

Lee & Benati 2007b

Means scores Pretest

SI vs. SI-E vs. C (SI = SI-E) > C Int. SI = .67 SI-E = .90 C = .57

Linguistic Treatment and feature/language results

Study

Table 5.2.  (cont.)

126  Structured Input enhancement and language development

The transfer-of-training effects for Processing Instruction

Chapter 6 What are the transfer-of-training effects for Processing Instruction? In this chapter we intend to present, interpret and analyze one line of research in the Processing Instruction agenda that has assessed the transfer-of-training effects of this approach to grammar instruction. In chapter two of this book we have referred to classroom-based studies which have investigated primary effects for Processing Instruction. Likewise the studies which have been conducted to assess primary effects for Processing Instruction, the studies carried out to measure the transfer-of training effects have used a similar research design (pre-test-instructional period-post-test). Within this new line of research we have made a distinction between two kind of transfer-of-training effects: secondary effects or cumulative effects. In Benati and Lee (2008) we defined secondary effects ‘‘the effects of the Processing Instruction treatment used to help learners to process a particular form affected by a specific processing principle that are transferred to another linguistic form affected by the same processing principle’’. However, we referred to ‘‘cumulative effects when we measure the transfer-of training effects of Processing Instruction on linguistic items that are affected by very different processing principles’’. This is a very new line of research as empirical studies measuring the effects of Processing Instruction have mainly focused on primary effects and have undoubtedly and consistently showed that Processing Instruction is the most effective approach to grammar instruction at improving L2 learner’s performance/knowledge of the targeted linguistic form in both interpretation and production tasks. This line of research into the transfer-of-primary effects of Processing Instruction generates from two hypotheses formulated by Lee (2004: 322): Hypothesis 1. Learners who receive training on one type of processing strategy for one specific form will appropriately transfer the use of that strategy to other forms without further instruction in Processing Instruction. Hypothesis 2. The cumulative effects of Processing Instruction will be greater than its isolated effects.

128 

The transfer-of-training effects for Processing Instruction

Five studies (see Table 6.1) have been carried out to provide support to the first hypothesis (Benati and Lee 2008) and in this chapter we will review empirical evidence that will show that L2 learners receiving Processing Instruction can transfer their training on one linguistic feature to other forms or linguistic features affected by a similar processing problem without further instruction on the other forms. Does training on Processing Instruction on a processing problem aid the acquisition of a different form affected by the same processing problem? In the first empirical study (Benati and Lee 2008) two linguistic items from Italian were selected: noun–adjective agreement and future tense verb morphology. The processing problems associated with these two forms are captured by two processing principles, specifically, the Preference for Non-redundancy Principle and the Lexical Preference Principle (VanPatten 2004). Considering that both forms are affected by the same processing principles the main purpose of this study was to measure the possible transfer-of-training effects of Processing Instruction. We intended to establish whether the Processing Instruction group can transfer the instructional training received on noun–­ adjective agreement to the acquisition of Italian future tense morphology. We hypothesized that learners receiving Processing Instruction on gender agreement in Italian will develop good intuitions (processing strategies) and will be able to process Italian future tense verb-final forms. Twenty-five subjects participated in this study. They were all native speakers of English who were learning Italian at undergraduate level. Subjects were randomly assigned to three groups: Processing Instruction, traditional output-based instruction, and a control group that did not receive instruction on the two linguistic items over the duration of the investigation. Participants were exposed to a two-day instructional treatment period. Two packets of instructional materials were developed for the noun-adjective gender agreement linguistic item: one for the Processing Instruction group and another for the group receiving traditional instruction. Pre- and post-tests for one interpretation and a production task were developed for measuring the primary effects of instruction on the first feature (noun-adjective gender agreement in Italian) and the secondary transfer-of-training effects on the second feature (future tense verb morphology in Italian). The main purpose of this study was to measure primary and secondary transfer-of-training effects of instruction,

Training Processing Instruction and acquisition Training Processing Instruction and acquisition 

129

in particular, by comparing the two different instructional groups of students. The instructional groups were taught over a period of four hours. The pre-test means for primary and secondary effects showed in Table 6.2 was less than 1% across the different pre-tests and the different instructional treatment groups. The statistical analyses conducted on the pre-tests indicated that the three groups were not statistically different before the beginning of the instructional period on interpretation and production of gender agreement and future verbal morphology forms. The mean scores of the posttest for the interpretation task measuring primary effects indicates that the Processing Instruction group outperformed the other two groups (mean scores for the Processing Instruction group = 57%; mean scores for the traditional instruction group = 14%; mean scores for the control group = 0.6%). The Processing Instruction group gained 51% on the interpretation task from pre- to post-testing whereas the traditional group gained only about 5% and the control group 1%. The mean scores of the posttest for the production sentence level task measuring primary effects shows that the Processing Instruction group (mean scores for the posttest = 47%) and the traditional instruction group (mean scores for the posttest = 53%) performed equally well and better than the control group (mean scores for the posttest = 0.6%) . In the production task the Processing Instruction group and the traditional instruction group significantly improved their performance from pre-test to post-test (see Table 6.2). The Processing Instruction group gained 44% from the pre-test to the posttest while the traditional instruction group gained 48%. The control group gained only about 2%. The results of the statistical analyses conducted in this study confirmed the overall results (see chapter three) of studies investigating the primary effects of Processing Instruction. The mean scores (see Table 6.2) of the Processing Instruction group (29%) in the interpretation task measuring secondary effects show a clear improvement form pre-test to posttest (25%). The traditional group (10%) increased about 4%. The statistical analyses used on the results of the interpretation task to measure secondary effects revealed that the Processing Instruction group was better than the traditional instruction group and the control group in improving learners’ interpretation of future tense morphology forms. The mean scores for the production task showed that both the Processing Instruction group (14%) and the traditional group (13%) made similar gains from pre-test to posttest compared to the control group. The improvement of the Processing Instruction group and traditional instruction group from the ­pre-test to

130 

The transfer-of-training effects for Processing Instruction

the post-test was minimal (the Processing Instruction group improves about 8% in producing future tense forms and the traditional instruction group improves about 9%). However, like in the case of the results on the interpretation task this significant improvement took place in the absence of direct instruction on the target form (Italian future tense morphology). The statistical analyses conducted on the raw scores of the production task showed that the Processing Instruction group was better than the control group and equal to the traditional group. The scores for the traditional and control group were not statistically different from each other. The main aim of this study was to measure primary and secondary effects of Processing Instruction. In terms of primary effects the results of this first study show that the performance of the Processing Instruction group on the interpretation task was statistically significantly superior to the performance of either the traditional or the control groups. In the production task the two instructional group made similar gains. In terms of measuring secondary effect, as previously said, the main findings of this study indicate that the performance of the Processing Instruction group on the interpretation task was statistically significantly superior to the performance of the other two groups. The main conclusion of this study is that Processing Instruction is superior to traditional instruction on the secondary transfer-of-training effects in both measured in both interpretation and production tasks. As pointed out by Benati and Lee (2008: 86) ‘‘these findings suggests that the previous training on Processing Instruction learners received has influenced the way these L2 learners approach the target language’’. A second empirical study which supported the same findings came from a research on the acquisition of English as a second language. The main question of this study was to provide evidence of whether L2 learners of English who received the Processing Instruction training on past tense ‑ed were able to transfer successfully that training to another word final temporal morpheme (-s third person singular). Both linguistic features are affected by The Lexical Preference Principle and The Preference for Non-redundancy Principle (VanPatten 2004). The main purpose of this second study measuring secondary effects was again to discover whether the positive and primary effects of Processing Instruction can be transferred/applied by L2 learners to other features affected by the same processing problem. Two groups of twenty-six Korean native speaker students participated to this experimental study. They were all studying English in a beginning-level class. After the pretests subjects were assigned through a randomization procedure to one

Training Processing Instruction and acquisition 

131

of the two groups. The two instructional groups consisted of the Processing Instruction group and the traditional instruction group. Immediate posttest (interpretation and production) were administered to the two groups after the end of instruction. The instructional treatments lasted for four hours, each and were delivered across two days of instruction. The Processing Instruction group was taught the English simple past tense marker -ed using a Processing Instruction treatment in order to help them alter their reliance on lexical items (Lexical Preference Principle). The traditional group was taught the English simple past tense marker ‑ed using a traditional treatment for which they will practicing producing the target in mechanical and meaningful output activities. We intended to establish whether L2 learners can transfer that training by applying it to a novel form (-s- third person singular). The pre-test mean scores for primary effects showed in Table 6.2 was similar across the two instructional groups (from 34% to 35% for the interpretation task and 22% to 24% for the production task. The Processing Instruction group gained about 46% on the interpretation task from the pre-test to the posttest whereas the traditional group lost 2%. Both the Processing Instruction group and the traditional group significantly improved their performance from pre-test to posttest on the production task. The Processing Instruction group gained about 42% while the TI group gained about 48%. The means of the posttest for the interpretation task measuring secondary effects showed that the Processing Instruction group outperformed the other group. The Processing Instruction group made a 22% improvement from pretest to posttest whereas the traditional group lost 7%. The mean scores of the posttest for the production task measuring secondary effects showed that the Processing Instruction group improved from pre-test to posttest of about 11% whereas there was no improvement in the traditional instruction group which lost 7%. Likewise in the previous study, in this study the statistical analyses conducted on the raw scores of the interpretation and the production tasks showed that the Processing Instruction group was better than the traditional group. The results of this study confirmed that the Processing Instruction group was significantly better than the traditional group and learners in this group were able to transfer their instructional training to a feature affected by the same processing principles. A third empirical study addressed (see Table 6.1 and 6.2) the question as to whether L2 learners after receiving Processing Instruction on the use of affirmative vs. negative present tense in Japanese, can process Japanese past tense morphology that is an inflection that appears in word final position. The

132 

The transfer-of-training effects for Processing Instruction

experiment was designed to make the results as objective as possible within the constraints of a private Japanese language programme in Italy. Beginners students of Japanese participated in this study. All subjects were Italian native speakers and were studying Japanese in a private school were they received four hours of instruction over two consecutive days every week. Twenty-four subjects constituted the final data pool in this experiment. Subjects were randomly distributed to three groups; Processing Instruction group; traditional instruction group; and control group. An interpretation task was used as a measure of knowledge gained at interpreting present tense (affirmative vs. negative) and past tense forms at sentence level. Written completion production tasks were developed and used to measure learner’s ability to produce sentences in the past forms and the use of positive and negative sentences in present forms. A pre and post test procedure was used with a control group and the two instructional treatments group. Participants received pre-test in both linguistics features before the beginning of the instructional period (on the use of present forms) and post-tests immediately after the end of the instructional treatment. Learners received the instructional treatments (four hours on two consecutive days) on the use of positive and negative present forms in Japanese. After the end of the instructional period, the groups were given post-tests on both the first instructed feature, and the second instructed feature (past tense forms) to measure for both primary and secondary effects. The pre-test mean scores for primary effects showed in Table 6.2 were similar across the two instructional groups. The Processing Instruction group gained about 61% on the interpretation task from the pre-test to the posttest whereas the traditional group gained only 1% and the control group lost 1%. The Processing Instruction group was significantly different than the traditional group and the control group; the TI group and the control group were not different. Both the Processing Instruction group and the traditional group significantly improved their performance from pre-test to post-test on the production task indicating an improvement of from 40% to 45%. Both treatments seem to bring about the same improved performance on producing Japanese present tense forms at sentence level. The means of the posttest for the interpretation task measuring secondary effects (use of Japanese past tense forms) indicated that the Processing Instruction group outperformed the other two groups. The Processing Instruction group made a 38% improvement from pre-test to posttest whereas the traditional group made a 5% improvement and the control 3%. The Processing Instruction group was significantly statistically different than the trad-

The cumulative effects of Processing Instruction The cumulative effects of Processing Instruction 

133

itional instruction group and the control group and there was no difference between the control group and the traditional group. Processing Instruction is a type of instruction that has secondary effects on the way learners process forms affected by the same processing problem. The mean scores of the posttest for the production task measuring secondary effects showed that the Processing Instruction group improved from pretest to posttest (20%) whereas there was little improvement in the traditional instruction group (6%) and the control group (7%). Once again the overall results of this study measuring secondary effects on Japanese linguistic features showed that Processing Instruction is a type of instruction that has secondary effects on the way learners interpret and produce forms affected by the same processing problem What are the transfer-of-training effects and cumulative effects of Processing Instruction? We present the results of a first study designed to investigate transfer-of-training effects of Processing Instruction in the acquisition of French. The study we firstly review examined secondary effects by measuring whether learners receiving Processing Instruction on the French imperfect tense, can transfer the instructional training they receive to the acquisition of other forms of French. More specifically, we sought to measure secondary transfer-oftraining effects on the acquisition of French subjunctive used for expression of doubt and cumulative transfer-of-training effects on French causative constructions with faire. The French imperfect and French subjunctive present second language learners a morphological processing problem captured in the Lexical Preference Principle (VanPatten 2004b) and The Sentence Location Principle. Because the processing problem is the same for the two forms, we refer to the transfer-of-training effects as secondary ones. From French imperfect to French causatives with faire is a move from morphology to syntax and involves two extremely different processing problems; the Lexical Preference Principle and the First Noun Principle (on the acquisition of French subjunctive used for expression of doubt and cumulative transfer-oftraining effects on French causative constructions with faire). The population of this study consisted of thirty undergraduate students enrolled in intermediate-level French course as part of their undergraduate degree. Subjects were assigned to three groups: Processing Instruction; traditional instruction and a control group. One interpretation task and one

134 

The transfer-of-training effects for Processing Instruction

p­ roduction task were developed for each linguistic feature. The interpretation task measured the ability for learners to interpret sentences of the target feature correctly. The production task measured the ability to produce correct sentences containing the target feature. The instructional treatment lasted for one class period of two hours and posttests were administered immediately after the end of the instructional treatment. As far as measuring primary effects (mean scores on French imperfect tense forms) the mean scores of the interpretation task showed that the Processing Instruction group gained about 58% from pre-test to posttest whereas the traditional group improved only 5% and the control group 2%. Both the Processing Instruction group and the traditional group significantly improved their performance from pre-test to post-test on the production task. The Processing Instruction group gained about 140% from pre-test to posttest while the TI group gained about 95%. The means of the posttest for the interpretation task measuring secondary effects (French subjunctive) indicated that the Processing Instruction group outperformed the other groups. The Processing Instruction group made a 19% improvement from pre-test to posttest whereas the traditional group lost 4% and the control group 5%. The mean scores of the posttest for the production task measuring secondary effects showed that only the Processing Instruction group and the traditional instruction group improved from pre-test to posttest improving (respectively 10% and 2%) whereas there was no improvement for the control group. Despite the fact that the two instructional groups improved, the Processing Instruction group was statistically superior than the other two groups. The mean scores of the posttests for both the interpretation and the production tasks measuring cumulative effects showed that only the Processing Instruction group made some significant gains (34% improvement in the interpretation task and 10% in the production task) from pre-test to posttests (see Table 6.2). In this study the results obtained in measuring primary effects showed the following: –– Processing Instruction is a more effective instructional treatment than traditional instruction in helping L2 learners at interpreting and producing sentences containing French imperfect form; –– Processing Instruction is successful at altering processing problems that affect the French imperfect forms (Lexical Preference Principle and Sentence Location principle).

The cumulative effects of Processing Instruction 

135

The results on secondary and cumulative effects indicated the following: –– the processing group was able to transfer the Processing Instruction training received for the French imperfect to another linguistic form in French (subjunctive) affected by similar processing problems; –– the processing group was able to transfer the Processing Instruction training received for the French imperfect to another linguistic form in French (causative) affected by different processing problems. In a partial replication of Benati, Lee and Laval’s study (2008), Laval (forthcoming) investigated whether L2 learners receiving Processing Instruction on French causative construction with faire were able to transfer that training to two French morphological forms (imperfect tense and subjunctive). The main purpose of this study was on one hand, to provide further evidence on the cumulative effects of Processing Instruction, on the other hand to address one of Lee’s hypothesis (Lee 2004: 320) ‘‘do learners pick up a second and third processing strategy more quickly and efficiently than they pick up the first one? Thirty-nine intermediate students studying French at University participated. The design, procedure, materials and assessment tasks where the same as the study previously reviewed in this chapter (­Benati, Lee and Laval 2008). However, the instructional groups (Processing Instruction and traditional instruction) received instruction on the French faire causative first. Cumulative effects were measured on the French imperfect tense and the French subjunctive. Learners received a training on a linguistics feature in French affected by the First Noun Principle and the effects of this training was also measured on different forms affected by different principles (Lexical Preference principle and Sentence Location Principle). According to Lee (2004: 320) ‘‘ Learners who receive Processing Instruction will develop better intuitions about the L2 than learners who receive other types of instruction’’. The mean scores on the interpretation and production of sentences containing French faire causative confirmed previous results on the effectiveness of Processing Instruction on primary effects. On the interpretation task the Processing Instruction group gained about 56% from pre-test to posttest whereas the traditional group improved only 16% and the control group lost 1%. Both the Processing Instruction group and the traditional group significantly improved their performance from pre-test to post-test on the production task. The Processing Instruction group gained about 114% from pre-test to posttest while the TI group gained about 78%. No gains for the control group.

136 

The transfer-of-training effects for Processing Instruction

The means of the posttest measuring secondary effects yielded interesting results. The Processing Instruction group was the only group that showed some improvement. The Processing Instruction group improved in the ability to interpret sentence containing subjunctive form of about 13%. The same group improved of about 21% in the interpretation task for French imperfect tense forms. In the production tasks for both French linguistic features, the three groups made no improvement from pre-tests to posttests (see Table 6.2). Conclusion Further research should continue to address the questions on transfer-oftraining effects posed in the present study. Five experimental studies have been reviewed in this section. They all indicated that Processing Instruction has primary effects on helping L2 learners at interpreting and producing forms correctly and accurately. In addition to that, these studied indicated that Processing Instruction has secondary effects as learners who received Processing Instruction were able to transfer that training to processing and producing another form on which they had received no instruction. Two of the studies reviewed in this chapter also showed a cumulative benefit of Processing ­Instruction in that learners transferred their training processing morphology to processing and producing a syntactic construction. Learners seems also to be able to transfer their training processing syntactic construction to processing and producing morphology.

Gender agree­ ment Italian

Benati and Lee 2008

Future tense Italian

Benati Lee with Laval 2008

French i­mperfect

French ­subjunctive Cumulative ­effects French causative with faire

Lexical Preference Principle Sentence Location principle The First Noun Principle

Principle Preference for Non-redundancy Principle

Lexical Preference

Lexical Preference Principle Preference for Non-redundancy Principle

Second linguis- Processing tic feature/ ­principles language

Benati, Lee -ed simple past -s third person with Dean in English singular Houghton 2008

First linguistic feature/ language

Study

Intermediate English ­native

Beginners Korean native

30

26

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate

Intermediate English ­native 25

Design

Subjects/L1 N

Interpretation (aural) Production (Written

Interpretation (aural) Production (written)

Interpretation (aural) Production (written)

Tests

Results for secondary e­ ffects

Prod. = (PI Prod. = PI= TI = TI) > C PI> C TI = C

Int. = PI > Int. =PI > (TI (TI = C) = C)

Prod. = (PI Prod. = P  I= TI = TI) > C PI> C TI = C

Int. = PI > Int. = PI > (TI = C) (TI = C)

Prod. = (PI Prod. = PI= TI = TI) > C PI> C TI = C

Int. = PI > Int. = PI > (TI = C) (TI = C)

Results for primary ­effects

Table 6.1.  Summary of secondary and cumulative effects studies for Processing Instruction (PI – TI- SI – EI)

Conclusion  137

Japanese present (affirmative and negative

Benati 2008

Japanese Past tense

Intermediate English native

First Noun Principle Lexical Preference Principle Sentence Location Principle

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate

24

39

Design

Interpretation (aural) Production (written)

Interpretation (aural) Production (written)

Tests

Results for secondary ­effects

Prod. = (PI Prod. = PI= TI = TI) > C PI> C TI = C

Int. = PI > Int. = PI > (TI (TI = C) = C)

Prod. = (PI Prod. = PI= TI = TI) > C PI> C TI = C

Int. = PI > Int. = PI > (TI (TI = C) = C)

Results for primary ­effects

PI = Processing Instruction; TI = traditional instruction; C = control; Int. = Interpretation test; Prod. = Production test

French imperfect French subjunctive

Italian Beginners students of Japanese

Subjects/L1 N

Lexical Preference Principle Sentence Location Principle

Second linguis- Processing tic feature/ ­principles language

Laval 2008 French causCumulative efative with faire fects

First linguistic feature/ language

Study

Table 6.1.  (cont.)

138  The transfer-of-training effects for Processing Instruction

Secondary effects Int. = PI > (TI = C) Prod. = PI > (TI =C)

Secondary effects Future tense Italian

Primary effects Benati, Lee and Dean Primary effects -ed English simple past Int. = PI > (TI = C) Houghton 2008 tense Prod. = (PI = TI) > C

Primary effects Int. = PI > (TI = C) Prod. = (PI = TI) > C

Primary effects Gender agreement Italian

Benati and Lee 2008

Treatment and results

Linguistic feature/ language

Study

Prod. PI = 42% TI = 48%

Prod. PI = 6.6 TI = 7.0 Prod. PI = 2.4 TI = 2.2

Prod. PI = 8% TI = 9% C =   2%

Prod. PI = 1.4 TI = 1.3 C = 0.3

Prod. PI = 0.6 TI = 0.4 C = 0.5

Int. PI = 46% TI =   2%

Int. PI = 25% TI = 4% CI = 0% PI = 2.9 TI = 1.0 C = 0.3

Int. PI = 0.4 TI = 0.6 C = 0.3

Int. PI = 8.0 TI = 3.3

Prod. PI = 44% TI = 48% C = 2%

Prod. PI = 4.7 TI = 5.3 C = 0.6

Prod. PI = 0.3 TI = 0.5 C = 0.4

Int. PI = 3.4 TI = 3.5

Int. PI = 51% TI = 5% C = 1%

Improvement

Int. PI = 5.7 TI = 1.4 C = 0.6

Means Scores Posttest

Int. PI = 0.6 TI = 0.9 C = 0.5

Means scores Pretest

Table 6.2.  Summary of the improvement from pretest to posttest(s) of studies investigating secondary and cumulative effects studies for Processing Instruction

Conclusion  139

Secondary effects Int. = PI > (TI = C) Prod. = PI > ( TI =C)

Secondary effects -s- third person singular

Secondary effects French subjunctive

Secondary effects Int. = PI > (TI = C) Prod. = PI > ( TI =C)

Primary effects Int. = PI > (TI = C) Prod. = (PI = TI) > C

Treatment and results

Linguistic feature/ language

Benati, Lee and Laval Primary effects 2008 French imperfect

Study

Table 6.2.  (cont.)

Prod. PI = 140% TI =   95% C =    6% Int. PI = 19% TI =   4% CI =   5% Prod. PI = 10% TI =  2% C =   0%

Prod. PI = 14.1 TI = 15.1 C = 1.6 Int. PI = 3.6 TI = 1.7 C= 1.1 C = Prod. PI = 1.0 TI = 0.2 C=0

Prod. PI = 0.1 TI = 0.6 C= 1.0 Int. PI = 1.7 TI = 2.1 C = 1.6 C =Prod. PI = 0 TI = 0 C=0

Int.14 PI = 58% TI =   5% C =   2%

Prod. PI = 11% TI =   7%

Int. PI = 8.1 TI = 3.5 C = 2.8

Prod. PI = 4.4 TI = 2.5

Prod. PI = 3.3 TI = 3.2

Int. PI = 22% TI =   7%

Improvement

Int. PI = 2.3 TI = 3.0 C = 2.6

Int. PI = 5.4 TI = 2.7

Means Scores Posttest

Int. PI = 3.2 TI = 3.4

Means scores Pretest

140  The transfer-of-training effects for Processing Instruction

Benati 2008

Secondary effects Japanese past

Secondary effects Int. = PI > (TI = C) Prod. = PI > ( TI =C)

Primary effects Primary effects Int. = PI > (TI = C) Japanese present (affirmative and nega- Prod. = (PI = TI) > C tive)

Comulative effects Comulative effects French faire causative Int. = PI > (TI = C) Prod. = PI > ( TI =C)

Int. PI = 61% TI =   1% C =   1% Prod. PI = 40% TI = 45% C =   2% Int. PI = 38% TI =   5% CI =   3% Prod. PI = 20% TI =   6% C =   7%

Prod. PI = 4.3 TI = 4.7 C = 0.4 PI = 4.2 TI = 1.1 C = 0.4 Prod. PI = 2.3 TI = 1.7 C = 0.8

Prod. PI = 0.3 TI = 0.2 C = 0.6 Int. PI = 0.4 TI = 0.6 C = 0.1 Prod. PI = 0.3 TI = 1.1 C = 0.1

10% 2% 0%

Int. PI = 6.8 TI = 0.9 C = 0.3

C = Prod. PI = 1.0 TI = 0.2 C=0

C =Prod. PI = 0 TI = 0 C=0

34%   5%   3%

Int. PI = 0.7 TI = 0.8 C = 0.4

Int. PI = 4.6 TI = 0.3 C= 0.8

Int. PI = 1.2 TI = 0.8 C = 1.1

Conclusion  141

Comulative effects Int. = PI > (TI = C) Prod. = PI > ( TI =C)

Comulative effects French subjunctive

Int. PI = 3.2 TI = 1.2 C = 1.3 Prod. PI = 0.4 TI = 0 C = 0.1

Prod. PI = 1.0 TI = 0.3 C = 0.5

Prod. PI = 0.2 TI = 0 C=0

Prod. PI = 0.6 TI = 0.4 C = 0.1 Int. PI = 1.9 TI = 1.2 C = 1.9

Int. PI = 4.0 TI = 2.1 C = 1.9

Prod. PI = 11.4 TI =   7.8 C= 0

Prod. PI = 0 TI = 0 C=0 Int. PI = 1.9 TI = 2.5 C = 2.2

Int. PI = 5.8 TI = 2.3 C = 0.8

Means Scores Posttest

Int. PI = 0.2 TI = 0.7 C = 0.9

Means scores Pretest

Prod. PI =   6% TI =   3% C =   4%

Int. PI = 13% TI =   0% CI =   6%

Prod. PI =   4% TI =   4% C =   1%

Int. PI = 21% TI =   4% CI =   3%

Prod. PI = 114% TI =   78% C =    0%

Int.5 PI = 56% TI = 16% C =   1%

Improvement

PI = Processing Instruction; TI = traditional instruction; C = control; Int. = Interpretation test; Prod. = Production test

Secondary effects Int. = PI > (TI = C) Prod. = PI > ( TI =C)

Comulative effects French imperfect tense

Primary effects Primary effects French faire causative Int. = PI > (TI = C) Prod. = (PI = TI) > C

Laval 2008

Treatment and results

Linguistic feature/ language

Study

Table 6.2.  (cont.)

142  The transfer-of-training effects for Processing Instruction

The short- and long-term effects of Processing Instruction

Chapter 7 Are the effects of Processing Instruction durative (short-term) and longitudinal (long-term)?

In this chapter we examine the body of research that has examined the durative and longitudinal effects of Processing Instruction. We use the term ‘­durative effects’ to refer to posttesting that occurs beyond the time frame of examining the immediate effects of instruction. General consensus exists that the effects of instruction will fade, wear off, or otherwise diminish in size over time. Norris and Ortega (2000) provided a meta-analysis of studies on instructional effects in second language acquisition. They noted that in such studies, the effect of instruction appears durable although the effect size does decrease over time. We expect the effects of instruction to fade but the value to learners of an instructional intervention whose effects are not manifest even a week later would be questionable. What is, then, the time frame over which we examine durative effects? The precise answer to this question is difficult to give. In the context of the existing Processing Instruction research, various scholars have examined the durative effects of Processing Instruction one week, two weeks, three weeks, 24 days, and four weeks after administering the first posttest immediately following the instructional intervention (see Table 7.1). Durative effects are, therefore, those that we will discuss as being in the short term, a time period from one to four weeks following instruction. An effective instructional intervention should provide learners short-term or durative effects. Longitudinal effects are those that we will discuss as being in the long term. General consensus exists that the effects of instruction will diminish over time, but that the results of effective instruction should not disappear completely. That is, learners should, in the long-term retain some, if not all, of what they learned. How far out from instruction does an effect have to hold for it to be considered a long-term effect? We can offer no precise time frame but, reasonably, examining the effects of instruction six months or more beyond the instructional intervention should meet anyone’s criterion for examining a long-term effect. All research areas of instructed SLA, including Processing Instruction research, sorely lack research on longitudinal effects. For Processing Instruction, only VanPatten and Fernández (2004) have

144 

The short- and long-term effects of Processing Instruction

­ anaged the task of tracking a group of learners 8 months out from the time m of instruction. They did so with great difficulty as we will see in a later section of this chapter. Are the effects of Processing Instruction durative in the short-term? All Processing Instruction research has used a pretest to posttest design to assess the effects of instruction. The pretest, administered at what we call Time 1 (T1 in our tables), serves a variety of purposes. First, we use the pretest to screen the potential participants. Learners who already demonstrate a high level of knowledge on the target structure do not need instruction on that structure. Typically in Processing Instruction research, the data from learners who score over 60% on the pretest(s) are usually eliminated from the statistical analyses. The second function of the pretest scores is to match the knowledge level of the target structure across the different instructional treatment groups. The pretest provides us the mechanism to assign learners to different treatment groups or to insure that intact groups begin at the same place. That is, the groups must have equivalent knowledge of the target prior to receiving (or not in the case of control groups) an instructional intervention. The scores of all groups must be equivalent prior to treatment in order to assert that any changes in the posttest scores are due to the instructional treatments, not to learners’ differential prior knowledge. All Processing Instruction researchers have conducted an immediate posttest. We call this Time 2 (T2 in our tables). Immediately following instruction, we assess the learners’ knowledge. The general expectation is that at this time, the learners’ knowledge of the target would be at its highest; their knowledge is at its freshest. In the present analyses, we then refer to Time 3 (T3), a delayed posttest, delayed meaning that it occurred sometime after instruction finished and the immediate posttest was administered. In Table 7.1 we present a summary of the findings of the original Processing Instruction studies (VanPatten and Cadierno 1993a, 1993b; Cadierno 1995). These three studies included a second delayed posttest (T4). The first delayed posttest (T3) took place one week after instruction and the second (T4) took place one month after instruction. Being the first Processing Instruction studies, VanPatten and Cadierno were concerned with detailing the effects of the treatments, hence the two delayed posttests. For example, if they had found that after one month, the effects of Processing Instruction diminished, then they would have wanted to know at what point did the effects begin to fade.

The long-term effects of Processing Instruction The long-term effects of Processing Instruction 

145

Yet, as the results of the variable of time show (final column of the table), the effects of Processing Instruction did not diminish in the month following instruction; the Processing Instruction posttest scores were not significantly different from each other (VanPatten and Cadierno 1993a; Cadierno 1995). Another point to make about the findings is that on the interpretation test, Processing Instruction learners significantly outperformed TI learners on each of the posttests (VanPatten and Cadierno 1993b). Likewise, on the production test, Processing Instruction learners performed equally as well as the TI learners on each of the posttests (VanPatten and Cadierno 1993b). These results for interpretation and production would be repeated time and again in subsequent Processing Instruction research. In Table 7.2 we present the results of this same research but from the point of view of improvement. That is, by what percent do the treatment groups improve from Time 1 to Time 2 and by what percent is the improvement sustained over time? As we show in Table 7.2, Processing Instruction learners improve from 32% to 69% from T1 to T2 on the interpretation test. The rate of sustained improvement is from 36% to 65% after one week and from 36% to 63% after one month. How much improvement there was depends, in part, on how much prior knowledge learners possessed at the start of the treatments. The interpretation pretest scores for Spanish preterit tense morphology (Cadierno 1995) were much higher than the pretest scores for Spanish object pronouns. Hence the rates of improvement from preterit tense were from 32% to 36% whereas there were from 63% to 69% for object pronouns. The important point to make is that learners sustained the degree of improvement one month after instruction. Traditional instruction shows its least impact on the interpretation test scores from a low of 12% to a high of 34%. The type of form-focussed language production practices that typify Traditional instruction do not allow learners to listen for and interpret the meaning of the target forms. Both Processing Instruction and TI demonstrate significant effects on the production posttests. The immediate effects of Processing Instruction on form production range across the studies from 54% to 67% whereas the immediate effects of TI range from 42% to 58%. This degree of improvement is sustained over the time of the studies. The improvement across the posttests range from 52% to 65% for Processing Instruction and from 42% to 63% for TI. These results show us that not only are the findings statistically significant but that the improvement is substantial. Other Processing Instruction research has also sought to demonstrate the delayed effects of instruction. We present a summary of the findings of Processing Instruction research that includes an immediate posttest followed

146 

The short- and long-term effects of Processing Instruction

by only one delayed posttest in Table 7.3. As can be seen in column four of this table, Time 3 could have occurred one week, two weeks, three weeks, or 24 days after instruction. What determines how long after instruction a delayed posttest is administered? The answers may vary from one researcher to the next, but curricular considerations are often the motivating factor. All Processing Instruction research to date has been conducted in regular classrooms meaning that the learners are following a course programme of study. Our research and data gathering are scheduled into this programme. The delayed posttest often must be scheduled around curricular considerations. Let’s take one example. Aguilar-Sánchez and McNulty present the timeline for their data collection (Lee, Benati, Aguilar-Sánchez and McNulty 2007: 77). They designed the instructional treatments to occur when the textbook for the course introduced the target structures. The delayed posttest of one week occurred because the participants were to take a regularly scheduled examination that included the two target structures one week after the scheduled delayed test. No further delayed posttesting was possible because the participants began to study on their own the structures and their instructors began to review the structures for the exam. The studies whose findings we summarize in Table 7.3 are presented chronologically. These studies have been reviewed in previous chapters for their research foci be it types of instruction compared or the effective component of Processing Instruction, for example. In the present chapter, we focus on durative effects of Processing Instruction and its component SI (structured input). To that end we have prepared Table 7.4 that provides an overview of the durative effects of Processing Instruction and SI across the 12 studies examined. In how many of the 12 studies and 24 tasks learners performed, did Processing Instruction or SI yield a significant improvement in score from T1 (pretest) to T2 (immediate posttest) on an interpretation test? The answer is in all 9 of the 9 studies that performed an aural interpretation test or 100% of the time. In how many of the 13 studies and 24 tasks learners performed, did Processing Instruction or SI yield a significant improvement in score from T1 (pretest) to T2 (immediate posttest) on an oral or written production test? The answer is in all 14 of the 14 studies that performed production tests or 100% of the time. Are the Time 3 delayed posttest scores significantly higher than the pretest scores? In all nine of the nine studies that performed an aural interpretation test, the answer is yes. In all 14 of the 14 studies that performed production tests, the answer is yes. In how many studies does the delayed posttest score significantly decrease from the immediate posttest score? In only one of the nine studies that performed an aural interpretation task do we find

The long-term effects of Processing Instruction 

147

a statistically significant drop in score at the three week later delayed posttest (Cheng 2002). In only one of the 14 studies that performed production tests do we find a statistically significant drop in score between the immediate posttest and the delayed posttest for Processing Instruction (Toth 2006). Toth also found the same drop in his output group. All in all, we take these findings as overwhelming support that not only does Processing Instruction, and SI (Benati 2004; Farley 2004b) provide learners with immediate effects of instruction on interpretation and production, but that these dual effects are also durative. The research base has documented significant improvement in the overwhelming number of studies that have examined delayed effects. Now let us examine the degree of improvement. We present in Table 7.5 a summary of the improvement from T1 to T2 (immediate improvement) and again from T1 to T3 (sustained improvement). The overall results of the investigation are presented in column 3. These figures show us that in only two of the 25 analyses, did the immediate effects of instruction significantly diminish at the delayed posttest, that is, a situation in which T2 > T3. Cheng’s (2002) investigation of Spanish copula posts both findings, one for the interpretation scores in the overall analysis of the dual copula choice and one for the sentence production scores for the specific analysis of only one copula verb. In the other 23 analyses, T2 = T3, meaning that effects of instruction were found to endure for one week, two weeks, three weeks, 24 days and four weeks beyond the immediate effects of instruction in an overwhelming number of investigations. The figures presented in columns 7 and 8 show the relative degree of improvement from the pretest to the two posttests. This percentage varies according to the starting point of the learners on the particular pretest they took as well as with the end point they achieved on the posttest. For example, production scores for Processing Instruction were as low as .03/1 for Spanish anticausative clitic se (Toth 2006) and as high as 3.4/10 for Spanish copula (Cheng 2002). These scores show us different starting places. The percent of improvement that we have calculated attempts to demonstrate a relative gain across the different studies, linguistic items targeted, tasks performed, and scoring procedures used. The improvement documented on interpretation scores from T1 to T2 for Processing Instruction learners (column 7) varies from a low of 17% (Cheng 2002) to a high of 59% (Morgan-Short and Bowden 2006); the average improvement across the eight studies that use an aural interpretation test is 39%. That degree of improvement we would consider to be not only statistically significant but a “real” increase in how

148 

The short- and long-term effects of Processing Instruction

learners ­processed the targeted linguistic items. The improvement documented on production scores from T1 to T2 for Processing Instruction learners (column 7) varies from a low of 19% (Cheng 2002) to a high of 70% (Farley 2004b); the average improvement across the 10 different production tests administered was 45%. We again consider this number to reflect a “real” not just a statistical increase in how accurately learners produced the target forms immediately after receiving Processing Instruction instruction that did not include practice producing the target forms. The sustained improvement figures are given in column 8. The improvement documented on interpretation scores from T1 to T3 for Processing Instruction learners (column 8) varies from a low of 8% (Cheng 2002) to a high of 54% (Morgan-Short and Bowden 2006); the average sustained improvement across the 8 studies that use an aural interpretation test is 35%. The improvement documented on production scores from T1 to T3 for Processing Instruction learners varies from a low of 18% (Cheng 2002) to a high of 58% (Farley 2001b, 2004a, 2004b); the average sustained improvement across the 10 different production tests administered was 38%. This percentage reflects a real improvement in learners’ knowledge of how to produce the various target forms. That Cheng’s (2002) investigation of Spanish copula yielded the lowest degree of improvement on both interpretation and production scores deserves comment. Cheng investigated the use of two Spanish copular verbs (ser and estar) used with predicate adjectives. She isolated a linguistic context in which overall sentence semantics would contrast depending on the use of the copula. The copular ser is used with predicate adjectives in the passive construction with an overt or implied agent. The copula estar is used in this linguistic context to denote a resultant state. The intermediate level learners she investigated had previously studied forms and other uses of the verb ser as well as forms and other uses of the verb estar. They had also studied some contrastive uses of ser and estar. Cheng’s improvement scores range from 8%–21% for interpretation and 18%–29% for production. The knowledge base of this set of learners was very different than those Benati (2001, 2004) investigated. He examined beginning level learners who received their initial exposure to Italian future tense morphology. His improvement scores range from 43%–56% for interpretation and from 42%–51% for production. We find additional support for our argument that learners’ previous knowledge contributes to low improvement scores in the findings from Lee, Benati, Aguilar-Sánchez and McNulty (2007). In our tables, we present their findings on Spanish preterite–imperfect distinction separately from their findings on

Are there longer term effects for Processing Instruction? 

149

Spanish negative informal commands. In their presentation of the research they examined the same group of subjects learning both linguistic items such that Linguistic Item was a variable in their statistical analyses. They found a significant interaction between Linguistic Item and Time. The results showed that learners possessed significantly different levels of knowledge of the two items at the time of pretesting: 46% for preterite–imperfect distinction and 6% for negative informal commands. And, then, after treatment, the learners possessed different levels of knowledge of the two items but in a different direction. The scores for negative informal commands were significantly higher than those for preterite–imperfect distinction. In columns 7 and 8 of Table 7.5, we note improvement on preterite–imperfect distinction ranging from 5%–18% whereas improvement on negative informal commands ranges from 62%–79%. The differences in knowledge at the outset were due to previous study. Learners has studied the forms and uses of the preterite as well as forms and uses of the imperfect. The treatment presented the aspectual distinction between the two past tenses. This scenario is similar to the situation in Cheng (2002). Learners went into the treatments not with specific knowledge of the focus of the treatment but with other knowledge about the items in focus. We do not discount that different linguistic items may yield different types of results. That is, the semantics involved with copular verbs and aspectual distinction may be more difficult to teach and/or more difficult to learn than future tense morphology, subjunctive mood morphology or object pronouns. But at this point in time, linguistic item and learners’ potential supporting knowledge of the linguistic item are intertwined arguments. Are there longer term effects for Processing Instruction? In order for researchers to conduct longitudinal studies, they must have continued access to their participants. Continued access to participants can be a major issue in longitudinal research. Researchers who examine the effects of instructional interventions on the acquisition of particular linguistic forms must also be concerned about additional exposure to the target form over the period of investigation. For Processing Instruction we have two studies that have examined longer term effects, VanPatten and Fernández (2004) and Marsden (2006). We present a summary of the findings of these two studies in Table 7.6. VanPatten and Fernández (2004: 278) note that the number of eligible participants who took the prestest, sat for the intervention, and took the im-

150 

The short- and long-term effects of Processing Instruction

mediate posttest numbered 163. Their second posttest took place 8 months later. In the interim, they lost eligible participants. “Because our purpose was to observe the long time effect of a one-time treatment, three entire classes containing participants had to be eliminated. The instructors of these classes inadvertently reviewed and practised object pronouns [the targeted linguistic form] outside of the study (no other instructors did).” (VanPatten and Fernández 2004: 278). This event reduced the pool from 163 to 94. Of those 94 only 54 enrolled in a Spanish course the following semester. And then, due to participants withdrawing from the course or being absent on the day of testing, the final pool on whom they collected longitudinal data was 45. Only 28% of the eligible participants made it to the third posttest. Marsden (2006) examined the effects of instruction on 13–14 year old children enrolled in a secondary school. The amount of time required for the study was 25 weeks, with the experimental interventions going on for a period of seven weeks. She focussed her investigation on one class, dividing its members into two instructional treatments. The total number of participants was 27, but not all participants took all the posttests. In Table 7.6 we present a summary of the findings of the two studies that have examined longer term effect of Processing Instruction. Marsden (2006) performed delayed posttesting 14 weeks after finishing the instructional intervention and immediate posttesting. VanPatten and Fernández (2004) performed delayed posttesting 8 months after finishing the instructional intervention and immediate posttesting Marsden (2006) found that the effects of Processing Instruction were stable for a period of 14 weeks (or 3.5 months). She found that they were stable for the five measures she took: reading, listening, writing, speaking at sentence level, and speaking at discourse level. VanPatten and Fernández (2004) found that the results of Processing Instruction diminished after 8 months, but that learners still knew more than they did prior to instruction. The effects of Processing Instruction instruction faded but did not disappear. In Table 7.7 we present a summary of the improvement learners experienced in these two studies. The amount of improvement Marsden (2006) documents is modest, albeit statistically significant. In our previous discussion of Lee et al.’s (2007) results with Spanish aspectual distinction as well as Cheng’s (2002) results with Spanish copula choice we suggested that learners came to the task with more knowledge that the tests tested for. We suspect something similar with Marsden’s results. She notes that her subjects had received 180 hours of instruction in French prior to participating in her study. One of her targeted linguistic forms was present-tense verb morphology in

Conclusion 

151

all its forms (person/number). It is likely that the learners had some exposure to the present tense in their previous 180 hours of instruction. Additionally, Marsden is testing simultaneously for the effects of Processing Instruction on two forms, the present indicative and the present perfect. The combination of forms may have contributed to the modest improvement. Conclusion The number of studies documenting the durative effects of Processing Instruction are far greater than the two that document longer term effects. Processing Instruction has been found to be an effective intervention who effects endure one week, two weeks, three weeks, 24 days, and even 3.5 months after immediate posttesting. These findings apply to a wide variety of posttest: interpretation, form production, grammaticality judgment, form selection tests, reading, listening, writing, and speaking. The effects of Processing Instruction were found to diminish eight months after immediate posttesting but the effects were still felt. We conclude that Processing Instruction is effective instruction not only immediately, but also in the shorter and longer term.

61 intermediate PI vs. TI vs. C English natives T1 = pretest T2 = immediate posttest T3 = 1 week later T4 = 1 month later

49 intermediate PI vs. TI vs. C English natives T1 = pretest T2 = immediate posttest T3 = 1 week later T4 = 1 month later

Spanish directobject pronouns First Noun

VanPatten & Cadierno 1993b

Cadierno 1995 Spanish preterit tense Lexical Preference

80 intermediate PI vs. TI vs. C English natives T1 = pretest T2 = immediate posttest T3 = 1 week later T4 = 1 month later

Spanish directobject pronouns First Noun

VanPatten & Cadierno 1993a

Treatment

Linguistic Subjects: feature and number and Processing Principle L1

Study

PI > TI = C

PI = TI PI = C TI > C

interpretation (aural)

production (written)

PI > TI = C

PI = TI > C

production (written)

interpretation (aural)

PI > TI = C

overall: T1 < T2 = T3 = T4 PI: T1 < T2 = T3 = T4 TI: unanalyzed C: unanalyzed

T1: PI = TI = C* T2: PI = TI > C T3: PI = TI = C T4: P  I = TI PI > C TI = C

T1: PI = TI = C* T2: PI > TI PI > C T3: PI > TI C > TI T4: PI > TI; PI > C

overall: T1 < T2 = T3 = T4

overall: T1 < T2 = T3 = T4

Treatment results Time variable results

interpretation (aural)

Tests

Table 7.1.  Summary of the findings of the original PI research that includes two delayed posttests

152  The short- and long-term effects of Processing Instruction

PI = TI > C

overall: T1 < T2 = T3 = T4 PI: T1 < T2 = T3 = T4 TI: T1 < T2 = T3 = T4 C: unanalyzed

*These comparisons were made on gain scores not raw scores. T1 = Time 1 or pretest; T2 = Time 2 or immediate posttest; T3 = Time 3 or delayed posttest; T4 = Time 4 or second delayed posttest; PI = Processing Instruction; TI = Traditional (output) Instruction; C = Control group

production (written)

Appendix Appendix  153

Cadierno 1995 Spanish preterit tense

VanPatten & Cadierno 1993b Spanish direct object pronouns

VanPatten & Cadierno 1993a Spanish direct object pronouns Prod. PI = 8.9 TI = 8.5 C = 3.4 Int. PI = 9.4 TI = 4.6 C = 4.5 Prod. PI = 8.3 TI = 9.3 C = 4.2 Int. PI = 7.4 TI = 5.1 C = 5.0

Int. PI vs. TI Int. vs. C PI > TI PI = 2.5 =C TI = 2.6 C = 1.9

Prod. Prod. PI = TI PI = 1.8 PI = C TI = 3.6 TI > C C = 2.1

PI vs. TI Int. Int. vs. C PI > TI PI = 4.2 =C TI = 3.9 C = 4.3

Int. PI = 7.8 TI = 5.3 C = 4.3

Prod. PI = 8.0 TI = 7.9 C = 4.2

Int. PI = 32% TI = 12% C = na

Int. PI = 36% TI = 14% C = na

Prod. PI = 62% TI = 43% C = na

Prod. PI = 65% TI = 57% C = na Prod. PI = 8.3 TI = 8.3 C = 4.2 Int. PI = 7.8 TI = 5.1 C = 4.6

Int. PI = 50% TI = 13% C = na

Int. PI = 69% TI = 20% C = na

Prod. PI = 60% TI = 63% C = na

Int. PI = 55% TI = 21% C = na

Int. PI = 8.8 TI = 6.0 C = 4.2

Prod. PI = 67% TI = 58% C = na

Prod. PI = 8.1 TI = 8.1 C = 4.6

Prod. PI = 8.2 TI = 9.0 C = 3.4 Int. PI = 7.5 TI = 3.9 C = 5.6

Int. PI = 64% TI = 17% C = na

Int. PI = 7.4 TI = 3.9 C = 2.2

Int. PI = 7.2 TI = 3.4 C = 2.3

Int. PI = 36% TI = 12% C = na

Prod. PI = 65% TI = 47% C = na

Int. PI = 63% TI = 34% C = na

Prod. PI = 59% TI = 54% C = na

Int. PI = 57% TI = 26% C = na

Means scores Means scores Mean scores Improvement Improvement Improvement T2 = Posttest T3 = posttest T4 = posttest immediate intermediate sustained T1 vs. T2 T1 vs. T3 T1 vs. T4

Prod. Prod. PI = TI PI = 2.2 TI = 2.7 >C C = 1.6

Means scores T1 = Pretest Int. PI = 8.1 TI = 3.0 C = 1.4

Results

PI vs. TI Int. Int. vs. C PI > TI PI = 1.7 =C TI = 1.3 C = 1.1

Study and Treat­ linguistic fea- ment ture/language

Table 7.2.  Summary of improvement of PI research that includes three posttests

154  The short- and long-term effects of Processing Instruction

Prod. PI = 7.8 TI = 6.0 C = 3.6

Prod. PI = 7.7 TI = 6.0 C = 3.3

Prod. PI = 7.6 TI = 6.0 C = 3.2

Prod. PI = 54% TI = 42% C = na

Prod. PI = 53% TI = 42% C = na

Prod. PI = 52% TI = 42% C = na

Int: = interpretation task; Prod: = production tasks; T1 = Time 1 or pretest; T2 = Time 2 or immediate posttest; T3 = Time 3 or delayed posttest; T4 = Time 4 or second delayed posttest; PI = Processing Instruction; TI = Traditional (output) Instruction; C = Control group

Prod. Prod. PI = TI PI = 2.4 >C TI = 1.8 C = 2.4

Appendix  155

Italian future tense 39 beginners Lexical Preference English natives

Benati 2001

Subjects: number and L1

Linguistic feature and Processing Principle

Study

PI vs. TI vs. C T1 = pretest T2 = immediate post test T3 = three weeks later

Treatment

PI > TI > C

PI = TI > C

PI = TI > C

production (written)

production (oral)

Treatment results

interpretation (aural)

Tests

Table 7.3.  Summary of the findings of PI research that includes one-delayed posttest (T3)

T1: PI = TI = C T2: PI = TI > C T3: PI = TI > C

PI: T1 < T2 = T3 TI: T1 < T2 = T3 C: T1 = T2 = T3

T1: PI = TI = C T2: PI = TI > C T3: PI = TI > C

PI: T1 < T2 = T3 TI: T1 < T2 = T3 C: T1 = T2 = T3

T1: PI = TI = C T2: PI > TI > C T3: PI > TI > C

PI: T1 < T2 = T3 TI: T1 < T2 = T3 C: T1 = T2 = T3 PI-T2 > TI-T2

Time variable results

156  The short- and long-term effects of Processing Instruction

Spanish subjunctive in noun clauses Nonmeaningful forms

Spanish subjunctive in noun clauses Nonmeaningful forms

Farley 2001a

Farley 2001b 2004a

PI vs. MOI T1 = pretest T2 = immediate posttest T3 = 2 weeks later

PI vs. MOI T1 = pretest T2 = immediate posttest T3 = 2 weeks later

29 intermediate English natives

50 intermediate English natives

PI > MOI

PI = MOI

PI = MOI

PI = MOI

interpretation (aural)

production (written)

interpretation (aural)

production (written)

T1: PI = MOI T2: PI = MOI T3: PI = MOI

PI: T1 < T2 = T3 MOI: T1 < T2 = T3

T1: PI = MOI T2: PI = MOI T3: PI = MOI

PI: T1 < T2 = T3 MOI: T1 < T2 = T3

T1: PI = MOI T2: PI = MOI T3: PI = MOI

PI: T1 < T2 = T3 MOI: T1 < T2 = T3

T1: PI = MOI T2: PI > MOI T3: PI > MOI

PI: T1 < T2 = T3 MOI: T1 < T2 = T3

Appendix  157

Spanish copular 109 intermediate English natives verb choice ser vs. estar Lexical preference

Cheng 2002

Subjects: number and L1

Linguistic feature and Processing Principle

Study

Table 7.3.  (cont.)

PI vs. TI vs. C T1 = pretest T2 = immediate posttest T3 = 3 weeks later

Treatment

PI = TI > C

PI = TI > C production (written, sentences)

Treatment results

interpretation (aural)

Tests

PI: T1 < T2 T1 < T3 T3 = T2 TI: T1 < T2 T1 < T3 T3 = T2 C: T1< T2 T1 = T3 T2 = T3

T1: PI = TI = C T2: PI = TI PI > C TI = C T3: PI = TI PI = C TI > C

PI: T1 < T2 T1 < T3 T3 < T2 TI: T1 < T2 T1 < T3 T2 = T3 C: T1 = T2 = T3

Time variable results

158  The short- and long-term effects of Processing Instruction

Cheng 2002

Spanish copula 109 intermediate estar - only English natives Lexical preference

PI vs. TI vs. C T1 = pretest T2 = immediate posttest T3 = 3 weeks later

PI = TI > C

PI = TI > C

production (written, discourse)

interpretation (aural)

T1: PI = TI = C T2: PI = TI PI > C TI = C T3: PI = TI = C

PI: T1 < T2 T1 < T3 T2 = T3 TI: T1 < T2 T1 < T3 T2 = T3 C: T1 < T2 T1 < T3 T2 = T3

T1: PI = TI = C T2: PI = TI > C T3: PI = TI > C

PI: T1 < T2 = T3 TI: T1 < T2 = T3 C: T1 = T2 = T3

T1: PI = TI = C T2: PI = TI PI > C TI = C T3: PI = TI PI > C TI > C

Appendix  159

Study

Table 7.3.  (cont.)

Linguistic feature and Processing Principle

Subjects: number and L1

Treatment

Treatment results

PI = TI = C

PI = TI > C

Tests

production (written, sentences)

production (written, discourse)

T1: PI = TI = C T2: PI = TI > C T3: PI = TI = C

PI: T1 < T2 = T3 TI: T1 < T2 = T3 C: T1 < T2 = T3

T1: PI = TI = C T2: PI = TI PI > C TI > C T3: PI = TI PI > C TI = C

PI: T1 < T2 = T3 TI: T1 < T2 T1 < T3 T3 < T2 C: T1 = T2 = T3

Time variable results

160  The short- and long-term effects of Processing Instruction

PI vs. TI vs. C T1 = pretest T2 = immediate posttest T3 = 3 weeks later PI vs. SI T1 = pretest T2 = immediate posttest T3 = 2 weeks later

83 intermediate Spanish copula contextual uses of English natives estar Lexical preference

54 intermediate English natives

Cheng 2004

Farley 2004b

Spanish subjunctive in noun clauses Nonmeaningful forms

PI vs. SI vs. EI T1 = pretest T2 = immediate posttest T3 = 4 weeks later

Italian future tense 38 beginners morphology English natives Lexical Preference

Benati 2004a

PI = SI > EI

production (written)

interpretation of subjunctive forms (aural)

PI > SI

production PI = TI > C (written discourse)

PI = SI > EI

interpretation (aural)

T1: PI = SI T2: PI > SI T3: PI > SI

PI: T1 < T2 = T3 SI: T1 < T2 = T3

T1: PI = TI = C T2: PI = TI > C T3: PI = TI > C

PI: T1 < T2 = T3 TI: T1 < T2 = T3 C: T1 = T2 = T3

T1: PI = SI = EI T2: PI = SI > EI T3: PI = SI > EI

T1: PI = SI = EI T2: PI = SI > EI T3: PI = SI > EI PI: T1 < T2 = T3 SI: T1 < T2 = T3 EI: T1 < T2 = T3

PI: T1 < T2 = T3 SI: T1 < T2 = T3 EI: T1 < T2 = T3

Appendix  161

Linguistic feature and Processing Principle

Spanish direct object pronouns First noun

Study

Morgan-Short & Bowden 2006

Table 7.3.  (cont.)

45 beginners native language not specified, presumably English

Subjects: number and L1

PI vs. MOI vs. C T1: pretest T2: immediate posttest T3: 1 week later

Treatment

PI > SI

PI = SI

production of subjunctive forms (written)

production of indicative forms (written)

PI = MOI > C

PI > SI

interpretation of indicative forms (aural)

interpretation (aural)

Treatment results

Tests

T1: PI = MOI = C T2: PI = MOI > C T3: PI > MOI > C

PI: T1 < T2 = T3 MOI: T1 < T2 = T3

T1: PI = SI T2: PI = SI T3: PI = SI

PI: T1 = T2 = T3 SI: T1 = T2 = T3

T1: PI = SI T2: PI > SI T3: PI > SI

PI: T1 < T2 = T3 SI: T1 < T2 = T3

T1: PI = SI T2: PI > SI T3: PI > SI

PI: T1 < T2 = T3 SI: T1 = T2 = T3

Time variable results

162  The short- and long-term effects of Processing Instruction

Toth 2006

Spanish anti­ causative clitic se 80 beginners English natives

production (written)

grammaticality PI vs. CO vs. C judgment T 1 = pretest T2 = immediate posttest T3 = 24 days later

production (written)

CO > PI > C

CO = PI > C

PI = MOI > C

T1: PI = CO = C T2: CO > PI > C T3: CO > PI > C

PI: T1 < T2 T1 < T3 T3 < T2 CO: T1 < T2 T1 < T3 T3 < T2 C: T1 = T2 = T3

T1: PI = CO = C T2: PI = CO > C T3: PI = CO > C

PI: T1 < T2 = T3 CO: T1 < T2 = T3 C: T1 = T2 = T3

T1: PI = MOI = C T2: MOI > PI > C T3: PI = MOI > C

PI: T1 < T2 = T3 MOI: T1 < T2 = T3

Appendix  163

PI classroom vs. form selection in PI computer vs. context PI  ybrid T1 = pretest T2 = immediate posttest T3 = 1 week later

PI classroom vs. form selection in PI computer vs. context PI  ybrid T1 = pretest T2 = immediate posttest T3 = 1 week later

Spanish negative 25 beginners informal English natives commands Lexical preference

Lee, Benati, Aguilar-Sanchez & McNulty 2007

Tests

Spanish preterit/ 25 beginners imperfect English natives aspectual distinction Lexical preference

Treatment

Lee, Benati, Aguilar-Sanchez & McNulty 2007

Subjects: number and L1

Linguistic feature and Processing Principle

Study

Table 7.3.  (cont.)

PI classroom = PI computer = PI hybrid

PI classroom = PI computer = PI hybrid

Treatment results

T1: PI cl = PI co = PI hy T2: PI cl = PI co = PI hy T3: PI cl = PI co = PI hy

PI cl: T1 < T2 = T3 PI co: T1 < T2 = T3 PI hy: T1 < T2 = T3

T1: PI cl = PI co = PI hy T2: PI cl = PI co = PI hy T3: PI cl = PI co = PI hy

PI cl: T1 < T2 = T3 PI co: T1 < T2 = T3 PI hy: T1 < T2 = T3

Time variable results

164  The short- and long-term effects of Processing Instruction

Japanese past 26 beginners tense marker in Italian natives sentence final position Lexical Preference Sentence Location production (written)

SI vs. SI-E vs. C interpretation T1 = pretest (aural) T2 = immediate posttest T3 = 1 week later SI = SI-E > C

SI = SI-E > C

T1: SI = SI-E = C T2: SI = SI-E > C T3: SI = SI-E > C

SI: T1 < T2 = T3 SI-E: T1 < T2 = T3 C: T1 = T2 = T3

T1: SI = SI-E = C T2: SI = SI-E > C T3: SI = SI-E > C

SI: T1 < T2 = T3 SI-E: T1 < T2 = T3 C: T1 = T2 = T3

Significant improvement from T1 to T2

9/9 14/14 2/2 1/1

Time Task

aural interpretation oral or written production form selection grammaticality judgment

9/9 14/14 2/2 1/1

Significant improvement from T1 to T3

8/9 13/14 2/2 1/1

Sustained improvement from T2 to T3

Table 7.4.  Overview of the durative effects of PI and SI across the research base

1/9 1/14 0/2 0/1

Significant decrease from T2 to T3

T1 = Time 1 or pretest; T2 = Time 2 or immediate posttest; T3 = Time 3 or delayed posttest; PI = Processing Instruction; SI = Structured Input; EI = Explicit Information; TI = Traditional (output) Instruction; MOI = Meaning-based Output Instruction; CO = Communicative Output tasks; C = Control group

Lee & Benati 2007

Appendix  165

PI vs. MOI T3 = 2 weeks

Farley 2001b 2004a Spanish subjunctive

Int. PI = MOI T1 < T2 = T3

Int. (9) PI = 3.2 MOI = 3.6

Int. PI = 6.5 MOI = 7.0

Prod.: PI = 6.2 MOI = 6.3

Prod.: (12) PI = .88 MOI = .33

Prod.: PI = MOI T1 < T2 = T3

PI vs. MOI T3 = 2 weeks

Farley 2001a Spanish subjunctive Int. PI = 6.8 MOI = 5.5

Prod. oral: PI = 7.0 TI = 7.2 C = 2.2

Prod. oral: PI = 2.7 TI = 2.8 C = 2.0

Prod. oral: PI = TI > C T1 < T2 = T3) Int. (12) PI = 3.2 MOI = 2.8

Prod. written: PI = 7.2 TI = 7.3 C = 2.7

Prod. written: PI = 2.3 TI = 2.3 C = 2.3

Prod. written: PI = TI > C T1 < T2 = T3

Int. PI > MOI T1 < T2 = T3

Int. PI = 8.1 TI = 5.5 C = 3.7

Int. PI = 3.7 TI = 3.8 C = 3.6

Int. PI > TI > C T1 < T2 = T3

PI vs. TI vs. C T3 = 3 weeks

Means scores T2 Posttest

Benati 2001# Italian future tense

Means scores T1 = Pretest

Results

Study and Treatment linguistic feature/ language

Int. PI = 6.8 MOI = 6.7

Prod.: PI = 6.2 MOI = 5.3

Int. PI = 6.7 MOI = 3.9

Int. 9 PI = 37% MOI = 38%

Prod.: PI = 44% MOI = 50%

Int. PI = 30% MOI = 23%

Prod. oral: PI = 43% TI = 44% C = na

Int. PI = 40% MOI = 34%

Prod.: PI = 44% MOI = 41%

Int. PI = 29% MOI = 9%

Prod. oral: PI = 42% TI = 43% C = na

Prod. written: PI = 47% TI = 49% C = na

Prod. written: PI = 49% TI = 50% C = na

Prod. written: PI = 7.0 TI = 7.2 C = 2.5 Prod. oral: PI = 6.9 TI = 7.1 C = 2.0

Int. PI = 43% TI = 16% C = na

Improvement sustained T1 vs. T3

Int. PI = 44% TI = 17% C = na

Improvement immediate T1 vs. T2

Int. PI = 8.0 TI = 5.4 C = 3.5

Means scores T3 posttest

Table 7.5.  Summary of the improvement from pretest to posttests of studies that include one delayed posttest

166  The short- and long-term effects of Processing Instruction

PI vs. TI vs. C T3 = 3 weeks

Cheng 2002 # PI vs. TI vs. C Spanish copula, T3 = 3 weeks estar only

Cheng 2002 # Spanish copula choice ser vs. estar

Int. PI = 4.2 TI = 3.6 C = 3.2

Prod. (discourse): Raw scores not supplied. Arcsine transformed scores were presented, which do not lend themselves to a discussion of improvement.

Prod. (discourse): PI = TI > C T1 < T2 = T3

Int. PI = 2.5 TI = 2.4 C = 2.3

Prod. (sentences): PI = 6.3 TI = 6.5 C = 5.0

Prod. (sentences): PI = 3.4 TI = 3.8 C = 3.6

Prod. (sentences): PI = TI > C T1 < T2 = T3

Int. PI = TI > C T1 < T2 = T3

Int. PI = 6.7 TI = 6.0 C = 5.4

Int. PI = 4.6 TI = 4.5 C = 4.2

Prod. written: (6) Prod. written: PI = .17 PI = 4.3 MOI = .33 MOI =4.2

Int. PI = TI > C T1 < T2 > T3

Prod. written: PI = MOI T1 < T2 = T3

Int. PI = 17% TI = 12% C = na

Int. PI = 19% TI = 8% C = na

Prod. (sentences): PI = 25% TI = 22% C = na Prod. (sentences): PI = 29% TI = 27% C = na

Prod. (sentences): PI = 5.9 TI = 6.0 C = 4.4

Int. PI = 3.4 TI = 3.2 C = 3.1

Int. PI = 8% TI = 10% C = na

Int. PI = 21% TI = 15% C = na

Prod. written: 6 Prod. written: PI = 69% PI = 58% MOI = 64% MOI = 58%

Int. PI = 5.4 TI = 5.5 C = 4.6

Prod. written: PI = 3.6 MOI = 3.8

Appendix  167

Benati 2004a Italian future tense

PI vs. SI vs. EI T3 = 4 weeks

Study and Treatment linguistic feature/ language

Table 7.5.  (cont.)

Prod. (sentences) PI = 1.5 TI = 2.0 C = 2.1

Prod. (sentences): PI = TI = C T1 < T2 > T3

Int. PI = SI > EI T1 < T2 = T3

Int. PI = 2.7 SI = 2.1 EI = 2.6

Int. PI = 8.3 SI = 7.8 EI = 4.6

Prod. (sentences) PI = 3.4 TI = 4.1 C = 2.9

Means scores T2 Posttest

Prod. (discourse):Prod. (discourse): PI = TI > C Raw scores not T1 < T2 = T3 supplied. Arcsine transformed scores were presented, which do not lend themselves to a discussion of improvement.

Means scores T1 = Pretest

Results

Int. PI = 7.7 SI = 7.5 EI = 3.8

Prod. (sentences) PI = 3.3 TI = 3.5 C = 2.5

Means scores T3 posttest

Int. PI = 56% SI = 57% EI = 20%

Prod. (sentences) PI = 19% TI = 21% C = na

Improvement immediate T1 vs. T2

Int. PI = 50% SI = 54% EI = 12%

Prod. (sentences) PI = 18% TI = 15% C = na

Improvement sustained T1 vs. T3

168  The short- and long-term effects of Processing Instruction

Prod. PI = 7.3 SI = 6.5 EI = 3.1

Prod. PI = 6.5 SI = 4.8 EI = 2.5

Prod. PI = 51% SI = 46% EI = 13%

Prod. PI = 43% SI = 29% EI = 7%

Int. indicative PI = 1.4 SI = 1.6

Int. indicative PI = 2.1 SI = 1.5

Int. indicative PI = 2.1 SI = 1.5

Int. indicative PI = 8% SI = -1%

Int. indicative PI = 8% SI = -1%

Prod. PI = 2% MOI = .6% C = 3%

Prod. PI = MOI > C T1 < T2 = T3

Int. PI = 59% MOI = 63% C =na Prod. PI = 29% MOI = 69% C = na

Int. PI = 62% MOI = 46% C = 36% Prod. PI = 26% MOI = 32% C = 16%

Int. PI = 67% MOI = 71% C = 14% Prod. PI = 31% MOI = 70% C = 13%

Prod. PI = 24% MOI = 31% C=

Int. PI = 54% MOI = 38% C = na

Prod. indicative Prod. indicative Prod. indicative Prod. indicative Prod. indicative Prod. indicative PI = SI PI = 2.5 PI = 2.4 PI = 2.5 PI = -2% PI = 0% T1= T2 = T3 SI = 2.8 SI = 2.6 SI = 2.7 SI = -3% SI = -2%

Prod. subjunctive Prod. subjunctive Prod. subjunctive Prod. subjunctive Prod. subjunctive Prod. subjunctive PI = .22 PI = 4.4 PI = 3.7 PI = 70% PI = 58% PI > SI T1 < T2 = T3 SI = .48 SI = 2.4 SI = 2.3 SI = 32% SI = 30%

Int. indicative PI > SI T1 < T2 = T3

Int. subjunctive Int. subjunctive Int. subjunctive Int. subjunctive Int. subjunctive Int. subjunctive PI > SI PI = 3.1 PI = 6.4 PI = 6.7 PI = 37% PI = 40% T1 < T2 = T3 SI = 3.9 SI = 4.4 SI = 5.0 SI = 6% SI = 12%

Int. PI = 8% MOI = 8% C = 10%

PI vs. SI T3 = 2 weeks

Prod. PI = 2.2 SI = 1.9 EI = 1.8

Morgan-Short & PI vs. MOI vs. C Int. T3 = 1 week PI = MOI > C Bowden 2006 T1 < T2 = T3 Spanish direct object pronouns

Farley 2004b Spanish subjunctive

Prod. PI = SI > EI T1 < T2 = T3

Appendix  169

form selection: PI cl = PI co = PI hy T1 < T2 = T3

form selection: PI cl = PI co = PI hy T1 < T2 = T3

Lee, Benati, Aguilar-Sánchez & McNulty 2007 Spanish nega­ tive informal commands

Prod. CO > PI > C T1 < T2 = T3

PI vs. CO vs. C gram. judg.: T3 = 24 days PI = CO > C T1 < T2 = T3

Results

PI classroom vs. Lee, Benati, Aguilar-Sánchez PI computer vs. & McNulty 2007 PI hybrid Spanish preterit– T3 = 1 week imperfect distinction

Toth 2006 Spanish anticausative clitic se

Study and Treatment linguistic feature/ language

Table 7.5.  (cont.)

form selection: PI cl = 68% PI co = 81% PI hy = 85%

form selection: PI cl = 69% PI co = 78% PI hy = 65%

form selection: PI cl = 60% PI co = 64% PI hy = 54%

form selection: PI cl = 60% PI co = 57% PI hy = 51%

form selection: PI cl = 46% PI co = 46% PI hy = 46%

form selection: PI cl = 6% PI co = 6% PI hy = 6%

Prod. PI = .24 CO = .50 C = .11

gram. judg.: PI = 1.4 CO = 1.9 C = .8

Means scores T3 posttest

Prod. PI = .47 CO = .63 C = .00

gram. judg.: PI = 1.6 CO = 2.2 C = .6

Means scores T2 Posttest

Prod. PI = .03 CO = .04 C = .05

gram. judg.: PI = .3 CO = .9 C = .7

Means scores T1 = Pretest

form selection: PI cl = 63% PI co = 72% PI hy = 59%

form selection: PI cl = 14% PI co = 18% PI hy = 8% form selection: PI cl = 14% PI co = 9% PI hy = 5%

form selection: PI cl = 62% PI co = 75% PI hy = 79%

Prod. PI = 21% CO = 46% C = na

gram. judg.: PI = 11% CO = 13% C = na

Improvement sustained T1 vs. T3

Prod. PI = 44% CO = 59% C = na

gram. judg.: PI = 16% CO = 16% C = na

Improvement immediate T1 vs. T2

170  The short- and long-term effects of Processing Instruction

int: SI = .67 SI-E = .90 C = .57 Prod. SI = .33 SI-E = .50 C = .43

SI vs. SI-E vs. C int: T3 = 1 week SI = SI-E > C T1 < T2 = T3

Prod. SI = SI-E > C T1 < T2 = T3

Prod. SI = 4.6 SI-E = 5.3 C = .57

int: SI = 5.8 SI-E = 6.0 C = .71 Prod. SI = 4.1 SI-E = 4.7 C = .29

int: SI = 5.0 SI-E = 5.2 C = .43 Prod. SI = 43% SI-E = 48% C = na

int: SI = 51% SI-E = 51% C = na Prod. SI = 38% SI-E = 42% C = na

int: SI = 43% SI-E = 43% C = na

# The means scores were read from a graph, rather then a table, and are, therefore, approximate. Int.: interpretation tasks; Prod. = production tasks; gram. judg.: = grammaticality judgment task; T1 = Time 1 or pretest; T2 = Time 2 or immediate posttest; T3 = Time 3 or delayed posttest; PI = Processing Instruction; SI = Structured Input; SI-E = Structured input with textual/ aural enhancements; EI = Explicit Information; TI = Traditional (output) Instruction; MOI = Meaning-based Output Instruction; CO = Communicative Output tasks; C = Control group

Lee & Benati 2007

Appendix  171

27 English natives French presentand perfect-tense after 180 hrs. of verb inflections French instruction 13–14 yr olds

Tests

PI > EnI PI > EnI PI > EnI PI > EnI

reading writing (sentences + narration combined) speaking (sentences only) speaking (discourse: narrative + guided conversation)

PI: T1 < T2 = T3 EnI: T1 = T2 = T3

PI: T1 < T2 = T3 EnI: T1 = T2 = T3

PI: T1 < T2 = T3 EnI: T1 = T2 = T3

PI: T1 < T2 = T3 EnI: T1 = T2 = T3

PI: T1 < T2 = T3 EnI: T1 = T2 = T3

T1 < T2 T1 < T3 T3 < T2

na

PI > EnI

T1 < T2 T1 < T3 T3 < T2

Time variable results

na

Treatment results

listening PI vs. EnI T1 = pretest T2 = immediate posttest T3 = 14 weeks later

PI interpretation T1 = pretest (aural) T2 = immediate posttest production T3 = 8 months later (written)

Treatment

T1 = Time 1 or pretest; T2 = Time 2 or immediate posttest; T3 = Time 3 or delayed posttest. PI = Processing Instruction; EnI = Enriched Input

Marsden 2006

45 intermediate English natives adults

Linguistic feature Subjects: number and processing and L1 principle

VanPatten & Spanish directFernández 2004 object pronouns

Study

Table 7.6.  Summary of the findings of PI research that included longer term effects

172  The short- and long-term effects of Processing Instruction

PI T3 = 8 months

PI vs. EnI T3 = 14 weeks

VanPatten & Fernández 2004 Spanish direct-object pronouns

Marsden 2006 French present- and perfect-tense verb inflections

PI: 9% EnI: −4%

PI: 23% EnI: −7% PI: 28% EnI: 18%

PI: 42% EnI: 15% PI: 19% EnI: 22%

speaking (discourse: narrative + guided conversation)

PI: 6% EnI: −1% PI: 8% EnI: −1% PI: 21% EnI: 14% PI: 23% EnI: 14%

PI: 15% EnI: 15%

speaking (sentences)

PI: 10% EnI: 0% PI: 7% EnI: 3%

PI: 7% EnI: 10%

writing (sentences + narration combined)

PI: 17% EnI: 10%

PI: 37% EnI: 29%

PI: 25% EnI: 27%

reading PI: 14% EnI: 13%

PI: 49% EnI: 42%

PI: 49% EnI: 44%

PI: 40% EnI: 42%

listening

PI: 9% EnI: 1%

prod: PI = 24%

PI: 12% EnI: 2%

prod: PI = 49%

int: PI = 14%

PI: 34% EnI: 28%

prod: T3 = 3.4

int: PI = 42%

PI: 9% EnI: 0%

prod: T2 = 5.9

prod: T1 = 1.0

prod: T1 < T2 T1 < T3 T3 < T2

int: T3 = 3.3

PI: 9% EnI: 2%

int: T2 = 6.1

int: T1 = 1.9

Means scores Means scores Means scores Improvement Improvement T1 = Pretest T2 Posttest T3 posttest immediate sustained T1 vs. T2 T1 vs. T3

int: T1 < T2 T1 < T3 T3 < T2

Results

T1 = Time 1 or pretest; T2 = Time 2 or immediate posttest; T3 = Time 3 or delayed posttest; PI = Processing Instruction; EnI = Enriched Input

Treatment

Study and linguistic feature/language

Table 7.7.  A summary of the improvement made in PI research that included a longitudinal measure

Appendix  173

Measuring the effects of Processing Instruction

Chapter 8 How have the effects of Processing Instruction been measured?

In the previous chapters we have reviewed studies which have investigated the relative effects of Processing Instruction when compared to traditional instruction and other output-based approaches to grammar teaching. This growing body of research measuring the effects of Processing Instruction has given this approach significant support. The positive effects of Processing Instruction have been found for a variety of romance (Spanish, French, Italian) and non-romance languages (German, English and Japanese), on a variety of morphological, syntactic and semantic linguistic items and among learners from a different first language (English, Italian, Greek, Chinese and Japanese). Previous research has focused on five important issues. A body of research has compared the effects of Processing Instruction on to those of other types of instruction. This work affirms the superiority of Processing Instruction over other types of output-based instruction (see chapter three). A second large body of research has examined the roles of explicit explanation and structured input practices in order to determine the source(s) of the effects of Processing Instruction on language development (see chapter two). This work affirms the importance of structured input activities. A third line of research has measured the effects of Processing Instruction when delivered via computer terminals (see chapter four). This work affirms that Processing Instruction is effective no matter the way of delivery. A fourth line of investigation has examined whether learners improve if they receive enhanced structured input activities (see chapter five). The work affirms the main role for structured input activities as the causative factor in Processing Instruction. A fifth line of research has investigated the secondary and cumulative effects of Processing Instruction (see chapter six). This work affirms that learners who have been exposed to a processing strategy do pick up a second strategy more quickly and efficiently than they pick up the first one. A number of conclusions can be drawn regarding the type and mode of assessment tasks in the Processing Instruction studies reviewed in the previous chapters:

Discourse-level production tasks Discourse-level production tasks 

175

1. All Processing Instruction studies provide evidence that learners who received this type of instruction performed significantly better on interpretation sentence-level tasks than learners receiving other type of instruction (e.g. traditional instruction, output-based instruction); 2. All Processing Instruction studies provide evidence that Processing Instruction can cause equal improvement compared to output-based approaches in learner’s performance in different sentence-level production tasks (oral and written mode);

Two questions have been raised:

1. Can we measure the effects of Processing Instruction on discourse-level production tasks? 2. Can we measure the effects of Processing Instruction discourse-level interpretation tasks? What are the effects of Processing Instruction on discourse-level ­production tasks? DeKeyser, Salaberry, Robinson, and Harrington (2002) argued that any true system wide effects will be revealed more clearly in the discursive-level tasks and not sentence-level tasks. Norris and Ortega (2000) underscored the fact that most research investigating the effectiveness of L2 instruction has used, to date, measures that require the application of explicit declarative knowledge under control conditions and not the use of spontaneous, fluent and contextualised language. As previously argued, in Processing Instruction studies, the effects of Processing Instruction, traditional instruction, meaning output-based instruction and other approaches to grammar instruction have been measured on a variety of tasks and on different modes. However, only few studies (see Table 8.1 and 8.2) have measured the effects of instruction on discourse-level tasks. VanPatten and Sanz (1995) set out to investigate whether the effects of Processing Instruction observed in on the sentence-level tasks (interpretation and production) on two previous studies (VanPatten and Cadierno 1993; Cadierno 1995) could be obtained on discourse-level production tasks. The subjects pool consisted of four classes of forty-four students of Spanish in their third year of a university programme. The subjects were assigned to two

176 

Measuring the effects of Processing Instruction

Processing Instruction groups and two control groups. The focus of instruction was the same as in VanPatten and Cadierno’s study (1993): preverbal objects pronoun in Spanish. The effects of instruction were measured on three different tasks: 1. the same sentence-level tasks used by VanPatten and Cadierno (1993); 2. a structured question-answer interview; 3. a video narration task. The data collection procedure was very similar to the other empirical research reviewed in previous chapters and consisted of two days of instruction. The findings of VanPatten and Sanz’s study showed that Processing Instruction is still effective even when measured on less controlled and discourse type of tasks. The mean scores of the two groups on the pre-test for the interpretation sentence-level task were similar at about 5.7%. On the post-test interpretation task the mean scores of the Processing Instruction group were much higher than the control group showed clearly an improvement from pre-test to posttest (5.4% for the Processing Instruction group and no improvement for the control group). The mean scores of the two groups on the pre-tests for both written and oral sentence completion tasks were also similar (see Table 8.2). However the Processing Instruction group improved in both tasks from pretests to post-tests (6% in the written sentence-level completion task and 10% in the oral sentence-level completion task) whereas the control group made no improvement. On the pre-tests of the written and the oral interview tasks the two groups obtained similar results (see Table 8.2). However, even in this case the performance of the Processing Instruction group improved in both tasks from pre-tests to post-tests (3.5% in the written sentence-level completion task and 2.6% in the oral sentence-level completion task) whereas the control group made no improvement. The mean scores of the two groups on the pre-test for the production video narration test were similar (see Table 8.2). On the post-tests for both the oral and written mode of this task the Processing Instruction group made some minimal gains (9% for the written and 3% for the oral) whereas the control group made no gains. Although the results of this study seems to emphasise the fact that altering the processing strategies used by L2 learners when they are processing input leads to a change in knowledge which is available for use in different types and modes of production tasks, the effects of Processing Instruc-

Discourse-level production tasks 

177

tion were more significant in more controlled oral tasks (completion task) rather than less controlled ones (video narration task). This study provided further evidence on the positive effects of Processing Instruction in syntax (object pronouns an word order) and the fact that the effects are observable in more communicative and discourse-oriented types of tasks rather than only ­sentence-level tasks. A second study on production discourse-level tasks is the one carried out by Cheng (2004) and reviewed in chapter three. Cheng measured the effects of Processing Instruction on a different kind of linguistic item of the Spanish linguistic system: namely ser and estar. Her study was conducted to find evidence of the effects of Processing Instruction on the acquisition of two copular verbs in Spanish (ser and estar). Processing Instruction was related to two of VanPatten’s principles (1996: 14): P1(a). Learners process content words in the input before anything else. P1(c). Learners prefer processing ‘more meaningful’ morphology before ‘less’ or ‘non meaningful morphology’. The second principle was particularly relevant for this study as copular verbs in Spanish are of low communicative value for L2 learners and redundant features of Spanish. The two main research questions addressed by this study were: 1. whether Processing Instruction and traditional instruction had the same effects on the acquisition of ser and estar; 2. whether the effects are short or long-lasting. Eighty-eight subjects studying Spanish in their third year of University college level course were involved in this study. Cheng used a very similar research design to that of VanPatten and Cadierno (1993a) comparing three groups in a pre-post-test format including three type of tasks: 1. interpretation 2. sentence completion 3. guided composition The results of this further study on the effects of Processing Instruction showed that students receiving Processing Instruction outperformed those receiving no instruction and traditional instruction. Once more the effects of

178 

Measuring the effects of Processing Instruction

Processing Instruction are observable on output tasks at both sentence and discourse-level, whereas the effects of the traditional type of instruction are not observable on interpretation tasks. Sanz (2004) investigated the effectiveness of Processing Instruction when implemented with the use of CALL. One of the purpose of this study was to address the role of explicit and implicit feedback in CALL. Twenty-six students enrolled in a Spanish language programme at intermediate level participated in this study. They were assigned to two groups receiving Processing Instruction via CALL with + and – feedback (the groups received no explicit information). An interpretation and production tasks at sentence level were used. In addition to that, a production discourse-level task (video retelling) was also used. The mean scores of the two groups on the pre-test for the interpretation sentence-level task were similar at about 26%. On the post-test interpretation task the mean scores of the structure input practice group + feedback and the structured input practice – feedback were similar and between 62% and 66% showing a clear improvement from pre-test to post-test of 36% to 40%. The mean scores of the two groups on the pre-tests for the sentence completion tasks were also similar (see Table 8.2) at about 42%. Again the two groups made an improvement from pre-tests to post-tests (structure input practice group + feedback = 41% and the structured input practice – feedback = 24%). The mean scores of the two groups on the pre-test for the production retelling task were similar (ranged between 22% to 31%). On the post-tests both groups made some gains (structure input practice group + feedback = 26% and the structured input practice – feedback = 24%). The results showed that both groups increased in their ability to comprehend and produce correct Spanish object pronouns. Structured input practice was responsible for the improvements of both groups in all tasks no matter whether the feedback received was implicit or explicit. VanPatten and Sanz (1995), Cheng (2004) and Sanz (2004) have demonstrated that Processing Instruction is effective not only at the sentence level but at the discourse level production tasks. What are the effects of Processing Instruction discourse-level ­interpretation tasks? Lee (2004) has argued that Processing Instruction has been investigated in various linguistics features affected by one or a combinations of processing

Discourse-level interpretation tasks Discourse-level interpretation tasks 

179

problems and in different languages. These effects have been measured on interpretation and production sentence-level tasks and production discourselevel task. As he has suggested (Lee 2004: 319) “because Processing Instruction affects discourse-level production, I can hypothesize that it will also affect discourse-level interpretation”. In an recent experiment which is still in progress Hikima (forthcoming) has measured the effects of Processing Instruction on discourse-level interpretation tasks. Hikima (forthcoming) has measured the relative effects of Processing Instruction in the acquisition of Japanese passive forms. Participants were English native speakers learning Japanese. The subjects pool consisted of seven learners. A one-shot experimental design was used with only one group Processing Instruction on passive form of Japanese. A pre-test and a post test procedure was used. The passive form in Japanese was chosen for two main reasons. The first reason is that a combination of principles might affect/delay the acquisition of this form (First Noun Principle, Sentence ­Location Principle and Primacy-of-Content-Words Principle). The second reason is that the passive is one of the most difficult grammatical items in Japanese. The Processing Instruction packets contained explicit information and structured input practice (referential and affective). The explicit information component provided learners with linguistics information about the target feature and the processing problems learners might encounter when processing this form. The assessment tasks consisted of sentence and discourse level interpretation tests and sentence and discourse level production tests. The sentence-level tasks were very similar to the ones used in previous studies reviewed in this book. The discourse level interpretation task consisted of a dialogue and seven questions. In the dialogue, two people talked about a TV program. Subjects listened to the dialogue and had to determine who did what to whom, and then tick the correct answer. If subjects were not sure they had to choose “I’m not sure”. There is no repetition of the dialogue so that the test would measure whether or not the subject comprehend the target form and meaning in natural conversation. The discourse level production test consisted of creating a story with a sequence of six story pictures. In the discourse level production test, subjects looked at the sequence of six pictures and had to write a story with Japanese letter or Roomaji. The mean scores of the Processing Instruction group in the interpretation (improvement = 6%) and production (improvement = 2%) sentence level tasks showed that Processing Instruction made an improvement from pre-test

180 

Measuring the effects of Processing Instruction

to post-test. The mean scores from pre-test to post-tests of the Processing Instruction group showed again an improvement in the discourse-level interpretation task (6%) and in the discourse-level production task (4%). The preliminary results of this study clearly indicate that the Processing Instruction group makes measurable gains not only in the interpretation sentence-level task but also in interpretation discouse-level task confirming Lee’s hypothesis (2004).

Spanish objectpronouns

Hikima forthc. Japanese passive forms

Sanz 2004

Spanish copula

Spanish objectpronouns

VanPatten & Sanz 1995

Cheng 2004

Linguistic feature/ language

Study

Table 8.1.

Intermediate English native

First noun Intermediate principle, English native Sentence location principle and Primacy of content words principle

First Noun Principle

7

28

83

Beginners Primacy of meaning Principle English native Preference for Nonredundancy

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate

Pre-test Post-tests Immediate & delayed

Pre-test Post-tests Immediate

Number Design

44

Subjects/L1

Intermediate English native`

First Noun Principle

Processing principle

Resutlts

Interpretation (Sentence and Discourse) Production (Sentence and Discourse)

Interpretation (aural) Production Sentence Production video retelling

Interpretation (aural) Production Sentence and discourse level

Int. = PI improves Pro. = PI improves

Int. = SIA (-E –F) = (-E + F) Pro. = SIA (-E –F) = (-E + F)

Int. = PI > (TI = C) Prod. = (PI = TI) >C

Interpretation Int. = PI>C (aural) Prod. = PI > C Production Sentence and discourse level Oral and Written mode

Tests

Discourse-level interpretation tasks  181

Linguistic feature/ language

Spanish object pronouns

Study

VanPatten & Sanz 1995

Table 8.2. 

Prod. (sentence-oral) PI = 9.94% C = − 0.7% Prod. (video narration-written) PI = 9.83% C = − 0.3%

Prod. (sentence-oral) PI = 1.25 C = .21 Prod. (video narration- written) PI = 1.12 C = .27 Prod. (video narration- oral) PI = .64 C = .15 Prod. (interview-written) PI = 0.4 C=0

Prod. (sentence-oral) PI = .29 C = .28 Prod. (video narration-written) PI = 0.29 C = .30 Prod. (video narration- oral) PI = .25 C = .12 Prod. (interview-written) PI = 0.4 C=0

Prod. (interview-written) PI = − C=−

Prod. (video narration- oral) PI = 3.9% C = 0.3%

Prod. (sentence-written) PI = 5.5% C = -0.12%

Prod. (sentence-written) PI = 1.22 C = .44

Prod. (sentence-written) PI = 0.67 C = .56

PI >C

Improvement Int. PI = 5.29% C = 0.3%

Means Scores Posttest Int. PI = 11.07 C = 5.74

Int. PI = 5.78 C = 5.71

Means scores Pretest

PI >C

Treatment and results

182  Measuring the effects of Processing Instruction

Spanish object pronouns

Hikima (2008) Japanese passive forms

Sanz (2004)

Pro. = SIA Prod. (sentence) (−E –F) = PI = 0.3 (−E + F) Prod. (discourse) PI = 0.3

Int.(sentence) PI = 0.2 Int (discourse) PI = 0.2

Prod. (sentence) PI = 0.5 Prod. (discourse) PI = 0.7

Prod. (sentence) PI = 2% Prod. (discourse) PI = 4%

Int. (sentence) PI = 6% Int (discourse) PI = 6%

Int. (sentence) PI = 0.8 Int (discourse) PI = 0.8

Int. = SIA (−E –F) = (−E + F)

Prod (video retelling) SIA + F = 36% SIA – F = 23%

Prod (video retelling) SIA + F = 58.15 SIA – F = 55.75

Prod (video retelling) SIA + F = 22.69 SIA – F = 32.38

Prod (sentence) SIA + F = 41% SIA – F = 24%

Prod (sentence) SIA + F = 83.62 SIA – F = 66.00

Pro. = SIA Prod (sentence) (−E –F) = SIA + F = 42.15 SIA – F = 42.44 (−E + F)

Int SIA + F = 40% SIA – F = 36%

Prod. (interview-oral) PI = 2.6% C = -0.3%

Int SIA + F = 66.67 SIA – F = 62.50

Prod. (interview-oral) PI = .26 C=0

Int SIA + F = 26.15 SIA – F = 26.25

Int. = SIA (−E –F) = (−E + F)

Prod. (interview-oral) PI = 0 C = 0.3

Discourse-level interpretation tasks  183

Conclusion

Chapter 9 Conclusion: What does the research on Processing Instruction tell us?

In the previous eight chapters we have evaluated and reviewed the effects of Processing Instruction through a large database of research. What does the research on Processing Instruction tell us? Our intention was to present and summarise research on Processing Instruction so that readers can formulate an overall view about the role of this approach to grammar teaching in instructed second language acquisition. We began with an account of ­VanPatten’s ­theory of input processing (VanPatten 1996, 2004, 2007). This theory is the one most closely aligned with Processing Instruction. From there, our evaluation addressed the following series of questions: –– What makes Processing Instruction effective? –– How does Processing Instruction compare to other types of –– instruction? –– Can Processing Instruction be delivered effectively online as well as in classrooms? –– Can you increase the positive effects of Structured Input on language development by enhancing it aurally and/or textually? –– What are the transfer-of-training effects for Processing Instruction? –– Are the effects of Processing Instruction durative (short-term) and longitudinal (long-term)? –– How have the effects of Processing Instruction been measured? We now present a synopsis of what the research on Processing Instruction tells us about each of these areas of research. The original question addressed in Processing Instruction research was: Is Processing Instruction more effective than traditional instruction? In chapter 3 of this book we reviewed the results of this research. Overall the main findings of the studies reviewed consistently indicated that Processing Instruction is better than traditional instruction. The results of all the studies contrasting Processing Instruction and traditional instruction suggest the following:

Conclusion 

185

1. Processing Instruction is an effective approach to grammar instruction in different languages (French, Italian and Spanish) and grammatical forms/ structures; 2. Processing Instruction is an effective form of grammar instruction for both beginners and intermediate L2 learners (English native speakers); 3. Processing Instruction is an effective form of intervention in altering inappropriate processing strategies (First Noun Principle; Lexical Preference Principle; Preference for Nonredundancy Principle) and instilling appropriate ones; 4. L2 learners receiving Processing Instruction performed better that L2 learners receiving traditional instruction in an interpretation sentencelevel task. The results of research described in chapter three showed that L2 learners exposed to Processing Instruction made an improvement from 32% to 63%, depending on the individual study, from pre-test to post-tests (see Table 3.2); 5. L2 learners receiving Processing Instruction performed in an equal way to L2 learners receiving traditional instruction in a production sentence-level task (written or oral). Their improvement from pre-test to post-tests ranged from 25% to 79% depending on the individual study (see Table 3.2). A second question generated from the Processing Instruction research agenda and a direct consequence of the first original question was: Is Processing Instruction more effective than meaning-based output instruction? In chapter three the studies investigating the effects of Processing Instruction and meaning output-based instruction were reviewed and the following conclusion drawn: 1. Processing Instruction is an effective approach to grammar instruction not only in romance languages (French, Italian and Spanish) but also in English; 2. Processing Instruction is an effective form of grammar instruction for both beginners and intermediate L2 learners no matter their L1; 3. Processing Instruction is an effective form of intervention in altering inappropriate processing strategies (First Noun Principle; Lexical Preference Principle; Sentence Location Principle) and instilling appropriate ones; 4. the majority of the studies reviewed in chapter three showed that L2 learners receiving Processing Instruction performed better than L2 learners receiving meaning output-based instruction in an interpretation sentencelevel task. The results of research described in chapter three showed that

186 

Conclusion

L2 learners exposed to Processing Instruction made an improvement which ranged from 28% to 60%, depending on an individual study, from pre-test to post-tests (see Table 3.4); 5. L2 learners receiving Processing Instruction performed equally to L2 learners receiving meaning output-based instruction in a production sentence-level task (written). Their improvement from pre-test to post-tests ranged from 27% to 67% depending on the specific study (see Table 3.4). A much discussed question in the instructional literature is the role of technology in second language learning. A natural question to have posed in this context was: Is online delivery of Processing Instruction an effective means of delivering PI? In Chapter 4 of this book we reviewed the results of this research. The main findings of the two studies reviewed converge, indicating that the online delivery of Processing Instruction is an effective means of delivering it. The results of the studies comparing and contrasting the online and classroom delivery of Processing Instruction show the following: 1. online and classroom delivery of Processing Instruction are equally effective means of delivering this type of grammar instruction; 2. the effects of online and classroom delivery of Processing Instruction endure for one week after the treatment; 3. online and classroom delivery of Processing Instruction can address equally well different processing problems (Lexical Preference, Redundancy, Nonmeaningfulness); 4. L2 learners receiving online delivery of Processing Instruction improve to the same degree as learners receiving classroom delivery, from 14%–18% for one grammatical item and 59%–79% on another (see Table 4.2). A second question generated from this branch of Processing Instruction research is: How does the online delivery of PI compare to the online delivery of MOI? In chapter 4 the handful of studies investigating the effects of mode of delivering PI and MOI were reviewed and the following conclusions can be drawn: 1. Processing Instruction is an effective form of grammar instruction for both beginning and intermediate L2 learners be it delivered online or in classrooms; 2. L2 learners receiving Processing Instruction performed better than L2 learners receiving meaning-based output instruction in a sentence-level

Conclusion 

187

interpretation task with PI learners improving from 54%–67% compared to MOI learners improving from 5%–33%; 3. L2 learners receiving Processing Instruction performed equally to L2 learners receiving meaning-based output instruction in a sentence-level production task (written) with PI learners improving from 54%–66% compared to MOI learners improving from 49%–61%; 4. Processing Instruction is superior to meaning-based output instruction as measured by a sentence-level interpretation task for both online and classroom delivery modes. 5. Processing Instruction is equal to meaning-based output instruction as measured by a sentence-level production task for both online and classroom delivery modes. A much discussed and investigated question in research on input is whether enhanced input is an effective means of drawing learners’ attention to forms in the input. As discussed in Chapter 2, structured input is the heart of Processing Instruction. In chapter 5 we examined the research on enhanced versus unenhanced structured input. One question addressed was: Will enhanced and unenhanced structured input help learners process non-meaningful and redundant forms? The non-meaningful and redundant forms investigated were noun–adjective agreement and subjunctive mood morphology in Italian. The main findings of this line of investigation follow: 1. enhanced and unenhanced structured input are equally effective in helping learners process non-meaningful and redundant forms as measured by an interpretation task with L2 learners who received enhanced input improving 47%–57% and L2 learners who received unenhanced input improving 44%–57%; 2. enhanced and unenhanced structured input are equally effective in helping learners process non-meaningful and redundant forms as measured by an production task with L2 learners who received enhanced input improving 38%–39% and L2 learners who received unenhanced input improving 37%–38%; 3. unenhanced structured input is sufficient for helping learners process redundant and nonmeaningful forms. A second question addressed in this line of research was: Can enhanced structured input help push learners to process a meaningful form over a lexical item? The meaningful form under consideration was Italian future tense verb morphology.

188 

Conclusion

1. enhanced and unenhanced structured input are equally effective in helping learners process meaningful forms as measured by an interpretation task with L2 learners who received enhanced input improving 42% and L2 learners who received unenhanced input improving 49%; 2. enhanced and unenhanced structured input are equally effective in helping learners process meaningful forms as measured by an production task with L2 learners who received enhanced input improving 41% and L2 learners who received unenhanced input also improving 41%; 3. unenhanced structured input is sufficient for helping learners process a meaningful form. A third question explored in this line of research was: Will enhanced and unenhanced structured input help direct learners’ attention to meaningful morphology in sentence final position? The form investigated was the Japanese past tense marker. This marker appears in verb-final position and all verbs in Japanese appear in sentence-final position. 1. Processing Instruction is an effective type of grammar instruction, not only for native speakers of English but also for native speakers of Japanese; 2. Processing Instruction is an effective type of grammar instruction for typologically distinct languages, Romance languages (French, Italian, Spanish), English, and Japanese; 3. enhanced and unenhanced structured input are equally effective in helping learners process a meaningful form in sentence final position as measured by an interpretation task with L2 learners who received enhanced input improving 51% and L2 learners who received unenhanced input also improving 51%; 4. the effects of enhanced and unenhanced structured input endure as measured by an interpretation test one week after treatment with L2 learners who received enhanced input improving 43% and L2 learners who received unenhanced input also improving 43%; 5. enhanced and unenhanced structured input are equally effective in helping learners process non-meaningful and redundant forms as measured by an production task with L2 learners who received enhanced input improving 48% and L2 learners who received unenhanced input improving 43%; 6. the effects of enhanced and unenhanced structured input endure as measured by an production test one week after treatment with L2 learners who received enhanced input improving 42% and L2 learners who received unenhanced input also improving 38%;

Conclusion 

189

7. unenhanced structured input is sufficient for helping learners process a meaningful form in sentence final position. By and large the research on Processing Instruction has examined its primary effects. Primary effects refer to learners being taught a particular form and then tested on that form and only on that form. In chapter six we presented the results of this line of research which was aimed at assessing nonprimary or transfer-of-training effects of Processing Instruction. The question is: What are the transfer-of-training effects for Processing Instruction? The main findings of this new line of research were: 1. Processing Instruction is an effective approach to grammar instruction and this primary effect can be measured in different languages (French, Italian, English and Japanese) and different grammatical forms/structures; 2. Processing Instruction is an effective form of grammar instruction for both beginners and intermediate L2 learners from different native languages (e.g. Italian and Korean); 3. Processing Instruction is an effective form of intervention in altering inappropriate processing strategies (First Noun Principle; Lexical Preference Principle; Sentence Location Principle; Preference for Nonredundancy Principle) and instilling appropriate ones; 4. Processing Instruction has secondary effects as learners who received Processing Instruction were able to transfer that training to processing other forms affected by the same processing principle. In the studies reviewed in chapter six, Learners who received Processing Instruction on a specific form were able to improve from pre-test to post-test on a second form affected by the same principle without further instruction. Their improvement ranged from 19% to 38% in an interpretation task; and from 9% to 20% in a production task (see Table 6.2). A further question, directly linked to the necessity to explore secondary effects for Processing Instruction, Is Processing Instruction effective at helping learners to pick up a second/different strategy? The studies reviewed in chapter six have clearly shown a cumulative benefit of Processing Instruction. The two studies conducted in this new line of research have indicated the following: 1. L2 learners who received Processing Instruction were able to transfer their training processing on different types of strategies. The cumulative

190 

Conclusion

effects have been measured in both interpretation and production tasks (minimal effects). Their improvement ranged from 13% to 21% in the interpretation task (see Table 6.2). The original PI research also addressed the question of whether the effects of Processing Instruction would be found not only immediately following instruction, but also at a later time. Would the effects endure, diminish, or disappear? We examine the research on durative effects in Chapter 7. The first question we addressed was: Are the effects of PI durative in the shortterm? The findings of these studies converge in that the effects of Processing Instruction are durative over the short term. The results of these studies suggest the following conclusions. Short-term durative effects for Processing Instruction have been found for: 1. a period of one week, two weeks, three weeks, 24 days, and one month after treatment as measured by a variety of assessment tasks (see Table 7.3). These results show that the effects of Processing Instruction are undiminished over the short-term in all but one study (see Table 7.4); 2. linguistic items that reflect different processing problems as captured by different processing principles; 3. morphological, syntactic, and semantic linguistic structures; 4. beginning and intermediate level learners; 5. native speakers of English and Italian. An important question to consider is: Are there longer term effects for PI? Only two studies provide us insights into this question. The results of these studies lead to the following conclusions. 1. the effects of Processing Instruction are undiminished 14 weeks after instruction; 2. the effects of Processing Instruction are diminished but present 8 months after instruction; 3. longer term effects are evident for morphological and syntactic linguistic items; 4. longer term effects are evident for both adults and younger learners. As reviewed in this book and examined particularly in chapter eight, most of the studies investigating the effects of Processing Instruction have measured L2 learner ‘s ability to interpret and produce forms/structures at sen-

Final comments 

191

tence level. Few studies have measured the relative effects of Processing Instruction on production discourse level tasks. Lee (2004: 319) has indicated that there is a limit in the current database as no study has yet addressed the following question: is Processing Instruction effective at improving L2 learners ability in interpreting forms/structure at discourse-level? The studies investigating the effects of Processing Instruction at discourselevel have been briefly reviewed in chapter eight. The main findings indicated the following: 1. Processing Instruction is an effective form of intervention in altering inappropriate processing strategies (Primacy of meaning principle; First Noun Principle; Sentence Location Principle; Primacy of Content Words Principle: Preference for Nonredundancy Principle) and instilling appropriate ones; 2. L2 learners receiving Processing Instruction improved from pre-test to post-tests in sentence-level interpretation and production tasks (see Table 8.2). 3. L2 learners receiving Processing Instruction improved form pre-test to post-tests in discourse-level production tasks (oral and written). 4. L2 learners receiving Processing Instruction improved form pre-test to post-tests in interpretation discourse-level tasks (see Table 8.2). Final comments To end this evaluation of what the research on Processing Instruction tells us, we return to the questions that form our chapter titles. In Table 9.1 we provide answers-in-brief to these questions. In addition to our answers-in-brief, we also offer the following commentary around each question. 1. What makes Processing Instruction effective? In that the research points to structured input practices, we must critically evaluate the role of explicit information in instructed SLA. Under what circumstances does explicit information have an impact on learner improvement? 2. How does Processing Instruction compare to other types of instruction? Collentine (2004) suggested that Processing Instruction did not, in the future, need to be compared to traditional instruction. VanPatten (2004c) concurred. Both assert that Processing Instruction had sufficiently established itself as an instructional type and that the way forward was to more closely investigate Processing Instruction itself. Processing instruction consists of explicit

192 

Conclusion

Table 9.1. Question

Answer-in-brief

What makes Processing Instruction effective?

Most research points to structured input activities as being the causative factor in learner improvement.

How does Processing Instruction compare to other types of instruction?

Processing Instruction has always been equally as effective as other types of instruction in helping learners produce the target forms. Processing instruction is usually superior to output-based instruction types in helping learners interpret target forms but some meaning-based output practices seem to help learners to interpret forms as well as Processing Instruction does.

Can Processing Instruction be delivered effectively online as well as in classrooms?

Online and classroom delivery of Processing Instruction are equally effective modes of delivering this type of instruction.

Can you increase the positive effects of Structured Input on language development by enhancing it aurally and/or textually?

Enhanced and unenhanced structured input are equally effective in helping learners improve. Unenhanced structured input is sufficient to bring about improvement.

What are the transfer-oftraining effects for Processing Instruction?

Transfer-of-training effects can be either secondary or cumulative; secondary to a linguistic form that shares the same processing problem as the primary form and cumulative to linguistic form that represents a completely different processing problem.

Are the effects of PI durative (short-term) and longitudinal (long-term)?

The effects of PI are both durative and longitudinal. Undiminished effects have been found up to 14 weeks after treatment but diminished effects were found 8 months after treatment.

How have the effects of Processing Instruction been measured?

The effects of Processing Instruction have been measured with both sentence- and discourselevel interpretation and production tasks as well as with written and oral production tasks.

Final comments 

193

information about form, information about processing strategies, referential structured input activities, and affective structured input activities. Which of these are the crucial factors that contribute to learner improvement? 3. Can Processing Instruction be delivered effectively online as well as in classrooms? Online delivery of Processing Instruction has proven to be an effective mode of delivery. Future research can certainly take for granted that mode of delivery is not an issue that needs testing. Rather, delivering Processing Instruction in computer labs can contribute to more effective research instrumentation. Classroom delivery involves using whole classes whereas online delivery may help randomize learners’ assignment to a particular research cell. 4. Can you increase the positive effects of Structured Input on language development by enhancing it aurally and/or textually? The research points rather strongly to the fact that unenhanced structured input is sufficient for bringing about improvement. The larger issue to be investigated is how to improve upon the results we have already found. How can learners make even greater gains as a result of receiving Processing Instruction? What can be done to boost performance from a 50% increase to a 75% increase? Is such an increase possible in a research setting? 5. What are the transfer-of-training effects for Processing Instruction? That Processing Instruction has demonstrated transfer-of-training effects is important to the theoretical underpinnings of Processing Instruction. This evidence is the strongest that Processing Instruction, by altering processing strategies, has an effect on the learners’ developing systems, not just on their metalinguistic knowledge. This line of investigation is among the newest lines to emerge and more research should be conducted in this area. 6. Are the effects of PI durative (short-term) and longitudinal (long-term)? The effects of Processing Instruction are both durative and longitudinal. The research has clearly established this point. Future research might begin to investigate other aspects of durative/longitudinal effects. For example, ­future research could compare two groups of learners who receive Processing Instruction. One group would receive just one treatment whereas the other would receive a second treatment at a later point in time. Then, eight months later, how do the two groups compare? Have the effects of PI diminished equally for both groups? 7. How have the effects of Processing Instruction been measured? ­Doughty (2004) commented that much of the Processing Instruction research includes assessments that appear to tap only metalinguistic knowledge. Yet, as our review shows, there is a body of research that has examined the effects of

194 

Conclusion

Processing Instruction using discourse-level instrumentation. This research is in the minority and certainly future work could well investigate how learners process discourse in which the target form has been embedded or produce discourse in which they use the target forms. The bibliography on Processing Instruction research begins with ­VanPatten and Cadierno’s (1993) work. The research has continued into the present day with ever newer lines of investigation presenting themselves. The future of Processing Instruction research looks rather robust.

References References Alanen, R. 1995 Allen, L. Q. 2000

Input enhancement and rule presentation in second language acquisition. In R. Schmidt (ed.), Attention and Awareness in Foreign Language Acquisition, 259–302. Honolulu: University of Hawaii. Form-meaning connections and the French causative. Studies in Second Language Acquisition 22: 69–84.

Anderson, J. 1983 The Architecture of Cognition. Harvard: Harvard University Press. Barcroft, J., and B. VanPatten 1997 Acoustic salience of grammatical forms: The effect of location stress, and boundedness on Spanish L2 input processing. In W. R. Glass and A. T. Pérez-Leroux (eds.), Contemporary Perspectives on the Acquisition of Spanish: Production, Processing, and Comprehension, 109– 121. Sommerville, MA: Cascadilla Press. Bardovi-Harlig, K. 1992 The use of adverbials and natural order in the development of temporal expression. International Review of Applied Linguistics 30: 299–320. Bardovi-Harlig, K. 2007 One functional approach to second language acquisition: The concept-oriented approach. In B. VanPatten and J. Williams (eds.), Theories of Second Language Acquisition: An Introduction, 97–113. Mahwah, NJ: Erlbaum. Bates, E., and B. MacWhinney 1989 Functionalism and the competition model. In B. MacWhinney and E. Bates (eds.), The Cross-Linguistic Study of Sentence Processing, 77–117. Cambridge: Cambridge University Press. Bavin, E. L., and T. Shopen 1989 Cues to sentence interpretation in Warlpiri. In B. MacWhinney and E. Bates (eds.), The Cross-Linguistic Study of Sentence Processing, 185–205. Cambridge: Cambridge University Press. Bernhardt, E. B. 1992 Reading Development in a Second Language. Norwood, NJ: Ablex. Benati, A. 2001 A comparative study of the effects of processing instruction and output-based instruction on the acquisition of the Italian future tense. Language Teaching Research 5: 95–127.

196 

References

Benati, A. 2004a

Benati, A. 2004b Benati, A. 2005 Benati, A. 2008

The effects of structured input and explicit information on the acquisition of Italian future tense. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 207–255. Mahwah, NJ: Erlbaum. The effects of processing instruction and its components on the acquisition of gender agreement in Italian. Language Awareness 13: 67–80. The effects of PI, TI and MOI in the acquisition of English simple past tense. Language Teaching Research 9: 67–113.

Japanese Language Teaching: A Communicative Approach. London: Continuum. Benati, A., and P. Romero Lopez 2004 The effects of structured input and explicit information on the acquisition of Gender agreement in Italian and the simple paste tense in Spanish. In M. Baynham, A. Deignan, and G.White (eds.), Applied Linguistics at the Interface, 29–46, London: Equinox. Benati, A., and J. F. Lee 2008 Grammar Acquisition and Processing Instruction: Secondary and Cumulative Effects. Bristol: Multilingual Matters. Benati, A., and J. F. Lee 2009 Processing Instruction and Discourse. London: Continuum. Benati, A., J. F. Lee, with S. D. Houghton 2008 Ch. 4. From Processing Instruction on the acquisition of English past tense to secondary transfer-of-training effects on English third person singular present tense. In A. Benati and J. F. Lee, Grammar Acquisition and Processing Instruction: Secondary and Cumulative Effects, 88–120. Bristol: Multilingual Matters. Benati, A., J. F. Lee, and C. Laval 2008 Ch. 5. From Processing Instruction on the acquisition of French imparfait to secondary tranfer-of-training effects on French subjunctive and to cumulative transfer-of-training effects with French causative constructions. In A. Benati and J. F. Lee, Grammar Acquisition and Processing Instruction: Secondary and Cumulative Effects, 121–157. Bristol: Multilingual Matters. Bever, T. G. 1970 The cognitive basis for linguistic structures. In R. Hayes (ed.), Cognition and Language Development, 279–362. New York: Wiley & Sons.

References 

197

Binkowski, D. D. 1992 The effects of attentional focus, presentation mode, and language experience on second language learner’s sentence processing. Ph.D. dissertation, University of Illinois at Urbana-Champaign. Bransdorfer, R. 1989 Processing function words in input: Does meaning make a difference? Paper presented at the annual meeting of the American Association of Teachers of Spanish and Portuguese, San Antonio. Bransdorfer, R. 1991 Communicative value and linguistic knowledge in second language oral input. Processing. Ph.D. dissertation, University of Illinois at Urbana-Champaign. Cadierno, T. 1995 Formal instruction from a processing perspective: An investigation into the Spanish past tense. The Modern Language Journal 79: 179–193. Carroll, S. 2004 Commentary: Some general and specific comments on Input Processing and Processing Instruction. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 293–309. Mahwah, NJ: Erlbaum. Carroll, S. 2007 Autonomous induction theory. In B. VanPatten and J. Williams (eds.), Theories of Second Language Acquisition: An Introduction, 155–173. Mahwah, NJ: Erlbaum. Carroll, S. 2002 The effects of processing instruction on the acquisition of ser and estar. Hispania 85: 308–323. Cheng, A. C. 2004 Processing instruction and Spanish ser and estar: Forms with semantic-aspectual value. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 119–141. Mahwah, NJ: Erlbaum. Chomsky, N. 1965 Aspects of the Theory of Syntax. Cambridge, Mass: MIT Press. Collentine, J. G. 1998 Processing Instruction and the Subjunctive. Hispania 81: 576–587. Collentine, J. G. 2004 Commentary. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 119–141. Mahwah, NJ: Erlbaum. Corder, P. 1981 Error Analysis and Interlanguage. Oxford: Oxford University Press. DeKeyser, R. 2007 Skill acquisition theory. In B. VanPatten and J. Williams (eds.), Theories of Second Language Acquisition: An Introduction, 97–113. Mahwah, NJ: Erlbaum.

198 

References

DeKeyser, R. M., and K. J. Sokalski 1996 The differential role of comprehension and production practice. Language Learning 46: 613–642. DeKeyser, R., R. Salaberry, P. Robinson, and M. Harrington 2002 What gets processed in Processing Instruction? A commentary on Bill VanPatten’s “Processing Instruction: An update” Language Learning 52: 805–823. Doughty, C., and J. Williams (eds.) 1998 Focus on Form in Classroom Second Language Acquisition. Cambridge: Cambridge University Press. Echevarría, M. S. 1978 Desarrollo de la comprensión infantil de la sintaxis española. Concepción, Chile: Universidad de Concepción. Ellis, N. 1994 Implicit and Explicit Learning of Languages. London: London Academic Press. Ellis, N. 2007 The associative-cognitive CREED. In B. VanPatten and J. Williams (eds.), Theories of Second Language Acquisition: An Introduction, 77–95. Mahwah, NJ: Erlbaum. Ellis, R. 1991 Grammar teaching practice or consciousness raising? In R. Ellis (ed.), Second Language Acquisition and Second Language Pedagogy, 232–241. Multilingual Matters. Ellis, R. 2001 Form-focused instruction and second language learning. University of Michigan: Blackwell. Ervin-Tripp, S. 1974 Is second language learning like the first? TESOL Quarterly 8: 111– 127. Farley, A. P. 2001a The effects of processing instruction and meaning-based output instruction. Spanish Applied Linguistics 5: 57–94. Farley, A. P. 2001b Authentic processing instruction and the Spanish subjunctive. Hispania 84: 289–299. Farley, A. P. 2004a The relative effects of processing instruction and meaning-based output instruction. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 143–168. Mahwah, NJ: Erlbaum. Farley, A. P. 2004b Processing instruction and the Spanish subjunctive: Is explicit information needed? In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 227–239. Mahwah, NJ: Erlbaum.

References  Farley, A. P. 2005 Gass, S. 1989 Gely, A. 2005 Glisan, E. 1985 González, N. 1997

199

Structured Input: Grammar Instruction for the Acquisition-Oriented Classroom. New York: McGraw-Hill. How do learners resolve linguistic conflicts? In S. Gass, and J. Schacter, (eds.), Linguistic Perspectives on Second Language Acquisition, 183–199. Cambridge: Cambridge University Press. Output-based Instruction versus Processing Instruction on the acquisition of the French imperfect tense. MA thesis, University of Greenwich, London. The effect of word order in listening comprehension and pattern repetition: An experiment in Spanish as a foreign language. Language Learning 35: 443–472. A parametric study of L2 acquisition: Interpretation of Spanish word order. In A. T. Pérez-Leroux and W. R. Glass (eds.), Contemporary Perspectives on The Acquisition of Spanish, Volume 1: Developing Grammars, 133–148. Somerville, MA: Cascadilla Press.

Harrington, M. 2004 Commentary: Input processing as a theory of processing input. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 79–92 Mahwah, NJ: Erlbaum. Heikel, E., and S. Fotos 2002 New Perspectives on Grammar Teaching in Second Language Classroom (eds.). Mahwah, NJ: Erlbaum. Hikima, N. forthc. Processing instruction and traditional instruction in the acquisition of Japanese passive forms. Houston, T. 1997 Sentence processing in Spanish a background knowledge. In W. R. Glass and A. T. Pérez-Leroux (eds.), Contemporary Perspectives on the Acquisition of Spanish: Production, Processing and Comprehension, 123–134. Somerville, MA: Cascadilla Press. Issidorides, D. C., and J. H. Hulstijn 1991 Comprehension of grammatically modified and nonmodified sentences by second language learners. Applied Psycholinguistics 13: 141–171. Jourdenais, R., M. Ota, S. Stauffer, B. Boyson, and C. Doughty 1995 Does textual enhancement promote noticing? A think-aloud protocol analysis. In R. Schmidt (ed.), Attention and Awareness in Second Language Learning, 183–216. Honolulu: University of Hawaii.

200 

References

Klein, W. 1986 Krashen, S. 1982 Laval, C. Forthc. Lee, J. F. 1987 Lee, J. F. 1990 Lee, J. F. 1998 Lee, J. F. 1999 Lee, J. F. 2002 Lee, J. F. 2003

Lee, J. F. 2004

Second Language Acquisition. Cambridge: Cambridge University Press. Principles and Practice in Second Language Acquisition. London: Pergamon. Measuring processing instruction secondary effects on the acquisition of French. Manuscript. Comprehending the Spanish subjunctive: An information processing perspective. Modern Language Journal 71 (1): 50–57. Constructive processes evidenced by early stage non-native readers of Spanish in comprehending an expository text. Hispanic Linguistics 4: 129–148. The relationship of verb morphology to second language reading comprehension and input processing. Modern Language Journal 82 (l): 33–48. On levels of processing and levels of comprehension. In J. GutiérrezRexach and F. Martínez-Gil (eds.), Advances in Hispanic Linguistics, 42–59. Somerville, MA: Cascadilla Press. The incidental acquisition of Spanish future tense morphology through reading in a second language. Studies in Second Language Acquisition 24: 55–80. Cognitive and linguistic perspectives on the acquisition of object pronouns in Spanish. In B. Lafford and R. Salaberry (eds.), Spanish Second Language Acquisition: State of the Science, 98–129. Washington, D.C.: Georgetown University Press.

On the generalizability, limits, and potential future directions of processing instruction research. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 311–323 Mahwah, NJ: Erlbaum. Lee, J., T. Cadierno, W. Glass, and B. VanPatten 1997 The effects of lexical and grammatical cues on processing past temporal reference in second language input. Applied Language Learning 8: 1–27.

References 

201

Lee, J., and B. VanPatten 1995 Making Communicative Language Teaching Happen. New York: McGraw-Hill. Lee, J., and B. VanPatten 2003 Making Communicative Language Teaching Happen, 2nd ed. New York: McGraw-Hill. Lee, J., and A. Benati 2007a Second Language Processing: An Analysis of Theory, Problems and Solutions. Continuum: London. Lee, J., and A. Benati 2007b Delivering Processing Instruction in Classrooms and Virtual Contexts: Research and Practice. Equinox, London. Lee, J. F., A. Benati, with J. Aguilar-Sánchez, and E. M. McNulty 2007 Ch 4. Comparing three modes of delivering processing instruction on preterite/imperfect distinction and negative informal commands in Spanish. In J. Lee and A. Benati (eds.), Delivering Processing Instruction in Classrooms and Virtual Contexts: Research and Practice, 73–98. Equinox: London. Lee, J. F., and R. Rodriguez 1997 The effects of lexemic and morphosyntactic modifications on L2 reading comprehension and input processing. In W. Glass and A. Perez Leroux (eds.), Contemporary Perspectives on the Acquisition of Spanish, 135–157. Somerville, MA: Cascadilla Press. Lee, J. F., and A. Rossomondo 2004 Cross-experimental evidence for the incidental acquisition of Spanish future tense morphology. Paper presented at the Annual Symposium on Hispanic Linguistics. Minneapolis, MN. Leeser, M. 2004. The effects of topic familiarity, mode, and pausing on second language learners’ comprehension and focus on form. Studies in Second Language Acquisition 26: 587–615. Leow, R. 1997 The effects of input enhancement and text length on adult L2 readers’ comprehension and intake in second language acquisition. Applied Language Learning 8(2): 151–182. Leow, R. 2001 Do learners notice enhanced forms while interacting with the L2? An online and offline study of the role of written input enhancement in L2 reading. Hispania 84: 496–509. Leow, R., T. Egi, A. M. Nuevo, and Y. Tsai 2003 The roles of textual enhancement and type of linguistic item in adult L2 learners’ comprehension and intake. Applied Language Learning 13(2): 1–17.

202 

References

Lightbown, P., and N. Spada 1993 How Languages Are Learned. Oxford: Oxford University Press. Liu, M., Z. Moore, L. Graham, and S. Lee 2002 A look at the research on computer-based technology use in second language learning: A review of the literature from 1990–2000. Journal of Research on Technology in Education 34: 250–273. LoCoco, V. 1987 Learner comprehension of oral and written importance of word order. In B. VanPatten, T. Dvorak, and J. F. Lee (eds.), Foreign Language Learning: A Research Perspective, 119–129. Rowley, MA: Newbury House. Long, M. 1983 Does second language instruction make a difference? TESOL Quarterly 17: 359–382. Long, M. 1991 Focus on form: A design feature in language teaching methodology. In K. de Bot (ed.) Long, M. 1991 Foreign Language Research in Cross-Cultural perspectives. Amsterdam: John Benjamins. Long, M. H., and P. Robinson 1998 Focus on form: Theory, research and practice. In C. Doughty and J. Williams (eds.). Long, M. H., and P. Robinson 1998 Focus on Form in Classroom Second Language Acquisition, 15–41. Cambridge: Cambridge University Press. Malovrh, P. A. 2006 L2 sentence processing of Spanish OVS word order and direct object pronouns: An analysis of contextual constraints. In Nuria Sagarra and Almeida Jacqueline Toribio (eds.), Selected Proceedings of the 9th Hispanic Linguistics Symposium, 169–179. Somerville, Mass.: Cascadilla Proceedings Project. Mangubhai, F. 1991 The processing behaviours of adult second language learners and their relationship to second language proficiency. Applied Linguistics 12: 268–297. Marsden, E. 2005 Exploring input processing in the classroom: An experimental comparison of processing instruction and enriched input. Language Learning 56: 507–566. MacDonald, J. L., and L. K. Heilenman 1992 Changes in sentence processing as second language proficiency increases. In R. J. Harris, (ed.), Cognitive Procesing in Bilinguals, 325–336. Elsevier.

References 

203

Mitchell, R., and F. Miles 2004 Second Language Learning Theories, 2nd. ed. London: Hodder Arnold. Montgomery, C., and M. Eisenstein 1985 Real reality revisited: An experimental communicative course in ESL. TESOL Quarterly 19: 317–333 Morgan-Short, K., and H. W. Bowden 2006 Processing instruction and meaningful output-based instruction: Effects on second language development. Studies in Second Language Acquisition 28: 31–65. Mangubhai, F. 1991 The processing behaviours of adult second language learners and their relationship to second language proficiency. Applied Linguistics 12: 268–297. Musumeci, D. 1989 The ability of second language learners to assign tense at the sentence level: A cross-linguistic study. Ph.D. dissertation, University of Illinois at Urbana-Champaign. Nam, E. 1975 Child and adult perceptual strategies in second language acquisition. Paper presented at the annual TESOL Convention, Los Angeles. Nassaji, H., and S. Fotos 2004 Current developments in research on the teaching of grammar. Annual Review of Applied Linguistics 24: 126–145. Norris, J.M., and L. Ortega 2000 Effectiveness of L2 instruction: A research synthesis and quantitative meta-analysis. Language Learning 50: 417–528. Ortega, L. 2007 Second language learning explained? SLA across nine contemporary theories. In B. VanPatten and J. Williams (eds.), Theories in Second Language Acquisition, 225–250. Mahwah, NJ: Erlbaum. Overstreet, M. 1998 Text enhancement and content familiarity: The focus of learner attention. Spanish Applied Linguistics 2: 229–258. Paulston, C. 1972 Structural patterns drills: A classification. In H. Allen and R. Campbell (eds.), Teaching English as a Second Language, 129–138. New York: McGraw-Hill. Peters, A. 1983 The Units of Language Acquisition. Cambridge: Cambridge University Press.

204 

References

Peters, A. M. 1985 Peters, A. M. 1985 Pica, T. 1985

Language segmentation: Operating principles for the perception and analysis of language. In D. Slobin (ed.). The Crosslinguistic Study of Language Acquisition: Theoretical Issues. Lawrence Erlbaum Associates. Linguistic simplicity and learnability: Implications for syllabus design. In K. Hyltenstam and Piennemman (eds.), Modeling and ­Assessing Second Language Acquisition, 137–152. Clevedon, UK: Multilingual Matters.

Pienemann, M. 1984 Psychological constraints on the teachability of languages. Second Language Acquisition 6:186–214. Pienemann, M. 1998 Language Processing and Second Language Development: Processability Theory Amsterdam: John Benjamins Plèh, C. 1989 The development of sentence interpretation in Hungarian. In B. MacWhinney and E. Bates (eds.), The Cross-Linguistic Study of Sentence Processing, 158–184. Cambridge: Cambridge University Press. Qin, J. 2008 The effect of processing instruction and dictogloss tasks on acquisition of the English passive voice. Language Teaching Research 12: 61–82. Rosa, E., and O’Neill, M. 1998 Effects of stress and location on acoustic salience at the initial stages of Spanish L2 input processing. Spanish Applied Linguistics 2: 24– 52. Rossomondo, A. E. 2007 The role of lexical temporal indicators and text interaction format on the incidental acquisition of Spanish future tense morphology. Studies in Second Language Acquisition 29: 39–66. Radford, A. 1990 Syntactic Theory and the Acquisition of English Syntax: The Nature of Early Child Grammars of English. Oxford: Blackwell. Rutherford, W., and M. Sharwood Smith 1985 Consciousness-raising and universal grammar. Applied Linguistics 6: 274–282. Rutherford, W., and M. Sharwood Smith 1988 Grammar and Second Language Teaching. Newbury House.

References 

205

Salaberry, M. R. 1997 The role of input and output practice in second language acquisition. The Canadian Modern Language Review 53: 422–451. Sanz, C. 2004 Computer delivered implicit versus explicit feedback in processing instruction. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 241–255. Mahwah, NJ: Erlbaum. Sanz, C., and B. VanPatten 1998 Processing the content of input processing and processing instruction research: A response to DeKeyser, Salaberry, Robinson, and Harrington. Language Learning 52: 825–831. Savignon, S. 1972 Communicative Competence: An Experiment in Foreign Language Teaching. Philadelphia Center for Curriculum Development Schmidt, R. 1990 The role of consciousness in second language learning. Applied Linguistics 11: 129–158. Schmidt, R. 2001 Attention. In P. Robinson (ed.), Cognition and Second Language Instruction, 3–32. Cambridge: Cambridge University Press. Sharwood Smith, M. 1986 Comprehension versus acquisition: Two ways of processing input. Applied Linguistics 7: 239–274. Sharwood Smith, M. 1991 Speaking to many minds: On the relevance of different types of of language information for the L2 learner. Second Language Research 7: 118–132. Sharwood Smith, M. 1993 Input enhancement in instructed SLA: Theoretical bases. Studies in Second Language Acquisition 15: 165–179. Shook, J.D. 1994 FL/L2 reading, grammatical information, and the input to intake phenomenon. Applied Language Learning 5: 57–93. Slobin, D. I. 1966 Grammatical transformations and sentence comprehension in childhood and adulthood. Journal of Verbal learning and Verbal Behaviour 5: 219–227. Slobin, D. I. 1973 Cognitive prerequisites for the development of grammar. In C.A. Ferguson and D. I. Slobin (eds.), Studies of Child Language Development.  New York: Holt, Rinehart & Winston.

206 

References

Spada, N. 1987 Spada, N. 1997 Terrell, T. D. 1991 Todd, L. 1974 Toth, P. D. 2006

Relationship between instructional differences and learning outcomes: A process-product study of communicative language teaching. Applied Linguistics 8: 137–161. Form-focused instruction and Seccond Language acquisition: A Review of Classroom and Laboratory research. LanguageTeaching 30: 73–87. The role of grammar instruction in a communicative approach. The Modern Language Journal 75: 52–63. Pidgins and Creoles. Boston: Routledge.

Processing Instruction and a role for output in second language acquisition. Language Learning 56: 319–385. Van Naerssen, M. 1981 Generalising second language hypotheses across languages: A test case in Spanish as a second language. Ph.D. dissertation, University of Southern California. VanPatten, B. 1984a Processing strategies and morpheme acquisition. In F. R. Eckman, L. H. Bell, and D. Nelson (eds.), Universals of Second Language Acquisition, 88–98. Rowley, Mass.: Newbury House. VanPatten, B. 1984b Learners’ comprehension of clitic pronouns: More evidence for a word order strategy. Hispanic Linguistics 1: 57–67. VanPatten, B. 1985a The acquisition of ser and estar by adult learners of Spanish: A preliminary investigation of transitional stages of competence. Hispania 68: 399–406. VanPatten, B. 1985b Communicative value and information processing in second language acquisition. In E. Judd, P. Nelson, and D. Messerschmitt (eds.), On TESOL ’84: A Brave New World, 88–99. Washington, D.C.: TESOL. VanPatten, B. 1990 Attending to content and form in the input: An experiment in consciousness. Studies in Second Language Acquisition 12: 287–301. VanPatten, B. 1993 Grammar instruction for the acquisition rich classroom. Foreign Language Annals 26: 433–450.

References  VanPatten, B. 1996 VanPatten, B. 2000 VanPatten, B. 2002 VanPatten, B. 2003 VanPatten, B. 2007

207

Input Processing and Grammar Instruction: Theory and Research. Norwood, NJ: Ablex. Processing instruction as form-meaning connections: Issues in theory and research. In J. F. Lee and A. Valdman (eds.), Form and Meaning: Multiple Perspectives, 43–68. Boston: Heinle & Heinle. Processing instruction: An update. Language Learning 52: 755– 803. From Input to Output: A Teacher’s Guide to Second Language ­Acquisition. New York: McGraw-Hill.

Input processing in adult second language acquisition. In B. VanPatten and J. Williams (eds.), Theories in Second Language Acquisition, 115–135. Mahwah, NJ: Erlbaum. VanPatten, B. (ed.) 2004a Processing Instruction: Theory, Research, and Commentary. Mahwah, NJ: Erlbaum. VanPatten, B. (ed.) 2004b Input processing in second language acquisition. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 5–31. Mahwah, NJ: Erlbaum. VanPatten, B., and T. Cadierno 1993 Explicit instruction and input processing. Studies in Second Language Acquisition 15: 225–243. VanPatten, B., and C. Fernández 2004 The long-term effects of processing instruction. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 273– 289. Mahwah, NJ: Erlbaum. VanPatten, B., and T. Houston 1998 Contextual effects in processing L2 input sentences. Spanish Applied Linguistics 2: 53–70. VanPatten, B., and G. Keating 2007 Getting tense: Lexical preference, L1 transfer, and native and non native processing of temporal reference. Manuscript available from the authors. VanPatten, B., J. F. Lee, and T. Ballman 2006 Viztazos: Un curso breve. New York: McGraw-Hill. VanPatten, B., and S. Oikennon 1996 Explanation vs. structured input in processing instruction. Studies in Second Language Acquisition 18: 495–510.

208 

References

VanPatten, B., and C. Sanz 1995 From input to output: Processing instruction and communicative tasks. In F. R. Eckman, D. Highland, P. W. Lee, J. Mileham, and R. R. Weber (eds.), Second Language Acquisition Theory and Pedagogy, 169–185. Mahwah, NJ: Erlbaum. VanPatten, B., and J. Williams 2007a Theories in Second Language Acquisition: An Introduction. Mahwah, NJ: Erlbaum. VanPatten, B., and J. Williams 2007b Introduction: The nature of theories. In B. VanPatten and J. Williams (eds.), Theories in Second Language Acquisition: An Introduction, 1–16. Mahwah, NJ: Erlbaum. VanPatten,B., J. Williams, and S. Rott 2004 Form-meaning connections in second language acquisition. In B. VanPatten, J. Williams, S. Rott, and M. Overstreet (eds.), FormMeaning Connections in Second Language Acquisition, 1–26. Mahwah, NJ: Erlbaum. VanPatten, B., and W. Wong 2004 Processing instruction and the French causative: Another replication. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 97–118. Mahwah, NJ: Erlbaum. Williams, J., and J. Evans 1998 What kind of focus and on which forms? In C. Doughty and J. Williams (eds.), Focus on Form in Second Language Classroom Acquisition., 139–155. Cambridge: Cambridge University Press. White, J. 1998 Getting the learners’ attention: A typographical input enhancement study. In C. Doughty and J. Williams (eds.), Focus on Form in Second Language Classroom Acquisition, 85–113. Cambridge: Cambridge University Press. Wong, W. 2003 Textual enhancement and simplified input: Effects on L2 comprehension and acquisition of non-meaningful grammatical form. Applied Language Learning 13(2): 17–47. Wong, W. 2004a The nature of processing instruction. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 33–63. Mahwah, NJ: Erlbaum. Wong, W. 2004b Processing instruction in French: The roles of explicit information and structured input. In B. VanPatten (ed.), Processing Instruction: Theory, Research, and Commentary, 187–205. Mahwah, NJ: Erlbaum.

References  Wong, W. 2005

209

Input Enhancement: From Theory and Research to the Classroom. New York: McGraw-Hill. Wong, W., and B. VanPatten 2003 The evidence is IN: Drills are OUT. Foreign Language Annals 36: 403–423.

Author index Alanen, R. 114 Aguilar-Sanchez, J. 100–104, 109, 111, 146, 148, 164, 170 Allen, L. Q. 21, 36 Anderson, J. 89 Ballman, T. 101 Barcroft, J. 20 Bardovi-Harlig, K. 32 Bates, E. 35 Bavin, E. L. 35 Bernhardt, E. B. 9, 10 Benati, A. 109.110, 3, 6, 17, 36, 39–41, 56–57, 61–67, 79, 83–86, 89–96, 100–108, 111–121, 124–130, 135– 141, 146–148, 156, 161, 164–171 Bever, T. G. 20, 33, 35 Binkowski, D. D. 34 Bowden, H. W. 85, 86, 94, 97, 147, 148, 162, 169 Bransdorfer, R. 18, 33 Cadierno, T. 14, 32, 77, 78, 91, 92, 101, 119, 144, 145, 152, 154, 175, 176, 177, 194 Carroll, S. 9, 13, 25 Cheng, A. 79–80, 91, 93, 147–150, 158–159, 161, 167, 177–178, 181 Chomsky, N. 68 Collentine, J. G. 81, 191 Corder, P. 70 DeKeyser, R. M. 36, 175 Doughty, C. 38, 70, 89, 193 Echevarría, M. S. 21

Egi, T. 114 Eisenstein, M. 69 Ellis, R. 69, 70, 74, 89 Ervin-Tripp, S. 20, 34 Farley, A. P. 2, 39, 45, 60, 63–65, 82–86, 94–96, 105, 113, 118, 147–148, 157, 161, 166, 169 Fernández, C. 149, 150, 172, 173 Fotos, S. 69, 70 Gass, S. 34, 35 Gely, A. 84, 94, 97 Glass, W. 14, 32 Glisan, E. 34 González, N. 21 Graham, L. 101 Harrington, M. 24, 25, 175 Heilenman, L. K. 34, 35 Heikel, E. 68 Hikima, N. 40, 179, 181, 183 Houghton, D. 137, 139 Houston, T. 23 Hulstijn, J. H. 35 Issidorides, D. C. 35 Jourdenais, R. 114 Keating, G. 5 Klein, W. 9, 31 Krashen, S. 68, 89 Laval, C. 135, 137, 138, 140, 142 Lee, J. F. 3, 6–8, 11–21, 26, 32–43, 61, 64, 67, 70–72, 82, 85–90, 95,

212 

Author index

100–121, 124–130, 135–140, 146– 148, 164–165, 170–171, 178–180 Lee, S. 101 Leeser, M. 114 Leow, R. 114 Lightbown, P. 31, 89 Liu, M. 101 LoCoco, V. 20, 21, 22, 34 Long, M. 38, 68, 69 MacDonald, J. L. 34, 35 MacWhinney, B. 35 Malovrh, P.,A. 23, 24 Mangubhai, F. 9, 31 Marsden, E. 86, 98, 99, 149, 150, 172, 173 McNulty, M. 100–104, 109, 111, 146, 148, 164, 170 Mitchell, R. 2, 26 Miles, F. 2, 26 Montgomery, C. 69 Moore, Z. 101 Morgan-Short, K. 85, 86, 94, 97, 147, 162, 169 Musumeci, D. 15, 16, 32 Nam, E. 20, 34 Nassaji, H. 70 Norris, J.M. 36, 70.143, 175 Nuevo, A.M. 114 Oikennon, S. 55, 57, 58, 59, 63, 65, 89 O’Neill, M. 20 Ortega, L. 2, 36, 70, 143, 175 Overstreet, M. 114 Paulston, C. 71 Peters, A. 31 Pica, T. 32 Pienemann, M. 70, 89 Plèh, C. 35

Qin, J. 88, 98, 99 Radford, A. 31 Robinson, P. 38, 175 Rodriguez, R. 7 Romero Lopez, P. 41 Rosa, E. 20 Rossomondo, A. 8, 15 Rott, S Rutherford, W. 70, 113 Salaberry, M. R. 36, 175 Sanz, C. 36, 42, 43, 59, 63, 66, 175, 178, 181, 182, 183 Savignon, S. 69 Schmidt, R. 69 Sharwood Smith, M. 7, 31, 70, 75, 113 Shook, J. D. 114 Shopen, T. 35 Slobin, D. I. 5, 20, 33 Sokalski, K. J. 36 Spada, N. 31, 69, 89 Terrell, T. D. 75 Todd, L. 32 Toth, P. D. 87, 98, 99, 147, 163, 170 Tsai Y. 114 Van Naerssen, M. 33 VanPatten, B. 1–39, 42–49, 53–59, 62–65, 70–72, 75–82, 88–93, 100–105, 108, 113–115, 118–121, 128–130, 133, 144–145, 149–154, 172–178, 181–184, 191, 194 White, J. 114 Williams, J. 1, 2, 25, 26, 38, 70, 89 Wong, W. 2, 10, 21–22, 36–39, 42, 58–59, 64, 66, 73, 80, 91, 93, 102, 113–114

Subject index developing system 26, 74, 75, 87 First noun principle/strategy 4, 6, 11, 20–21, 23–25, 29–30, 35, 47, 53–54, 133, 135, 179, 185, 191 input 1–2, 4, 8, 14, 26, 37–39, 41–46, 48–49, 71–76, 100, 113–116, 119 input processing 1–3, 6–7, 13, 24–27, 36–37, 39, 43, 46, 76, 184 intake 2, 39, 50, 75 meaning-based output instruction 72, 81–86, 89–90, 106–107, 177, 185–187, 191 output 1, 26, 62, 75 Processing Instruction 1–2, 17, 36–40, 55–62, 68, 74–90, 100–108, 113, 116, 118–120, 127–128, 130–131, 133–136, 143–151, 174–179, 184–194 primacy of meaning principle/strategy 3, 4, 27, 31, 46 second language acquisition (SLA), 1–3, 5–7, 9, 20, 24–26, 68–69, 71, 143 structured input activities 36–37, 42– 46, 48, 55–56, 60–62, 72, 87–88, 116–117, 121–122 traditional instruction 71, 74–81, 83–84, 89–90, 145, 177