The Handbook of Psycholinguistics

Eva M. Fernández, Helen Smith Cairns, Eva M. Fernández, Helen Smith Cairns

• Publisher: Wiley-Blackwell
• Year: 2017
• Print ISBN: 9781119096528, 9781118829509
• eBook ISBN: 9781118829554
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Linguistics
• Index: Languages & Linguistics

Incorporating approaches from linguistics and psychology, The Handbook of Psycholinguistics explores language processing and language acquisition from an array of perspectives and features cutting edge research from cognitive science, neuroscience, and other related fields.

The Handbook provides readers with a comprehensive review of the current state of the field, with an emphasis on research trends most likely to determine the shape of psycholinguistics in the years ahead. The chapters are organized into three parts, corresponding to the major areas of psycholinguists: production, comprehension, and acquisition. The collection of chapters, written by a team of international scholars, incorporates multilingual populations and neurolinguistic dimensions. Each of the three sections also features an overview chapter in which readers are introduced to the different theoretical perspectives guiding research in the area covered in that section.

Timely, comprehensive, and authoritative, The Handbook of Psycholingui s tics is a valuable addition to the reference shelves of researchers in psychology, linguistics, and cognitive science, as well as advanced undergraduates and graduate students interested in how language works in the human mind and how language is acquired.

Part I
Production

1
Overview

FERNANDA FERREIRA

University of California, Davis

As a person speaks, a great deal of processing activity is taking place behind the scenes. In the situation of some sort of dialogue or conversation, a speaker will grab one of the many ideas activated in working memory and commit it to the processes of linguistic formulation. This thought must then be translated into specific concepts, the words that express those concepts must be retrieved, the words must be organized into a structure that communicates the thought effectively, and the entire plan must be converted to a phonological representation (either in speech or sign) that will allow the utterance eventually to be articulated. At the same time, speakers must take into account the needs of their conversational partners. They also need to keep track of how the communication is unfolding: That is, they must consider the knowledge their interlocutors bring to the conversation, how their common ground is incrementally being built up as the interaction proceeds, and how effective their contributions are given the goals of the exchange. The chapters in this section touch on many of these important ideas, spelling out in detail how syntactic structures are generated, how redundancy and givenness are conveyed, how multiple language systems are coordinated, and how conversations are managed. These chapters make clear the enormous progress that has been made over the last 50 years in uncovering the architecture of the production system as well as the systems with which it interacts.

Generating a syntactic structure

The syntactic level of representation takes center stage in research on language production. As Bock pointed out decades ago (Bock, 1982), it is at least coherent to ask whether people build a syntactic level of representation to mediate between form and meaning during comprehension, in part because, for many semantically constrained sentences, a “bag of words” approach will be sufficient to establish the underlying event structure specifying who did what to whom. But when it comes to production, you can’t fake syntax. If you attempt to speak a language with grammatical agreement, it will be painfully obvious when you err. If you do not understand the rules of word order, the result may be awkward at best and misleading at worst. For these and other reasons, theoretical models of production have always taken the problems of syntactic planning seriously. In speech error models of production like the ones proposed by Garrett and Fromkin (Fromkin, 1971; Garrett, 1975), the first stage of grammatical encoding (in Garrett’s model, the functional level) is the entry point into the language system proper, as distinct from the general‐purpose conceptual system that supports any type of perceptual‐motor encoding. Syntax cannot be ignored in theories of language production, and research has proceeded accordingly.

Franck thus opens up this production section with a careful, analytical discussion of the syntactic system in general, and then moves quickly to what she treats as the model process for syntactic planning, which is agreement. As Franck notes, agreement is genuinely syntactic: It is a formal mechanism for linking words across often widely separated sentence positions. Franck begins with the studies by Bock and colleagues that initiated this research program, which used agreement errors as a source of information about the nature of syntactic representations. As Franck also points out, this early work seemed to reinforce Garrett’s fundamental assumptions about the architecture of production, which mapped the modularity of representations onto strictly serial processing. Speakers were assumed to first plan the meaning of what they would say, then generate the appropriate syntactic forms, and then engage in phonological planning.

This early work was critical for fostering a discussion about the consequences of this architecture for people’s ability to make syntactic decisions online, and as one hopes to see in a field of inquiry in which concepts are specified in enough detail to be falsifiable, some of the key assumptions did not survive tough empirical scrutiny. Franck summarizes the problematic results, many of which came from her own research on agreement, done in collaboration with her colleagues. These challenges to what Bock and colleagues referred to as the Marking and Morphing model motivated the development of Franck and colleagues’ alternative Feature and Controller Selection model, which differs from the Marking and Morphing model in a number of key respects. Perhaps the most central is the separation of the stage responsible for selecting features relevant to agreement from the stage that identifies the relevant controller of agreement. Semantic, syntactic, and morphological features influence both stages, but in different ways. This model also tightly links syntactic structures and memory representations, since the selection of both features and controllers is strongly influenced by the availability of information in working memory. In addition, the model assumes that the more prominent the syntactic position of a word, the greater its accessibility in memory. Franck ends her chapter with a useful roadmap for future research on syntax in production.

But what does this model of agreement imply for the concept of incremental production, for example? In traditional models of language production, a key question concerned the planning units for any level of representation, including for syntax. Do speakers plan an entire clause before beginning to speak, or are planning and execution processes cascaded? In the early days of production research, many papers were published on this subject (e.g., Ford & Holmes, 1978), with evidence suggesting clausal planning units for grammatical encoding. Late twentieth‐century models such as Levelt’s (1993) moved away from this idea and toward incrementality with the suggestion that planning domains should be as small as possible, and perhaps no larger than a single word. But how do we reconcile this incremental approach with the facts concerning agreement, where, as Frank notes, the controller and the form with which it agrees could in principle be indefinitely separated (and in practice are often separated by several words)? In fact, in recent years, the pendulum has begun to swing back toward the view that planning units for syntax are probably at least phrasal (Allum & Wheeldon, 2007; Bock & Cutting, 1992; Ferreira, 2000; Martin, Crowther, Knight, Tamborello, & Yang, 2010), and that the size of those units are likely not architecturally determined, but instead vary depending on the goals of the speaker (Ferreira & Swets, 2002; Wagner, Jescheniak, & Schriefers, 2010). It would be interesting to know, then, how findings concerning agreement speak to this question of planning units for grammatical encoding in more detail.

Distributing information

Typically, the same idea can be linguistically conveyed in more than one way, which presents the production system with both an opportunity and a set of processing decisions. This issue of flexibility in production connects to the previous discussion concerning how syntactic structures are generated, because one of the tasks of the production system is to make syntactic choices such as whether to produce an active or passive sentence, or whether to include an optional element such as the complementizer that in a sentence. In addition, speakers vary the way they pronounce the same word depending on features such as familiarity as well as predictability. The chapter by Jaeger and Buz focuses on the phenomenon of reduction, and probabilistic reduction in particular, which they link to contextual predictability. The general idea is that the more expected something is, the more reduced will be its pronunciation.

Jaeger and Buz then link the phenomenon of reduction to three general accounts of production: one that emphasizes ease of production, another that emphasizes the facilitation of communication, and one that emphasizes representational issues. The first account they consider assumes that reduction occurs because it facilitates the job of the speaker. The second account links reductions to speakers’ attempts to make the task of the listener easier. And the third account attempts to connect phenomena of language change to online language production. Jaeger and Buz conclude by arguing that all three influences must play a role in explaining the robust, cross‐linguistically attested tendency on the part of speakers to reduce predictable forms. As they point out at the end of their chapter, an important question that remains to be answered is precisely how these three approaches mesh with one another. Another open question concerns omission of linguistic forms, which is also related to predictability. For example, in a null subject language, the likelihood that a speaker will produce an overt pronoun or leave the position null depends on the predictability of the corresponding referent. Is omission simply the extreme case of reduction, or does the speaker face a binary choice in cases such as these? This would seem to be an important question for future work, as Jaeger and Buz note.

Bilingualism, multilingualism, and signing

The majority of people live with more than one language system, and so a critical question for theories of language production concerns how these different databases of information are coordinated and managed. Paolieri, Morales, and Bajo’s chapter examines this issue in detail. They present the problem as follows: How do speakers choose between two forms from different languages that express the same idea? One class of models of bilingual production assumes selective access, so that the decision to speak in a given language effectively shuts off any other languages the speaker might know. In contrast, in nonselective models, forms across languages interact, potentially leading to competition and interference. Linking back to Frank’s interesting chapter, Paolieri et al. (this volume) go beyond standard evidence for lexical interference effects to highlight findings concerning grammatical gender interactions across languages as well. The so‐called gender congruency effect is observed when words in different languages happen t...

Table of contents

• Cover
• Title Page
• Table of Contents
• Notes on Contributors
• Prologue
• Part I: Production
• Part II: Comprehension
• Part III: Acquisition
• Index
• End User License Agreement