Aspects of Language Production

Linda Wheeldon

• Publisher: Psychology Press
• Year: 2013
• ISBN:
  9781317774587
• Edition: 1
• Topic: Psychology
• Subtopic: History & Theory in Psychology
• Index: Psychology

This volume represents major research issues in language production today, presenting readers with a picture of the breadth of current research in the field. Contributors have focused on models of visual word processing, aphasic speech, object recognition and language production in children. Many chapters highlight the need for psychological models of language production to learn from theoretical linguistics in order to become better informed about the structure of language itself. Therefore, this volume also includes chapters written by linguists for psychologists which serve to remind us of the complexity of structure and process in the languages of the world.

Chapter One

Introduction

Linda Wheeldon

School of Psychology, University of Birmingham, UK

It is usual for the introduction to a volume on language production to begin by bemoaning the lack of relevant research. Traditionally in the field of psycholinguistics, the study of language production has always played the poor relation to research on language comprehension and language acquisition processes. However, happily, I think that this is no longer the case. In terms of research endeavour, language production now boasts a dedicated, imaginative, and highly productive group of researchers, which continues to grow steadily. The first aim of this volume is, therefore, to present readers with a picture of the breadth of current research in the field. To this end, there are chapters investigating all levels of the language production system from the generation of the phonological and phonetic representations of speech to the coordination of spontaneous conversation.

Nevertheless, rather than break with tradition, I do want to start this introduction with a gripe. My complaint is not about the amount or quality of the research in this field but about the acknowledgement of the data and models of language production research in other related fields. Often, a substantial chunk of language production processes is assigned to one output arrow in cognitive models of processes involving speech output. Thus, a second aim of this volume is to draw links between different approaches to language research and to argue the need for the development of models informed by the wide range of language production data that already exists in the literature. In particular contributors have focused on models of visual word processing, aphasic speech, object recognition, and language production in children.

In addition, many chapters highlight the need for psychological models of language production to learn from theoretical linguistics in order to become better informed about the structure of language itself. Indeed, in recent years, psycholinguists have made increasing use of innovations in linguistic theory to explain psychological data. These efforts have met with no little success as a number of the chapters included in this volume will testify. Nevertheless, psycholinguistic models of language processing still have a long way to go before they can provide an account for the complexity of structure and process in the languages of the world. Therefore, this volume also includes chapters written by linguists for psychologists, which serve to remind us of this fact.

My approach to compiling this volume has been simply to invite some of the foremost researchers in the field to write about what they know best. The result is a diverse set of chapters that are representative of the major research issues in language production today. This diversity, while welcome, has foiled all my attempts to impose a clear structure on this volume. However, in general, the chapter topics run from the low level processes of sound form generation and lexical retrieval to high level processes such as the generation of syntactic and conceptual structure and the coordination of conversation. In the rest of this introduction I will attempt to plot a course for the reader through these issues by summarising the chapters in order.

In Chapter 2, Bock and Griffin tackle the lack of acknowledgement of speech production data head on. They argue that the fall of stimulus-response psychology in the early 1960s lead to an implicit avoidance of all processes that result in an overt physical response, e.g., speech. They reevaluate the impressively broad literature on word perception and comprehension in the light of recent advances in the modelling of spoken word production processes. They look at implicit memory, brain imaging, and visual word recognition. In all of these areas data from tasks involving the overt articulation of words have been used to inform models of language comprehension. However, the models that have been developed to explain these data fail to acknowledge the complexity of speech output processes. Bock and Griffin argue convincingly that an acknowledgment of the involvement of production processes in these tasks is not only desirable but essential if comprehension processes are to be accurately modelled and located in the brain. In the chapters that follow, the complexity of the processes by which we generate speech is amply illustrated.

In Chapter 3, Meyer focuses on the generation of the sound form of the words we speak. She describes and evaluates what she terms the “Standard Model” of word-form encoding. The Standard Model incorporates a set of claims that are consistent with most current models of word-form encoding and that have formed the framework for more detailed computational models. The Standard Model can be derived almost exclusively from speech error data. Meyer first reviews the classic speech error data and their standard interpretation. She then evaluates the standard interpretation in the light of new data drawn from both error corpora and experimental studies. What emerges from this exercise is a clear picture of the relative security of the main assumptions that form the foundation of current theories of word-form encoding. A computational approach to some of the same issues is taken by Roelofs in Chapter 4. Roelofs describes WEAVER++, a simulated theory of both lexical selection and word-form encoding. Unlike early models of spoken word production (e.g., Dell, 1986, 1988; Shattuck-Hufnagel, 1987), WEAVER++ has been developed to encode reaction time data rather than speech error generation. Roelofs provides us with a detailed comparison of WEAVER++ and the classical spreading activation model of Dell and colleagues and discusses a wealth of experimental data for which WEAVER++ provides an account.

In Chapter 5, Nickels and Howard examine the role of neuropsychological data has in the development and testing of cognitive models of language production. In particular, they address the relevance of acquired disorders of language production to models of spoken word production concentrating on the level of single word production. They take as their framework Levelt’s model of speech production (Levelt, 1989, 1992; Levelt & Wheeldon, 1994; Levelt, Roelofs, & Meyer, 1999; Roelofs, this volume) and examine it step by step, identifying predicted impairments and discussing the limitations of the model in accounting for particular patterns of breakdown. Two things rapidly become clear from this exercise. First, that models of spoken word production can and should be informed by patterns of breakdown in aphasic speech. And second, that detailed models of language production should allow questions concerning impairments in word production to be posed with much greater precision.

In Chapter 6, Humphreys, Price, and Riddoch also focus on the production of single spoken words. They examine one of the most frequently used tasks within speech production research—the naming of pictured objects. In particular, they examine the evidence that early stages of object recognition constrain later word retrieval processes. The evidence they consider is drawn from a wide range of disciplines including experimental and neuropsychology, functional imaging, and computational modelling. They discuss data that demonstrate that name retrieval for objects is strongly influenced by visual differentiation between category exemplars. These data, they argue, present a challenge to discrete stage models of spoken word production (e.g., Levelt, 1989; Levelt et al., 1999; Roelofs, this volume). In contrast, they propose a Hierarchical Interactive Theory of object naming, which allows continuous interaction between different processing levels.

Chapters 7 and 8 signal a complete change of emphasis. These chapters are written by linguists for psychologists and provide a much needed cross-linguistic look at the complexity and diversity of linguistic structures and processes. Lahiri outlines current cross-linguistic generalisations made in phonological theory which relate to the structures and processes required to represent the phonological system of a language. She provides us with a typology of phonological processes and motivates several layers of distinct linguistic constituents. Lahiri argues that an adequate model of language production must be able to provide an account for the cross-linguistic phonological generalisations she describes, and that this will require a more detailed description of phonological representations in the lexicon than exists in current models. Lahiri also discusses these generalisations in the light of current psycholinguistic models of phonological encoding for speech production. The focus is on Levelt and colleagues’ incremental model of phonological encoding, which postulates a strict left-to-right build-up of phonological structure with minimal look ahead (Levelt, 1989, 1992; Levelt & Wheeldon, 1994; Levelt et al., 1999; Roelofs, this volume). She presents two tonal phenomena, which require information from further down stream in the utterance than incremental models would allow. Phenomena such as these provide a strong challenge to strict incrementality.

Waksler takes a similar approach to morphological processes. Until recently morphology has received scant attention in production research and the few theories which have modelled the production of complex words are limited to the production of roots with single suffixes or prefixes. As Waksler demonstrates, such forms encompass only a small subset of the morphological structures used in the world’s languages. In this chapter we are provided with a typology of the morphological systems in the languages of the world. Waksler then goes on to discuss the potential consequences of different morphological systems for language production models, in particular, the consequences for the representation of lexical entries and for the relationship between semantic, syntactic, and morphological processes.

In Chapters 9, 10, and 11 we are provided with examples of the application of linguistic theory to psychological data. In Chapter 9, I examine the form encoding of connected speech. This chapter describes a wide range of phonological phenomena that occur during connected speech production and argues that some of these phenomena arise during the generation of the prosodic (or rhythmic) structure of an utterance. In the field of linguistics, the theory of the Prosodic Hierarchy divides an utterance into a nested hierarchy of prosodic units, which need not correspond to syntactic constituents. These prosodic units form the domains of application for many phonological rules. I evaluate data relevant to the production of prosody and argue that both psychological and linguistic theory must be combined in order to build adequate models of connected speech production.

Gerken and Ohala also make use of evidence from theoretical phonology to interpret data from child language production studies. To date most studies of child language production have focused on the ways in which child forms deviate from adult forms and thus how the child’s production system differs from that of the adult speaker. In contrast, Gerken and Ohala argue that remarkably similar phenomena exist in child and adult speech. They argue that children’s deviations do not reflect processes that are specific to immature talkers but are a reflection of certain properties of a single underlying production system. They examine the tendency towards consonant cluster reduction and weak syllable omission in child language production. Their explanation of these phenomena is based largely on evidence from theoretical phonology which points to optimal syllabic and prosodic forms both across and within languages. They show that children’s omissions tend to result in forms resembling frequently occurring cross-linguistic patterns. They suggest that adherence to optimal forms may play a role in deviations in both child and adult language (i.e., speech errors or aphasic speech) and argue that we must begin to approach child and adult language production as part of a single developing system.

In Chapter 11, Ferreira reviews the experimental findings relevant to the grammatical encoding of sentences for speech. She introduces an account for these data based on an approach to syntax from the field of computational linguistics. This approach, known as tree adjoining grammars, involves the constrained combination of primitive syntactic trees. Ferreira argues that such an approach can provide an account for the available experimental data. Crucial to her argument is that prototypical trees correspond roughly to a simple clause consisting of a verb and its argument positions. The verbal concept activates its verbal lemma, and all possible elementary trees compatible with a verbal lemma are accessed. Crucially, as Ferreira demonstrates, such theoretical work has a direct relevance to experimental work in speech production in that it naturally suggests a number of testable hypothesis with regard to perennial issues in language production research such as the scope of planning during grammatical encoding.

In Chapter 12, Smith investigates the relation between the conceptual and grammatical structures generated during language production. He begins by ana-lysing rationalist models of speech production (Garrett, 1982; Roelofs, 1992) in which it is claimed that the speaker is compelled to produce a single grammatical structure that precisely replicates both the content and form of the information contained within the conceptual structure. Smith points out some of the theoretical problems entailed by such a view and marshals a broad range of linguistic and psycholinguistic evidence in support of the opposing claim that conceptual and grammatical structures contrast strongly in terms of the information they contain. Smith further argues that this contrast arises because the information within a conceptual structure possesses such a rich, analogue complexity that it cannot be fully preserved within the grammatical structure without incurring excessive formulation costs. On such a view, then, speakers minimise formulation costs by producing grammatical structures that provide only a partial representation of the conceptual structures that underlie them and thus entail much processing effort on the part of the hearer who has to reconstruct something approximating the original conceptual structure on the basis of this incomplete representation.

Finally, in Chapter 13, Fox Tree examines the factors influencing language production during spontaneous conversation. The vast majority of the speech we produce is conversational; however, this form of speech receives scant attention in the production literature. One reason for this is that spontaneous speech is messy—filled with paralinguistic phenomena such as stops and starts, ums, and long pauses. Unlike slips of the tongue, such dysfluencies have been perceived as unfortunate and uninformative elements of the speech stream. In contrast to this view, Fox Tree argues that these phenomena form an integral part of the communicative enterprise. She argues that certain paralinguistic phenomena are used to get around some of the problems inherent to the communication medium of spontaneous speech. They are necessary to coordinate conversation, achieve the grounding criterion, indicate turn units, create a coherent discourse, and warn listeners of upcoming dysfluencies and production trouble.

In summary, the chapters in this volume address a wide range of issues relevant to the cognitive modelling of language production processes. Current theories of phonological encoding, word retrieval, sentence formulation, conceptual structure, and spontaneous speech are evaluated. In addition, the relevance of these models for other areas of language research is addressed, including language comprehension, neuropsychology, language development, and linguistic theory. Finally and most importantly, what emerges from the chapters in this volume is a new set of research questions that should keep us busy well into the third millennium.

References

Dell, G.S. (1986). A spreading activation theory of retrieval in sentence production. Psychological Review, 93, 283–321.

Dell, G.S. (1988). The retrieval of phonological forms in production: Tests of predictions from a connectionist model. Journal of Memory and Language, 27, 124–142.

Garrett, M.F. (1982). Production of speech: Observations from normal and pathological language use. In A. Ellis (Ed.), Normality and pathology in cognitive functions (pp. 19–76). London: Academic Press.

Levelt, W.J.M. (1989). Speaking: From intention to articulation. Cambridge, MA: MIT Press.

Levelt, W.J.M. (1992). Accessing words in speech production: Stages, processes and representations. Cognition, 42, 1–22.

Levelt, W.J.M., Roelofs, A., & Meyer, A.S. (1999). A theory of lexical access in speech production. Behavioural and Brain Sciences, 22, 1–75.

Levelt, W.J.M., & Wheeldon, L.R. (1994). Do speakers have access to a mental syllabary? Cognition, 50, 239–269.

Roelofs, A. (1992). A spreading-activation theory of lemma retrieval in speaking. Cognition, 42, 107–142.

Shattuck-Hufnagel, S. (1987). The role of word onset consonants in speech production planning: New evidence from speech error patterns. In E. Keller & M. Gopnik (Eds.), Motor and sensory processing in language. Hillsdale, NJ: Lawrence Erlbaum Associates Inc.

Chapter Two

Producing words: How mind meets mouth

Kathryn Bock

Department of Psychology and th...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Table of Contents
• List of contributors
• Series preface
• 1. Introduction
• 2. Producing words: How mind meets mouth
• 3. Form representations in word production
• 4. WEAVER++ and other computational models of lemma retrieval and word-form encoding
• 5. When the words won’t come: Relating impairments and models of spoken word production
• 6. On the naming of objects: Evidence from cognitive neuroscience
• 7. Phonology: Structure, representation, and process
• 8. Morphological systems and structure in language production
• 9. Generating prosodic structure
• 10. Language production in children
• 11. Syntax in language production: An approach using tree-adjoining grammars
• 12. Conceptual structures in language production
• 13. Coordinating spontaneous talk
• Author index
• Subject index