Psycholinguistics

Introduction

Neobehaviorists and linguists differed in their views of language development. Neobehaviorists took words as the prime units of language, focused upon its repetitive aspects, and concluded that language is learned according to stimulus-response principles. Linguists took sentences as the unit, focused on their novelty, and concluded that language is innate. Today, most cognitive psychologists believe at least that language is specific to the species Homo sapiens. Primate research may yet modify this conviction.

The Species-Specific Nature of Language The reasons for believing in the species-specificity of language are its:

Universality Among Humans

Relationship to Physiological Factors

Relationship to Other Measures of Intelligence

Developmental Regularities

Animal Communication Language is distinguished from communication, and the evidence concerning communication among bees, dolphins, and apes is considered. The evidence does not damage the claim that language is specific to people, but it fits neatly into an evolutionary view of language and cognition.

Neobehavioristic and Information-Theory Approaches to Language Contemporary accounts of language are built on reactions to both the behaviorist and information theory approaches.

Neobehaviorist Accounts

Classical Conditioning Osgood (1953) offered an explanation of word meaning in terms of classical conditioning.

346 Instrumental (Operant) Conditioning Skinner (1957) and Staats and Staats (1963) gave book-length accounts of how operant learning principles might account for word and sentence meanings.

Mediation Conditioning theory was applied to various aspects of language by extending the concepts of stimulus and response to covert events that might mediate complex relations between observable stimuli and responses.

Linguists' Reactions to Neobehaviorist Accounts of Language Linguists invented many sentences thai embarrassed neobehaviorist theory, and psycholinguists brought some data to the argument against neobehaviorism (Brown, Cazden, & Bellugi, 1969). The data were probably less important in the demise of neobehaviorism's approach to language than were judgments based on pretheoretical beliefs.

Information-Theory Approaches to Language Following World War II, information theory was used as a guide to psychological research on language. The amount of information in linguistic strings was shown to influence memory and perception. However, the information metric could not be applied precisely to sentences or longer strings, and a more attractive alternative to neobehaviorist accounts came from new developments in linguistics.

The Influence of the Linguist Noam Chomsky

While he was a student of linguistics, Chomsky studied both behaviorism and information theory. He rejected both as approaches to language. Neither dealt satisfactorily with the fact that sentences are hierarchically organized, since both depended on left-to-right chaining. His arguments for an alternative view produced a new paradigm for linguists and gave them new theories.

Chomsky's Paradigm Chomsky distinguished language performance from language competence. Performance is what people do with their language; competence is what they know about their language. Chomsky argued that linguistic theory should explain competence, not performance. He took sentences as the units of linguistic theory, and he took the construction of sentences from smaller elements as the fact to be explained by theory. He offered generative grammar as the appropriate sort of theory, and advanced the criteria of elegance and agreement with speakers' intuitions for evaluating theory. He gave syntax the primary emphasis in theory, arguing that it could be understood without reference to the meaning conveyed in the sentences under study.

Some Adjunct Importations From Linguistics In order to test Chomsky's ideas, psychologists had to learn not only his concepts but also the notational conventions and style of argumentation used by linguists. Pertinent ones of these for psychologists are described.

Notational Conventions

Major Concepts

Chomsky's Theoretical Positions

Phrase Structure Grammar Chomsky's transformational grammars were built on a simpler model, called phrase structure grammar, which is described here.

Transformational Grammar Chomsky built two transformational grammars, both of which were intended to do more justice than phrase structure grammar to intuitions of native speakers about their language. The first transformational grammar included the notions of kernel 347sentence and transformation rule, both of which prompted considerable psychological experimentation. The second transformational grammar substituted the notions of deep structure and surface structure for the notion of kernel sentence.

The Psychology of Syntax

The Psychological Reality of Grammar Psychologists designed experiments to show that syntax and grammaticality influence memory and perception.

The Psychological Reality of Phrase Structures Psychologists did research to see whether phrase structure units are used to organize sentences for perception and memory.

The Psychological Reality of Grammatical Transformations We describe studies designed to show that the operation of transformation rules in Chomsky's theory corresponds to psychological processes. The experiments were intended to show that a sentence is stored in memory as a kernel plus transformational information, that these are forgotten independently, and that transformations take cognitive time and use up memory.

The Psychological Reality of Deep Structures It was proposed that deep structures are the storage mode of sentences, while surface structures are maintained only long enough to allow deep structure extraction. A number of memory studies show that deep structure information lasts longer than surface structure information. Experimenters also used deep structure diagrams to predict which words from a sentence would best prompt its recall; their results were mixed.

Semantics and Pragmatics

While studying syntax, psychologists discovered that many experimental results can only be explained by reference to extralinguistic factors such as context, subjects' expectations, and their knowledge. The effects of syntax are overridden by context and subject focus. Imagery was proposed as an alternative to deep structures as a storage mode for sentences, and experiments have not ruled it out.

Case Grammar

Fillmore, a linguist, proposed a transformational-generative grammar called case grammar as an alternative to the standard theory. Case grammar uses a semantic form of deep structure representation. It is attractive to psychologists, who have incorporated it in several models of comprehensive and semantic memory. Despite its intuitive appeal, it has resisted empirical test.

Summary

We emphasize conclusions that seem well-established after 20 years of psycholinguistic research.