A BRIEF HISTORY
Interest in the mind and language both date back for millennia, with a documented history of language study going back 2,500 years and spread across many cultures (including India, China, Mesopotamia, and Greece). Documented interest in the mind and knowledge—the foundations of what we consider to be psychology—also dates back, in one form or another, at least as early as language study, and perhaps earlier. However, modern versions of both linguistics and psychology are much more recent, with modern psychology tracing back to Wilhelm Wundt’s lab in Leipzig in the late 1800s and modern linguistics tracing back to roughly the same time. Both fields have undergone some revolutions in even that relatively short time, with both fields experiencing some major shifts in the mid1900s that are still felt today.
From the beginning of psychology, there has been an interest in language. Wilhelm Wundt, for example, published a book on language (die Sprache) in 1900. This book, with 1,367 pages by its 1913 edition, covered a number of topics that are still very much relevant in current psycholinguistics, including child language acquisition, sign language, language perception, and grammatical structure. The interest between the domains of language and psychology was mutual and, as Blumenthal (1987) discussed, many linguists of the day were also interested in Wundt’s work and attended his lectures at the University of Leipzig, including several influential language researchers such as Leonard Bloomfield, Ferdinand de Saussure, and Edward Boas. For example, Bloomfield’s approach to analyzing the structure of language is important to almost all modern theories of grammar. De Saussure made a critical distinction that is still part of how language researchers think about language: langue, which is the knowledge of a language system that exists collectively among speakers of a language, and parole, which is the use of that system.
While Wundt was an experimentalist, he also acknowledged the importance of internal mental states and viewed language as importantly reflecting mental representations. He viewed the sentence as a key unit of language, and sought to show how universal characteristics of human information processing, like attention and memory, would influence its production and comprehension. This is really not so different from what modern psycholinguistic researchers are doing, though the route from Wundt’s experiments to the present day is not entirely direct.
Wundt was not the only researcher at that time interested in language, of course, and in fact there was some conflict between Wundt’s approach and others, including approaches that disagreed with the idea of a unified mental representation and other types of “introspective” approaches. In fact, at the turn of the century there were quite a number of competing perspectives on psychology, which led one linguist of the day, Bernard Delburck, to suggest that linguists might do better to part ways with psychology. This is largely what happened, and for the next 30 to 40 years linguistics focused instead on the formal aspects of language—sound systems, grammatical structures, word formation rules—without much reference to the mental processing needed for their actual use. This approach still forms a core aspect of linguistic study today.
Although Wundt’s work clearly foreshadows modern views and topics on psycholinguistics, his influence declined following the First World War. Bloomfield, once a proponent of Wundt’s approach, had turned to behaviourism instead by 1933 when he published one of his major contributions to linguistics, a book simply called Language. In psychology, behaviourism was a movement in which the study of mental states was more or less rejected, and the idea that one could account for human behaviour in terms of mental states or representation was discounted. In linguistic terms, this meant a stronger focus on descriptive accounts of language rather than studying language as a window onto human mind. Perhaps the most famous attempt to account for language processing in a behaviourist tradition comes from B.F. Skinner’s 1957 Verbal Behaviour in which language is not a complex mental construct with rules and representations, but instead is reduced to, well, verbal behaviour. As such, it can be explained, according to Skinner, in terms of the same conditioning theory that applied to other behaviours, such as classical and operant conditioning, in which links between stimuli and outcomes are formed and shaped by experience. For example, a child saying “I want milk” may result in the child receiving a glass of milk, and this reinforces (or conditions) the use of this verbal behaviour.
The trouble was that there are a number of aspects of language that cannot be explained by classical and operant conditioning. In a famous critique of Skinner’s book, Noam Chomsky (1957) successfully argued against verbal behaviourism with several key points. Crucially, as we shall see in chapter 2, language is recursive and can produce an infinite number of sentences from a finite set of systematic rules and representations. The complexities of language that these rules create are difficult (if not impossible) to account for in simple stimulus–response terms. As part of his arguments, Chomsky reintroduced the idea of mental representations back to the study of language. He also drew an important distinction between the knowledge that one has about a language, called “competence” and the use of the language, “performance” (similar to the distinction of langue and parole drawn by Ferdinand de Saussure roughly 60 years earlier). Chomsky’s influence on modern linguistics and psycholinguistics is profound, and his focus on competence (as opposed to performance) drew linguistics heavily in this direction. On the other hand, psychology continued to be quite interested in the concept of language performance. Nonetheless, several of Chomsky’s proposals about the nature of syntactic structure, and in particular his work on transformation grammar, prompted experimentation by psychologists in the 1960s to see whether the linguistic processes proposed were psychological processes. For example, one could reasonably ask whether structures that were proposed to be more complex linguistically would cause longer processing times. The results were mixed—research showed that there was an important relationship between linguistic structure and psychological processing, but didn’t support the particular relationship proposed by transformational grammar, despite initial successes. Also, it became increasingly clear that this distinction between competence and performance was not trivial, and that the competence theories proposed by linguists could not simply be transferred to performance. Another difficulty was that linguistic theories were changing rapidly and that made it more difficult for psychologists to test them.
As a result, there was relatively little interaction between the study of psychology and linguistics for the next couple of decades. Psychologists were still interested in language—very much so—but focused more on issues of performance, such as the processes by which syntactic structures are constructed in real time, how ambiguities in language are resolved, and how word knowledge is accessed upon encountering a word. Linguists were still interested in language as a mental phenomenon, but focused on issues of competence—what knowledge of a language entails, and formulating theories that could apply to all languages, regardless of the apparent differences among them.
The separation between these two fields was not to last. Starting in the late 1980s and early 1990s, there was a renewed interest in psycholinguistics as a joint venture between linguists and psychologists, and these days many researchers have been trained in both disciplines (as well as other fields). Substantial advances in related disciplines, including neuroscience, computer science, and cognitive science have given researchers interested in language processing a huge new set of resources, both in terms of knowledge and tools. We now have researchers working on computational models of language processing, informed by current knowledge about how the brain works. We have researchers working at the intersection of language and other cognitive abilities, including not just working memory, but things like scene perception and reasoning.
Psycholinguistics is concerned with the relationship between language and the mind. This distinguishes it from sociolinguistics, on the one hand, where the focus is on the social dimension of language, and stylistics, on the other, where it is on the expressive functions of language. Psycholinguistics explores the psychological processes involved in using language. It asks how we store words and syntactic structures in the brain, what processes of memory are involved, and how we understand and produce speech. These are all of considerable practical importance when it comes to understanding language disorders. But above all, psycholinguists are interested in the acquisition of language: with how children learn.
Many linguists feel that if we can understand the internal mechanism which enables children to learn language so quickly we shall have penetrated one of the deepest secrets of the mind. To what extent are humans programmed from birth to acquire language? Is there such a thing as a language gene? Or is it simply that we have a general cognitive, or mental, ability that enables us to pick up language quickly? All of these issues are part of an ongoing debate within linguistics. Currently, the genetic view of language ability holds the field. Book by the psycholinguist Steven Pinker, entitled The Language Instinct (1995), makes a strong case for considering the elements of linguistic knowledge to be innate. This fits in very neatly with the Chomskyan concept of universal grammar: the idea that there is a common underlying structure to every language, the knowledge of which we are born with.
Psycholinguistics, then, is at the theoretical cutting edge of linguistics and, as such, is pretty heady stuff. So the question is, how can we begin studying it? First of all, we can be encouraged by the fact that much of the recent literature on the subject is very accessible. There is a strong tradition within linguistics of popularising the results of research in ways that demand little previous knowledge. The work of Pinker, mentioned above, and, in particular, Jean Aitchison, provide excellent ways into the subject. In these works one can find discussions of the various methods by which psycholinguists gather their evidence and how they set about analysing it. Secondly, as with sociolinguistics, one can carry out simple observational tasks oneself.
The most effective way to do this is to observe and monitor the speech of one or two young children over a period of time. You need to have in mind, of course, what you are looking for and the purpose of the activity. Your initial concern is to identify distinctive usages, either in sounds, syntax, or word meanings. You will be surprised in doing so how much of children’s speech you have hitherto taken for granted. The next stage is to establish what kind of rule your informants are following in producing these usages. Psycholinguistics proceeds on the principle that children’s use of language is rule-governed. You could start with observing how children form the plural and the past tense. These probably comprise the most conspicuous ‘errors’ in childhood speech. Young children will frequently say tooths and mouses, instead of teeth and mice, and holded and finded, instead of held and found. These are examples of over-generalisation the extension of a rule beyond its proper limits. In these cases the child knows the regular rule for forming the plural and the past tense but doesn’t know that these particular words are irregular. Having established the presence of this phenomenon, you can then test to see whether all irregular forms are regularised or only some, and how long it takes a child when corrected to acquire the correct form. It’s on the basis of experiments like these that psycholinguists form hypotheses about how children memorise forms and self-correct.
Over-generalisation is a frequent phenomenon in language development. It can be found not only in syntactic usage but also in word meanings. Many young children will sometimes refer to all animals as dogs or call all vehicles cars, and perhaps more disconcertingly, all men, dad. Discovering the limits of these words, what they do, and do not, apply to, is a useful way of penetrating the child’s semantic system. It can take time, for example, for children to learn that words can refer to separate things. When a child refers to milk, for instance, does s/he mean the whole process of pouring it into a mug and placing it down, or does it have the restricted meaning we are used to? Children also under-generalise; indeed, undergeneralisation is probably a more frequent phenomenon than its counterpart. A child may often only be able to use words in a particular context. It’s not uncommon for children to call their own shoes shoes but not know what someone else’s are called.
An initial way into psycholinguistics is to carry out some field research of your own into the acquisition of language, using a couple of basic concepts as your guides. On the basis of this, you can speculate about the kinds of lexical, syntactic, and semantic knowledge which your informants possess. If you do this it will enrich your understanding of the linguistic literature which you read. You will also find that it adds to your knowledge of how language changes; because all of us under- and over-generalise. Over-generalisation is one of the processes behind the loss of inflections from Anglo-Saxon times; we used to have many more irregular forms then than we do now. The morphology of modern English has developed as a consequence of generalising particular ways of forming the plural and past tense into regular paradigms. And it is also a key process in dialectal change. People who say I loves him are generalising the rule for the third person singular to cover all forms of the present tense. And in using a word like deer with its modern meaning we are under-generalising it: its Anglo-Saxon original, deor, meant an animal.
Having begun in a fairly simple way you can extend the process and consider more complex aspects of language acquisition: the formation of the negative, for instance. It takes some time for children to acquire the specific rule about attaching the negative to the auxiliary verb. Initially they will tend to put it at the beginning of the word string: no Jenny have it. Later the child decides to put the negative after the first noun phrase: cat no drink; he no throw it. The interesting thing about these rules is that the child cannot have acquired them from listening to adult discourse. They have been generated from scratch. And yet they are commonly followed by most children. Are they then a representation of some internal grammar in the child’s brain? Does the child start out with a set of possibilities for the formation of the negative and narrow them down as s/he encounters confirmation or disconfirmation from the speech of others? Questions like these form the basis of much psycholinguistic enquiry. It’s impossible to see directly into the brain so all we have is the second-hand evidence of language to work on. Over the years psycholinguists have amassed a good deal of observational data and case history analysis, all of which you can work through in time, but it is no substitute at the outset for making your own observations, and for using your linguistic knowledge to speculate about how we manage what is, arguably, the most amazing learning feat of our lives.