Theories of language

Joan Phillip

Title Page
Introduction | Ch 1
Behaviourism & Psycholinguistics | Ch 2 Sociolinguistics | Ch 3 Interactionism | Ch 4 Social learning theory |
Ch 5 Schools & communities

Chapter 1 - Behaviourism & psycholinguistics in brief


Skinner's theory - language as verbal behaviour

Skinner's seminal work, Verbal Behaviour (1957), begins with a chapter called, "A functional analysis of verbal behaviour". However, you should be aware that his theory is very far from the functional, or sociocultural, approach to language, which is followed in this subject. You will also become aware that the antecedents of the sociocultural approach to language which underpin this subject, preceded the work of B. F. Skinner by several decades. Nevertheless, this section begins with an introduction to B. F. Skinner's theory of language as 'verbal behaviour' (1957). This is partly because his learning theory was transposed into language teaching methodologies prior to that transposition of the work of linguistic anthropologists and linguists to language pedagogies; and partly because Skinner's theory has had such definite, and enduring, influences on language teaching. The residual echoes of his theory can be heard every time one of us mentions 'positive reinforcement' (or 'negative reinforcement', for that matter) and his theory is operational every time one of us includes a teaching practice which begins with drills and grammar study decontextualised from meaning. Skinner rejected the very idea of 'meaning'. Skinner's view of 'meaning' can be seen in his comment which follows:

As Jespersen [a significant linguist and grammarian whose major work, Language, was published in 1922] said many years ago, "The only unimpeachable definition of a word is that it is a human habit." Unfortunately, he felt it necessary to add, "an habitual act on the part of one human individual which has, or may have, the effect of evoking some idea in the mind of another individual." Similarly, Betrand Russell asserts that "just as jumping is one class of the word 'dog' is [another] class," but he adds that words differ from other classes of bodily movements because they have "meaning". In both cases something has been added to an objective description (Skinner 1957: 13).
I draw your attention to: It is probably fair to say that Skinner did not deny the existence of "meaning"; he certainly accepted that we do ask, "What do you mean?" (Skinner 1957: 9); and he acknowledged that 'the answer is frequently helpful'. He accepted that seeking clarification, has 'an important place in every sort of intellectual endeavour'. But he went on to argue:
But the explication of verbal behaviour should not be allowed to generate a sense of scientific achievement. One has not accounted for a remark by paraphrasing "what it means" (Skinner 1957: 9).
Trying to sort out Skinner's theory of 'meaning' is very difficult because he would only consider meaning in terms of observable behaviours which might be used to account for the interconnections of variables which established and maintained the behaviour. He wrote:
The extent to which we understand verbal behaviour in a "causal" analysis is to be assessed from the extent to which we can predict the occurrence of specific instances and, eventually, from the extent to which we can produce or control such behaviour by altering the conditions under which it occurs. In representing such a goal it is helpful to keep certain specific engineering tasks in mind. How can the teacher establish the specific verbal repertoires which are the principal end-products of education? How can the therapist uncover latent verbal behaviour in a therapeutic interview? How can the writer evoke his [or her] verbal behaviour in the act of composition? how can the scientist, mathematician, or logician manipulate his verbal behaviour in productive thinking? (Skinner 1957: 3).
Skinner certainly acknowledged these questions were complex and created many problems; but I suggest that even in the acknowledgement he took a reductionist position, and one which quite explicitly places the teacher in the role of manipulator and controller. For example, he reduced learning a language to the concept of a "verbal repertoire", as though language itself is a limited, circumscribable task to be used in a performance. Then he specified someone outside of the learner (the teacher is specifically mentioned, along with "we") who will "produce or control" behaviour by "altering conditions under which it occurs". Skinner did acknowledge that, "Stimulus control,..., is never perfect", because, "Verbal behavior is probably never completely independent of the condition of a particular speaker"....nor, he seemed to suggest,  is it independent of listeners (Skinner 1957: 147), who act as potential 'reinforcers'.  In fact, Skinner suggested the "stimulus control" might be "distorted", because of some states of the speaker or audience. In that case he described the "response" as "subjective," "prejudiced," biased," or "wishful" (Skinner 1957: 147). 'Reinforcers' might be understood as 'consequences'. He began Verbal Behaviour (1957) with the paragraph:
Men [and women] act upon the world, and change it, and are changed in turn by the consequences of their action. Certain processes, which the human organism shares with other species, alter behaviour so that it achieves a safer and more useful interchange with a particular environment. When appropriate behaviour has been established, its consequences work through similar processes to keep it in force. If by chance the environment changes, old forms of behaviour disappear, while new consequences build new forms (Skinner 1957: 1). (Emphasis added.)
Sometimes undergraduate students conflate Skinner's notion of language as "verbal behaviour" with the sociocultural view of language as a "mode of action", or as "social and cultural practice". Admittedly, there are some odd overlaps between the two positions; for example, Skinner did not believe there are pre-existing meanings which we express in language. That is, Skinner did not subscribe to language as a "conduit" which expresses something outside of language. That is a view also held by sociocultural theorists. But, nevertheless, the two theoretical positions are far, far apart. Thus: It is important to know that Skinner's ideas derived from the school of behavioural psychology, which explained behaviour in terms of Stimulus (S) Response (R) and Reinforcement (R) whereby an organism was conditioned to act in a certain way. That school of psychology was, in many ways, based on a rejection of another school of psychology which drew on the processes of introspection. But even more importantly, much of the work of behavioural psychologists was to do with shaping the behaviours of animals (in confined conditions); the outcomes of which were then extrapolated to humans. Even Skinner's theory of 'verbal behaviour' was based on the incremental changes to animal's behaviour in what is sometimes called 'a Skinner box'. In his study of animal 'learning', Skinner developed the idea of operant conditioning whereby a creature, rat or pigeon, might make a chance move which was rewarded, or 'reinforced' so that it moved incrementally towards the response its controllers (my word) were targeting. So, a pigeon might eventually learn to peck the food button; a rat might learn to press more heavily on a lever to gain food because random actions which approximated the desired actions were reinforced. There was obviously a great deal of studied patience on the part of the researchers and tiny incremental changes in the creatures' behaviours.

Skinner applied these studies to 'verbal behaviour'. He wrote:

the basic processes and relations which give verbal behaviour its special characteristics are now fairly well understood. Much of the experimental work responsible for this advance has been carried out on other species, but the results have proved to be surprisingly free of species restrictions. Recent work has shown that the methods can be extended to human behaviour without serious modification. Quite apart from the possibility of extrapolating specific  experimental findings, the formulation provides a fruitful new approach to human behaviour in general, and enables us to deal more effectively with that subdivision called verbal (Skinner 1957: 3). (Emphasis added.)
The following text from an Internet site developed by an American Public Broadcasting Service (PBS) is a fair summary of Skinner's theories and how they have been extrapolated to teaching language:
With pigeons, he developed the ideas of "operant conditioning" and "shaping behavior." Unlike Pavlov's "classical conditioning," where an existing behavior (salivating for food) is shaped by associating it with a new stimulus (bell ringing), operant conditioning is the  rewarding of a partial behavior or a random act that approaches the desired behavior. Operant conditioning can be used to shape behavior. If the goal is to have a pigeon turn in a circle to the left, a reward is given for any small movement to the left. When the pigeon catches on to that, the reward is given for larger movements to the left, and so on, until the pigeon has turned a complete circle before getting the reward. Skinner compared this learning with the way children learn to talk -- they are rewarded for  making a sound that is sort of like a word until in fact they can say the word. Skinner believed other complicated tasks could be broken down in this way and taught. He even developed teaching machines so students could learn bit by bit, uncovering answers for an immediate "reward." They were quite popular for a while, but fell out of favor.  Computer-based self-instruction uses many of the principles of Skinner's technique. 

Skinner expressed no interest in understanding the human psyche. He was as strict a behaviorist as John Watson, and he sought only to determine how behavior is caused by  external forces. He believed everything we do and are is shaped by our experience of punishment and reward. He believed that the "mind" (as opposed to the brain) and other such subjective phenomena were simply matters of language; they didn't really exist (PBS/WGBH 1998  [Accessed 5 May 2001]

Structural linguistics and behaviourism

Closely related to behaviourist approaches to language are the approaches derived from Structural linguistics. Structural linguists sought to describe the unique structural elements of individual languages.  They rejected the idea of grammar as prescription and the application of the grammar of one language, for example, Latin, or German, to another, such as English (Lamberts 1960 in Wilson (Ed) 1967: 3-9).  Structural linguists aimed to identify phonological, morphological and vocabulary classes and functions of individual languages, seeking to codify the unique structural features of each language at the smallest unit of meaning. Their attention was the fine details of linguistic structures and, as such fitted easily with the "bottom-up" approaches of Behaviourist approaches to language teaching.

Structural linguists have given us a detailed account of phonological, morphological structures of languages. Although the structural elements of a  language are not the whole story, and are quite inadequate as a basis for a language teaching pedagogy (Savignon 1983), they are, nevertheless very important to language teachers. The knowledge of  phonology and morphology is one tool for identifying students' learning progress and needs in spoken and written modes.
Phonology is treated in three traditional parts:

Phonetics  -  the study of the actual vocal sounds used by speakers.

Phonemes are the smallest units of meaning in the sound systems of a language. They are the sounds which we identify with different meanings, so the bi-labial plosives [p] and [b] are formed in exactly the same way by blocking of the breath by closed lips and, but, in one is "voiceless" and the other is "voiced". If you say each sound with your fingers on your throat you will be able to feel the vibrations as you say [b]. In English the difference means a difference in meaning. However, the fricative [l] at the beginning of "light" is quite different from the fricative [l] at the end of "dull" are said in quite different ways in the mouth. In the first the tongue tip is against the ridge of the teeth, in the other the tongue is lifted up and back and pressed towards the soft palate. Say both words! One is a "light" [l] and the other is a "darl" [l]. In other languages such  difference might discriminate meanings, In English they do not.

Morphemes are the smallest units of meaning in a language. There are "free morphemes" and "bound morphemes".
Free morphemes are base words, as in "cat", "dog", "mouse", "house", "goose", "jump", "write", "sing", "walk". These words are each one morpheme.
Bound morphemes are grammatical inflections, showing, for example, plurality and tense, as in "cat-s", "dog-s", "m-i-ce", "house-s", "g-ee-se". These words are two morphemes. In irregular forms the morpheme is usually called an  infix.
Morphemes are also involved in the development of compound words and are a representation of the historical origins of words.

Such knowledge will help you identify which syntactic and phonological rules students have internalised and which they need to give further conscious attention. For example, just as in first language learning we learn the morphological pattern for forming the past tense verb, ("-ed") and regular plural forms ("-s") (and overgeneralise them) before we consistently use the irregular forms, so, too, in second language learning. Such knowledge helps to make sense of errors as stages in "acquisition" and is also useful for raising students' consciousness of which features it will be productive to attend to. However, there are also some patterns in the order in which morphemes are learnt, which are significant. (Nunan 1995: 194-196). While we might them assume that we should teach morphemic structures in the order in which we are able to use them, this seems not to be the case.


Behaviourism, language pedagogy and beyond

Behaviourism has been highly influential in language teaching. It has been used as a rationale for extreme "bottom-up" pedagogies based on the atomisation of phonology, syntax and grammar in classroom exercises. It has also been associated with heavy drilling of syntactic patterns: Skinner identified "habit" as an important aspect of learning. One of the dominant methodologies associated with Behaviourism in language teaching is 'the audio-lingual method', which is explained by Knight in your readings.



Chomsky's theory - Psycholinguistics

Twice now you have read a small quotation from Chomsky's reflections on children's aptitude for learning language. He argued that children, well before they go to school, learn language from what he believed to be exposure to language which was often fragmentary and incomplete  (Chomsky in Cashdan et al (eds) 1972:132).  He observed that the children seemed to have an innate ability to learn language, and that what they did was deduce linguistics rules, which they would over generalise on their way to mastery of the language. We have all heard the examples which caught Chomsky's attention, such as 'mouses', 'mices', 'runned' 'digged'; but while we may have just thought these utterances were a rather sweet feature of early childhood language Chomsky postulated they revealed something about the nature of the human mind's structure. His explanatory metaphor, or model, for these patterns in children's language development was the "Language Acquisition Device" (LAD) (McNeill 1972: 146).


The Language Acquisition Device (LAD)

Chomsky's theory was influenced by the work of Wilhelm von Humboldt. Chomsky wrote:

His [von Humboldt's] theory of speech perception supposes a generative system of rules that underlies speech production as well as its interpretation. The system is generative in that it makes infinite use of finite means. He regards a language as a structure of forms and concepts based on arrangement, and organization. But these finite materials can be combined to make a never-ending product (Chomsky in Cashdan et al (eds) 1972:129).
Chomsky also drew on neurophysiological studies of other species, such as studies of the eye structures of cats and the auditory systems of frogs which identified specific "cortical centers" related to cat's eye and frog's ear functions. Chomsky wrote:
These and other studies [of, for example, studies of human infants responses to faces] make it reasonable to inquire into the possibility that complex intellectual structures are determined narrowly by innate mental organization (Chomsky in Cashdan et al (eds) 1972:130).
Chomsky was fascinated by the speed at which most children "acquire" language.  He saw the child's process as "a kind of theory construction". Chomsky argued that the child's language acquisition could only be possible  if humans possess an "innate mental organization" devoted specifically to language. His conclusions are based on:
Linguistic Universals

The "innate mental organization" which Chomsky was to call the "deep grammar" of the language is the basis for all languages. That is, the "innate mental organization" is a universal human inheritance, just the "surface grammar", the syntactic and phonetic structures will be different.

He argued that:

A language involves a set of semantic-phonetic percepts, of sound-meaning correlations, the correlations being determined by the kind of intervening syntactic structures just illustrated [whether, for example, the language is English or Japanese] (Chomsky in Cashdan et al (eds) 1972:131).
Of course, if Chomsky's theory of an innate Language Acquisition Device (LAD) is right, then obviously  it applies to all humans, to all languages, and we must be able to identify linguistic universals (Chomsky in Cashdan et al (eds) 1972:134). The existence, or not,  of "linguistic universals" is an issue in psycholinguistic theory.

There are a number of problems with the idea of language universals (Hockett 1966), nevertheless there are a number of grammatical and phonetic universals which Hockett identifies, a sample of which are listed below:

These "universals of language" would seem to support Chomsky's contention about the innate language structures of the mind. However, there are other explanations of the phenomenon which will be considered later. For example, Whorf (1983 in Hodge (ed.)) argued that linguists find language universals because the language they study "are all Indo-European dialects cut to the same basic plan, being historically transmitted from what was long ago one speech community". Whorf turned to the indigenous North American languages and went so far as to argue that the "noun"- "verb" categories do not always exist. However, there have been more recent studies of languages such as Hopi and Nootka which dispute aspects of Whorf's grammars (Hockett 1963; Pinker 1995). Whorf's arguments were to do with his theory of the relationships between language and culture and the unique features of thought which the each language engendered.



In any case "the innate mental organization" which Chomsky argued made possible the extraordinary language learning achievements of young children are "the system of rules that underlies the normal use of language" (Chomsky in Cashdan et al (eds) 1972:134). As he developed his theory of these underpinning linguistic rules, Chomsky used native speakers' judgements about the grammatical, or non-grammatical nature, of utterances. He tested the existence of grammatical awareness by gathering responses to a small number of grammatical and ungrammatical sentences. The most famous grammatical-ungrammatical pair is:

1.    Colorless green ideas sleep furiously (G)
2.    Furiously sleep ideas green colorless (U) (Hill 1961 in Wilson (Ed) 1967: 281).

According to Chomsky it is possible to deduce the innate linguistic structures of the mind through developing a system of finite rules from which an infinite number of grammatical sentences might be generated. He was searching for rules of transformation. He wrote:


I will consider a language to be a set (finite or infinite) of sentences, each finite in length and constructed out of a (sic) finite sets of elements. All natural languages in their spoken or written form are languages in this sense, since each natural language has a finite number of phonemes...and each sentence is representable  as a finite sequence of these phonemes...though there are infinitely many sentences....The fundamental aim in the linguistic analysis of a language L is to separate the grammatical sequences which are the sentences of  L from the ungrammatical ones. One way to test the adequacy of a grammar proposed for L is to determine whether or not the sentences that it generates are actually grammatical, i.e.,  acceptable to a native speaker (Chomsky 1957 in Wilson (Ed) 1967: 276).  (Emphasis added.)


Generative-Transformational Grammar

Another way of thinking about Chomsky's search is a search for the rules that link what he called "deep structures", or "deep grammar", with which he postulated we are born, and the "surface structures", or "surface grammar" which we use. Chomsky came to the conclusion that there are four simple, declarative, sentences which are the kernel sentences of the grammar. By applying a set of very specific transformational rules to these sentences an infinite number of grammatical sentences might be generated. These transformational rules Chomsky identifies as the grammar of the language.

Thomas explains the "cognitive processes" in the transformations of the kernel sentences:


we must consider his definition of grammar: a grammar is a device for generating the sentences of a language....   
A kernel sentence is "simple, active, declarative" and Chomsky feels that "all other sentences" are derived from kernel sentences by means of "transformations". Roughly, a "transformation" is a rule that either introduces new elements into kernel sentences (e.g., adjectives, negatives), or rearranges the elements of a kernel sentence (e.g., to produce an interrogative sentence), or both (e.g., to produce a passive sentence)...Chomsky's "kernel sentence" bears a strong resemblance to the simple "subject-verb-complement" sentence of traditional grammar. He states that a kernel sentence is composed of a "noun phrase plus a verb phrase." A "noun phrase (symbol: NP) consists simply of an article (T) plus a noun (N), and the presence of the article is optional. A "verb Phrase" (VP) consists of an auxiliary (Aux) plus a main verb (V) plus a noun phrase (and this last "Noun Phrase" is of course similar to the traditional "complement") These transformations...are invariable. derive a passive sentence, we first need a kernel "string" containing the following elements: a noun phrase (NP), an auxiliary (Aux) a verb (V), and second noun phrase (NP). These might be represented as follows:   

[NP1] + [Aux]+[V]+[NP2]   

To transform this string into a "passive string," the four basic elements are invariably added as follows:   

[NP] + {Aux]  + be + en +[ V] + by +[NP1]   

(The "en" which is added is the so-called "past participle morpheme") (Thomas 1962 in Wilson (Ed) 1967: 197-198).

That is, as Wilson argues:

transformational-generative grammar assumes that within an English sentence, however complex, we can observe a deep structure represented by one or more of a finite list of kernel sentences, all of which are simple and declarative (Wilson in Wilson (Ed) 1967: 275). 


Competence and performance

From the point of view of language teaching pedagogy a major influence of Chomsky's theory derived from its stress on our innate ability to learn language. Lechte writes:

[Chomsky] brought about a reconsideration of language learning by arguing that language competence is not acquired inductively through a behaviourist stimulus-response conditioning, but is the consequence of an innate cognitive capacity possessed by humans  (Lechte 1994 in Bucsescu 1996 Naom Chomsky. (Accessed 19 February 2001)
This innate ability to generate and recognise grammatical, or "well formed"  sentences Chomsky identified as our linguistic competence. This position was in contrast with Behaviourist learning theories and was to have far reaching effects on language curriculum and syllabus directions. The concept related to competence is performance. Performance, or actual language usage, is the corpus of  material upon which Chomsky based his theory.

That is, language for Chomsky is a cognitive capacity (competence) and usage (performance) is the evidence of that capacity. Chomsky's theory of an innate language learning device (LAD) has had far-reaching, problematic, influences on language teaching practices, however it should be noted that Chomsky did not specify any kind of educational theory. His theory was a philosophical exploration of the relationship between language and the mind, not as a source of thought but as a neurological organisation for cognitive transformations of linguistic structures.


Psycholinguistics - influences on language pedagogy

Negative influences

Ruqaiya Hasan (1999 in Christie (ed) 1999:13) classifies "Chomskian linguistics" as an "endotropic" theory, a theory isolated from others "where language as a mental organ is simply a variety of biological phenomenon having no connection with anything else". Hasan's description is, perhaps, harsh; Chomsky certainly identified a function of "creativity" in the transformational rules which humans use to generate unknown, unheard, grammatical utterances, but his theory remains cognitivist. His data was extremely limited; he did not even spend time studying how parents really interact with their babies and young children, simply because it was not his interest: he was bound by his interest in the "mind" as some kind of organ separate from the person's embodiment as a social being. From Plato onwards in Western thought there has been a discourse which represents the body and the mind as dichotomous. People have written about their body as though they were somewhere else; as though they were a real presence elsewhere. It is very recently that philosophers have grappled with this oddness, and have explained the way language constructs our embodied subjectivities firmly within culturally and socially specific situations (Gatens 1986; Threadgold ed. 1990; Grosz 1994).

But at the time Chomsky's theory had profound influences on language teaching. Perhaps the most dramatic influence can be seen in the removal of explicit grammatical knowledge from language teaching curricula: if children have an innate creative ability to learn the grammatical rules of a language without any explicit teaching of grammatical categories, then such knowledge was thought unnecessary. In NSW that meant twenty years (1974-1994) without the teaching of any kind of metalanguage.

The period of Chomskian influence, is associated with the implementation of a "progressive pedagogy" (Christie 1992; Martin 1993; Rothery 1996; Martin 2000). This is a pedagogy which focuses on the individual, who will learn through immersion in classroom activities, supported by the teacher as guide and facilitator. The child is seen as the possessor of a mental capacity for language acquisition/development, including choices of "expressive" activities. This was the limitation of the "progressive", "process" and "whole language" pedagogies. In spite of the fact that the models acknowledge the importance of "meaning" and "communication" their underlying "nativist" view of learning, which can be traced partly to Chomsky's views of "innate mental organizations",  limit direct teaching of language issues. Furthermore, emphasis on the individual's capacity for language learning hides the potential significance of different social and cultural backgrounds in language learning outcomes.

Rose (2000) identifies the problems of "progressive pedagogies" in indigenous language teaching in Northern Australia with Chomsky's concept of "competence". He describes a "competence" pedagogic model as about "actualizing innate competencies" through "invisible" pedagogic practices. Rose opposes this kind of pedagogy to the one he favours which he classifies as a "performance" pedagogy. This pedagogy emphasises "acquiring skills for text production", it is "visible" and its modes are "vocational/professional". This latter view accords with those of Martin Nakata (Chapter 5 in your text, Cope and Kalantzis 2000).

Chomsky's theory, with its rejection of structuralist linguistics and Behaviourist pedagogies with their narrow emphasis on habit formation and carefully sequenced drill exercises, also encouraged almost phobic judgements of any activity that seemed like a memorisation task. These tasks were generally judged to be rote learning. And yet, as Biggs (1994) suggests of Chinese students, what looks like rote learning might rather be seen as rehearsal for prompt retrieval. There is no way we can learn a second, or other, language without some memorisation facility. While the material of the memorisation needs to be embedded in meaningful language situations the memorisation activities are a shortcut for the multitudinous linguistic experiences from which L1 speakers develop their lexicosyntactic knowledge. As you may know, or shall see in your readings there are many differences between L1 and L2/3... learning.


Positive influences

On the other hand there are some positive applications of psycholinguistics on language teaching pedagogies. One of these applications is to do with drawing the distinction for students between the processes of "learning" and "acquiring" a language (Krashen 1988; Gee 1992). This distinction can be a positive explanation for students who feel frustrated or ashamed of inevitable errors. It can be supportive for students to see "learning" a language as a conscious learning of patterns of the language, across which unconscious "acquisition" processes inevitably take place as their "innate mental organization" extrapolates and applies the grammatical rules of the language, bringing periods marked by specific, and virtually universal, error patterns. In support of this position some research has indicated a number of morphological rules appear to be learnt in a specific order (Nunan 1995: 147). It is possible that the order of acquisition is to do with innate cognitive processing to do with memory constraints and the complexity of the transformation involved, as has been argued is the case in L1 development (Brown and Bellugi 1972: 142-144; Brown 1973: 295-325). Although some of the "findings" may also be to do with the kinds of instruments used in the research (Nunan 1995: 147) "subsequent research has provided substantial evidence that certain grammatical items appear in predetermined sequences" (Nunan 1995: 147).

Such information may support students by showing them that systemic errors are signs of development; the information may also be applied to help them focus upon their particular developmental errors, as a kind of consciousness raising (CR) (Nunan: 1995: 149-151) process.

Furthermore, if transformations are significant, then designing some quick games which require transformational processing can provide useful cognitive activities which also energise classrooms at transition points.