Integrationism: a very brief introduction

   •Roy Harris

1. Historical background.
2. Language as a faculty.
3. Integration.
4. Language and context.
5. Speech and writing.
6. Languages and rules.
7. Language and other faculties.
8. Aims of integrational linguistics.
9. A personal view.


1. Historical background.

1a. There is a long tradition in Western civilization of studying language for reasons other than its intrinsic interest. This tradition began in antiquity. Plato was interested in language mainly as a means of access to eternal truths. Aristotle was concerned with language primarily as a necessary tool for winning arguments. Dionysius Thrax focussed on describing Greek grammar for the benefit of learners and students of literature.

1b. In the age of Descartes, the scholars of Port-Royal studied language as a manifestation of human reason (Grammaire générale et raisonnée, 1660). This kind of thinking about language lasted well into the 20th century, especially in academic philosophy. (It is still evident, for example, in Bertrand Russell’s Inquiry into Meaning and Truth, 1940.)

1c. In the 19th century, the Neogrammarian school identified the study of language with the study of languages; in particular, the Indo-European languages, their ‘dialects’ and their history. This kind of approach also lasted well into the 20th century, and was given canonical status by the ‘general linguistics’ of Saussure, who accorded priority to the study of la langue over le langage (Cours de linguistique générale, 1916). Generative linguistics, as developed by Chomsky and his followers 50 years later, was primarily an attempt to find a niche for the study of la langue (misleadingly renamed ‘linguistic competence’) under the new aegis of ‘cognitive science’ (see 1e). Their main innovative move was to construe individual languages as quasi-mathematical systems of linguistic ‘rules’ or algorithms (see 1e and 7b), unconsciously mastered by speakers.

1d. Under the growing academic influence of anthropology and sociology, it became fashionable in the course of the 20th century to regard languages and their development as public barometers of the development and structure of society. Scholars interested more particularly in these aspects of speech and writing imported into their studies classifications and categories borrowed from social theory (such as the social class, income and occupation of speakers) and tended to equate the study of language with the study of its forms of social differentiation. The umbrella term for these and related studies was sociolinguistics.

1e. After WW2, psychologists announced the advent of a linguistic component in what was declared to be ‘the mind’s new science’ (H. Gardner, The Mind’s New Science, 1985). This ‘new science’ (baptised ‘cognitive science’ by its practitioners) was based on treating the digital computer as a model of the mind. Thinking in all its forms, linguistic and non-linguistic, was alleged to rely on ‘programs’ analogous to those by which a computer engages in ‘information processing’.

1f. All the approaches listed above reflect aspects of speech and writing singled out for attention by scholars who were coming from different intellectual directions, but who did not have as their main concern analysing the way language makes human life what it is for the individual human being. Rather, these different approaches take that for granted, instead of throwing any light on it. Consequently, they tend to obscure – even to divert attention from – the features that, for the overwhelming majority of people, characterize the lay exercise of one’s linguistic abilities in daily affairs.

1g. There is another approach, which should not be mistaken for any of those mentioned above, and which also emerged in the late 20th century. It was in part a continuation of, in part a reaction against, the antecedent tradition. It was designed to avoid conflating the study of language with that of logic, or of languages, or of social differentiation, or of hypothetical psychological mechanisms. This is the study of language as it features in the various modes of human interaction; in other words, as the faculty that makes available for us the characteristically human forms of communication. This approach is pursued today in integrational linguistics. It aims to free the understanding of language from the many popular misconceptions and biases about it that are current in the modern world.

2. Language as a faculty

2a. Language is the faculty that underlies both speech and writing. It may be considered one part or facet of a more comprehensive faculty: that of sign-making (for which there is no general term in common use). If we adopt the term sign, however, it must be clearly understood that for the integrationist a sign is not a form which carries its own meaning permanently around with it. A sign acquires a meaning only when occurring in a specific context. (See 4.)

2b. Language is often described as ‘the use of words’ or the capacity for ‘the use of words’. But that phrase hardly advances matters; in fact, it is no more perspicuous an expression than the term language itself. Perhaps less so, in certain respects. For it comes with the accumulated intellectual baggage acquired from centuries of use by grammarians, who were preoccupied with their own criteria for identifying the ‘units’ out of which ‘sentences’ are constructed. Furthermore, use of words is at best a clumsy phrase, because it is not at all clear what I am claiming to know if I claim to know that speech and writing both consist of words. (See 8c.)

2c. A linguistic act, in any case, does not necessarily require the utterance or inscription of words. (See 3c.)


3. Integration.

3a. The term integrational alludes to the recognition that the linguistic sign alone cannot function as the basis of an independent, self-sufficient form of communication, but depends for effectiveness on its integration with non-verbal activities of many different kinds. These include all those that do not depend in any way on being able to speak or write; i.e. most of the basic activities needed for everyday living (eating, drinking, bodily movement, standing up, lying down, walking, fetching and carrying, avoiding obstacles, using elementary tools, paying attention to objects and happenings in the immediate environment, etc.). This ubiquitous prelinguistic substrate of behaviour is a prerequisite for the emergence and maintenance of verbal communication in all its forms. Not a single example has ever been found of a system of words or sentences which is useful and meaningful although totally unrelated to any non-verbal human activity. Even the ‘abstract’ language of mathematics has its roots in everyday integrated procedures of counting and measuring, as practised from time immemorial in markets all over the world.

3b. The integrated character of linguistic and non-linguistic practices is so fundamental for human beings as to make it difficult to separate out any purely linguistic component. However you try to tackle the problem, there is no universal dividing line between the linguistic and the non-linguistic. This becomes comprehensible when it is realized that the human child has no way of learning what are in retrospect regarded as its ‘first words’ other than as extensions of pre-linguistic vocal activities. Nor, having mastered these ‘first words’, coached by its parents, does the child immediately abandon the use of all earlier non-linguistic signs. On the contrary, the verbal signs become ever more complexly integrated into a whole range of other signs and activities that gradually assume importance in the child’s life. Thus the notion that any episode of linguistic communication can be reduced to its verbal component, without reference to the prelinguistic activities involved, must be rejected.

3c. The same is true of adult communication. To take a trivial example, raising a hand and pointing to a certain building supplies an answer to the question ‘Which is the Town Hall?’. In that respect it is as much a linguistic ability as being able say, without pointing, ‘That building opposite is the Town Hall’. Again, when you ask someone to shut the window, what you expect (or hope for) is a non-verbal response in the form of actions taken by the person you are addressing. In order to make sense, your question has to relate to an existing non-verbal situation where there is a window open and your addressee is in a position to shut it. In order to understand your question, the person you are speaking to has to understand not only the words you uttered but also what a window is and how to shut it. By shutting the window, your addressee gives a contextually integrated response to your question. Thus there is a sense in which that response is no less a linguistic act than your utterance, since what makes it the right response is determined by the language of your question. Shutting the window in those circumstances is an intrinsic move in the communication process. If you then say ‘Thanks!’ when the window is closed, you have just taken part in an integrated sequence of activities involving you and someone else, of which the crucial component (getting the window shut) did not involve uttering words at all.


4. Language and context.

4a. Language, then, is the human capacity for communication by integrating signs into series of activities, some of which involve speech or writing or both. Traditionally, only spoken and written signs are counted as ‘linguistic’ signs. But that assumption is challenged in integrational studies (see 3c above), which focus on the communicational function of the sign in its context.

4b. That is why integrationism pays far more attention to contextualization than any other approach to language. There is no linguistic topic on which more naive and simplistic ideas abound than about context. There are no context-free signs, whether verbal or non-verbal. Contextualization is a complex activity, still too often neglected and poorly theorized. It is not just a function of the immediate situation, but of the entire communicational experience of the participants. The act of contextualization is the act by which the sign is identified as a sign. No contextualization, no sign. This is a basic assumption of integrational linguistics.

4c. Contexts are not ‘given’: they are constructed by the participants in particular communication situations. How exactly this is done – how the distinctions are drawn between what is relevant and what is not – no one has yet explained. Integrational research aims to explore this problem.

4d. The complexity of contextualization is one of the reasons why misunderstandings are common in human communication. Individuals contextualize differently from one another, depending on the personal experience they bring to bear on dealing with a given situation. Not even in the case of identical twins do two individuals share the same history of communicational experience.

5. Speech and writing.

5a. Traditional Western education inculcates the view that speech can be ‘represented’ in written form. Alphabetic letters are thought of as ‘representations’ of individual spoken sounds. Ideally, there is assumed to be a one-to-one correspondence between the elements of the spoken and the written word. That is why a word such as bed is said to be ‘written as it is pronounced’ or ‘pronounced as it is written’: the three separate ‘sounds’ of the spoken word are matched by the three separate letters of the written word, and these three letters are said to represent ‘the same sounds’ in other words (bad, red, etc.). On the other hand, chew is regarded as a misleading or ‘irregular’ spelling, in that the spoken form does not contain any separate sounds corresponding to the four separate letters c-h-e-w. English writing is said to be full of these ‘irregularities’, and that is alleged to be a major source of difficulty for foreigners learning English.

5b. All these judgments are based on the ethnocentric prejudice of regarding alphabetic writing as the only ‘rational’ or the ‘best’ system of writing, while non-alphabetic writing is ignored or treated as inferior. (Dr Johnson, when asked on one occasion why he regarded the Chinese as barbarians, replied: ‘Sir, they have not an alphabet’.) This kind of prejudice is what gives rise to all the confusions listed in 5a. They have no place in integrational linguistics. Collectively, they make it virtually impossible for anyone who has had a monolingual education in a Western school to look at language from an integrational point of view.

5c. The further we move from a fixation on alphabetic spelling towards recognizing the existence and antiquity of non-alphabetic forms of writing, the more we are led to question the notion of writing as a ‘representation’ of speech. Speech and writing are indeed complexly integrated activities, but each deploys a totally different inventory of signs. There is nothing in common between the two. The signs of speech are based on the continuum of sound, whereas the signs of writing are based on the continuum of space (not on vision, as is commonly supposed). One form of language is dynamic and ephemeral: the other is static and relatively permanent. Neither ‘represents’ the other, although either may be substituted for the other up to a point, depending on the specifics of the communication situation. The first thing that has to be understood about language is that whatever makes speech possible as a form of human communication also makes possible the integration of that form of communication with forms of communication based on signs of an entirely different kind.

5d. A literate community is a community which has developed ways of integrating the use of spoken signs with the use of written signs for purposes of communication. A preliterate community relies on the former exclusively. That is why a literate person’s view of language can never coincide with that found in communities where speech is its sole form of verbal expression.

5e. All communities – as far as is known – are potentially literate communities. There are no documented cases of communities where the introduction of writing proved to be impossible, i.e. where no writing system was ‘learnable’. However, that universal potential has not always had the happiest consequences for the study of language.

5f. Many of the conceptual muddles that are pervasive in contemporary linguistics can be traced to the fact that literate linguists – even when dealing with preliterate communities – typically start by ‘reducing’ utterances to writing (commonly called ‘phonetic’ or ‘phonemic’ transcriptions), and then analysing the transcriptions as if they were analysing sequences of spoken forms. This is also the procedure followed in many student textbooks on linguistics. It betrays a complete failure to appreciate how literacy can mask one’s view of the integrational character of language. Speech is being treated as faithfully ‘represented’ by a chosen sequence of written forms, which then act as surrogates for purposes of linguistic analysis. No attention is paid to features of speech which cannot be captured at all in writing, given that the two types of linguistic sign are intrinsically different. It is as if, as Saussure once remarked, it were thought better to study a person’s photograph in preference to studying the face, in order to gain a better understanding of what that person looks like. This objection to studying speech on the basis of written forms is one that university linguists continue to ignore, without realizing that this is a tacit admission of the methodological incoherence of their own discipline. The reason for this incoherence points to a failure to recognize the nature of the disparity between oral and written communication. Even when analytic work in the classroom is based on ‘live’ recordings of speech, the analyst is at a loss to suggest how oral units can be identified for metalinguistic discussion without overt or tacit reference to some hypothetical transcription. (It soon becomes tedious to have to identify sounds by constant reference back to their ‘position’ in ‘utterances’ that now exist only on tape.)

6. Languages and rules.

6a. Traditional European education also inculcated the doctrine that every language has its own ‘rules of grammar’ which must not be infringed. (‘Never use the pronoun we after a preposition’, ‘Plural verbs must have plural subjects’, etc.) These alleged ‘rules’, along with the alleged ‘parts of speech’ on which they were based, were the inventions of grammarians, whose main aim in formulating them was to simplify morphology and syntax for the benefit of learners. The ‘rules’ supposedly governed the ‘correct’ forms of speech, i.e. the forms approved by the ‘best’ speakers (usually educated members of the upper class).

6b. In modern linguistics, these traditional ‘rules’ have been taken over by linguistic theorists, partially renamed and greatly complicated. However, they are now passed off as ‘describing’ or ‘underlying’ the actual linguistic practice of the majority of speakers of the language, or at least of speakers of the ‘standard language’. This is another invention of theorists, aided and abetted by governments wishing to impose uniformity on the linguistic usage of the populations under their control. Since the Renaissance, speaking ‘standard English’, ‘standard French’, etc. has been promoted as a kind of badge of national identity. Apologists for standard languages commonly appeal to the necessity for distinguishing between ‘descriptive’ and ‘prescriptive’ grammar, but without apparently realizing that ‘standard’ and ‘standardization’ are inherently prescriptive notions.

6c. From an integrational perspective, there are no ‘rules of grammar’ and there are no ‘standard languages’. These notions are pedagogic fictions, maintained chiefly for educational and political reasons, and serving to disguise the endless variety of integrational patterns that are to be found in everyday human interaction. Linguistic usage is subject to constant innovation and experiment (new words, new constructions, new applications), as intelligent observers can notice for themselves almost from day to day, if they keep their eyes and ears open.

6d. What particular language (or variety of a language) individuals regard themselves or others as speaking is a question open to empirical research. This is research into the popular use of language-names and descriptions (such as English, Glaswegian, Cockney, slang, etc.). The answer will vary in different cases. It cannot be answered in advance by postulating that every such designation corresponds to some specific system of linguistic ‘rules’, of which the speakers themselves may be only dimly or unconsciously aware. That is neither a ‘scientific’ nor even a plausible assumption.

6e. In brief, the way the term rule is used in contemporary linguistics is another example of conceptual confusion perpetrated in the name of linguistic theory. Rules do not describe anything, either in language or any other sphere of human activity: rules lay down prescriptions in the name of some authority. The Highway Code does not describe how motorists actually behave on the road, but prescribes how they have to behave in order to keep within the law. To imagine that rules are just descriptive generalizations of some kind (e.g. of linguistic usage, or of postulated brain processes, or of a corpus of texts) is to confuse rules with regularities.

7. Language and other faculties.

7a. Words are commonly said to be ‘stored’ in the memory, along with many other things. But this is little more than a picturesque metaphor. It seems self-evident, however, that language and memory are connected in some way, perhaps in many ways, since ordinary conversation would presumably be impossible if one speaker immediately forgot what the other speaker had just said.

7b. Not only memory but other faculties are obviously involved in activities such as speech and writing. It is sometimes said that a full understanding of our ‘linguistic knowledge’ (or, alternatively, a ‘scientific’ understanding of language) will be impossible until advances in the study of the brain reveal exactly how the language faculty and other faculties are related. This is held out as one of the hopes for future ‘cognitive science’.

7c. Thinking of language in this way, however, rests on a misunderstanding. The mistake is analogous to supposing that the explanation of why a clock keeps good time must be that inside it there is a set of instructions for time-keeping. Research into brain mechanisms is interesting in its own right. But the fact that linguistic communication has already come to play such a central role in civilization without relying so far on any such research suggests that whatever human beings already know about language from their own experience is quite adequate for an ‘understanding’ of the relevant phenomena. Integrational linguistics is primarily concerned with elucidating this lay knowledge, and is sometimes described for that reason as being ‘lay-oriented’.

8. Aims of integrational linguistics.

8a. It should be evident from the above that integrational linguistics does not aim to provide improved ‘alternatives’ to the misleadingly compartmentalized studies that are offered in the name of ‘linguistics’ in many university courses. Integrationists are not attempting to set up such rival studies as ‘integrational phonology’, ‘integrational morphology’, ‘integrational syntax’ or ‘integrational semantics’. All these alleged subdisciplines – phonology, morphology, syntax and semantics – are artifacts of the methods employed to compartmentalize them. They do not correspond to anything in the actual linguistic experience of day-to-day communication. They are convenient fictions of the classroom, intellectual hangovers from the centuries during which Western education was based on the copying and studying of approved written texts.

8b. The integrational approach to language seeks to liberate linguistic thought from these classroom shibboleths and provide a way forward for those seeking a clearer analytic understanding of their own communicational experience. Every episode of communication, however trivial, necessarily involves creative activity by the participants, including their own interpretation of the situation in which it occurs. Every utterance is a new utterance, no matter how many times someone may have ‘said it before’. Words are not temporal invariants (as dictionaries like to present them). The static abstractions listed by lexicographers thus impose yet another linguistic misconception on the unfortunate student. The lexicographer’s view of language is a case of trying to impose a normative straitjacket on an open-ended flux of relations between linguistic acts and the world.

8c. What role, then, do words play in these relations? Words may be thought of as components of linguistic acts, and linguistic acts are events in people’s lives. They belong in the same temporal continuum as the other events that go to make up all we do or are called upon to deal with in the course of a day. The main reason which prevents people from realizing this simple fact is that they tend to focus on the overt form in which the word appears and remains in the mind. The written forms visible in a text such as the one you are now scanning are not words, but preserved traces of prior acts of writing. You are in the process of transforming these traces into words on your own account by integrating them into a context other than that in which these traces originally appeared. You yourself have to supply this contextualization: no one else, not even the writer can supply it for you. And the result of the process is what is commonly called ‘reading’. So words do not have some unique time-track of their own, or a status that sets them apart from other events. You may think you can read ‘the same text’ tomorrow, but the reading will not be the same. For one thing, you will bring to bear on it the experience of having read it before. Yesterday’s words cannnot be re-used again today, any more than the meal we ate yesterday can be eaten over again today, or the goal scored yesterday can be scored once more today. Those who have learnt this fundamental integrationist lesson may then come to see themselves not as language-users but as language-makers, which is indeed their natural role in the never-ending evolution of human communication. They may also come to advance their understanding of language beyond the conceptual oversimplifications that have for so long surrounded it.

9. A personal view.

Why is it worth bothering about any of this? Certainly not for any academic motive, such as adding a few footnotes to the long history of linguistics. It is worth bothering for one reason only: because we all live under the ancient injunction inscribed on the temple of Apollo at Delphi: ‘know thyself’. This Carlyle regarded as an ‘impossible precept’, but I regard it as an indispensable precept. Self-knowledge is the key to understanding other people and our relations with them. No progress towards self-knowledge can be based on misconceptions about language and the role it plays in our lives, because that role is too central in its contribution to our humanity.


The author is grateful to Adrian Pablé and Marc Haas for comments on an earlier version of this text.

Back to Top
© Roy Harris, Emeritus Professor of General Linguistics, Oxford, 2010-2015