Language not universal?
The sense in which this term is currently used in English
How big a deal is Heryanto’s concern?
Have we been wrong about the main function of language?
The prototypical function of language: Three arguments


I’d seen various references to Ariel Heryanto’s article, “The making of language: Developmentalism in Indonesia” (Prisma 50: 40-53, 1990), but hadn’t actually seen the article itself until recently. In what is probably the most cited passage, he says, “Language is not a universal category or cultural activity. Though it may sound odd, not all people have a language in the sense in which this term is currently used in English.” (ibid. 41). But how could language not be a universal category?


Cases like this are particularly interesting because of the likelihood that the author has perceived a truth that we’ve been overlooking. In such cases it’s generally wise to follow the advice attributed to George Miller: "In order to understand what another person is saying, you must assume it is true and try to imagine what it could be true of." The appropriate question would therefore be: What truth is he trying to call attention to?

Are we to imagine that in saying “language is not a universal category” he’s challenging the assumption that all human societies have language? Is he denying that all normal humans are born with the capacity for language, or that this capacity is put to extensive use in all human societies? No, it’s clear from the context that that’s not what he’s saying. He is certainly not suggesting that the pre-colonial Malays and Javanese or anyone else lacked an innate capacity for language or that, having such capacity, they neglected to use it.

What he says next provides some clarification: “not all people have a [my emphasis] language in the sense in which this term is currently used in English.” Thus, what he is claiming is not that some peoples lack language, but that their language doesn’t come in the form of such an individual system.

But what does he mean by that? What characteristics does current English usage suppose a language to have? What does current English usage take a language to be?

As best I can determine from the article, what he has in mind is a particular form of standardized language that developed in Europe and the ideology that has developed around it. James Milroy has called this the “ideology of the standard language”. In Milroy’s (2001: 530) words
“Certain languages, including widely used ones such as English, French and Spanish, are believed by their speakers to exist in standardized forms, and this kind of belief affects the way in which speakers think about their own language and about 'language' in general."
Note that there are two parts to this ideology. First, English (and French, Spanish, etc.) speakers believe that English (and French, Spanish, etc.) exist in standardized forms, and second, that this belief affects the way they think about language in general.

There is a clear basis for the first belief: the belief (to take the example of English) that there is a correct form of this particular language with its specifiable inventory of words--each with definable meanings--and of specifiable grammatical rules. The clear basis is that there are dictionaries and grammars that are generally recognized as authoritative--as providing official specifications. These official rules, with relatively few national differences, are the same for English speakers everywhere. Thus, there clearly is a Standard English that exists--it exists in such measure that it’s probably what most English speakers have in mind when they speak of “the English language”.

But such uniform systems haven’t always been there, and none came into being spontaneously. In fact, at the time that the European “vernaculars” were taking over the role previously played by Latin, their lack of uniformity aroused a lot of concern. They were--or appeared to observers at the time to be--confusing arrays of differing varieties and idiosyncrasies.

Samuel Johnson, whose dictionary is regarded as one of the key landmarks in the standardization of English, described the situation that confronted him as follows:

"When I took the first survey of my undertaking, I found our speech copious without order, and energetick without rules: wherever I turned my view, there was perplexity to be disentangled, and confusion to be regulated; choice was to be made out of boundless variety, without any established principle of selection..." (Samuel Johnson, Preface to the Dictionary).

The definitions and rules that defined the official English language were prescribed in a variety of works, among which Johnson’s dictionary played a prominent role. I think of the creation of this “Standard English” as having been essentially accomplished by the end of the 18th century, but in any case it’s now well established. The populations of the English-speaking countries have been quite successfully brought not only to acknowledge, but in considerable measure actually to conform to, the official rules.

To repeat, then, there is a clear basis for the belief that languages such as English, French, and Spanish exist in standardized forms. But there’s no reason to suppose the Heryanto was questioning that belief.

It’s the other part of the ideology that he is presumably questioning: its “way of thinking about ‘language’ in general”. This other part of the ideology involves an assumption that the way the standard languages are supposed to work is the way language universally works. Language is seen as necessarily consisting of such lexicon-grammar systems for encoding and decoding messages (prototypically in the form of propositions). The implication is that communication depends on such systems and is only possible between individuals who know and obey the rules of the system.

But it’s not clear in what ways and to what extent this assumption is valid in situations where no standard language exists. It’s not clear how it might have applied to the pre-standardization European vernaculars. I think that all Heryanto really meant is that standard languages are not universal. One might object that that fact is obvious--not worth mentioning.

But I think not. I think he was pointing out something that is actually quite significant. I think we (not excluding professional linguists) do tend to assume that standard languages provide an adequate model for understanding language everywhere--that our analysis of standardized language is a sufficient source for questions to ask in other linguistic situations. The fact that some of these situations may more closely resemble, for instance, that of the pre-standardization European vernaculars is likely to have less effect on the choice of questions than on the frustrations experienced in getting the original ones answered. And questions such as how people communicate when they don’t know the same rules may not be thought of as linguistic questions at all.

Heryanto, Ariel. 1990. The making of language: Developmentalism in Indonesia. Prisma 50: 40-53.

Johnson, Samuel. (Multiple editions, of which the first was in 1755). A dictionary of the English language.

Milroy, James. 2001. Language ideologies and the consequences of standardization. Journal of Sociolinguistics 5: 530-555.


But why should standard languages not be representative of language universally? Why should they not be suitable models for understanding language everywhere? The answer is that they’ve been shaped by human intervention. They’re a particular kind of example of what Heinz Kloss has called “Ausbausprachen”. For English readers, Kloss explained

"The term Ausbausprache may be defined as 'language by development'. Languages belonging in this category are recognized as such because of having been shaped or reshaped, molded or remolded--as the case may be--in order to become a standardized tool of literary expression." (Kloss 1967: 29)
The cultural interventions that led to the kind of standard languages that we’re talking about came in at least two major stages. First was the invention of written language--these standard languages (or maybe standard languages of any kind) wouldn’t be conceivable in an oral world.

As I understand it, writing came into use as a mnemonic device before the possibility of adapting it to represent language was discovered, but its place in human affairs at that time was a relatively insignificant one. The association with language changed all that--the range of what could be recorded expanded immeasurably.

A number of scholars have pointed out the significance of the fact that writing produces lasting artifacts--texts. While it is true that speaking has always produced artifacts of a sort, those artifacts consisted of nothing more permanent than evanescent sound waves.

Although this doesn’t seem to be what Heryanto had in mind, writing itself suggests a distorted picture of what is universal in language. One fact that deserves a lot more attention than it has received is the fact that written language is selective. Written text represents a kind of transcription of the utterance--the artifact produced when language is used--but these transcriptions were always far from complete. They necessarily selected out for representation the parts of the utterance that could be accommodated with the available repertoire of symbols.

It may seem natural to literate people today (part of the ideology) to assume that the part selected out by writing is the most essential part, but there is no assurance that it would have seemed so before writing was invented. Consequently, we can’t be at all certain that written language--or the view of language that it has led to--fairly reflects the true nature of language.

In any event, there is the fact that written language produces concrete artifacts that persist. What’s more, these texts can be reproduced in multiple copies, any or all of which can be contemplated in isolation. This means that they may be interpreted without consideration of--even in total ignorance of--the context in which they were originally produced.

But how could one make sense of a text read in isolation, in the absence of any contextual clues? How could one know what the speaker/author meant? The only answer: the meaning had to be inferred from the text itself. Hence the slogan “the meaning is in the text”. And thus we arrive at the notion of autonomous text--the idea that it could be possible to compose texts whose meaning could be determined from their form alone.

The idea that meanings could be encoded in texts has opened the door to many institutions that are taken for granted today: written constitutions, rules, laws, contracts, treaties, and the like. Indeed, modern science couldn’t exist without such texts--some would say that the essential product of science consists in the statements presenting its results.

However, the languages of the pre-writing period weren’t adapted to producing autonomous text. After all, there would have been no texts available for people with no knowledge of the author or the original context to interpret.

One problem that the need for autonomous text presented was the previously mentioned of lack of linguistic uniformity. The required encoding and decoding could only work to the extent that it was based on a code that was known--and understood in the same way--by both the author and relevant readers. In other words, the writer and reader must use the same language.

This problem became particularly acute in Europe after printing was introduced and the role of written documents began to expand. During the same period (and probably in large part because printing made it commercially advantageous to reach larger audiences), the “vernaculars” rapidly began to assume the role previously played by Latin. However as mentioned above, these vernaculars were not the kind of uniform codes required for an instrument suitable for autonomous text, they were--or appeared to observers at the time to be--confusing arrays of differing dialects.

Anyway, standardization of the languages was required to ensure that writers and readers conformed in their interpretation of the meanings of words and grammatical constructions. This required an agreed-upon authority that governed essentially what could be said and what it could mean. This governing authority took the form of grammars and dictionaries.

However, there was a second problem that is less well understood. The languages that had been inherited from oral days weren’t adapted to producing autonomous text--to enabling one person to encode meaning into a text in such a form that it could be extracted by another person who might have no knowledge about either the author or the original context.

Confronting this problem led, originally in Europe, to the changes in the design of the standard languages that Havránek has called its “intellectualization”. In his (more precisely, his translator’s) words:

By the intellectualization of the standard language, which we could also call its rationalization, we understand its adaptation to the goal of making possible precise and rigorous, if necessary abstract, statements, capable of expressing the continuity and complexity of thought, that is, to reinforce the intellectual side of speech. (Havránek 1964: 6)
The resulting phenomenon has spread increasingly to much of the world--in more recent years often being advertised as “development” or “modernization”. I find Joseph 1987 to be a very helpful account of what is involved in creating such a language. Although he calls them simply “Standard Languages”, that term might be thought appropriate for any language that has some kind of official status. For that reason, I’ve (e.g. Grace 2002) proposed the label “modern standard language” as less ambiguous.

To summarize, then, I believe the modern standard language (MSL) is what Heryanto had in mind when he spoke of “a language in the sense in which this term is currently used in English”. And I believe also that language in that sense is indeed not universal.

The MSL is designed to pursue the goal of autonomous text. When MSLs come to be thought of as representative of language as a whole, it becomes easy to see the production of text that comes as close as possible to autonomous as being the main function of language.


Grace, George W. 2002. Collateral damage from linguistics? 2. The cultural evolution of language. Ethnolinguistic Notes, Series 4, Number 22. Internet World Web page at

Havránek, Bohuslav. 1964 [orig. 1932]. The functional differentiation of the standard language. In Paul L. Garvin (ed. & translator), A Prague School Reader on Esthetics, Literary Structure, and Style. Washington University Press, pp. 3-16.

Joseph, John Earl. 1987. Eloquence and power: The rise of language standards and standard languages. London: Frances Pinter.

Kloss, Heinz. 1967. "Abstand languages' and "Ausbau languages'. Anthropological Linguistics 9(7): 29-41.


All right, suppose Heryanto is right about language in other parts of the world being rather different from the MSLs we’re accustomed to. So what? What kinds of implications might this fact have for the way we as linguists go about our business? I don’t know the answer, but I think it’s a question well worth exploring. (In fact, it seems at least possible that the implications might lead ultimately to a re-examination of the basis for the autonomy of linguistics as a discipline).

Heryanto’s concern was with the extent of the changes that “development” was imposing on Indonesia. To quote him further:

As commonly understood in contemporary English or its translation in the many languages of the so-called Developing countries, development was not only unknown and unspoken of in the pre-colonial communities of what is now Indonesia. It was simply unthinkable. (1990: 40, emphasis mine)

Language is involved because “language development”--the goal of which is designing a national MSL--is, he argues, an inextricable part of “development”.

…the work of ‘Developing the nation’ has been persistently accompanied by the task of ‘Developing the national language’ (ibid.)

But what implications does this “development” have for the people and their language?

The following statement by S. Takdir Alisjahbana, who played a prominent role in the discussions of Malay/Indonesian development, seems relevant:

With the problem of modernization, we are confronted, in particular, with new emerging nations as they attempt to adjust the concepts, ideas and ways of thinking of their culture to the concepts, ideas and ways of thinking of the modern world through the coining of new words or terms…..
“It goes without saying, that this change of concepts and values orientation, manifesting itself in the acceptance of new ideas and sometimes even in a new worldview, is not only limited to the coining of new words, but is also extended to the more or less conscious change of the morphology and even of the phonemic system of the language.” (Alisjahbana 1965: 520, emphases mine)

There are two particular suggestions here that seem important. To take them in reverse order, they might be put this way:

(1) There are more than superficial effects on the language itself; the morphology and even the phonemic system may be affected. (So they can’t be dismissed as “just vocabulary”). I’ve discussed this elsewhere, most fully in Grace 2002. There, I pointed out that speakers of “modernized” languages report finding texts only a couple of generations old virtually unintelligible. I then tried to identify some of the structural changes that seem to be involved, but of course what I say there just barely touches the surface of a topic that deserves much more attention.

(2) More fundamental things are at stake--things that he describes variously as “ways of thinking”, values orientation”, and “worldview”. In short, he’s saying that this kind of “development” can alter very basic aspects of the lives and cultures of the peoples.

Still, if the purely linguistic changes are hard to pin down, the changes to ideology, ways of thinking, and/or worldview are certainly not easier. And yet I’ve encountered a number of observers who feel that some such thought-permeating changes are involved. I first got interested in this matter some thirty years ago. In fact, my short Ethnolinguistic Note written at that time (Grace 1978) still represents my best attempt at introducing the topic. Of course, it was anything but conclusive, and the whole idea that language might be connected with anything like worldview has--for reasons I’ve never fully understood--been persistently resisted by the linguistic establishment.

I think it was Foucault, especially in Foucault 1973, who first alerted me to the idea that an important change in language ideology played an essential part in the fundamental changes in European worldviews that marked the end of the Renaissance. At the same time I was only beginning to be aware of the changes that were being imposed on the languages themselves to fit them to the emerging MSL design.

Although, Foucault emphasizes the radical quality of the changes introduced by the new “episteme”, I can find no concise statement of their specifics. (I’ve had similar difficulties with other authors).

The sources I cited in Grace 1978 were mainly concerned with struggles involved in developing the MSL version of English. One idea that I didn’t discuss there (and haven’t discussed very fully anywhere) is that of Walter Ong that it was “Learned Latin” (his name for the Latin that existed from A.D. 550 or so on) that made the MSL possible. He points out that Learned Latin was no one’s native language but was effectively the property of an intellectual elite. He notes that:

This is a strange situation for a language. Latin was distanced--alienated--not from day-to-day life, for it was of the substance of daily life for lawyers, physicians, academic educators, and clergymen, but from the psychological and psycho-somatic roots of consciousness. It no longer in any sense belonged to mother, it did not come from where you came from. (Ong 1977: 27-28)

He believed that the changes in the European vernaculars that occurred under the influence of Latin were accompanied by more fundamental changes that affected what he calls the “state of consciousness”:

Studies of the matter are nonexistent, but one could argue as an initial hypothesis that the modern intellectual world and the modern state of consciousness could never have come into being without Learned Latin or something like it.(ibid. 36)

His “state of consciousness” sounds much like Alisjahbana’s “ways of thinking”. I’m also reminded of Stanley E Fish’s (1972: 374) reference to an “opposition of epistemologies” in seventeenth century England (that I talked about in Grace 1978) as well as, of course, Foucault’s change of “episteme”.

Thus people have spoken variously of changes in episteme, worldview, ways of thinking, state of consciousness, values orientation. It would be too much to assert that they are all referring to the same phenomenon. However, it does seem reasonable to suppose that many--quite possibly all--of the phenomena these terms are intended to describe are somehow related, and that there is something important involved. It also seems reasonable to suppose that all of the various descriptions they have used are in one way or another suggestive of its nature. And finally, one must also suppose that this phenomenon is very involved with language. If this is true, it’s easy to understand how Heryanto can have felt that something very fundamental was at stake.

And surely we need to understand more about it.


Alisjahbana, S. Takdir. 1965. New national languages: A problem modern linguistics has failed to solve. In G. B. Milner and Eugénie J. A. Henderson (eds.). Indo-Pacific Linguistic Studies, Part III, Descriptive Linguistics. Amsterdam: North-Holland Publishing Company, pp. 515-30.

Fish, Stanley E. 1972. Self-consuming artifacts: The experience of seventeenth-century literature. Berkeley: California.

Foucault, Michel. 1973. The order of things: An archaeology of the human sciences. [Translation of Les Mots et les Choses--1966]. New York: Vintage Books.

Grace, George W. 1978. Intellectualized language: Writing. Ethnolinguistic Note, new series, no. 6. Mimeograph. Also available online at

Grace, George W. 2002. Collateral damage from linguistics? 2. The cultural evolution of language. Ethnolinguistic Notes, Series 4, Number 22. Internet World Wide Web page at

Heryanto, Ariel. 1990. The making of language: Developmentalism in Indonesia. Prisma 50: 40-53.

Ong, Walter J. 1977. Interfaces of the word: Studies in the evolution of consciousness and culture. Ithaca NY: Cornell University Press.


I said above that according to the view that underlies the MSL, the main function of language is the production of text that is as close as possible to autonomous. But one immediate difficulty with seeing this as the main function, at least of human language in its entirety, is that text itself is dependent on writing. This makes it hard to imagine how, before the invention of written language, anyone might have conceived of autonomous text as an objective. In fact, the languages invoked by Heryanto--the Malay and Javanese of past centuries--were obviously written, but autonomous text appears clearly not to have been a goal for their speakers.

Still, although autonomous text itself is an artifact of writing, the principal function of language must surely be to convey factual information even though with something less than the fine-tuned efficiency of MSLs. Or so I thought and persistently argued for years. True, I was sometimes made aware that others had different ideas; I just found them incomprehensible.

As I look back now, I recall that as early as my fieldwork in 1955 I had begun to realize that the Melanesians I was working with didn’t have exactly the same picture of language that I did. But I certainly didn’t understand what theirs was (were?) like. I never had much time to try to find out, and what’s more, nothing in my linguistic training had provided any guide for investigating such questions.

Anyway, I didn’t worry too much about it at the time. After all (I thought) I was trained in linguistics, and that was the unique science of language--the source of the most sophisticated understanding of language that existed. The Melanesians’ views, whatever other interest they might have, represented no challenge to the validity of my own work.

Nevertheless, the mystery of how they did conceive of language continued to lurk in the back of my mind. How different could their picture be? Why would they not see in their languages the same thing we saw in ours? Of course, since their languages were not standardized, a certain amount of indiscipline--of sloppiness and imprecision--would not be unexpected. However, aside from that, how could their languages work, how could they perform their principal function, if they weren’t pretty much like I pictured them?

And that, I now realize, may be precisely where the problem lies--in my assumption as to what the principal function of language was. This assumption seemed secure enough. What kind of challenge to it could even be imagined? What grounds might such a challenge be based on?

But then, when you get down to it, what is actually being assumed? What does it mean to say that something is the “principal function” of a capacity like language, and how can we determine what function is principal? As I struggled with the first question, I came to realize that what I really had in mind might be better described as the prototypical function.

What I mean by saying that a function is “prototypical” is that it’s a key to whatever other functions may exist--that to understand it would provide a unique entry point for understanding the rest. Furthermore, this kind of logical priority must at least strongly suggest a chronological priority for the same function--that the others probably actually evolved from it. (I, myself, certainly was making this assumption about language).

And what was the prototypical function of language? The answer seemed self-evident; its prototypical function was the communication of factual information. That was surely beyond question. As early as Grace 1981 I was arguing that both the design and the use of languages supported this conclusion.

Surely, I argued, the characteristics of actual languages (as revealed in linguistic descriptions) seemed best explained as designed for--as presumably having evolved for--this function. In fact, this function seemed necessary to account for the fact that the capacity for language had evolved at all. What other advantage of any consequence did language offer? And in any case, it seemed obvious that the prototypical act of speaking is initiated by an idea and the intent to communicate it.

What’s more, all of the other functions I could think of seemed clearly derivative. They seemed to consist of nothing more than secondary uses that had been discovered for the already-existing communication device. (A metaphor I had in mind for these secondary adaptations was modulations on a basic carrier wave). The speaker was always saying something--making use of structures that seemed to me to be obviously designed for carrying factual information--whatever his/her immediate purpose in speaking. Of course language was used for other purposes as well, but that was hardly surprising. Indeed, it’s the opposite that would be surprising; it would be surprising if human ingenuity over the millennia had not devised ways to make use of the language instrument for a whole array of incidental functions.

It seemed to me that all of this reasoning confirmed that the information-carrying function was logically prior and, almost as certainly, the original function that had driven the evolution of language in the first place. It’s true that some colleagues--in particular Mike Forman and Ron Scollon--tried to shake my certainty, but nothing anyone said seemed to me to constitute a clear, or even understandable, alternative.

But other views were being proposed; I just couldn’t make sense of them because they clashed with my preconceptions. One often-cited alternative came from the anthropologist, Bronislaw Malinowski. For example, he had written:

The fact is that the main function of language is not to express thought, not to duplicate mental processes, but rather to play an active pragmatic part in human behavior. Thus in its primary function it is one of the chief cultural forces and an adjunct to bodily activities. (1935/1965: 7).

And again:

In its primitive uses, language functions as a link in concerted human activity, as a piece of human behaviour. It is a mode of action and not an instrument of reflection." (1923/1946: 312)

He was thus challenging the very foundation of my understanding of language, but I found it hard to see how he pictured language--how he pictured it as working. However, I’ve read other things that expressed views about language functioning that seemed in tune with his. For example, returning to Heryanto’s article, we find the following description of the (pre-Westernization) views of some Indonesian communities:

… in older Malay and Javanese communities, the term bahasa (or bhasa; basa) did not refer to something abstract and neutral. It was neither simply a tool of communication nor a system of codes or symbols that arbitrarily signified something else (a reality). It was a social activity. It was socially bound, constructed and reconstructed in specific settings, rather than scientifically and universally rule-governed. (1990: 43)

Of course, however much in tune with the speakers of these languages the writers may have been, they are still outside observers. It would be of great interest to hear the views of the speakers themselves--especially speakers of unwritten languages--but unfortunately they don’t usually philosophize about prototypical functions of language. Or at least, I’m not aware of published accounts of such philosophizing. The most likely clues that I can think of come from reports on how they teach language--the ways they conceive of what’s to be learned.

A number of reports on how language instruction is approached and conceived indicate that the objective is behavior appropriate to a variety of situations. For example, Bambi Schieffelin’s (1990) enlightening work on the Kaluli describes in detail the coaching through which the child learns to make appropriate verbal responses (utterances) to social situations. She explains that, according to Kaluli ideas,

…language was social, they were teaching sociality. (Schieffelin, personal communication).

Interestingly, second language instruction in two different parts of the Melanesia-New Guinea area has been described as following very similar patterns.

Don Laycock reports the following observations in the Sepik Province of Papua New Guinea:

Teaching proceeded by means of whole sentences and occasional individual lexical items, either volunteered by the Abau speaker or requested by [one of the learners]. (1979: 91)

He points out that this seems to be representative of a general pattern.

But it seems likely that this method of teaching by whole sentences of potential use--the phrase-book method--is the normal one in Papua New Guinea; my own informants commonly adopted this method during eliciting. (ibid. 92)

William Thurston reports a similar pattern in the region of northwestern New Britain where he worked,

The New Britain concept of language instruction is highly systematic in that the language taught follows the progression of social use parallel to the socialisation of the student into the linguistic group. ... At all stages, the language taught is governed by its use in actual social situations. (1987: 72-3)

All of these cases seem to indicate that the speakers of a number of non-MSL languages have pictures of the way language works that are quite compatible with Malinowski’s idea that its main function is “to play an active pragmatic part in human behavior”. In all of these cases, it seems that what is to be taught is appropriate behaviors in a variety of recurring situations with the status of language being that of verbal component in these behaviors.

There’s obviously a lot I don’t understand about this; there’s surely even a lot more that could be gleaned from the existing literature. What does seem clear, at least, is that there are other views of how language functions, and that many of them tend to converge on the idea that what’s to be learned is not a language system but verbal behavior. Furthermore, it also seems clear to me now that my own assumptions about the nature of language represented little more than the MSL ideology as it had been refined by linguistics.

The sources I’ve cited here suggest a way of looking at language that focuses on language use--that looks at it in terms of the verbal behavior that is an integral part of social interaction between and among humans. In fact, they suggest that this part isn’t even necessarily communicative in function; it may consist of nothing more than “phatic communion”.

Phatic communion is of course also a concept introduced by Malinowski: He defines it as:

...a type of speech in which ties of union are created by the mere exchange of words (1923: 315).

As he points out, we all practice it a lot.

Can it be that Malinowski was right after all--that the main function of language has always been “to play an active pragmatic part in human behavior”?


Grace, George W. 1981. An essay on language. Columbia SC: Hornbeam Press.

Heryanto, Ariel. 1990. The making of language: Developmentalism in Indonesia. Prisma 50: 40-53.

Laycock, Don. 1979. Multilingualism: Linguistic boundaries and unsolved problems in Papua New Guinea. In Stephen A. Wurm (ed.). New Guinea and neighboring areas: A sociolinguistic laboratory. The Hague, Paris, New York: Mouton, pp. 81-99.

Malinowski, Bronislaw. 1923/1946. The problem of meaning in primitive languages. Supplement to C. K. Ogden and I. A. Richards, The meaning of meaning. 8th edition. New York: Harcourt, Brace, and Company.

Malinowski, Bronislaw 1935/1965. Coral Gardens and their Magic. Vol. 2, The language of magic and gardening. Bloomington: Indiana University Press. (Originally published in 1935 by George Allen & Unwin, Ltd.).

Schieffelin, Bambi B. 1990. The give and take of everyday life: Language socialization of Kaluli children. Cambridge: Cambridge University Press.

Thurston, William R. 1987. Processes of change in the languages of north-western New Britain. Pacific Linguistics B99.


At the end of the previous segment I was asking whether Malinowski might not, after all, be right in claiming that the main function of language is “to play an active pragmatic part in human behavior”. But earlier in that segment I referred to three arguments, that I found compelling, that supported the hypothesis that the communication of factual information had to be its prototypical function. Now I no longer find any of these convincing; I should probably explain why.

1. One argument was that the characteristics of actual languages (as revealed in linguistic descriptions) seemed best explained as designed for this function--so exactly designed for it as to invite the inference that language had evolved specifically for it.

The picture presented by the descriptions of actual languages was of a device for representing logical relations such as different kinds of questions, negation, possession, subordination, etc.; for keeping the pronunciation of messages clear and distinct; and ultimately, for referring to whatever can be talked about. One only had to look at these descriptions to see how the structure of the languages was adapted to the function of conveying factual information.

But today I’m realizing that there is another consideration here: what descriptions tell us consists essentially of answers to questions the describers asked. And the questions the describer is trained to ask--as well as the possible acceptable answers--are essentially prescribed in advance. The catch is that these questions clearly derive from the assumption that the communication of factual information is the prototypical function, and their objective is to spell out how the particular language is designed to accomplish it. This objective recalls Kuhn’s description of normal scientific research as “a strenuous and devoted attempt to force nature into the conceptual boxes supplied by professional education” (1970: 5).

This, of course, raises the question of alternatives. What other kinds of conceptual boxes might the same language be forced into? What very different descriptions might be equally accurate, and what kinds of pictures of language might they yield. Some of the works cited in the preceding segment suggest that there might be people who would propose very different questions indeed.

These would certainly be very interesting questions to explore, but they’re not the main concern here. All I wanted to explain here is that I was wrong in giving weight to what we might call “the argument from linguistic descriptions”--that is, the argument that the characteristics of languages demonstrate that the prototypical function of language is to convey factual information. Now that I realize that the design of these descriptions was biased from the outset, I no longer see them as providing any significant evidence on language functions.

2. My second argument was that (it seemed to go without saying that) the prototypical act of speaking was initiated by an idea and the intent to communicate it. The first step was conceiving the intent to convey such and such; the next was encoding it in a particular language.

But, as I tried to explain in Grace 1997, this picture of the process appears to be untenable. As I put it there:

It seems to require two quite separate mental roles--almost as if there were two minds, each working independently of the other. The mind that possesses linguistic knowledge is the encoder. However, all it seems to be able to do is to encode messages provided to it by the other mind and to decode those received via ears or eyes--presumably transmitting the results to the other mind. The other mind is the thinker. It seems to be a pretty complete reasoning machine, a pretty complete human mind, except for its one big deficiency: it doesn't have access to language. It isn't able to encode and decode messages.

Just one kind of example that occurs to me of the problem with such separate components would be the following situation. Speaker A has made some remark; B recognizes that language has a formula that would make a perfect retort, and answers using it. But in order for B to have made the decision to speak at all, his/her “thinker component” had to know about the resources provided by the language--had to have some “linguistic competence”. Likewise, bilinguals in conversation with other bilinguals often resort to code switching in order to exploit resources offered by one language and not the other. Such decisions would be possible only if the “thinker” component had some knowledge of the languages.

It also makes no sense to imagine the “encoder” component being called upon to encode messages without access to any knowledge of the effect the message is intended to produce--what the purpose of the speech act is.

But a hypothesis that would duplicate the information--that would give each component access to the same information as the other--makes no sense. It now seems obvious that there is one continuous decision-making process. It extends from deciding whether or not to speak at all through the designing of any resultant speech act.

What’s more, with a minimum of introspection it’s apparent that the factual information encoded in the linguistic expression is very often in one or another degree incidental to the purpose of the speech act (cf. Malinowski’s “phatic communion” mentioned in the preceding segment). In fact, it’s been pointed out that when individuals who have no language in common are placed together, the most likely strategy for developing some kind of rapport involves talking. This, even though not a single word--not a bit of the factual content--will be decodable.

3. My third argument was that the fact-communication function seemed necessary to account for the fact that the capacity for language had evolved at all. It seemed to me that the ability of our species to exchange, and accumulate, factual information was a huge advantage, and I could think no other advantage of any consequence that language had to offer.

But is this anything more than a teleological argument? Wasn’t I reasoning rather like this: The kind of communication that language makes possible is probably the biggest advantage we enjoy over other species, so that must be what made it evolve? The problem with that argument, of course, is that what drove evolution had to be an immediate advantage offered at each stage, not an ultimate advantage that may not necessarily have emerged until a later stage. (In fact, the advantage we attribute to it may even not be as important universally as we suppose.)

I think now that if we want to explain why language evolved, the place we have to begin is with the question: Why us? Why did language evolve in the ancestors of our species while their nearest relatives remained unaffected? At the time when our ancestral line split from its nearest kin, both lines must have possessed those traits that--in our line--turned out to be the antecedents of language. Then something happened that led to a different evolutionary path. What might this have been? The two kinds of hypothesis that I think of, might be labeled respectively, “lucky miracle” and “ecological change” hypotheses.

Lucky miracle hypotheses hold that some change occurred within ourselves (or more accurately, our ancestors). I presume this would have to have been a mutation that permitted new advantageous behaviors. The particular advantageous behaviors might well have emerged as an adventitious by-product that they didn’t discover how to exploit (“exapt”) until long after they first became possible. I believe lucky miracle hypotheses generally take it more or less for granted that the advantages in question primarily concerned the communication of factual information.

The ecological change hypotheses hold that what first changed was our ecological adaptation. Most likely there was some change in the environment in which our ancestors lived, either because of climate change or because they had moved to a new area. Of course, it would also be possible that they had just stumbled onto a different means of exploiting the same environment. In any case, something would have happened that enhanced the relative value of some already-existing behaviors and reduced the value of others or even resulted in the invention of previously unknown ones.

Either of kind of hypothesis could account for the fact that we took off on a different path. However, I don’t like the lucky miracle kind. It seems too contrived--like invoking a deus ex machina.

What about ecological change hypotheses? What, specifically, might they propose? The one I’m familiar with is Robin Dunbar’s “social bonding” hypothesis. As I indicated earlier in this website (see Grace 2003), I find it quite attractive. It would be redundant to repeat my 2003 account here, but briefly Dunbar’s main argument is as follows:

Our ancestors moved into a more open environment (I understand that this is the generally accepted understanding); this environment made them more vulnerable to predators, and they responded by coming together into larger groups (I understand that this is a common response that has been observed in other species). Thus, their environment changed and they adapted by making some changes in behavior.

As group size increased, the alliances into which the individuals grouped themselves became larger, and much of the time alliance members were more widely dispersed about the countryside. These developments, in turn, affected the nature of interpersonal relations (if “interpersonal” is the right label for the interactions of the species in question). In particular, physical contact with other alliance members became more limited, the important activity of grooming became less available, and vocal signals came to be depended on more and more.

To review, the hypothesis proposes this scenario:

1. What set our line apart began with a change in habitat.

2. Adaptation to the conditions of the new habitat led to changes in the social organization of the group.

3. These changes in turn led to changes in the way individuals interacted. There were more individuals with whom one needed to communicate, and they were often farther away.

4. Vocal signals provided a partial solution. They made it possible to know where other individuals were (“contact calls”) and even to maintain some limited measure of intimacy. As these signals assumed greater importance in the functioning of the community, innovations that enhanced the repertoire came to be favored in natural selection.

I’m not arguing here for this specific hypothesis. The point of this segment is to review the validity of certain arguments that I’d previously suggested--arguments that I’d believed made a pretty convincing case that the communication of factual information was the prototypical function of language. The particular argument under discussion here was that the fact-communication function seemed necessary to account for the fact that the capacity for language had evolved at all.

Dunbar’s social bonding hypothesis seems clearly to invalidate that argument. It proposes an entirely different function as accounting for the fact that language evolved. Not only that, but I know of no other hypothesis that accounts for the crucial fact that our ancestral line was singled out in the first place.

To sum up, I’ve reached the conclusion that none of my three “compelling” arguments for believing that the communication of factual information had to be the prototypical function of language deserves any consideration at all.


Dunbar, Robin. 1996. Grooming, gossip, and the evolution of language. Cambridge MA: Harvard University Press.

Grace, George W. 1997. Linguistic Change: 5. The individual's knowledge of language. Ethnolinguistic Notes, Series 4, number 6. Internet WWW page at

Grace, George W. 2003. Robin Dunbar’s Social Bonding Hypothesis. Reflections on the evolution of human language, number 1. Internet WWW page at

Kuhn, Thomas S. 1970. The structure of scientific revolutions, second edition, enlarged. International Encyclopedia of Unified Science, Volumes I and II: Foundations of the Unity of Science, Volume II, Number 2. Chicago: University of Chicago Press.

To go to other places in this website, click on one of the cells below

Home Page The Ethnolinguistic Notes The Ethnolinguistic Notes, Series 1 and 2 Ethnolinguistic Notes, Series 3 The Ethnolinguistic Notes, Series 4 Reflections: Language Evolution
Reflections: Knowledge of Language Personal Page The Human Predicament Why Write Unpublishable Things? Modest Proposals Odds and Ends Pictures

Put on the Web 3 September 2007
Latest addition 6 November 2007

© Copyright 2006 George Grace, Honolulu Hawaii. All Rights Reserved