Comments Welcome
George W. Grace
University of Hawaii

Ethnolinguistic Notes

Series 4, Number 6

LINGUISTIC CHANGE

5. THE INDIVIDUAL'S KNOWLEDGE OF LANGUAGE

This Note along with the two that follow it may be thought of as constituting a single unit on the relation between changes in the KOL (knowledge of language) of individuals and the changes reported in linguistic descriptions. However, I've been trying to keep the notes in this Series short, and am therefore breaking the unit into three (although that still leaves them longer than I'd have liked).

The last note preceding this (Grace 1997) concluded that since all linguistic change must ultimately be realized as changes in individuals' knowledge of language, we can't expect ever to understand linguistic change as long as our work is based on misleading assumptions about the kind of knowledge that individuals have.

Therefore, this Note (the first of the three) presents a hypothesis about the kind of knowledge of language (KOL) that individuals possess.

However, our ultimate aim is to answer questions that are at present formulated in terms of "languages", and yet we've now had to conclude (Grace 1996a, 1996b) that what we call "languages" are actually fictions. Therefore, further inquiry was needed into what lies behind the fiction--into the connection between individual KOLs and the "languages" recognized in the literature. That is the subject of the second Note.

Then, the third Note brings us back to linguistic change, that is, to the kinds of change that are reported in linguistic descriptions. It deals with (1) how KOLs change, and (2) the connection between changes in the KOLs of the individuals who count as speakers of a language and changes in the états de langue that would be reported in normal linguistic descriptions.

QUESTION: The last note ended with the question of the individual's knowledge of language. What's the answer? What form does this knowledge take?

ANSWER: The answer most frequently given by linguists would go something like this: The knowledge consists of the grammar of a particular language (underlain by an innate knowledge of universal grammar).(1)

QUESTION: What do you mean, "grammar of a particular language"?

ANSWER: A "grammar" in this sense means what is covered in a traditional linguistic description: syntactic rules and a dictionary.

QUESTION: But how can the reference to "a particular language" be right? Haven't we just seen (in Grace 1996a) and 1996b) that what we recognize as languages are in fact fictions?

ANSWER: That's true. However, I think we need to permit ourselves to talk provisionally as if individual languages were real entities. The important thing is to be careful not to let this way of talking mislead us into reification.

QUESTION: But if they're not real, why should we maintain the pretense that they are?

ANSWER: To begin with, most of the phenomena that the concept of the language was designed to explain are quite real and relevant to our discussion, and we don't have any other established way of talking about them. To avoid using this concept would get us into circumlocution that would be very difficult to follow. However, one of our aims should be to give a better account, as soon as possible, of the reality that lies behind these fictive entities.

QUESTION: All right, I'll go along with that provisionally, but I intend to keep watching out for anything that smacks of reification.

However, with these reservations set aside, is it otherwise true that our knowledge of language consists of the grammar of a particular "language" (with an innate knowledge of universal grammar underlying it)?

ANSWER: I think, as I've tried frequently to explain (most recently in Grace 1995), that it clearly isn't true.

If we consider traditional linguistic descriptions as hypotheses about what speakers know, we must conclude that they conceive of the scope of linguistically relevant knowledge ("knowledge of language" as opposed to "knowledge of the world") much too narrowly, and furthermore, that they strive for a kind of economy that's fundamentally at odds with the economy of the human mind (I've discussed this latter point in Grace 1981: 13-21).

QUESTION: Why are these hypotheses so wrong? What's responsible for these misconceptions?

ANSWER: Well to begin with, these descriptions aren't based on any attempt to study the kinds of knowledge that speakers actually show evidence of. Traditional linguistic description represents a particular kind of structural analysis of what I've called "artifacts of speech"--i.e., either written texts or transcriptions of linguistic expressions from actual speech or from (purportedly) potential speech (i.e., what speakers say they would or could say).

QUESTION: Why would anyone believe that that kind of structural analysis of texts would tell us anything about speakers' knowledge of language? How did such an idea ever arise?

ANSWER: That would be a good subject for research. It's my impression that it comes from the translation method of language teaching. In that method, a great emphasis was placed on translating texts using a grammar and dictionary, with the ultimate goal being for the student to eliminate the need for the grammar and dictionary by gradually learning grammatical rules and vocabulary items.

QUESTION: All right, if speakers' knowledge of language (let's abbreviate that to KOL) isn't anything like what traditional linguistic descriptions describe, what is it like? What kinds of knowledge do speakers show evidence of?

ANSWER: There seems to have been no particular attempt to find out (since it's assumed that we already know).

QUESTION: We seem to be starting from square one. How do you propose to go about looking for the answer?

ANSWER: I propose to start by considering one aspect of that knowledge that seems to me to be of critical importance--namely, the knowledge that's required in order to prepare to speak.

QUESTION: What do you mean?

ANSWER: I mean, what kind of knowledge does a prospective speaker resort to in deciding what utterances to make: that is, in deciding whether or not to speak at all, and (although this is hardly likely to be a separate decision) in deciding what to say?

QUESTION: Wouldn't it generally be assumed that that has nothing to do with linguistics, that linguistic knowledge only comes into play when it's time to encode a message that's already been decided on?

ANSWER: I think that has been generally assumed, and that's the assumption that I find untenable.

QUESTION: What's wrong with it?

ANSWER: 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.

QUESTION: Your analysis may be right, but why do you think those assumptions can't be correct?

ANSWER: Because I don't see how the two minds--or components of the mind if that's what they should be called--can work independently of each other. For example, do you imagine that the "thinker" component could do its work without any idea of the linguistic resources available to the "encoder"? Can you imagine that decisions about what utterances to make are really made completely without regard to the resources afforded by the language to be used?

QUESTION: I'm supposed to be the one asking the questions, but I'll humor you and tell you my reaction. If there can't be any overlap in the knowledge possessed by the two components, then the two-component hypothesis makes it sound like the speaker is in the situation of trying to communicate in a strange language through an interpreter. That does sound like an awkward situation in which to have to decide when to speak and when to remain silent or, having decided to speak, how to settle on a message. However, can't that problem be solved by allowing the "thinker" component of the mind to have some knowledge of the linguistic resources available to the "encoder"?

ANSWER: Maybe, but I think the reverse would also hold--that the "encoder" would need access to information that would be thought of as belonging to the "thinker" component.

QUESTION: What do you mean?

ANSWER: For example, are we to believe that the encoder doesn't have some knowledge of the purpose of the speech act? Can we believe that such knowledge isn't required for encoding? But such knowledge would duplicate knowledge required also by the other component, the "thinker". Thus, it seems that each component would need some of the other component's knowledge.

QUESTION: Couldn't all of the difficulties that we've mentioned so far about the two-component hypothesis be solved just by allowing each component full access to the knowledge employed by the other component?

ANSWER: Maybe so, but positing separate components with identical resources seems to be multiplying entities beyond necessity (as enjoined against by "Ockham's razor" (2)).(3)

But even if one were prepared to accept the analysis into two activities, I still don't think this scheme would be very convincing.

QUESTION: What else would be wrong with it?

ANSWER: Well, I don't think it's clear how the "encoder" is supposed to work. It's not clear how, even with a grammar (cum dictionary) of the language, it would actually go about encoding the message.

QUESTION: What do you mean?

ANSWER: Although we don't know in what form the message is submitted, this sounds like a translation problem--translate the message from whatever form it's in into a target-language expression. But knowing the grammar (and presumably also whatever needs to be known about the code in which the message has been made available) doesn't tell one how to encode (=translate) the message. Some knowledge specifically about how to encode, how to apply the grammar to actual messages to give them adequate expression in the target language, would also be necessary. Therefore, it seems that the encoder component of the mind needs some knowledge beyond the linguistic knowledge that's been attributed to it.

QUESTION: Mightn't the kind of knowledge required to translate also qualify as "linguistic" knowledge?

ANSWER: In fact, I think it should, although it isn't included in the grammar of the language, which is what's supposed to constitute one's KOL.

QUESTION: Where does this leave us?

ANSWER: I think we need to try an entirely different hypothesis of the nature of the KOL

QUESTION: What do you suggest that should be?

ANSWER: The only one that I'm aware of that attempts to answer the question of what kind of knowledge of language is necessary for deciding what utterances to make is one we proposed in Grace 1993 as "Hypothesis 3". This hypothesis was (as stated in Grace 1993: 685-686):

"Individuals' knowledge of language is nothing more than a large memory store of experiences--with the emphasis on experiences in which language was used. For linguistic purposes, we may think of this as essentially a store of utterances. We interpret what's said by recalling other cases where the same thing or something similar was said. We decide what to say by recalling similar situations and what was said in them (and what the consequences were)."

QUESTION: Why do you think we should take that hypothesis seriously?

ANSWER: I think so for two reasons. First, I think it's essentially right. However, I need to qualify that statement. I think it represents the basic form of knowledge of language. By that I mean that it constitutes the basis of everyone's KOL even though many (conceivably even everyone) carry out some conscious or unconscious analysis and come to recognize formulas, lexical items, and even grammatical rules. I believe that the more verbal, the more articulate, and the more educated a person is, the more prominent a role such analysis is likely to have played in the development of his/her KOL. The converse is that the most natural KOLs should be found among the least linguistically-adept people, especially people who've had minimal exposure to an elaborated ideology of language.

QUESTION: Wait. What do you mean by "natural" KOLs?

ANSWER: That's a good point. I was using "natural" in contrast with "cultural". Thus, a natural KOL in this sense would be one that had developed without being influenced at all by cultural factors. It would be one that had developed spontaneously from the reaction of the human organism to the instances of language use experienced by the learner (with no linguistic ideology having been imparted along the way).

QUESTION: Your equivalent of a "language acquisition device" being exposed to raw data?

ANSWER: Yes, that's about what I had in mind.

QUESTION: But surely you aren't suggesting that such a "natural KOL" could actually exist?

ANSWER: It would certainly be very exceptional, and if it did exist, it couldn't be "natural" in another sense--the sense of being normal for humans. It's certainly not normal for human beings to grow up in a cultural vacuum.(4)

QUESTION: Anyway, you seem to be saying that the KOLs of uneducated people are more nearly natural (in your sense) than those of educated people.

ANSWER: As a general rule, that must surely be true, at least as far as the formal educational systems of the Modern World are concerned. One of the major aims of this formal education is to shape up the KOL--particularly by inculcating reading and writing skills, and most particularly the skills required for dealing with expository prose.

QUESTION: Then, none of the closest approximations that do exist to your "natural KOL" would be found anywhere in the Modern World?

ANSWER: The Modern World is certainly not a good place to look for minimally-tampered-with KOLs, but even in many modern societies there might be relatively isolated communities where the levels of verbal interaction are low and where there are people who have managed to escape schooling and have remained illiterate.

I should probably emphasize again that the hypothesis doesn't hold that there is anyone anywhere whose KOL takes precisely the form of what we're calling the "natural KOL". What it does hold is:

1. That it's also not necessarily the case that there is not anyone whose KOL takes precisely this form--that it isn't impossible that someone could have just such a KOL and be accepted as a normal speaker of a language. In other words, there's no minimum amount of grammar construction, of drawing of generalizations, or of any kind of analysis of linguistic expressions that is required in order to qualify as a normal speaker of a language.

2. That no matter how much analysis a person has done--no matter how much superstructure of memorized rules and formal definitions of words has been added--his/her speaking and understanding of speech rely primarily on the kind of memory store proposed by our hypothesis. What the analysis contributes is in the nature of fine-tuning, even though there may be a lot of it.

QUESTION: All right. You said your first reason for taking the hypothesis seriously is that you believe it to be true. What's the second?

ANSWER: It's that it's at the opposite pole from the received hypothesis. Thus, if our hypothesis errs, its errors must be in the opposite direction from those of the traditional view, and the truth must lie on the line connecting the two.(5)

QUESTION: In what sense is it at the opposite pole from the received hypothesis?

ANSWER: In the sense that the traditional view assumes that the ideal KOL incorporates all possible generalizations whereas our hypothesis assumes that in the ideal (i.e., "natural") KOL no generalizations at all have been drawn.

But note that I don't believe that our hypothesis is off the mark as long as it's understood as a hypothesis of the natural KOL, and not of the KOL of every actual individual. I don't question that the individual usually begins to draw generalizations as s/he begins to acquire culturally-prescribed skills.

QUESTION: I still have plenty of questions about this hypothesis. To begin with: Does it claim that we remember every utterance we ever hear--in fact, everything that happens to us (I think some people have maintained that every experience we've ever had is stored away somewhere and could be brought to consciousness if we knew how to access it.)?

ANSWER: No. It seems obvious that we don't remember all of the details of any utterance we've heard, and we don't remember every utterance individually.

QUESTION: Then what, according to the hypothesis, do we remember?

ANSWER: Let me begin by quoting a little further from what we said there.

"… as far as I can judge it seems likely that other things being equal, we remember particular utterances (or parts of utterances) better in proportion to: (1) how often we've heard them; (2) how recently we've heard them; and (3) the amount of emotional impact hearing them had on us. By "remember better" I mean either that the details are preserved more accurately, or that the entire utterance (or utterance part) is more accessible to ready recall, or (probably most often) both." (ibid. 686-87).

However, I'd like to add something more to that. First of all, understanding any experience is finding experiences in our memories (remembered experiences) that it can be associated with. These associations remain, i.e., they serve to position it (the newly understood experience) in the memory store. And if we think back subsequently on the experience, we may create further associations in the process. In the long run an experience may retain its positions in the system of associations even after its specific details have faded.

QUESTION: Was there something more?

ANSWER: Yes, a quite different point. We regard some people as being more suitable role models for ourselves than are others. Our memory stores will tend to give particular prominence to situations in which these role models found themselves and especially to their utterances in these situations. This point is important in understanding the basis for the notion of individual languages.

QUESTION: Very well, but I have another question. You said, "We decide what to say by recalling similar situations and what was said in them (and what the consequences were)." Does that mean that we can only speak when the situation we're in is sufficiently similar to one we can remember--where we can remember what the speaker who was in that situation said?

ANSWER: In a sense you could say that that is what it means, but I'd prefer to put it differently. I'd say that our ability to find an appropriate verbal response to a situation depends on our finding something familiar in it. That means that it has to have enough similarity--or rather, that it's up to us to find enough similarity in it--to one or more situations from our previous experience to furnish a basis for designing our response.

QUESTION: I still don't understand how we're supposed (by the hypothesis) to design this response.

ANSWER: Let me quote again from Grace 1993:

"According to [this hypothesis], what we know are examples, and our task is to use these examples as a basis for designing our own utterances. I don't know very much about how this is (to the extent that the hypothesis is valid) accomplished, but I would suppose that the principal basis is analogy. For a simple example of what I have in mind, given utterances A, B, C in the memory store, we decide that we want to produce an utterance that will be similar to A, but differ from it in just the way that B differs from C." (ibid. 687).

But now I'd like to add the following qualification. Analogy seems too mechanical a procedure. I think we should, following Mark Johnson (1987), recognize that humans regularly make (imaginative) use of metaphorical relations (and for that matter, of metonymy) in interpreting situations.

QUESTION: I'm beginning to get the idea that this hypothesis is still rather vague on a number of points.

ANSWER: That's absolutely true. Furthermore, it'll certainly remain more or less vague for a long time. Even in the most favorable circumstances--if the hypothesis should immediately attract the full attention of linguists, specialists in memory research, and everyone else that might turn out to have something to contribute--it would surely still take a long time before even all of the questions that are obvious to us now would be answered.

However, to refuse to consider the hypothesis further until these questions are answered would be analogous to postponing all comparative linguistic research on (say) the Austronesian languages until descriptive linguistics had completed its task and every one of the languages had been fully described. (That's not too far from what some linguists seemed to be advocating to me some years ago).

QUESTION: All right, we'll proceed on the basis of this hypothesis as to the form of the speaker's knowledge of language. Now, what does it tell us about how linguistic change occurs?

ANSWER: That's too big a question to take up here if we're going to keep these Notes to a reasonable length. We'll continue with it in the next Note.

NOTES

1. For example, Noam Chomsky says (1965: 8), "Obviously, every speaker of a language has mastered and internalized a generative grammar that expresses his knowledge of his language." Back up

2. Entia non sunt multiplicanda praeter necessitatem. Back up

3. As a more general comment, I might add that the idea of thinking in the absence of any knowledge of language is a strange one for linguists to entertain. In fact, some linguists have gone so far as to take the exact opposite position, asserting--maybe only casually--that thinking was nothing more than silent speech. Whether or not that's entirely true, the idea of analyzing the process of speaking as consisting of two separate activities, of which one involved thinking without any access to language, has always seemed very wrong to me. Back up

4. Of course, there's generally a lot more information available than what could be gained from the simple observation of language use. One can normally also observe the reactions of others to his/her own and others' use of language. He/she can probably even ask questions about usages and may quite possibly receive unsolicited advice as well. Back up

5. Actually, there's a third reason why I like the hypothesis, although it isn't relevant to our present concerns. It's that it doesn't require as great a discontinuity between human language and the communication systems of other species as do other hypotheses. In fact, the evolution appears to be quite continuous, and therefore doesn't require an appeal to either divine intervention or what Gould and Vrba (1982) call "exaptation". The most striking change is an enormous increase in complexity that is attributable to an enormously increased ability to respond distinctively to subtle differences in our external environment. In other words, the number of situations to which we can respond with distinctive vocalizations has multiplied immeasurably, and I'm proposing that this increase has been driven in very large part by more subtle discrimination among what I've called (for example, in Grace 1987, 1989) "objective situations". Back up

REFERENCES

Chomsky, Noam. 1965. Aspects of the theory of syntax. Cambridge MA: M.I.T. Press. Back up

Gould, Stephen J., and Elisabeth S. Vrba. 1982. Exaptation--A missing term in the science of form. Paleobiology 8:4-15. Back up

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

Grace, George W. 1987. "What they would say in the same situation". Ethnolinguistic Notes, Series 3, Number 31. Printout. Also (1996) Internet WWW page at http://www2.hawaii.edu/~grace/eln31.html. Back up

Grace, George W. 1989. The association of situations with linguistic expressions. Ethnolinguistic Notes, Series 3, Number 35. Printout. Also (1997) Internet WWW page at http://www2.hawaii.edu/~grace/eln35.html. Back up

Grace, George W. 1993. What are languages? Ethnolinguistic Notes, Series 3, Number 45. Printout. Also (1996) Internet WWW page at
http://www2.hawaii.edu/~grace/eln45.html. Back up

Grace, George W. 1995. Why I don't believe that language acquisition involves the construction of a grammar. Ethnolinguistic Notes, Series 4, Number 1. Internet WWW page at http://www2.hawaii.edu/~grace/elniv1.html. Back up

Grace, George W. 1996a. Linguistic change: 1. The "état de langue". Ethnolinguistic Notes, Series 4, Number 2. Internet WWW page at
http://www2.hawaii.edu/~grace/elniv2.html. Back up

Grace, George W. 1996b. Linguistic change: 2. The "état de langue" of Abstand languages. Ethnolinguistic Notes, Series 4, Number 3. Internet WWW page at http://www2.hawaii.edu/~grace/elniv3.html. Back up

Grace, George W. 1997. Linguistic change: 4. More on the locus of linguistic change. Ethnolinguistic Notes, Series 4, Number 5. Internet WWW page at http://www2.hawaii.edu/~grace/elniv5.html. Back up

Johnson, Mark. 1987. The body in the mind: The bodily basis of meaning, imagination, and reason: Chicago: University of Chicago Press. Back up

SIDE COMMENT (18 March 1997)

I have tried in this Ethnolinguistic Note to stick as close as possible to the main point, which in this and the two following Notes is to begin laying a more solid foundation for historical linguistics.

However, some of the assumptions that I've mentioned here as being made by some linguists make more sense when seen in the context of the larger ideology of which they are part. Therefore, I thought it might be useful to offer this sidetrack from the main course of this sequence of Notes for those who choose to take it.

First, the assumption that the individual's knowledge of language consists of a grammar (including a dictionary) of a particular language makes sense when one understands the underlying assumption about how language works. This assumption, that of the "governing system" (my term), is described in Grace 1992: 654 as follows:

"Thus, we have arrived at the idea of individual systems with (1) their store of vocabulary items, each of which would "pick out" aspects of the real world, and (2) their logical syntax--that is, their means of specifying logical relations among such real world phenomena. Each such system would govern the construction and interpretation of sentences. Through the grammar and dictionary of which it would consist, the governing system would authorize the construction of certain sentences and warrant the meaning(s) of each sentence so authorized."

The governing system assumption also explains (or at least provides a very significant part of the explanation for) the persistence of the fiction of individual languages.

Again I'll quote, this time from Grace 1993: 689,

"The "governing system" role that each [language] is called upon ..... to play seems to presuppose a tightly-structured system. It would seem that the possibility for variation within such a system would be strictly limited--that adding more varieties to a governing system would soon loosen it up to the point that it would lose the ability to govern anything (more and more rules becoming optional). The presupposition that a language must be a governing system implies that all of language is divided up into quite distinct languages, and that therefore, it must be possible to sort each individual speaker into one or another pigeon hole (or perhaps two or more in the case of multilinguals, but with each multilingual's KOL neatly partitioned among the pigeon holes)."

I believe that some light can also be thrown on the distinction between what I called the "thinker" and "encoder" components of the mind if we add to the notion of governing system also the assumption that the main purpose for using language is to produce expository prose. I myself gave enough support to this assumption when I said (in Grace 1981: 6) that "the straightforward communication of information" was "the ostensible primary function" of language that I've never been able to live it down with Mike Forman and Ron Scollon [protest as I might that "ostensible" meant only "apparent"]. Anyway, I've recovered from my error, but I think it is part of the ideology of our culture (and of mainstream linguistic theory) that this is the primary function of language--in fact, that the primary use of language is to encode "propositions".

If one conceives of the "thinker" role, not as that of devising appropriate responses to all kinds of situations, but more narrowly as just that of formulating propositions, then it (apparently) becomes more plausible to think of the performance of this role, and of the propositions themselves, as quite disconnected from the resources of any individual language. Consider, for example, Jerrold Katz's "effability principle": "Every proposition is the sense of some sentence in each natural language" (Katz 1976: 37), which openly asserts the total independence of one from the other.

But what I'm describing here is just the received linguistic ideology of our culture. As I think should be apparent from this Ethnolinguistic Note, I think it's fundamentally wrong. I think it's astonishingly ethnocentric, and (as some of you may already be too well aware) I think that its continued acceptance by the linguistic profession will someday prove highly embarrassing to that profession.

ADDITIONAL REFERENCES

Grace, George W. 1992. Another attempt to explain why I have misgivings about what we tell people about language Ethnolinguistic Notes, Series 3, Number 43. Printout. Also (1996) Internet WWW page at http://www2.hawaii.edu/~grace/eln43.html. Back up

Katz, Jerrold J. 1976. A hypothesis about the uniqueness of natural language. In Steven R. Harnad, Horst D. Steklis, and Jane Lancaster (eds.). Origins and evolution of language and speech. Annals of the New York Academy of Sciences, vol. 280, pp. 33-45. Back up

ADDED NOTE (15 April 1997)

Readers of this Note might be interested in an article by Mark S. Seidenberg (1997) which I've just read. Seidenberg reports on studies involving connectionist models in relation to "statistical and probabilistic aspects of language". These studies have led to a view of knowledge of language that calls into question the assumption that it includes a knowledge of grammatical structure. I don't really understand much about this work, but the fact that it emerges from an active research tradition suggests that it's likely to be taken seriously, and that alone would be auspicious. I also have a hunch that it will turn out to be generally confirmatory of what I've been trying to convey, but of course it'll be a while before anyone can be sure about that.

The reference is:

Seidenberg, Mark S. 1997. Language acquisition and use: Learning and applying probabilistic constraints. Science 275 (14 March), 1599-1603.