Links to pages 22, 25, 27, 30, 31, 34, 35, 36, 42

Comments Welcome

George W. Grace

University of Hawaii

September 28, 1981

ETHNOLINGUISTIC NOTES

Series 3, Number 2

DEFINING WHAT LINGUISTICS IS ABOUT (AND WHY

It is often said that linguistics (or mutatis mutandis, any other discipline) is (best defined as) what linguists do, and I must admit that such a definition has the advantages of both convenience and accuracy. It has some offsetting disadvantages, however, in that it is hardly very informative and may be perceived by people outside the discipline as not very forthcoming if not as outright arrogant. In any case, it does little to enhance the case that linguistics is of great value, actually or potentially, to society at large.

There are various signs at this particular time which suggest that it would be a good time for us linguists to attempt to explain ourselves better. One such sign is the view of the Reagan administration that a group of disciplines in which linguistics is apparently included are politically biased and that they accomplish so little of scientific value that their funding can be heavily cut without significant loss to society. Another sign is found in the enormous discrepancy between the importance of language in human affairs and the place accorded to linguistics--which is supposed to be the general science of language--in university curricula. Why, then, are we so unappreciated? I think that part of the answer is that we are misunderstood--that the world does not fully understand the historic mission of linguistics and how the activities of linguists are designed to contribute to that mission. If that is so, what can we do to encourage better understanding of what linguistics is about--of what linguists do and why?

To begin with I would suggest that it would be helpful for us to attempt to define our ultimate objective more clearly for ourselves. This definition might take the form of a central question such that the objective of the discipline may be said to be the search for the answer to that question. Such a question could serve as a reference point in showing how the discipline is relevant to particular human concerns and where it fits into the pursuit of understanding in general (including its proper place in an academic curriculum). It could also serve as a reference point in showing the relevance of particular lines of research within the discipline-- showing how they are conceived as contributing to an eventual answer to the central question. I think, therefore, that there are good reasons to regard it as desirable to have such an acknowledged central question. Chomsky, of all linguists, seems to have been particularly sensitive to this desideratum, and has made several attempts to define a central problem for linguistics. The proposal which I will put forward here, however, differs more or less from all of his.

I believe that there are two considerations which must be given particular weight in the attempt to identify our central question. First, I assume that we want linguistics to be generally acknowledged to be the general science of language. However, and this is the second consideration, it is important to keep in mind that a wide array of academic disciplines are concerned in fundamental ways with language use. If linguistics is to appear relevant to them--if it is to be accepted by them as satisfactorily filling the role of a general science of language--its central question must imply answers to the questions that they themselves have about language. Therefore, I believe that linguistics must seek to be the discipline to which other disciplines feel that they may turn when questions concerning the nature and functioning of language arise. If it fails to fill that role, I expect that sooner or later a competing science of language will emerge from somewhere.

Sociology can serve as an example of a discipline which is concerned with language. It is so concerned because virtually all social interaction is mediated by language, and because the very realities in which individuals function are constructed in social interaction and preserved in language. In fact, some sociological investigations have thrown considerable light on language use and norms. It is important for us to recognize, however, that in the very situations covered by those investigations there are problems about the functioning of language which are not dealt with. Those problems are directly concerned with the nature of language, and they, I believe, should be regarded as falling strictly within the domain of linguistics. To illustrate, consider the example of a team of sociologists working with English-speaking groups and then writing their report in English for an English- speaking audience. In such a case, the only linguistic information which they need to give may be provided by quoting the actual words used and giving a description of the context. They can assume that the readers will be able to tell for themselves what was being said with those words. That is, they can assume that everyone involved--the participants in the interaction, the observer/author, and the audience--all have available to them equivalent language- interpreting devices. We can all presumably understand what is going on without being told how the language functions in order to make this possible.

But suppose instead that the interaction had taken place in some language which is totally foreign to the expected audience. In this case, quoting the actual words and describing the context would be quite insufficient to enable the reader to understand what was going on. The reader, lacking the appropriate language- interpreting device, would simply not be able to make sense of the interaction in most cases. A practical device which our sociologist authors might well resort to in such a situation would be to quote the words in translation, i.e., to attempt to provide English equivalents for each of the utterances. That strategy would transform the report of the verbal interaction in such a way as to permit the English-speaking audience to interpret it with its English language-interpreting device. It would quite deliberately evade the question of how the language-interpreting device works.

In sum, we can observe what is happening in the interaction as long as we can tell what is being said, and we can tell what is being said as long as the interaction takes place in a language we know. When it takes place in a language which we do not know, we cannot tell what is being said, and we cannot tell what is happening. It seems to me that the question of what is happening in the interaction is a sociological question. However, any complete account of what happens in such an interaction must ultimately come to grips with what is being said and with how anyone (the participants as well as the sociologist) knows what is being said--i.e., with the nature of what I have called the "language-interpreting device" (1)--and that is a question that transcends sociology.

The language-interpreting device concerns many other disciplines in roughly the same degree as it concerns sociology. It concerns anthropology, where substantially different realities are encoded into the languages of different cultures. It concerns psychology where interaction with the external world in general and with other persons in particular is mediated by language and the realities which the language represents. It concerns philosophy in that all philosophical activity depends upon language and the accuracy with the realities which can be represented in language reflect the realities of the external world. In short, the nature of the language-interpreting device is of crucial importance to a wide array of disciplines, but its investigation does not fall within their scientific competence. On the contrary, in the academic division of labor it seems naturally to belong to a science specifically devoted to language. And that science, I am assuming, already exists and is named linguistics.

It seems, then, that linguistics is the discipline within whose domain the problem of what I have called the language- interpreting device properly falls. That is, linguistics is centrally concerned with the question of how it is that we can tell what is being said. And that question in turn leads to the question which I propose as the central question of linguistics: What is involved in saying something, anyway? In short, What is it to say something?

IN SUPPORT OF THE QUESTION: WHAT IS IT TO SAY SOMETHING?

A proposal that any particular question is the central question of some discipline is obviously something to be approached with some caution, especially for the practitioners of that discipline. I think of three general objections which may be raised to the particular question which I have proposed for linguistics. These objections would be: (1) that I am (ethnocentrically) exaggerating the importance of saying things--of transmitting information--as a function of language, (2) that there is no great difference--except in degree--between our ability to say things and what other species can do, (3) that the question is not a very interesting one because we already understand pretty well how saying things works, anyway. I will discuss each of these hypothetical objections in turn.

The first objection contended that it is ethnocentric to assume that the function of saying things--the transmission of information--is a centrally important function of language. After all, so the objection might go, we live in what has been characterized as an "information society", where the processing, storage, and transmission of information has grown to proportions which are probably unique in the history of human affairs. It may well be that people in our particular society would generally be inclined to recognize the transmission of information as the primary function of language, but people in other societies might have different ideas (in fact, there is evidence that people in some societies do profess to attribute primacy to other functions). Therefore, we must regard the question of what the principal functions of language are as still completely open, and must strive furthermore to maintain open minds on the matter. And members of our society must be particularly suspicious of any suggestion that saying things is a central function. So goes the objection.

I would attempt to answer this objection by saying, first, that I do not intend to make any claim about what the members of any society think is the primary function of language. And, more specifically, I do not intend to claim that the view that saying things is the most important function of language is at all widespread. As far as that goes, I think that even members of our own society have difficulty in appreciating the importance and uniqueness of this function. It is so intimate a part of humanness and so utterly familiar to us humans that we can not imagine how not to take it for granted. Much of this note will therefore be devoted to trying to convince my audience that saying things is as extraordinary a function of language as I think it is.

The second objection would, in fact, be that which I have just indicated. It would be objected that there is nothing very special about being able to say things. After all (the objection might go) many kinds of other animals might be said to be able to say things--at least we do often speak of an animal as "telling us" something or other or even of "saying" something or other. What can be done with human language (it might be argued) differs from what can be done with the communications systems of other animals only in the degree of its complexity.

My answer is that I believe that what we do when we say something is very different from what any other species in the history of the universe (as far as we are able to tell) has ever been able to do. Of course, we do not understand at all fully just what other animals can do and what they cannot. Most of their communicative behavior seems to consist of a kind of acting out-- e.g., indicating the presence of danger by acting frightened, threatening by acting out the preparation for attacking. In fact, it is difficult to distinguish the intention to communicate information from other motivations in much animal behavior, but I think there probably are grounds for assuming that some behavior is at least partly motivated by an intention to communicate.

Further below I will attempt to analyze in more detail what is involved in saying something, but let me try here to illustrate that a great difference exists between saying and what I will call "indicating". Some simple examples of saying of the sort that human language makes possible might be the following: "It's raining up in the valley", "The fish weren't biting today", "The mangos are almost ripe", "Be careful, the water suddenly gets deep just to your left." I find it difficult to imagine members of any other species saying any of these simple things or doing anything which even comes close to it. But acts of saying comparable to these are an entirely natural aspect of human social interaction. I would insist furthermore that the information which such acts transmit often is enormously significant, and that the ability to communicate in this fashion radically alters the conditions of life for the species.

The above examples were intended as illustrations of saying. To illustrate the nature of "indicating", I would like to quote a passage from Gregory Bateson (1972: 367),

"When your cat is trying to tell you to give her food, how does she do it? She has no word for food or for milk. What she does is to make movements and sounds that are characteristically those that a kitten makes to a mother cat. If we were to translate the cat's message into words, it would not be correct to say that she is crying "Milk!" Rather, she is saying something like "Mama!" Or, perhaps still more correctly, we should say that she is asserting "Dependency! Dependency!" The cat talks in terms of patterns and contingencies of relationship, and from this talk it is up to you to take a deductive step, guessing that it is milk that the cat wants. It is the necessity for this deductive step which marks the difference between preverbal mammalian communication and both the communication of bees and the languages of men."

This example of cat communication in Bateson's analysis seems much closer to the kind of human behavior that we call "hinting" than to what we call "saying". The cat has not said that it wants milk. In fact, we have no way of knowing just what it wanted, whether milk specifically, or food in general, or food of some kind of class recognized as acceptable by the cat. In fact, we have no way even of guessing how specifically the cat itself knew what it wanted. All we have to judge from is what it ultimately accepts.

The difference between what the cat does and saying in the human sense is further brought out when we consider trying to "translate" the cat's communicative behavior into sentences in English. Do we translate "I want milk", or (more politely) "Please give me some milk", or 'Please notice how dependent I am on you", or what? Did the cat "say" one of these more nearly than another? If so, which one? I think that it is obvious that by putting forward any of these sentences as representing what the cat said we are putting words into the cat's mouth. We cannot choose among them because we are setting ourselves the wrong problem. The cat was not saying any of them because cats do not say things. I propose to use the word "indicate" for what they (and other mammals) do when they communicate. They will be said to "indicate" something.

But I think we can and indeed, need to, go further. I think it would be quite wrong to imagine that any other kind of animal even gets so far as to have the experience of "having something to say" as we know it. For example, it is gratuitous to imagine that an animal in the presence of humans is comparable to a human put in a place where he/she does not know the language. This human might have all kinds of things to say (e.g., questions to ask, etc.), but be effectively unable to say them. I would suppose that the experience of "having something to say" in any form recognizable to us would be totally foreign to any other species of animal (in its natural state--I won't speculate on what signing apes seem to themselves to be doing). There are two reasons why I would suppose this. First, I think that since saying something is (as I believe) so very different from anything that any other species can do, it is hard to imagine how they might conceive of doing so. That is, the ability to formulate the objective (i.e., to have something to say) and the ability to put it into execution (i.e., to say it) are not independent of each other, but rather presuppose each other. (2) Second, I suppose that they don't have the experience because I cannot conceive of any species that knows the experience of "having something to say" in our sense not eventually figuring out some way, however imperfect, to get some of it said.

In short, I assume that the idea of saying something is an extremely extraordinary idea, an idea that has probably arisen independently only once in the history of the universe. Furthermore, I assume that the reason why we do not recognize how extraordinary it is is that we naturally take it for granted. If we cannot say specifically how much and in what ways human language differs from the communication systems of other animals, it is partly because we do not know what they can and cannot do, but it is also partly because we do not know what it is that we can do. And it is that-- what I take to be the crux of what language permits us to do- -that I am proposing as the central question of linguistics.

But the third hypothetical objection that I identified above would take the premise that we already know the answer to this proposed central question, at least in its general outlines. It is not surprising that we should think so, for, as Thomas Kuhn has pointed out (1970: 4-5) effective research in a discipline cannot get under way until the scientific community believes that it already knows the answers to all of the most significant questions about the subject matter of the discipline. It is hard to imagine any paradigm for a science of language which did not provide some kind of account of what is going on when someone says something-- of the nature and functioning of the linguistic sign as a whole.

Let us call the parts of the linguistic sign other than the sign vehicle the "meaning" and the study of meaning "semantics". The basic principles of semantics as supplied by our paradigm are supposed to be something like this: the meaning of a sentence is nothing more nor less than the class of conditions under which the sentence would be true. To utter a sentence is consequently to say that (at least) one condition of the class defined by the sentence is true. That, in essense, is what it is to say something according to the received semantic myth. It would be interesting to investigate this conception further and to trace its history. I have the impression that it derives primarily from logicians-- possibly above all from the work of Carnap, with Charles Morris having been infuential in making it accessible to linguists. I think the conception was strictly prescriptive in origin and that it deserves to be called a myth only when it is imagined to be descriptive of ordinary language.

I think that there is a quite general awareness among linguists that this semantic myth does not account for very many facts very well. Still, it is supposed to represent, in broad outline at least, what meaning is. Consequently, the feeling has grown that meaning is not very interesting or important and that saying things occupies a relatively trivial place among the functions of language (which takes us back to the first objection discussed above). (3) I have tried to defend my choice of question by answering some of the likely objections to it. I will now attempt some further explication of the question itself.

ON WHAT IT IS TO SAY SOMETHING

Saying in the strict sense can be illustrated by assertion (as one kind of example). In Principia mathematica, Whitehead and Russell used an "assertion- sign" (a vertical bar with a horizontal bar projecting to the right from its midpoint). In their words (Whitehead and Russell 1957: 8), "It is required for distinguishing a complete proposition, which we assert, from any subordinate proposition contained in it but not asserted". I take it that such assertion is an example of what we mean by saying--that the act of affixing the assertion-sign is a written equivalent of the act of saying. On the other hand, what Whitehead and Russell called the "proposition" in the quoted statement is an example of a sayable. However, I think that it is most useful to broaden the concept of saying to include acts other than asserting.

I propose to define the act of saying as what I have (Grace 1981b) called the "specification of the condition of instantiation" (abbreviate it SCI). It seems that we do normally (under the influence of our semantic myth, perhaps) take the ordinary (unmarked) SCI to involve an assertion. However, SCIs as I define them can include various alternatives to assertion, such as the questioning or denial of the sayable or the assertion that it is contingent upon a given factor or factors, etc. All of the following examples (and many more which one could devise) involve different sayings with the same sayable:

The dog bit the man.

Did the dog bite the man?

The dog didn't bite the man.

The dog will bite the man.

The dog would have bitten the man if ... etc.

This, then, is my characterization of the act of saying. As I said above, I am aware of no indication at all that animals of any other species are capable of anything of this sort.

It should be apparent further that the act of saying cannot occur except in conjunction with some kind of specification of a sayable. Any complete account of saying must also, therefore, give some consideration to the manner in which sayables are specified. (4) I believe that it is important to distinguish two different (i.e., functionally distinct) aspects of the specification of the sayable. I will take each up briefly in turn.

First, there is the reality construction aspect of language use which I mentioned in talking about the linguistic concerns of sociology and anthropology. Saying something involves the construction of what may be called a reality model. This model specifies a kind of event (5) (for example). The specification is made in terms of (for example) a kind of agent performing a kind of act upon a kind of patient. "Dog bites man" (to cite an example in headlinese) specifies an event of the kind in which an agent of the kind which we can call "dog" performs an act of the kind that we can call "biting" upon a patient of the kind that we can call "man". By my definition, of course, a headline of this sort does not actually say something, but rather indicates it. However, there is an implication of saying in actual newspaper headlines; that is, such a headline implies that the writer had something to say, and that he/she would be able to say it fully (in fact, probably has done so in the text of the article appearing below the headline).

It is in this aspect of speaking, or more precisely of specifying the sayable, that the "reality construction" aspect of language is most manifest. The models constructed are themselves little constructed realities, and the linguistic materials from which they are constructed are drawn from the store made available by the language. This store represents a more general reality--the "weltanschauung"-- (or, more exactly, an assemblage of realities).

Still, this aspect of the specification just provides us with an abstract model, and leaves it floating in space as it were. It characterizes a reality, but then tells us nothing whatever about it. The second aspect which I want to take up might be said to bring it down to earth. In this aspect the sayer specifies what he/she is talking about. Often this involves making it clear who is being talked about (who is being referred to). Our headlinese example might be made relevant for the audience, for example, by making clear the identity of the man bitten by the dog. How this specification can be made depends on the context--if the man in question has been the subject of immediately preceding discussion, it might be sufficient simply to use a definite article, or the third person pronoun might even suffice [Thus: (dog bites) the man, or (dog bites) him].

I have referred to this bringing down to earth as "connecting the model to the ken of the audience" (in Grace 1981b). In the example which we are discussing it could also be the dog which constitutes the point of connection (the "old information"), of course, but it might also be that neither identity is important. It is perfectly proper in English to say "A dog bit a man" (with all noun phrases indefinite). The connection to the ken of the audience for such a sentence would probably involve the context (time and place) in which the event occurred. We might speak of this second aspect of the specification of the sayable as that which indicates the relevance of the constructed reality to the particular context of the speech event.

It should be emphasized that the choice of kind of agent, patient, etc. is not necessarily made with the task of establishing reference in mind (as our semantic myth might suggest). Another motive for the choice is the way one wants to characterize the event. I have the impression that the characterization of the event is most often the main motive of choice in the case of indefinite noun phrases. The choice between "a dog bit a man" and "a dog bit a haole" seems most likely to involve primarily a choice in the characterization of the event--in the kind of event one wants to represent it as being. The choice makes a difference (how important a difference it seems to the audience is another matter) in the reality that has been constructed.

However, the case is more complicated when the noun phrase is definite. It is easy to imagine one choosing to say "A dog bit the haole" as opposed to "A dog bit the man" (or vice versa) on the basis of which choice is more likely to help the audience determine the identity of the patient (as, for example, in the case where they had just been talking about the person and referring to him as the "haole"). Still, I would think that such definite noun phrases as "that bastard" or "the old son of a bitch" are very rarely motivated by the task of establishing reference.

To review what has been said about what is involved in saying something, it is the specification of the condition of instantiation (SCI) that really constitutes saying in the strict sense. However, the specification of the sayable provides everything that is required to produce the kind of nominalization which corresponds in all other respects to a complete sentence. These two aspects suffice, for example, to produce "the biting of the man by the dog". However, these two aspects alone are not sufficient to produce an actual sentence, and to pronounce such a noun phrase is not to say something in the sense which I am concerned with.

There are two final points to be made here. First, although I see three distinct functions (specifying a condition of instantiation, reality construction, and specifying the relevance to the audience) involved in the overall process of saying something, and although the implementation of each function represents a separate decision and can be analyzed independently, I do not suggest that the formal syntactic apparatus of the language divides up along the same three lines. Rather, there is a creativity in language use that regularly exploits formal resources in new ways. This said, however, I do think there is some tendency for each of the functions to correspond to its own characteristic kinds of syntactic apparatus. I would find it interesting to see experiments in which syntax was examined in this framework.

Second, at this point it is possible to be a bit more specific about my contention that saying something is very different indeed from anything which is possible in the communication systems of other species. What I want to add now is that I do not believe that other species are capable of performing any of the three different functions which I have identified as required for a complete act of saying. As I have already pointed out, there seems to be nothing in other species which is comparable to the actual act of saying in the strict sense (the specification of a condition of instantiation). Secondly, I do not know of any evidence that would indicate that any other species is capable of specifying the identity of a person or thing being referred to or otherwise specifying how whatever is being communicated is relevant to what the audience may be presumed to have in mind or in view. Finally, although I imagine that some species are able to construct realities of two or more elements as part of their normal strategy for coping with their environments, I know of no evidence that such complex constructed realities have any part in their communication systems or that they have the capability of communicating them in any way. Thus, if I am right in all of this, there is quite a substantial gulf separating human language from the communication systems of any other species, and it must have required a whole series of evolutionary steps to cross it.

CONCLUSIONS

I have suggested a question: What is it to say something? as the central question motivating all purely linguistic activity. I believe that it, appropriately interpreted, can serve as the basis for an account of what we are about that meets the qualifications that I have sketched out.

To propose a single central question implies that linguistic research (except for some applied research) is all part of one great quest. And, of course, it is just my point that it is. For example, the saying of something is accomplished by means of linguistic signs, and the nature of linguistic signs is, therefore, a crucial part of the answer to the central question. Thus, all research concerned with the structure of sign vehicles [One would need to reword that statement slightly for those who would not consider the sentence to be a kind of linguistic sign (6) , but the point would remain] contributes to the solution of the central question.

I do feel strongly that our quest has been hampered by our inability to deal effectively with the rest of the linguistic sign. The major obstacle, I contend, has been what I have called our "semantic myth". However, (I am still using the term "meaning" to refer to the rest of the sign and "semantics" for the study of meaning in that sense) I do not believe that semantics is impossible to study. My recently completed manuscript, Grace 1981b, is an attempt to show that it is possible to investigate how meaning works in ordinary language and to show what kinds of questions might inform such investigation. A key point, I think, is refrain from assuming that we know in advance what the central questions of semantics are.

Once we acknowledge the linguistic sign as a whole to fall within the competence of linguistics, the relevance of linguistics becomes very conspicuous and reaches very far. Fruitful collaboration with a variety of discipines seems feasible and mutually advantageous. In my view, linguistics should have much to contribute to, and to gain from, various of the social sciences and the humanities and even biology (because, among other things, the problem of the evolution of language seems ripe for reopening).

NOTES

2. In saying this I am, of course, claiming a connection between language and thought. Back up

3. One unfortunate direction in which this seems often to lead is to an assumption that the main functions of language are sociological, and, since (as we saw) much of the sociological functioning of language can be studied without any specialized knowledge of the languages involved, ultimately to the idea that language per se is just not an interesting object to study. The principal alternative direction that it leaves us is to concentrate our attention on the structure of sign vehicles. Back up

4. What I call specifying the sayable here is essentially what I called construing an idea in Grace 1981a. A "sayable" would be an idea, generally a relatively fully-formed one. Back up

5. I do not intend "event" as a technical term. The realities of many models are of a nature such that it seems natural to refer to them as events. For others, labels such as "situation" or "state of affairs" might seem more appropriate. In Grace 1981b I spoke of the meanings of such verbal models as "emic situations". Back up

6. I find it useful to recognize "ad hoc" signs (such as most sentences [but not all, cf. "A stitch in time saves nine"] are) as well as conventional signs (such as most words are)-- cf. Grace 1981b. Back up

REFERENCES

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

Grace, George W. 1981b. Ordinary language. ms. Back up

Kuhn, Thomas S. 1970. The structure of scientific revolutions, 2nd ed. International Encyclopedia of Unified Science, vol. 2, no. 2. Chicago: University of Chicago Press. Back up

Whitehead, Alfred North and Bertrand Russell. 1957. Principia mathematica, 2nd ed. (2nd ed. originally published 1925) Cambridge: Cambridge University Press. Back up

Last updated 5 January 1996