From the Tree to the Labyrinth

17

Five Senses of the Word “Semantics,” from Bréal to the Present Day

The term “semantics” has a number of different meanings, several of which seem to be completely at odds with one another. This state of affairs is often a source of considerable embarrassment in dealing with our students, to whom we find ourselves having to explain that our discipline is a bit like the country where some people call “red” what others call “white” and vice versa. With the result that, every time we use the word “red,” we would have to assign it a superscript or subscript number, specifying that we mean “red₁, in such and such a sense.”

Still, although the term “semantics” may have a number of meanings, those meanings are less irreconcilable than might at first appear.

In 1883 Michel Bréal (Les lois intellectuelles du langage: Fragment de sémantique) defined semantics as the science of meaning, but when he came to publish his Essai de sémantique in 1897 he gave it the more general subtitle Science des significations, and only in chapter IX, in which he proposed to examine “by what causes words, once created and endowed with a certain meaning, are induced to restrict, to extend, to transfer this meaning from one order of ideas to another, to raise or to lower its dignity, in short to change it,” does he say “it is this second part which, properly speaking constitutes Semantics or the Science of Significations.”¹

Semantics, then, is the science of meanings, but, for Bréal, only insofar as they are subject to historical development. And this is not all. Each time Bréal has to deal with the meaning of a word he proves incapable of isolating it from the set of enunciates, or more extensive fragments of text, in which the word appears. To give but a single example, in the chapter on the laws of specialization, Bréal is less interested in defining the meaning of the French word plus than in the fact that it takes on different meanings in different expressions.

The notion of semantics, then, is born, historically speaking, in reference to that imponderable entity we label meaning, but only to a lesser extent is it concerned with the meaning of words, or, to put it differently, of terms in isolation. For this, what was needed was not a science but an empirical praxis, lexicography in its most hands-on sense, that is, the actual compilation of dictionaries. Still, we must not forget that the whole of lexicography is simply the description of a langue, and therefore of an abstract entity, and not of the practical use of parole by means of which the speaker “means” something.

17.1. Various Meanings of Semantics

I would argue that the more or less explicit semiotics of former centuries did not question the fact that terms expressed something, but they did not presume that a special science was needed to clarify what that something was. Knowing the signs implied knowing either the things they referred to or the ideas they brought to mind, or the definitions given them by common consent, according to which the Latin homo, for instance, signified “a mortal rational animal.” In any case, for Aristotle, providing correct definitions was a task either for logic (see the Analytics) or for the various natural sciences, as is seen in his definitions of animals.

If we examine Abelard’s use of terminology, we remark that a verbal expression (i) significat a mental concept, (ii) designat or denotat its definition or “meaning,” and (iii) nominat the thing.

What we have here are three notions of semantics: (i) as the study of cognitive processes, (ii) as the study of dictionary or encyclopedia definitions, and (iii) as the study of the truth conditions of sentences. Many of our current problems stem from these medieval perplexities (and they are indeed perplexities: what exactly does a vox significativa do—signify, denote, or name?). Furthermore, Abelard’s threefold division is missing a fourth dimension, not unknown to previous semiotics, that of the disambiguation of complex texts (see Augustine’s De doctrina christiana, which is concerned with the “meaning” of a text like that of Scripture). And, lastly, there is also a fifth dimension missing, whose absence in Abelard does not imply its absence in medieval thought. What is missing is what we would call today a structural semantics as a theory of content, already present in the binary system of the division of predicables as represented in the Arbor Porphyriana (see Chapter 1).

Let us go back then, or let us look forward, beyond Abelard and beyond Bréal, and observe that, in the course of the debates on meaning, five areas of investigation have been identified, sometimes proceeding independently of each other, sometimes contradicting each other, and sometimes one of them presupposing—however acritically—the other:

1. Semantics as the study of the meaning of terms removed from any context (for instance, Carnap’s theory of meaning postulates, much of componential semantics, and the various forms of semic analysis, not to mention lexicography of every kind and tendency).

2. Semantics as the study of content systems or structural semantics (Hjelmslev and structural approaches to semantic fields in general et similia).

3. Semantics as the study of the relation between term (or sentence) and referent, or as the study of reference (for instance, Morris, Ogden, and Richards, much of analytic philosophy, and in primis Kripke). Let me remind the reader, however, of the distinction I posited in Kant and the Platypus between (i) providing instructions to identify the possible referent of a term and (ii) the act of reference itself.

4. Semantics as the study of the truth conditions of propositions expressed by sentences.

5. Semantics as the study of the particular meaning that terms or sentences assume in context or in the text as a whole (this is a vast and variegated field that is concerned with the meaning of the same sentences in different contexts and circumstances, for which we may cite in first and foremost the later Wittgenstein, as well as the theory of different discursive isotopies, etc.).

Any student of semiotics is familiar with all these meanings of semantics, and yet it would be optimistic to claim that this awareness is shared by all students of language—not to mention the fact that the semioticians themselves, though well aware of the definition of semantics cited in 3 and 4 above, are often prone to reject it as nonpertinent, or to consider it as a single problem, whereas in fact it involves two quite different problems.

17.2. Encyclopedia Entries

Let us consider a few examples of how the different conceptions of semantics may fail to recognize one another. In the Dictionnaire raisonné des sciences du langage of Greimas and Courtés, “truth” is defined.² The concept of truth might be foreign to a semanticist in senses 1 and 2, but it is certainly central for anyone concerned with senses 3 and 4 and, as Greimas demonstrates, it cannot be sidestepped by someone concerned with sense 5:

Truth designates the complex term which subsumes the terms being and seeming situated on the axis of contraries within the semiotic square of veridictory modalities. It might be helpful to point out that the true is situated within the discourse, because it is the fruit of veridiction operations; this thus excludes any relation (or any homologation) with an external referent.

I suspect that an analytic philosopher would find this definition confused and troublesome. But it would also be troublesome for someone concerned with semantics in sense 3, while it would have to be viewed with some indulgence by someone concerned with sense 4. In fact a truth-conditional approach is not concerned with establishing whether a given proposition is true but rather with what inferences one might legitimately draw if the proposition were true. What is certain is that, for someone who subscribes to sense 4, these truth conditions are posited within a “corporate body,” or coherent set of assumptions. When, however, doubt is cast upon the existence of this holistic system, as it is in Davidson, we may say that the principle of charity leads us to assume that a proposition makes sense and is therefore true within a discursive exchange (even if the sentence that conveys it does not present itself as especially perspicuous). Are we then to say that Greimas’s definition is so foreign to the analytical koine? I am not so sure. Granted, if someone says to me on the freeway “Look out, there’s a train ahead,” and I am aware that what is in front of me is a trailer truck, given that I have certain convictions concerning the real world, and appealing to the principle of charity, I assume that the speaker meant to say that there was a big rig ahead, and I let the communicative interaction go forward without a hitch. But do I only exercise the principle of charity in cases where I am able to counter an ambiguous sentence with certain convictions based on experience (that is, on what I consider to be true in the outside world), or do I not also behave in the same way when I attribute to someone else convictions that coincide with those I hold on the basis of a shared system of assumptions?

Let us suppose that a student of astronomy were to say to me: “Assuming that, after Galileo, the sun revolves around the earth and not vice versa,” I understand perfectly well, based on a shared system of assumptions, that he is asserting something false, but, applying the principle of charity, I assume that what he meant to say was what I consider to be true, in other words, the exact opposite of what he actually said (and that what he said was a common or garden lapsus), and I go on listening to his argument. In such a case I would be considering as true not what is confirmed by my experience of the outside world but what is guaranteed to be true by the holistic system of our received assumptions. We have only to stretch a little the notion of “situated within the discourse” for the conversation between a Greimasian and a Davidsonian, provided each of them exercises a reasonable principle of charity, to lose its dramatic edge.

But let us proceed with our examples. In the Einaudi Enciclopedia, under the entry “Semantica,” Diego Marconi—after defining semantics as the study of meaning—assumes that standard semantics is concerned only with natural languages. He devotes the bulk of his article to model semantics (sense 4), the semantics of possible worlds (senses 3 and 4), the dictionary vs. encyclopedia discussion (sense 1), but only in his introduction does he mention the existence of another tendency known as componential analysis (which is certainly concerned with meaning 1 and, for the purpose of establishing theoretically the number of components, presupposes sense 2).

The fact is that Marconi, at the time of writing, was an orthodox analyst and shared the conviction, common to many of his persuasion, that these problems were part and parcel of lexicography (senses 1 and 2); and for the analysts lexicography had nothing to do with semantics.

In the Encyclopedic Dictionary of Semiotics edited by T. A. Sebeok (1986), the entry “Semantics,” written by Bierwisch, at first tries to elude the often mortal embrace between semantics and the study of natural languages. After defining semantics as the study of meaning, Bierwisch excogitates the formula “A interprets B as representing C,” in which B is an object or an event, which permits it to be understood as something different from a phonation or a verbal enunciate. The author lists all the problems and takes into account the various positions, but in the body of the article he gives his own personal solution to these problems. Thanks to a happy decision by the editorial board, an entry on “Seme” (by Schogt) follows, in which we find a broad investigation into other linguistic positions, for the most part structural in their orientation (Lyons, Lamb, Pottier, Apresjan, Coseriu, Buyssens, Prieto, Greimas, etc.). Unfortunately, there is no mention of truth-conditional semantics—for some scholars the only semantics worthy of the name.

On the other hand, the Greimas-Courtés Dictionnaire, through a number of different entries, provides a review of the various lexicographical theories in the semantic and notional fields. They examine the componential theory, taken to the Hjelmslevian extreme of its own ambitions (how from “a score of binary semic categories, considered as the taxonomical basis of possible combinations” one can succeed “in producing several million sememic combinations”). They assume as prerequisites of any semantics that it be at once generative (recognizing too the work of the post-Chomskyan generativists), syntagmatic (attempting to overcome the limits of linguistic taxonomism and come to grips with a semantics of the sentence and finally of the text, sense 5), and general; semantics is not limited to the investigation of linguistic meanings but must address the semantics of the natural world insofar as it is made manifest by the various semiotics. I would say that senses 1, 2, and 5 are covered, and at this point we cannot expect Greimas to give his attention to model semantics or to the semantics of possible worlds (given that his treatment takes no account whatsoever of modal logic) or to truth-conditional semantics, considering the position he has staked out with regard to truth. The dictionary proceeds, then, with entries dedicated to Discursive semantics, Fundamental semantics, Generative semantics, Narrative semantics, Seme and Sememe, all, however, in a strictly Greimasian key.

The most balanced treatment is the one appearing in John Lyons’s two-volume Semantics, which represents a tolerant approach to all of the relevant traditions. Being on the one hand a linguist, exposed to the sirens of European theory, and on the other an insular Briton, did not prove a disadvantage for Lyons. But he does not formulate a theory, he expounds what has previously been said on the subject, and he can therefore afford to be ecumenical. Ecumenical, but hardly systematic.

Is ecumenism a pis aller, a necessity for the popularizer, or may it also be a theoretical choice? I would choose the second option. The problem of meaning is so complex that it is preferable to be syncretistic rather than a dogmatist and fundamentalist on the subject.

17.3. Does the Notion of Meaning Still Have a Meaning?

I remarked earlier that it is still a moot point whether or not semantics is concerned with the meaning of words (sense 1). It would appear that all we need do is pass from sense 1 to the subsequent meanings to accept that semantics is still concerned with meaning. It is arguable, however, whether the notion of meaning still enjoys citizenship rights in sense 3 (for some—for instance Quine—meaning can be shelved as long as one has a good theory of reference). Most interestingly, it is also debatable whether the notion of meaning (at least in the sense of a meaning conventionally agreed—sense 1) still has citizenship rights in sense 5. So long as, apropos of sense 5, we have in mind Greimas, in whom a generative semantics of texts is preceded by a structural semantics, there is no reason for this suspicion. But deconstructionists and Davidsonians, or those like Sperber and Wilson who subscribe to the theory of inference, can also be subsumed under sense 5.

Here meaning itself is called into question. In the case of Derrida, the denial of so-called transcendental meaning seems to be directed rather at the single meaning of a text (which he certainly calls into question) while sense 1 is not in question. In his De la grammatologie (Of Grammatology) he declares that, without the tools of criticism and traditional philology (including, I presume, dictionaries), interpretation could take off in any direction and consider itself authorized to say whatever it liked. Only he adds that this indispensable guardrail protects but does not initiate a reading, and he is evidently convinced that existing grammars and dictionaries are sufficient to protect a reading.

In the case of Davidson and the various theories of inference, he chooses to ignore the fact that terms have meanings fixed by the community (the ones provided by dictionaries) because what counts is that I take for granted that anyone speaking to me sees the world as I see it and intends to say what I would say in the same circumstances. It would therefore seem irrelevant that a boat be designated as a “boat,” because if someone were to say to me “Let’s get on that wagon,” pointing to a boat, I would understand, through the principle of charity, that he meant to refer to the boat and I don’t go splitting hairs about the “conventional” meaning of the terms.

The example, however, presupposes that all there is in front of us a boat, and not a boat and a wagon, and that the direction in which I am pointing is unambiguous. In the latter case, and in the absence of any further circumstantial indicators, if the speaker says “Let’s get on that wagon,” I understand that he wants to get on the wagon and not on the boat. This is a consequence of the fact that social and linguistic conventions assign two different meanings to the words “boat” and “wagon” independently of any context or act of charity. Of course, out of a principle of charity that, under the circumstances, would be tantamount to a “principle of malevolence,” I could always assume that the person speaking had a selection disturbance and said “wagon” whenever he meant “boat,” but we do not usually push the malevolence principle that far. We assume that there is a semantics in sense 1 involved, in which words have a certain meaning independently of any specific context.

Note that not even Davidson denies this evidence—see “Communication and Convention” (Davidson 1984b)—in which he asks himself if we need a convention to tie every word to a fixed meaning for all speakers, and assumes, as a condition of the existence of a convention, the position of Lewis, which is clearly more valid for poker than for languages. At this point Davidson realizes that we can even understand terms we are not familiar with and decides that all conventions are useful but not necessary. The argument is that we simply tend to speak like everyone else—and this would shift the problem of the existence of a code to that of consistency of usage.

To be quite frank, this strikes me as merely playing with words. Saying that we regularly associate the word “boat” with a floating vessel and saying that the code establishes that a boat is a floating vessel doesn’t change much. In fact when linguists speak of a code they are speaking of a statistical extrapolation from common usage: the code de la langue that De Saussure talked about is a fiction based on consistency of usage.

Otherwise, it would be like saying that it is not true that the penal code establishes that whoever kills someone else must serve x number of years in prison, but that “usually” (that is, as a rule) whoever kills someone else gets x number of years in prison. If this were the only difference between Roman Law and Common Law, what we would have is identical conventions. The difference is that, in order to decide what is customary, Common Law, has recourse, not to a rule fixed once and for all, but to the precedent set by a previous case.

Now, Davidson does not deny that there are conventions according to which a “boat” always signifies a floating vessel; he simply decides that this is a marginal or obvious case (obvious because it is marginal and marginal because it is obvious), and he prefers to give his attention to the more dramatic cases. The dramatic cases are when we use the word “boat” to indicate something other than a floating vessel. The most convincing example, of course, is that of metaphor (think of the example of “sauce boat”). But we cannot build a theory of a language on its use of metaphor, unless it be to say that the meaning of all linguistic terms is originally metaphorical—and I do not believe this was Davidson’s intent.

The confusion lies in demonstrating that what is dramatic is normal and what is normal marginal, whereas in science the dramatic cases are always used as marginal examples to demonstrate that the normal cases are not as simple as we think. True, the principle that the exception confirms the rule is scientifically infantile (something proved by Popper’s falsificationist theory, according to which an exception calls the rule into question), but to state, as humorist Achille Campanile does, that “rules made up entirely of exceptions are rules fully confirmed,” is to state a paradox, and it is equally paradoxical to claim that the exception constitutes the rule. The rule for defining a rule, in the human sciences, is that it must allow for a number of exceptions, but that they must be controllable, that is to say, predictable. In the physical sciences, either all bodies fall according to the laws of gravity or, if only one body does not, the laws of gravity must be called into question. In the human sciences, on the other hand, the statistical rule is that the majority of human beings come together in heterosexual congress in order to procreate (otherwise our number would not have increased from two to six billion in a matter of fifty years), but this does not exclude the fact that some human beings choose not to procreate, which allows us to include Catholic priests and homosexuals among human beings.

Were it true that there is no such thing as meaning in the sense of sense 1, we would have no end of trouble understanding each other, and in fact, Davidson, though through gritted teeth, has never defended this thesis. What we may be sure of is that, in terms of sense 5, the principle of charity theory must be taken very seriously. It is then that we discover that Davidson, by suggesting that he is contesting sense 1 (which he nevertheless presupposes) and seeming to place in discussion, for purely academic reasons, senses 3 and 4, was in fact proposing the principles of the semantics in sense 5—in other words, a semantics not of terms, or of sentences, but of texts. From the lexicographic point of view, Davidson seems to be denying the evidence, but from the point of view of a theory of textual interpretation he is a fairly sane person, or—though he is not aware of it—someone with something serious to say about the interpretation of the meaning of texts (texts that, on top of everything else, produced as they are in complex situations, are always multimedia; made up, in other words, of words, demonstrative and deictic gestures, paralinguistic elements, and maybe even hypoiconic supports).

Allow me to remind you of a well-known example of Ducrot’s. The expression je suis le rognon (“I am the kidney”), uttered by a human being, is false (from the point of view of senses 3 and 4), but, when said in the context of a restaurant, accompanied by a gesture first pointing to the dish in the waiter’s hand and then to the speaker himself, it signifies unequivocally that the speaker is affirming that he is the one who ordered the kidney and not the one who ordered the sirloin steak.

17.4. The Identification of Meaning and Synonymy

In order to deny that semantics makes sense in sense 1, to affirm, that is, that words do not have meanings agreed upon by convention, it is customary to employ a quite fallacious argument. Meaning is identified with synonymy. Philosophers of language are more responsible for this fallacy than lexicographers. No sensible person versed in languages can believe that there are two synonyms that really do mean the same thing (and even the authors of dictionaries of synonyms offer their alternatives as possibilities faute de mieux, as stylistic variants to be evaluated on a case-by-case basis, not as absolute equivalents). Once meaning and synonymy have been identified, however, the tendency is to demonstrate that, since there is no such thing as absolute synonymy (so much is obvious), there can be no meaning (which is not so obvious). This is the argument of those who hold that translation is an impossibility.

Let us consider a fairly curious text of Quine’s, “The Problem of Meaning in Linguistics,” which appears in his From the Logical Point of View. In it Quine speaks of lexicography and lexicographers, saying that lexicographers seem to be interested in the problem of meaning, and he is certainly correct. After which he opines that a lexicographer “differs from the so-called formal linguists only in that he is concerned to correlate linguistic forms with one another in his own special way, namely, synonyms with synonyms” (Bréal 1900). If by “lexicographer” we understand the author of a dictionary of synonyms, this is certainly what he does, though with all due caution, as we said before. But Quine seems to think that the only thing the lexicographer is interested in is deciding what linguistic forms are synonymous, that is, “alike in meaning.” This is not true. The lexicographer’s first task is precisely to establish why the same expression may have different meanings in different contexts. Rather than cultivate the myth of synonymy, a good lexicographer contests it.

If by “lexicographer” we understand someone who is writing a dictionary for tourists, and if he tells us “steak” is a synonym of bistecca, he may be taking advantage of the work of other lexicographers, but he is deliberately impoverishing it, though he may make it possible for an English-speaking tourist to order a bistecca if he happens to be in Italy. The good lexicographer is the one who explains that cagna is not a synonym of “bitch,” except in a few rare cases, so that in Italian I may define a lousy singer as a cagna, however impeccable her morals, but I could not call her a “bitch” in English without suggesting that she is of easy virtue (though her singing may be divine).

So lexicographers, real lexicographers, are indeed semanticists in sense 1, in that they endeavor to establish on what common bases we may legitimately use a word, but they are above all semanticists in sense 2 when they try to decide the nature of our lexicographic conventions, not in the caricatural terms of synonymy and homonymy, but on the basis of an inspection of the systems of content and basing their findings on a broad survey of previous texts and their meaning (sense 5).

When Quine says that lexicographers do not hold the monopoly on the problem of meaning, he appears to have in mind the authors of pocket dictionaries for tourists, rather than lexicographers who are scholars of structural semantics.

Just how debatable Quine’s ideas about lexicography are can be seen from the paragraphs which follow, in which he equiparates the work of the lexicographers to that of the phonologists who decide whether two phonemes are different according to whether or not the meaning of the word changes if we substitute one for the other within the same language. It is true that the phonologist decides that a given phoneme is different from another because, if we substitute one for the other within the same syntagm, we obtain two words with two different meanings (ship and sheep, for instance), but, when he does so, the phonologist is not concerned with the notion of meaning. He simply assumes that the native speaker (of whom he himself is a reliable sample) perceives a variation in meaning in the passage from one phoneme to another. He merely registers a fact, he does not remotely presume to define what a ship or a sheep is. The lexicographer on the other hand takes as a given the proof of substitution provided by the phonologist and is concerned with defining the difference between a ship and a sheep.

17.5. Truth-Conditional Semantics

Let us go on to the fourth sense of semantics. It goes without saying that if we have such an impoverished notion of lexicography and meaning as that of synonymy, we are free to experiment with phenomena such as the substitution of apparently synonymous terms in opaque contexts (and clearly, someone who believes that Aristotle wrote the Metaphysics does not automatically believe that Alexander’s teacher wrote the Metaphysics). These are exercises of considerable importance for the study of logic, but not very important for understanding the way we speak. No speaker in his right mind, once it had been affirmed that “Giorgione” has three syllables, would affirm that “Barbarelli” too had three syllables.³ I am one of the first to admit how many things we would have failed to understand if we had not performed exercises of this kind, but they have nothing at all to do with at least four of the five senses of semantics that I am talking about.

Let us come now to the differences between sense 3 and sense 4. It is my conviction that a truth-conditional semantics has nothing to do with the problem of reference. The problem of reference has to do with our ability to designate objects and states of the world, to reach an understanding on this act of designation (and hence it has something to do with sense 1), and—eventually (but this is not a semantic but an epistemological and gnoseological problem)—to say whether the object or the state of the world we referred to exists or is taking place to the extent that we referred to it. In simple terms, if I say that it is raining today, we have to be agreed on the meaning of “rain,” we have to grant that the speaker is saying that water is falling from the sky, and that (another problem) water is actually falling from the sky.

Let us take a look at Tarsky’s truth criterion. Its concern is with how to define the truth conditions of a proposition, but not with how to establish if the proposition is true when used for acts of reference. And saying that understanding the meaning of a sentence means knowing its truth conditions (that is, on what conditions the proposition expressed would be true) it is not the same thing as proving the sentence to be true or untrue.⁴

Agreed, the paradigm is nowhere near as homogeneous as is usually maintained, and there are those who tend to interpret Tarsky’s criterion according to a correspondentist epistemology. But, whatever Tarsky may have thought,⁵ it is hard to read in a correspondentist sense his famous definition:

The sentence [i] “snow is white” is true if, and only if, [ii] snow is white.

We are in a position to say what kind of logical and linguistic entity [i] is—it is a sentence in an object language L that conveys a proposition—but we have no idea what [ii] is. If it were a state of affairs (or a perceptive experience) we would be extremely embarrassed: a state of affairs is a state of affairs and a perceptive experience is a perceptive experience, not a sentence. If anything, a sentence is produced to express a state of affairs or a perceptive experience. But if what appears in [ii] is a sentence about a state of affairs or a perceptive experience, it cannot be a sentence expressed in L, since it must guarantee the truth of the proposition expressed in [i]. It must be, then, a sentence expressed in a metalanguage L₂. But in that case Tarsky’s formula ought to be translated as follows:

[i] The proposition “snow is white,” conveyed by the sentence (in L) snow is white

is true if, and only if,

[ii] the proposition “snow is white” conveyed by the sentence (in L₂) snow is white is true.

It is clear that this solution is destined to produce a sorites of infinite sentences, each one expressed in a new metalanguage. If we assume Tarsky’s example in a naïve fashion, we find ourselves in the same boat (or wagon) as the editors of De Saussure who represented the relationship between signifier and signified with an oval split into two superimposed halves, in which the word arbre is contained in the bottom half while the top half contains a drawing of a tree. Now, the signifier arbre is certainly a word, but the drawing of the tree is not intended to be and cannot be the signified or a mental image (because what it is, if anything, is another nonverbal signifier that interprets the word below it). Seeing that the design excogitated by De Saussure’s editors had no formal ambition, only a mnemonic function, we can forget about it. But the problem with Tarsky is different.

We could of course interpret the definition in a strictly behaviorist sense: snow is white if, when confronted by the stimulus snow, every speaker reacts by saying it is white. But, apart from the fact that we would find ourselves up to our necks in the difficulties of radical interpretation, I do not believe this was what Tarsky had in mind, and, even if this was his intention, this would not be a way to decide whether a proposition is true, because it would simply tell us that all speakers are guilty of the same error of perception, just as the fact that for thousands of years all speakers declared that the sun sinks into the sea in the evening does not prove that the proposition was true.

It seems more convincing to admit that, in Tarsky’s formulation, [ii] stands conventionally for the assignment of a truth value to [i]. The Tarskyan state of affairs is not something we can check in order to recognize the proposition that expresses it as true; on the contrary, it is what a true proposition, or indeed anything that is expressed by a true proposition, corresponds to (cf. McCawley 1981: 161), in other words its truth value. In this sense Tarsky’s notion does not tell us whether it is more true to say that a cat is a cat than it is to say that a cat is a mammal.

17.6. Meaning, Referent, Reference

This node, between truth-conditional semantics, reference semantics, semantics of the sentence, and textual semantics compels us to revise a few concepts, something I attempted to do in my Kant and the Platypus.

1. Meaning of a term like cat. It may contain categorial elements like mammal and feline, but it also contains instructions to define the referent. But the referent of cat is not some individual cat, but cats in general. In this sense, terms that denote nonexistent objects may contain reference instructions (for instance, “unicorn”). It is possible to transmit and understand instructions for identifying the referent without having had and without ever having the occasion of referring to something.

2. References to cats (My cat is in the kitchen). Reference as a linguistic act, to be negotiated. The reference, however, is completely unconnected either to the empirical truth of the proposition (if I am lying or mistaken, the cat is not there) or to any discussion of its truth value. To be sure, if there is a cat in the kitchen, there is a feline mammal in the kitchen (sense 4), I am certainly referring to a specific cat, mine (sense 3), but it might not be true that it is in the kitchen. Senses 1 and 2 are presupposed by semantics of the third type, and can be taken into consideration by semantics of the fourth, but the converse is not true. Let us not forget that Morris (1938) reminded us that semantics is concerned with the relationship between a sign and its designata, that a semantic rule establishes under what conditions a sign is applicable to an object, but that the notion of designatum has nothing to do with the existence of the object; the designatum of a sign is something the sign may denote—but establishing whether there really are objects of this kind goes beyond the competence of semiotics.

Let us now suppose that I were to visit a culture (with a language adequate to express it) in which only two animals are known: the cat, hairy, smaller than a human being, domesticated, and harmless, and the crocodile, usually bigger than a human being, and scaly. For the members of that culture, based on such an elementary system of oppositions, which constitutes the full extent of their classification of the animal kingdom (a cat is everything a crocodile is not, and vice versa), if a dog were to show up, given that it was hairy, domesticated, and friendly, it would be defined as a cat (however unusual its appearance) and certainly not as an unusual crocodile. Let us suppose again that I realize that there is a boa constrictor behind my native interlocutor’s back. I wouldn’t be able to tell him that it was a boa because there is no adequate term in his language, and I couldn’t describe the strange and unusual animal without wasting precious time. I would therefore have to tell him that there was a crocodile behind him, assuming that, since in that culture animals are divided into harmless and hostile, I would thus be informing him that he was in a dangerous situation. This example is not chosen at random because in some medieval encyclopedias, not knowing how to define a crocodile (since the author had probably never seen one), they were content to call it a serpens acquaticus.

If I succeed in causing my interlocutor to be concerned, as was my intention, and if I obtain his consent to my proposition (he turns around, gives a start and concedes that the animal, obviously not a cat, is indeed a crocodile), I will have behaved according to certain methodological principles of sense 2, to make a successful reference in the sense of sense 3, obtaining his consent in terms of sense 4.

But in fact all this is because I am basing myself on the principles of sense 5, according to which it is the text and the context that have the last word in defining the meaning of terms.

This whole discourse will no doubt lead someone to opine that there is no semantics that does not need to be backed up by a pragmatics. I can only agree, as indeed I always have, from my A Theory of Semiotics (1976) to Kant and the Platypus (2000).

Paper presented at the symposium “La semantica fin de siècle: Dalla fondazione di Michel Bréal all’attualità della ricerca” (“Fin de siècle semantics: From Michel Bréal’s foundation to contemporary research”), held at the Center for Semiotic and Cognitive Studies of the University of San Marino in November 1997. It was later published with the title “Cinque sensi di ‘semantica’ ” [“Five Meanings of ‘Semantics’ ”] (Eco 2001).

1. English translation: Bréal (1964: 99).

2. English translation: Greimas and Courtés (1982: 432).

3. [Translator’s note: The Venetian painter Giorgione was born Giorgio Barbarelli. The two names are used as a paradoxical example in Quine 2004.]

4. “To understand a proposition means to know what is the case if it is true” (Wittgenstein 1922: 4.024). Nevertheless, we know the truth conditions of the proposition: “At twelve noon on August 2, 1810, all living cats made up an odd number,” but it is unlikely that either today or in the future anyone will ever be able to say whether it is true or false.

5. In any case, here is what Tarsky thought in 1944: “We may accept the semantic conception of truth without giving up any epistemological attitude we may have had; we may remain naive realists, critical realists or idealists, empiricists or metaphysicians—whatever we were before. The semantic conception is completely neutral toward all these issues.”