4.6.4 Vanville
All this obliges us to offer some reflections on the concept of truth. Is there a difference between saying it is true that something is an egg and saying it is true that something is a mammal? Or between saying it is true that something is a mountain and saying it is true that something is a MOUNTAIN? If the continuous oscillation (between structural organization and interpretation in terms of experience) I was talking about before did not exist, the answer would be easy: to say that something is a MAMMAL or a MOUNTAIN can only be true within a language L, while saying that something is an egg or a mountain is true in terms of experience. Yet we have seen that even in order to recognize an egg, we cannot elude the restraints imposed by a language L, the same one by virtue of which it is decided that BIRDS are such insofar as they lay eggs (but not all animals that lay eggs are BIRDS).
There is a definition of truth in the Dictionnaire by Greimas-Courtes (1979) that seems tailor-made to irritate any upholder of a truth-functional semantics, not to mention every supporter of a correspondence theory of truth:
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, for it is the fruit of the veridiction operations; this thus excludes any relation (or any homologation) with an external referent.
Perhaps the Dictionnaire has discovered the most complicated way to say something that is by no means simple but has nonetheless been said before: and that is that the concept of truth should be seen within the context of a system of content; that the propositions that the receiver already deems guaranteed within the framework of his own cultural model are “true”; and that the interest of the analysis has shifted from the protocols for defining an assertion as a true one (the logical neopositivist stance, and that of the young Wittgenstein) to the analysis of the discursive strategies that present something as true.
This position was less scandalous and less impermeable to the (apparently opposite) discourses of analytic philosophy. Greimas’s position is based on a Hjelmslevian version of the structuralist para digm, and the Hjelmslevian version anticipated (and when it was not anticipating, it was on a parallel course: the dates speak volumes)17 the development of that internal criticism of logical neopositivism and analytic philosophy that goes by the name of holism, the challenging of the difference between analytic and synthetic, the principle of warranted assertion, internal realism, and the individuation of scientific paradigms as incommensurable structures (or, in any event, as structures that do not admit of simple translation from one to the other). Even though indirectly, Hjelmslev’s version also influenced, rather than anticipated, the criticism of knowledge as Mirror of Nature, and Rorty’s (1979) idea that every representation is a mediation and that we must drop the notion of correspondence and see propositions connected with other propositions rather than with the world.
The only difference is that from what is known as the holistic standpoint there is in any case a tendency to define in what sense something can be assumed as true, albeit in terms of “warranted assertion,” while in the semio-structuralist school, of which Greimas perhaps represents the most radical wing, the thrust of research was directed at understanding how a discourse makes one believe that something is true.
The limitations of the semio-structural approach lie in the fact that, to be able to say whether and how people accept something as true, and to make them believe it is true, we must also assume that there is a naive concept of truth, the same one that authorizes us to say that the sentence It’s raining today is empirically true—within the context in which it is uttered. I do not think that this criterion exists within the structuralist paradigm.
The trouble is that it does not exist within the truth-functionalist paradigm either. In any event, it is not provided for in Tarski’s criterion of truth, which concerns the way in which the truth conditions of a proposition are defined but not how to establish whether or not the proposition is true. And to say that understanding the meaning of a sentence means knowing its truth conditions (that is to say, understanding under what conditions it would be true) does not amount to proving whether or not the sentence is true.
Agreed, the paradigm is by no means so homogeneous as it is usually held to be, and some also tend to interpret Tarski’s criterion in accordance with a correspondence theory of truth. But, whatever Tarski thought,18 it is hard to read in a correspondentist sense the canonical definition:
(i) “Snow is white”
is true if and only if
(ii) snow is white.
We are able to say which type of logical and linguistic entity (i) is—it is a statement, or a sentence in a language L, which conveys a proposition—but we still have no idea of what (ii) is. If it were a state of affairs (or a perceptual experience), we would be extremely puzzled: a state of affairs is a state of affairs, and a perceptual experience is a perceptual experience, not a statement. If anything, a sentence is produced to express a state of affairs or a perceptual experience. But if what appears in (ii) is a sentence regarding a state of affairs or a perceptual experience, it cannot be a sentence expressed in L, given that it must guarantee the truth of the proposition expressed by the sentence (i). It must therefore be a sentence expressed in a metalanguage Lr But then Tarski’s formula must be translated as
(i) The proposition “snow is white,” conveyed by the sentence (in L) S now is white, is true if and only if
(ii) the proposition “snow is white,” conveyed by the sentence (in L,) Snow is white, is true.
This solution is clearly destined to produce a series of infinite sentences, each expressed in a new metalanguage.19
Unless we understand the definition in a strictly behaviorist sense: snow is white if—when confronted with the stimulus of snow—each of the speakers reacts by saying that it is white. Apart from the fact that we would find ourselves up to our necks in the difficulties of radical translation, I don’t think this was what Tarski thought, and even if it were, this would still not be a way of deciding whether a statement is true, because it would simply tell us that all speakers make the same perceptual error, just as the fact that for thousands of years all speakers said that in the evening the sun fell in the sea is not proof that their statement was true.
It seems more convincing to admit that, in Tarski’s formula, (ii) conventionally stands for the assignation of a truth value to (i). The Tarskian state of affairs is not something we can check in order to acknowledge that the proposition it expresses is true; on the contrary, it is that thing to which a true proposition corresponds, or everything that is expressed by a true proposition (see McCawley 1981:161), in other words, its truth value. In this sense the Tarskian notion does not tell us if it is truer to say that a cat is a cat or that a cat is a mammal.
Which brings us back once more to the question of whether there are truth criteria for observation sentences that are different for nonobservation sentences.
Since such questions were debated in exemplary fashion in Quine’s “Two Dogmas of Empiricism,” I shall recycle a story I re-counted in 1990 in the course of a conference on Quine himself.20 I specify this, because otherwise it would not be possible to understand the names of the streets and the localities I use (all refer to famous examples taken from Quine’s works)—neither the name Vanville attributed to the city (Van is how Willard Van Orman Quine was known to close friends) nor the passing reference to a brick-built house on Elm Street, a typical example of an observation sentence used in Quine 1951.
Figure 4.1 shows the map of Vanville, a little town that first grew up north of the river Gavagai in the days of the first pioneers. Vanville is made entirely of wooden buildings, including the Presbyterian church, with the exception of the Civic Center, where at the beginning of the century they constructed three masonry buildings with cast-iron columns. The map also shows a house made of bricks on Elm Street, but this was built in 1951, and we shall have more to say about it later.
As can be seen, Tegucigalpa Street, Pegasus Street, and Giorgione Street run perpendicular to Elm, Orman, and Willard streets as well as to Riverside Drive. A sort of Broadway, called Tully Road, tells us that Vanville is not necessarily a Roman castrum, but that its development was inspired by a certain Anglo-Saxon empiricism. Then there are Midtown Place and Uptown Square, and between Midtown Square and Elm
