Possible Language Semantics
Possible Language Semantics
The theory of meaning is supposed to concern the semantics of our language: in virtue of what does human language have the meaning it has? But do all possible languages share the same semantics? Might a semantic theory that is incorrect for our language correctly describe the semantics of some alien language? If so, what are the implications of that fact? This bears on the question of whether meaning is a natural kind or is more like an artifact (a human construct). I shall urge the latter view: meaning can be multiply realized, has no deep real essence, and can be varied at will—like furniture or marriage ceremonies.
Let’s start with something simple: the semantics of names. It has been argued at length that names in our common human language are not synonymous with descriptions, at any rate certain types of descriptions. Let’s accept that. But does it imply that in other languages the description theory of names has to be false? Clearly not: there are such things as descriptive names, stipulated to be so. It might be that speakers always have in mind a synonymous description for any name they use; they use names simply to save breath. This is what they have decided—or what they have become accustomed to. It is up to them what they mean by their words. Similarly, speakers of Martian might expressly deny that their definite descriptions obey Russell’s theory; they subscribe to Strawson’s theory (they may be incompetent with quantifiers). Or what if a positivistic population subscribed to a verification conditions theory of meaning, viewing anything else as meaningless? They refrain from saying anything unverifiable, on pain of imprisonment for repeat offenders. Being good positivists, they never utter metaphysical sentences; perhaps their brains prevent them from constructing unverifiable sentences. Then, wouldn’t it be true that their language obeys a verificationist semantics? Or suppose the Venusians are deeply wedded to possible worlds, talking of them quantificationally all the time; their linguists are possible worlds semanticists to a man. Wouldn’t it be reasonable to attribute this semantics to their modal words? It may well be that they consciously intend such an ontology in their modal talk. What about a nation of Platonists who fervently believe their predicates denote Platonic universals—wouldn’t it be correct to attribute this semantics to their words (whether or not there really are such things)? Or consider an imaginatively rich population who have mental images for every word they utter (unlike us): couldn’t these images form the meaning they attach to their words? So long as the assigned entities can be combined according to syntactic rules to form meanings of whole sentences, almost anything would do, even ordinary objects. It’s the form that matters not the substance. What if an emotional group spoke only expressively—no assertions of truth-evaluable sentences? Couldn’t they do this, so that the semantics of their language is purely expressive? It’s a free country and all that. A certain population might restrict themselves to a Tarski-Davidson semantics—and good luck to them. This bunch would speak a language for which Tarski’s semantics is true. There might be Montague languages or Carnap languages or Austin languages or Russell languages or Grice languages or Tractatus languages or Investigations languages. The more the merrier. Maybe we don’t speak these languages, but they do—this is what they mean (and their sentences too).
Thus, the nature of meaning can vary from one possible language to another, as a matter of principle. Syntax can vary from one language to another, and so can phonology, and so can pragmatics—so why not semantics? Because languages are artifacts, like musical forms or fashions or dwellings. So long as the words can be used to communicate (serve this function), they have meaning. They employ different means to the same end. Combinatorial power needs to be added too, but it doesn’t much matter what gets combined. The correct semantic theory is a contingent matter, resulting from history, biology, decision, stipulation, and happenstance. Meaning is like the law or means of conveyance or types of dance. The Tractatus and the Investigations might both be true—of different populations. So long as the language is logically possible, someone could speak that way; and no one denies that such languages are logically possible, just incorrect for our language. Aliens might communicate using electromagnetic waves—so what? As long as the words convey the right signals into another brain we have meaning. It’s not like water and H2O, where the former is necessarily the latter.
It is therefore misleading to speak of the theory of meaning, as if all conceivable meaning must be identically composed. In some languages there might be just one sense for any given reference, or meaningful words that are never used, or mental images corresponding to logical connectives, or languages that are never employed to communicate but only to think with. No doubt there will be some common features such as combinatorial potential or meaningfulness, but the variety of existing semantic theories might be actually applicable somewhere—from image theories to direct reference theories. It may even be that our language obeys different theories in different parts, as has been contended by some theorists: some parts are all about truth conditions, other parts are purely expressive, yet other parts are merely punctuational (Wittgenstein thought something like this, language being a family resemblance concept for him). There is no one thing that every instance of meaning has to be. Meaning is essentially heterogeneous. The various theories may not be true of our language (as a whole anyway), but they are true in some possible world of some languages. Are all languages even referential—mightn’t some speakers use them just for inner reverie, referring to nothing, or just to soothing shades of blue? Meanings are made not found, and they come in great variety (See The Varieties of Meaning by Professor John Smith, PhD). In some possible languages sentences really do denote truth-values (a la Frege), while in others words stand only for imagistic ideas (a la Locke). It all depends on the speakers. Meaning is protean.[1]
[1] One can’t help feeling that theorists of meaning, though discussing only human languages (European ones at that), really had their sights set on all possible languages; they never say, “Of course, this only applies to human language—other possible languages might have quite different kinds of meaning”. They tacitly assume semantic uniformity. How does the linguistic turn fare once we recognize that meaning is diverse and heterogeneous across possible worlds? How philosophically significant is the philosophy of language if meaning is as variable as life itself? Different types of mind produce different types of language. Human language reflects human nature, not Martian or Venusian. What about octopus semantics?
Languages are systems of categories that speech communities have constructed over time to help them make sense of the world. The question is, by what mechanism did the categories for this language (specifically the ones that are the result of efforts to make sense of the world) come to be what they are at this point, and what are the specific differences between the categories of this language and those of some other language? I’ve noticed that philosophy rarely focuses attention on linguistic systems, as opposed to instances of systems in acts of language use.
But you spoke of “further significance”, so I’m wondering what the prior significance is and how meaning is further than it. I find the phrase obscure.
I.e., the act of producing an “acoustic blast” has a significance (for people making these blasts) beyond just the production of funny noises. The question is, where does that “further significance” come from? The difference between the effect of an insult vs a praise on an addressee is not due to the noises.
That seems perfectly reasonable and what philosophers assume.
Sorry for the obscurity. Trying to give brief comments. For me the term ‘utterance’, as in “natural language utterance” or “utterance act” refers to only the perceivable aspect of the sign (or the production of the “acoustic blast” part of the sign). (Searle, e.g., was not as strict in his use of the term ‘utterance’.)
It doesn’t seem to me that the phrase allows us to dispense with the concept of meaning.
Well, the phrase (“further significance”) was meant to be only a temporary expressive device, not a theoretical category. It’s intended to have an open-ended denotation, with the only principle being “connected to the event, just not the “acoustic blast”. Then I could narrow down the intended referent by describing what I took to be the phenomenon of interest by other language, not using the term ‘meaning’. In fact (to connect with your comment on the other post), I have an example of “remote physical effects of an acoustic blast”, as an example of what needs to be excluded. I haven’t dispensed with the concept of meaning and I don’t want to dispense with it. I’m pretty sure I’ll use the term ‘meaning’ again eventually (when I’ve arrived at a relatively clear idea of what the object of interest is), but it’s difficult to continue to use it uncritically in trying to be accurate in referring to that puzzling intentional object that I’m interested in. So I would like to clarify the special usage of that term, to sharpen its connection with the world, to be clear at least in my own mind what I’m actually referring to. BTW, I prefer to use the term ‘category’ (as opposed to ‘concept’ (and I think Kant should have done that too)) to refer to objects like (what are usually called) meanings of lexemes, but I’d better stop so far.
Thank you for the pushback, because if I ever try to publish something on it I’ll need to deal with issues like this.
Yes, indeed.