Referential Deflation and Internal-Mind Reference
Philosophy 362: Language and Thought
The problem of reference impacts all other areas of philosophy. On the traditional view of reference, in which there is a mediated relation between word and object, many metaphysical issues arise. That is, if words reference objects, then the truth-making of any given sentence should be satisfied by something “out there.” For instance, when I say “That bucket is sitting there,” this sentence is true only if there is actually a bucket sitting there. This is unproblematic enough (although issues of vagueness arise). The more problematic feature of this view comes with the truth-making of such sentences as “It’s possible for it to have rained today.” On the traditional view, there would need to be something in the world, or “out there” making this sentence true. Metaphysically, we should now posit a “possible world” in which, counterfactually, it did rain today. This possible world is what would make the modal claim of this sentence true.
Truly, there are several different conceptions of exactly what a possible world is, but any of them entail something like a mind-independent “way” the world could be. Now clearly, there are mind-independent conditions which pertain to our example sentence, but they do not satisfy its truth-conditions. Nomologically (that is, by the laws of nature), the weather could possibly have done any variety of things today. But this nomological fact cannot satisfy the truth-conditions of our example sentence above. “It’s possible for it to have rained today” is much too specific to be satisfied by any truth-making features of nomology. That is, the nomology of weather sets facts about what the world does in actuality, not what it might do in counterfactual situations. For instance, it’s just a law of nature that it might rain tomorrow. However, “it might rain tomorrow” is not satisfied by this law of nature. The truth-making of this sentence relies on it actually raining tomorrow (if this sentence refers to the outside world) either in this world or some other one. If it rains in “some other world” then the truth-making of this sentence is satisfied by the conditions in the other world, and if it rains in this world, then the truth of the sentence is satisfied by the conditions in this world. But in neither case does nomology fix what actually (or counterfactually) happens. Nomology is not a truth-maker. Thus, it seems that either there are really possible worlds (a generally accepted, but nonetheless bizarre result), or the truth-making of our example sentence is not mind-independent.
I said that “possible worlds” is a bizarre result. Allow me to clarify. It’s not an incorrect picture of how truth-making works because it’s weird. It’s an incorrect picture because it’s un-parsimonious. In philosophy, parsimony is always a desideratum of whether or not something should be posited. Parsimony is the idea that no new entities should be posited to explain phenomena, if the entities we already have will do the trick. That is, if something would add qualitative entities to the world beyond what’s necessary, it should not be posited. “Possible worlds” are certainly new qualitative entities, so what needs only to be shown is that they are not necessary posits. Truly any desiderata is something of a methodological bias, but the desiderata of parsimony is what restricts us to positing natural forces to explain gravity rather than demons or alien devices, for example. Well, in order to show that possible worlds are not necessary posits, one needs a different view of reference and truth-making – the correct one.
Before I offer other potentially correct views of reference, allow me to make an additional point which counts against the traditional view of reference. Consider the sentence “Johnny does not exist” or better yet “Hercules does not exist anymore.” On the traditional view of reference, there would need to be something “out there” making each true. Well, to say that there’s something “out there” which does not exist is a contradiction. If it doesn’t exist, then it’s not “out there” at all; it’s nowhere. This is the traditional problem of non-being, which linguistic realists have been wrestling with for quite some time. There is really no end in sight for this problem, however, for those who subscribe to the traditional view. This is not to say that their reason for subscribing to the traditional view is unreasonable.
The verb “refer” applies strictly to two-place relations. If there’s no relation occurring when we use words, then “refer” and “reference” should not be used in connection with language. This is generally pointed out as the reason that reference must involve a relation between two things. Because the word-object picture is the only theory of reference put forth, usually, which involves a two-place relation, it is defended with some zest. Some two-place relation must be going on in word reference in order for words to refer. Something must be going on here. There must be some relation that words have to something else, right? Wittgenstein didn’t think so. On the contrary, he argued that if words really do have a relation, then the relation is always indeterminate. That is for any word, w, it could refer to (or relate to) an infinite number of objects. For example, “bobsled” might refer to a bobsled today, but there’s really no way of knowing what it’s going to refer to tomorrow. The basic point of the argument is that the reference of a word involves interpretation, and interpretations can be infinite. And even when the reference of the word isn’t interpreted in innumerable ways, even if the reference of the word is fixed somehow, it could potentially apply to anything out there given interpretations. Interpretations are potentially infinite, and interpretations fix reference. While meaning may not be, and, I think, is not infinite, reference and meaning are two entirely separate things. An argument is needed here.
The difference between reference and meaning is crucial for Wittgenstein, and is one of his more popular points. To jump slightly ahead, Wittgenstein needs for words not to bear relations at all (in which case, do they really refer?). If meaning and reference are synonymous, then he’s just written himself out of the running for any word theory whatsoever. But I digress: for Wittgenstein, language simply has meaning (albeit ever-changing), and no reference to either mental or physical things. That is, reference for Wittgenstein can only be determined behavioristically (or dispositionally). It can be determined, but it is not a relation. It can be determined, but “refer” for Wittgenstein is no longer a verb. When I refer to something using a word (somehow in a non-two-place-relational way) on this view, it is only through my indicating of a thing or an idea through my other observable actions related to that thing. For instance, when I say “Bob,” the only way anybody knows what I’m talking about is how I behave toward other mentions of “Bob” and the person himself. Reference, then, is fixed for any instant in which I react to “Bob” or Bob. But it is not fixed for any instant after that. This is where meaning comes into play. Reference, essentially, is the what we’re reacting to, and meaning is the how we, and others, react to words. Both reference and meaning are fluid, and ever-changing, on this view. An important difference though is that Wittgenstein thinks we just strictly fail to refer when we use any word or phrase which expresses non-being (i.e. “Without” or “Bob doesn’t exist”). This seems intuitive enough, as we don’t normally think of ourselves as referring to anything when we mention a non-existent thing, but we do certainly mean something. Wittgenstein thinks so too. That is, we react to words expressing non-being in similar ways, and thus the communal meaning of, say, “Without” is fixed (for that moment) by how it’s used. In short, words reference our dispositions towards a thing or idea, and meaning is the dispositions themselves. This is a general overview of the conception, but the steps leading up to it need some fleshing out.
To begin with, though it seems self-evident from our preliminary discussion that words do not reference objects or anything “out there”, it’s not exactly plain that words do not, instead, reference our ideas or concepts. The traditional picture states that the mind has a relation to an object, and this is how word-reference is determined. This has already been rejected. But this does not mean that word-reference is not just internal to the mind, independent of any object-relations. This issue is addressed in passing by Wittgenstein. To begin with, if the mind is just an internal physical state, then, it is argued, mental states can only refer to other internal physical states. The general idea is that, on a physical monist conception of mind (physicalism), the only thing that internal physical states can stand in immediate causal relation to are other internal physical states. Thus, for Wittgenstein, we would only be able to refer to our internal mental states and never reference the outside world on a completely internal-mind theory of reference. There are some problems with this argument, but putting those aside for now, the other alternative suggested by Diane Proudfoot and Jack Copeland (speaking for Wittgenstein) is that of non-physical mental states which could then stand in immediate causal relations to anything whatsoever, and thus would not limit reference to mental states (Proudfoot & Copeland, 2002, pg 4). The supposed “price” of accepting this alternative view is dualism. Dualism is not explicitly argued against in this paper, and hence it’s not clear why this is such a sacrifice. In any case, this is where Wittgenstein’s argument against an internal-mind theory of reference falls apart. This is not to say to say that his theory of intra-linguistic reference falls apart at this point; that will be saved for later.
It is not at all clear why mental states, if they are just internal physical states, even need to reference the outside world at all. That is, even if internal physical states just reference other internal physical states, this is the not the same as saying that the outside world cannot causally interact with internal physical states. It is just saying that the outside world is not the thing ever referenced. So, the outside world on this view would stand in a non-immediate causal relation with internal physical states such that when reference occurs it is referencing the internal physical state acted upon by the outside world, but it is not directly referencing the outside world. For example, suppose there is some fox standing in front of me. The image of the fox is then imprinted on my mind, and then when I go to reference the fox, I say “fox”, and I am referring to what is imprinted on my mind, rather than the fox itself. This avoids the problem of non-being by way of stipulating that there need exist no external thing satisfying the reference of any word. But this is not immediately clear.
There is a certain reading on which the physicalist internal-mind theory of reference does succumb to the problem of non-being. Suppose that when I say “absence” or “lack,” I am referencing my idea or concept of such things. For some it may seem as though there would need to be an actual lack of absence contained in my idea of “lack” or “absence” for the word to refer correctly. This would be very odd, and, I think, unacceptable. Luckily, the idea referenced need only be about “lack” for the word to refer; it does not need to contain lack, whatever that is. This is so because, if internal physical states do interact so causally with other internal physical states, then there also need be no causal relation with the world in the case of most references made. Instead, ideas and thoughts generate new ideas and thoughts which can be referenced apart from any external, empirical experience. This is plausible enough that dualism need not be accepted, even though this would not be such a tragic result, prima facie. Dualism, however, might set the internal-mind theory of reference back to the problem of non-being.
If the mind is immaterial then it does bear immediate causal relations to the physical world. In which case, even when we reference an idea like “fox”, that reference stands in some immediate causal relation to the outside world. This is essentially the traditional view of reference in disguise, in this case. The whole advantage of a physicalist internal-mind theory of reference is that the reference of an idea does not stand in an immediate causal relation to objects, thus bypassing the whole non-being problem. Thus, though dualism is not really a bad thing, it does nothing to help this burgeoning theory of reference. The physicalist internal-mind theory of reference (hereafter PIM) also has other advantages over dualism, which skirt the objections by Wittgenstein.
Relations, says Wittgenstein, are either mysterious or indeterminate. Let us first talk about mysterious relations. Relations are supposedly very weird connections between the mind and object, which are formed in either an immaterial or a material way. In either case, material or immaterial, it’s very odd to think of their being an invisible line called a relation that exists between a mind and an object. With a dualistic internal-mind theory of reference, there still exists immaterial connections between the mind and an object. I suppose this doesn’t really count as an objection, but because it’s so odd, Wittgenstein wants to say that this cannot be how it works. PIM, thankfully, does not commit to any mind-object relations at all. Any relations PIM does have are entirely mind-mind. Reference on this view works like this: the mind generates a word which references back to an idea in the mind. These relations are not very odd at all. They don’t need to be either immaterial or material or even “things” at all. Indeed, it’s very odd to think of a relation as being a thing at all, which may mean that dualism also escapes this Wittgensteinian objection. Even supposing that relations are “things”, the PIM relations would be material relations: electrical neuron exchanges. I doubt very much that Wittgenstein could deny the existence of electrical neuron exchanges. So, relations really aren’t that mysterious. But there remains the problem of indeterminacy.
It’s very easy to say that referential relations are theoretically indeterminate, but it’s quite another thing to say that they actually work indeterminately. The most plausible version of the objection that referential relations are indeterminate is that, if reference is really a relation, then it wouldn’t work. That is, if words bear relations, then we could never be certain which thing(s) any given word referred to. This is supposedly unacceptable. The reality of the situation, however, is that we don’t have certainty of what thing(s) we’re referring to, and we don’t need certainty. We have our ideas of things, and we can reference those perfectly well, even if our idea of the thing is imperfect or uncertain. Certainty is not required for relations to function. Additionally, for whatever reason, we seem able to discern which idea of a thing we want to reference. That is, while the mind-mind relation of words and ideas isn’t exactly determinate, we’re able to figure out enough to at least make it function. So, while theoretically the mind-mind relation of words and ideas should be indeterminate, it’s instead semi-determinate. Why exactly it does not turn out to be indeterminate is worth some investigating, but at this juncture it’s enough that it’s, thankfully, not. It is worth nothing, though, that on the dualist conception of IM reference, the immaterial relations between mind and object would be, if not indeterminate, at least less determinate than those suggested in PIM. However, perhaps even those less determinate relations could function. Well, Wittgenstein’s elimination of the competition in terms of theory of reference has failed. But it remains to be shown that his intra-linguistic theory of meaning and reference is not plausible.
The central objection I’d like to raise to Wittgenstein’s intra-linguistic theory is one mentioned in passing by Proudfoot & Copeland. “On Wittgenstein’s view, thought is dependent upon the existence of language” (2002, pg. 17). This is un-controversially accepted, and is really the view that Wittgenstein expresses. But the contention that thought is dependent on language is patently false. This would entail that non-human animals or pre-linguistic humans (i.e. infants) lack thought altogether. This is falsified just by the fact that both non-human animals and infants react to stimuli, and not only that, but react in distinct ways. The logical entailment of Wittgenstein’s idea is that we have no thought before we learn language. In which case, there is no background language being translated into natural language when we use it. But this is falsified by introspection. We all have the experience of having something to say, but not knowing how to word it. Thus, thought and language cannot be synonymous. Also, occasionally thought occurs in mediums other than natural language such as images, sounds, even tastes, scents, and feelings. Wittgenstein would have to deny that any of this occurs, which is empirically falsified. Additionally, Wittgenstein is also committed to saying that there are no mental states. I’d like to ask, if this is the case, how is it then possible for the experience of emotions to occur? It seems that mental states, even when they don’t refer to the outside world, can be brought on or altered by emotional states. Does Wittgenstein now want to deny that there are states of the brain? We can empirically look into a brain and see how the flow of neurons is altered, disrupted, and that areas of the brain fire while others do not. These are mental states! Phenomenologically, we all experience our own inner states independent of language, and this I think is proof enough that language functions in some other way than intra-linguistic or behavioristically. I clearly have priviledged access to my own knowledge, apart from any external confessions of thought. Perhaps if Wittgenstein was to soften his theory to include the possibility of internal self-confessions of thought, then he would allow for mental states recognized linguistically and all would be well. In the meantime, I suggest that PIM is the most feasible of the three.
PIM avoids the problem of non-being altogether by stipulating that nothing need exist in the world to satisfy the conditions of non-being. Nothing need “exist” in the mind at all, or anywhere for that matter to satisfy the conditions of non-being. There is just the thought that is referenced. The thought does not even need to be of anything. Secondly, the physicalist relations it’s committed to are verified by neurology (i.e. neuron energy exchange). Thirdly, and most importantly, it’s compatible with everything phenomenologically experienced (i.e. emotions, mental-states, and pre-linguistic or non-linguistic thought). This is not to say that PIM is ultimately correct however. I believe that elements of all three theories of reference discussed account for specialized datum of neurology and linguistics, relevant to each theory’s strength. A compatibilist approach is recommended in approaching theories of reference and meaning.
Proudfoot, Diane; Copeland, Jack; “Wittgenstein’s deflationary account of reference” in “Language and Communication” 2002, published at www.elsevier.com/locate/langcom by Pergamon. Pgs. 4 and 17.
Wittgenstein, Ludwig; “Philosophical Investigations” (Blackwell Publishing Ltd., Malden, Ma, 2001)
[All rights reserved. You may not copy, reprint, or use my ideas without my express permission. If permission is obtained, you must use proper citation and notify me of publication.]