Three points are not clear to me about the relations between semantic externalism (Kripke, Putnam) and holism (Quine):

  1. Is there a way according to which externalism and holism can be held together or are they inherently contradicting each other? It seems that holism implies that the meaning of a term depends upon its linguistic description, its conceptual role or its relations to other terms within the web of language. According to externalism the meaning is (at least partially) external to the mere language. Does this mean that the two approaches contradict?

  2. Externalism about meaning says that there are two different kinds of sentences that can be said about an object - contingent and necessary. "Kripke is the son of X" will be necessary because it's essential to "Kripke". "Kripke is a logician", on the other hand, will be contingent. Quine says that there is no difference between Analytic and synthetic sentences. Now, the Analytic / Synthetic division is not the same as the necessary / contingent one, but yet, they are somehow connected. Does the fact that sentences can be divided in general into two categories create some problems to Holism?

  3. If the two do contradict, it seems to me that Holism would be a better choice only because it goes deeper. I can say that the sentence "Kripke is the son of X" seems to be necessary only because the predicate "son of" in the my specific language bares a linguistic relation to "necessary". In other words - it's only because my web of words if organised this way and not the other that some sentences seem necessary and some seem contingent. Therefore, this metaphysical characteristic of sentences is only an illusion. Holism says - meaning is in the language. Externalism says - given a language (or "in the current world") this and that are correct (water is H2O and not XYZ, for example). Hence, it seems to me like Holism goes deeper and can explain why externalism fails. Am I missing something?


3 Answers 3


Holism is an epistemological position, and externalism is a semantic one. Of course, some degree of interaction is to be expected, but not only is it possible to hold them together, it is not particularly challenging. The appearance of incompatibility comes from the misleading use of the word "meaning". In the holism, especially Quine's and Davidson's, "meaning" is an obscurity best left out of the vocabulary. There is no such "thing" as "meaning", sentences involving it rephrase facts about inference, relating concepts to one another, or acting based on them. Quine explicitly endorses Wittgenstein's "meaning is use", see Quine and Kripke’s Wittgenstein by Jónsson. It is not too far off to say that on Quine's (and Wittgenstein's) view the externalism/internalism distinction is altogether moot, there is no internal "there" (traditionally hypostatized as "meaning") there. The issue is pressing only for realists more robust than Quine.

In its turn, externalism explains how language user relates to her referents, not whether the reference is established atomistically or holistically. The chief opponents of externalists are not holists, but theorists of content, like Dretske and Fodor, who advocate the computational theory of mind creed that "meaning" inheres in some sort of materially coded "information" in the brain, and then proceed to explain how this "information" relates to "reality" (the problem of intentionality). It is true that Kripke gravitates towards realism, even essentialism, this is needed to ground his modal logic in the face of Quine's objections to its interpretability. There are also disagreements over Kripke's metaphysics vs Quine's empiricism, intensionality vs extensionality, de re vs de dicto, etc., contingent a priori, necessary a posteriori are also related to this. Tuboly's Quine and Quantified Modal Logic is a good review of these controversies from a historical perspective. It is even true that Putnam finds Quine's holism to be too much, and advocates the reality of natural kinds.

But all these issues are separate from externalism. One can perfectly well hold that we pick up semantics semi-holistically a la Davidson (by applying charity to the totality of other users' linguistic behavior), but that what we pick up is still external-context-dependent, and so the "meanings" (i.e. conceptual and operational roles) are different on Earth and Twin Earth. I suppose one could argue that holism undermines realism (Quine's claims to be a "realist" notwithstanding), and hence the distinctions that externalists uphold. But even anti-realists have to take ontology at least as a temporary make believe reflecting the best of current science, and within it the question re-emerges in a more pragmatic way: is it more effective to represent "meaning" in our theories as an engram in the head, or as spread over the environment. I suspect that Quine's physicalism and behaviorism would tip him towards the latter. The causal theory of reference, which is considered to imply externalism, has a distinct holistic flavor, reference is picked up opportunistically by imitating linguistic behavior of others.

  • 1
    Thanks! yet, I have a few more questions: 1. "meaning in use" and "meaning in inference" are more or less the same, aren't they? Yet Quine endorses the former and not the latter. How come? 2. conceptual role semantics is supposed to answer the problem of intentionality and is considered as a holist approach. I don't know if it's Quinean but it seems to be a holistic position about semantics rather than epistemology. 3. a hidden assumption I have in (2) that might be incorrect: the problem of intentionality is the problem of meaning, hence a semantic one. is it true? Is intentionality meaning?
    – Amit Hagin
    Commented Mar 7, 2017 at 15:43
  • @AmitHagin Inferentialism is one form of semantic pragmatism (use), some pragmatists are inferentialists (Brandom), others not (Dreyfus). Davidson's semantics is both pragmatist and holist, but truth conditional, not inferential. Pragmatist approach to intentionality is a 2-step: analyze "meaning" into more fundamental notions (inference, social pragmatics, habits of action), tell a story of how these come to be and relate to the world, more here.
    – Conifold
    Commented Mar 7, 2017 at 21:02
  • 2 more things: It seems from your comment that holism can be a semantic theory as well (not only epistemological). So is Holism as a semantic theory contradicting externalism a a semantic theory? Also - if semantic holism is true, then the meaning of a word doesn't need to refer to the world but to other linguistic terms. Hence, I think intentionality can be explained without any relation to the real world, isn't it the case? (I feel that I have semantic "powers" but I actually relate to nothing external to me).
    – Amit Hagin
    Commented Mar 8, 2017 at 11:41
  • and actually 2 more things (quite methodological): 1. Why would we make a sharp distinction between an epistemological theory and a semantic theory. It seems logical to believe that the way knowledge is established has a lot to do with the meaning words are given. 2. what exactly a semantic theory should answer to and what should a theory of intentionality answer to? I'm trying to understand the relations between epistemology, meaning and intentionality.
    – Amit Hagin
    Commented Mar 8, 2017 at 11:53
  • @AmitHagin I think most generally holism is a methodological approach that can be applied to different areas, Quine is best known for epistemological holism. "Semantic holism" would simply mean that semantics is not established word by word but at the level of sentences or texts or even languages as wholes. I do not see how it affects internalism/externalism though, that is about where it is "stored". Not even coherentists like Davidson claim that semantics/reference can be established without a relation to the world, but this relation is not seen as semantic, for Davidson it is purely causal.
    – Conifold
    Commented Mar 8, 2017 at 19:51

Conifold's answer seems to contain all the relevant materials, but I'll try to arrange them a bit differently.

So yes, there seems to be a certain collision between semantic externalism and holism. The reason is not because holism implies internalism (as your first point suggests) but because holism disputes the very distiction between internal and external (this is closer to your second point). To put it another way: externalism is indeed "about meaning". Holism is "about meaning" mainly in a negative sense: it rejects the traditional concept of meaning. As to grounds, externalism was based on a consideration of how a linguistic reference is fixed (one of Putnam's assumptions, in the Twin Earth argument for externalism, was that an item cannot be a piece of "meaning" if it does not serve to fix a reference). Holism was based on a consideration of the analytic / synthetic distinction, while abstracting from and mostly ignoring the very idea of linguistic reference.

As to your third point, I agree that there is a sense in which holism is a deeper alternative. Holism views things from a more flexible, less committed point of view. On the other hand holism is more abstract, and there is not much guidance how to apply it in new particular cases. So that one might "get lost" more easily with holism.


Consider this analogy:

Take a functionalist position in the Philosophy of Mind that says:

(F) The content of each mental state is determined by its relation to other mental states, and the input-output relations it has to entities outside the mind.

This is both an externalist and a holist position. And both the holistic and the externalistic theses it combines are ontological, not epistemological. They are about what is, not about how we can know, believe, etc. This does not mean that this functionalist thesis cannot have epistemological consequences when combined with a particular epistemology, but that is beside the point.

Combining semantic holism with semantic externalism is analogous. There is no prima facie reason why they are not compatible.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .