Fodor, Psychosemantics, Chapter 3:

Meaning Holism

The Issue: Fodor introduces Meaning Holism as a doctrine that “... looks to be entirely destructive of the hope for a propositional attitude

psychology” (p. 56).

The Background to the Problem:

- Fodor introduces the notion of ‘epistemic liaisons’

- “When an intentional system takes the semantic value of P to be relevant to the semantic evaluation of Q, I shall say that P is an epistemic liaison of Q.”

- epistemic liaison is a psychological notion, not an epistemological one.

- Fodor introduces “Meaning Holism” (MH)

- “MH is the idea that the identity - specifically, the intentional content - of a propositional attitude is determined by the totality of its epistemic liaisons.”

- The problem outlined:

- “... one way that intentional psychologies achieve generality is by quantifying over all the organisms that are in a specified intentional state... But now, if - as surely is the case - people quite generally differ in their estimates of epistemic relevance, and if we follow MH and individuate intentional states by the totality of their epistemic liaisons, it’s going to turn out de facto that no two people... ever are in the same intentional state... So no two people will ever get subsumed by the same intentional generalizations. So intentional generalizations won’t, in fact, succeed in generalizing. So there’s no hope for an intentional psychology” (p. 57, paragraphs 1 & 2).

Outline of Chapter Three:

Given the nature of MH, Fodor sees two available options for intentional psychology :

A) Option #1: Live with MH and look for ways to mitigate its consequences for intentional psychology

B) Option #2: Attempt to undermine MH

A) Mitigating the Consequences of MH

1. Graded Notion of ‘Same Intentional State’

2. The Platonic Reply

B) Undermining MH

Fodor argues against MH from each of the three ways that philosophers have tried to establish it.

1. The First Way: From Confirmation Holism to MH (a.k.a. epistemology via the assumption of confirmation holism - p. 60)

2. The Second Way: From Psychofunctionalism to MH (a.k.a. from the philosophy of mind via the assumption of ‘psychofunctionalism’ - p. 60)

3. The Third Way: From Functional-Role Semantics to MH (a.k.a. from the philosophy of language via a ‘functional role’ theory of meaning - p. 60)

Before arguing against MH, Fodor outlines the general form of arguments for MH (each argument is a specialized form of this sort of argument):

The Ur-Argument for Meaning Holism (p. 60)

Step 1: Argue that at least some of the epistemic liaisons of a belief determine its intentional content.

Step 2: Run a ‘slippery slope’ argument to show that there is no principled way of deciding which of the epistemic liaisons of a belief determine its intentional content. So either none does or they all do.

Step 3: Conclude that they all do (1,2: modus tollens).

A1: Graded Notion of ‘Same Intentional State’ (p. 56 - 57)

- Fodor claims that with a graded notion of same intentional state, two people could more or less believe the same thing (ex. that Callas is better than Tebaldi).

- having a graded notion of ‘same intentional state’ could presumably also allow two people to be subsumed by an intentional generalization, which would “...buy back the predictive power of intentional psychologies.”

- setting up the objection: Fodor states that there are two senses of ‘more or less believing that P’

i) the commonsensical ideas that two people can differ in degree of their epistemic commitment to P

ii) whereas you can grade epistemic commitment, it is far from commonsense that you can grade propositional identity - the belief that P would be unequally the object of our attitudes.

A2: The Platonic Reply (p. 58 - 59)

- the appeal to Platonic constructs

- first admit, and then ‘idealize away from’ variations in estimates of epistemic relevance

- imagine the generalizations of an intentional psychology to be strictly satisfied only in the case of communities of ideally homogeneous believers

- the predictive power of the theory increases asymptotically as the domain that it’s applied to approaches this ideal

- objection: “The problem isn’t that MH forces us to scientific idealization; it’s rather that MH makes any old idealization seem just about as good as any other”

- “Which ‘ideally homogenous’ community of intentional generalizations should we idealize to when we specify the domain of intentional generalizations about people who believe that....”

- MH says we must choose, but gives us no idea of how we are to do so

p. 59: “So much for trying to reconcile a substantive Intentional Realism with a holistic account of the individuation of the attitudes.”

B1: The First Way: From Confirmation Holism (CH) to MH (p. 62 - 67)

- a.k.a. from epistemology via the assumption of confirmation holism

- CH: the belief system is the minimal unit of confirmation

- MH: the belief system is the minimal unit of intentional content

- Verificationism: the doctrine that the content of a belief is identical to the means of its confirmation

- Fodor’s Argument: you can get to MH from CH, but to offer CH as an argument for MH is ill advised for two reasons:

1) you would need verificationism as a premise, and verificationism is false

2) an inference from CH to MH courts circularity because MH is presupposed by the standard arguments for CH

- Quine’s argument against verificationism (the Positivist account of confirmation) from “Two Dogmas of Empiricism”

B2: From Psychofunctionalism to MH (p. 67 - 71)

- a.k.a. from the philosophy of mind via the assumption of ‘psychofunctionalism’ (p. 60)

- a meaning holist has to make a case that the content of a belief is determined by its epistemic liaison; that what determines the propositional object of a belief is its place in a belief system

- Fodor describes the positions of the Logical Behaviourists and the Central State Identity Theorists: both have problems in regards to the attitudes (see page 69)

- enter psychofunctionalism

- since psychological particulars are physical, they can enter into whatever causal relations physics acknowledges; which makes psychofunctionalism better than behaviourism

- since psychological properties are relational, they can be exhibited by non-neural particulars; which makes Psychofunctionalism better than the Central State Identity Theory

- psychofunctionalism supplies step 1 of the Ur-argument: epistemic liaisons are among the relations that a belief state has essentially (so there is a route from psychofunctionalism to MH)

- Fodor’s Argument: The trouble is with the premise of the argument: our best grounds for Psychofunctionalism don’t justify a version of that doctrine nearly as strong as what the inference to MH requires

- psychofunctionalism does not underwrite the claim that the belief that P, being a belief that has a certain content, is a matter of having the right connections to inputs, outputs, and other mental states (lack of specificity?)

- Fodor: “... many of the most powerful generalizations don’t care about content per se; what they care about is only relations of identity and difference of content”

- Fodor: “It looks as though our best arguments for Psychofunctionalism do not entail that the sorts of (relational) properties that make a thing a belief are also the sorts of properties that make a thing a belief that P.”

B3: From Functional-Role Semantics to MH

- a.k.a. From the Philosophy of Language via a ‘Functional Role’ Theory of Meaning (p. 60)

- Fodor: you get Functional-Role Semantics by a process of reaction

- a relatively unsophisticated view of meaning (denotational theory) appears to fail - a certain diagnosis of the failure comes to be widely accepted - the unsophisticated theory went wrong because it ignored the contribution of functional role to the determination of content

- the denotational theory of meaning: for a mental entity to have content is just for it to have a denotation (the denotation of a thought is whatever it is about the world that makes - or would make - the thought true)

- Following Sections of the Essay:

- B3.1. a case that makes trouble for denotational theory

- B3.2. treatment of the case by functional-role semantics

- B3.3. argument that there are serious problems for the functionalist program in semantics

- C. survey of some more standard objections to the denotational theory

B3.1. A Case That Makes Trouble for Denotational Theory (p. 73 - 74)

- classical problem for denotational semantics: denotational theories slice meanings too thick

- if you identify contents with denotations, you fail to distinguish between contents that are in fact distinct

- case of Oedipus: “wanted to marry Jocasta” was true of him, “wanted to marry Oedipus’ mother” was false of him - but “Jocasta” and “Oedipus’ mother” have the same meaning - so how could one be true of him, and not the other?

B3.2. Treatment of the Case by Functional-Role Semantics (p. 74 - 76)

- functional-role theories of meaning begin to seem plausible in the case of codenotational mental states

- although both of Oedipus’ thoughts have the same truth conditions, they are kept semantically separate because they have different roles to play in Oedipus’ mental economy

- most notably: they differ in their epistemic liaisons

B3.3. Argument That There are Serious Problems for the Functionalist Program in Semantics (p. 76 - 83)

- The hard problem for functional role theory: functional-role semantics says that content is constituted by function. Very well, then, just how is content constituted by function?

- the vocabulary that is required for the individuation of contents is, by assumption, not available for the individuation of functional roles. - if the theory is to be other than question begging, functional roles must be picked out nonsemantically and nonintentionally.

- psychofunctionalism: two networks - one generated by causal interrelations among mental states, and one generated by the semantic interrelations among propositions

- the basic idea for a functionalist solution to the hard problem is that, given the two networks just described, we can establish partial isomorphisms between them; and that under such an isomorphism, the causal role of an attitude mirrors the semantic role of the proposition that is its object.

- two-factor theory of semantic functionalism

- it looks unavoidable that two-factor theories are going to assign satisfaction conditions to a mental state not only via its causal connections to the world, but also via the propositional interpretation of its functional role. And, as previously noted, the theory has no mechanism at all for keeping these two assignments consistent.

One-factor functional-role semantics is out because of the Twin cases; and two-factor functional-role semantics is out because of the problem of coordinating factors.

C. Survey of Some More Standard Objections to the Denotational Theory (p. 83 - 93)

- Denotational Account of Names

- Fatness of Slice

- Philosophical Objections:

- A purely denotational semantics breaks the connection between content and consequence. If the content of your belief is independent of its functional role, then believing that P is compatible with believing practically anything else: even ~P

- A purely denotational semantics breaks the connection between content and behaviour

- Purely denotational semantics doesn’t solve the problem about individuating contents; it only begs them

Fodor: None of the three main ways philosophers have argued for MH has seemed as convincing as routine belief/desire explanations are.

Scotch Verdict: Not Proven (Neither Guilty, Nor Innocent)