The following is an excerpt from my senior thesis on intentionality and Wittgenstein’s philosophy of language (“A Wittgensteinian Critique of Analytic and Post-Analytic Theories of Intentionality and Mental Content” by David Price, Goucher College philosophy department, 2007).
In this section, I discuss a particular section in The Blue Book where Wittgenstein explicitly critiques Russell’s philosophy of mind. Please excuse any formatting errors. I tried my best to correct them all but it is entirely possible I missed a few.
In The Blue Book Wittgenstein characterizes Russell’s logical approach to meaning and language in terms of his analysis of predicates that name intentional states like wishing, perceiving, thinking, et cetera. The subsequent rejection of Russell’s analysis involves a sort of litmus test for Russell’s necessary conditions for the proper use of the verb ‘to wish’. The litmus test comes in the form of an introduction of a certain grammatical distinction which, when introduced, shows that the meaning of the word does not play solely by the rules Russell’s analysis provides for, indeed necessitates. Wishing, according to Wittgenstein’s interpretation of Russell’s theory, means something like
…a [sort] of hunger.-It is a hypothesis that a particular feeling of hunter will be relieved by eating a particular thing. In Russell’s way of using the word ‘wishing’ it makes no sense to say ‘I wished for an apple but a pear has satisfied me’. But we do sometimes say this…
Russell’s interpretation of intentional verbs like ‘to wish’ requires that those verbs be used transitively (i.e. the grammatical use of ‘wishing’ presupposes that there is some definite thing that is the object of that state’). Wittgenstein’s example of using ‘wishing’ in a different sense whereby the speaker may say “I wished for an apple but a pear has satisfied me” shows that the grammatical requirement of a direct object of an intentional state can be satisfied in a multitude of ways. We might say that “in order to wish something, there must be something wished” (echoing Socrates, Brentano, and virtually any subsequent account of intentionality), but saying that does not exclude the possibility of constructing expressions that seem to oppose the sense of ‘to wish’ as a sort of hypothesis about the thing that will satisfy that wish. Contrary to Searle’s position that an intentional state is self-referential with respect to its ‘conditions of satisfaction’, the predication of an intentional state like ‘to wish’ may not include its ‘direct object’ in the ideal sense that leads philosophers to say that ‘in order to wish there must be some determinate object of that wish’. In logical accounts of intentionality, what can be said of intentional states like ‘wishing’, ‘perceiving’, ‘thinking’, et cetera is that in every case their use presupposes a direct object which can be thought of as that element in Searle’s theory that is self-referential with respect to its conditions of satisfaction. This way of thinking is susceptible to formal notation:
S wishes that p ↔
(i) there is some definite direct object c that is the condition of satisfaction for that particular state for S
(ii) the meaning of ‘S wishes that p’→ it is not the case that (‘S’s wish that p’) could be satisfied by some other object d
Formally speaking, Wittgenstein’s example of the proposition P2: ‘S wished for an apple but a pear has satisfied S’ violates the first and the second condition of the definition above. The confusion is this: the phrase ‘wish for an apple’ contains the condition under which S’s wish would be satisfied (namely, obtaining an apple) while the phrase ‘the pear has satisfied that wish’ seems to imply that somehow S did not know what would satisfy S’s wish when S wished it. The dilemma, as Wittgenstein shows, could be solved in two ways. The first is via recourse to intransitivity:
There are certainly cases in which we say, ‘I feel a longing, though I don’t know what I’m longing for’…or ‘I feel fear but I’m not afraid of anything in particular’….we may describe these cases by saying that we have…sensations not referring to objects [think: intentional inexistence]…if in characterizing such sensations we use verbs like ‘fearing’, ‘longing’, etc., these verbs will be intransitive [my italics]; ‘I fear’ will be analogous to ‘I cry’. We may cry about something, but what we cry about is not a constituent of the process of crying; that is we could describe all that happens when we cry without meaning what we are crying about [my italics].
The meaning of the ascribed mental state does not require the presence of a direct object: Notice the similarity between the intransitive sense of mental predicates and Brentano’s notion of intentional inexistence; that is to say, Brentano’s description is not at all accidental, it reflects a way we express thinking, longing, wishing, et cetera in various languages. Notice also that in this sense, a description of the mental state ‘to cry’ does not include what exactly ‘the crying is directed towards’ in Brentano’s sense of the ‘intentional relation’ between thought and what the thought is directed at. Depending on how the process is described, it may or may not strike an accord with a particular characteristic of intentional content in philosophical discourses.
The second part of the grammatical distinction is via recourse to transitivity:
Suppose now that I suggested we could use the expression ‘I feel fear’, and similar ones, in a transitive way only. Whenever before we said ‘I have a sensation of fear (intransitively) we will now say ‘I am afraid of something, but I don’t know of what’. Is there an objection to this terminology? We may say: ‘there isn’t, except that we are then using the word ‘to know’ in a queer way’.
Wittgenstein is showing here how language and meaning are not bounded to logically perfect (i.e. grammatical) determinations. To characterize the sense of a mental predicate transitively, that is, to signify the meaning of the predicate ‘to feel fear’ in terms of that state’s condition of satisfaction (i.e. the object that satisfies what the mental state calls for), seems to presuppose not only some particular object that stands for the condition of satisfaction for that state, but fixes how other related predicates like ‘to know that’ are used. If someone were to say, “When I use the expression ‘to feel fear’ I mean the verb in a transitive sense, that it is the fear requires a direct object to be grammatical” then it would appear odd to express in conjunction with that definition the phrase ‘…but I don’t know of what’. We think of transitivity as a sort of rule for making sense of a proposition. The law might said like this: ‘to feel fear (transitively) presupposes knowledge of the object to which that fear is directed’. The transitive sense of ‘to feel fear’ requires an object of that fear. The intentional content represents the object that one is fearful with respect to, so fixing the intentional content as representative of a direct object in conjunction with failure to know that direct object seems grammatically confused.
This is precisely Wittgenstein’s point: grammatical distinctions do not resolve difficulties philosophers face when attempting to clarify under what conditions something can be said. There is no necessary determination of the verb ‘to know’ in the transitive characterization of a particular mental predicate like ‘to feel fear’. Wittgenstein’s point is that we tend to think that we are duped into thinking that transitivity within a proposition must be consistent: if one intentional state is defined in a transitive sense, then another intentional state must also be transitive. We think of transitivity as distributable among ‘types of expression’ like ‘expressions containing mental predicates’.
To illuminate this sort of deception, Wittgenstein asks us to consider an instance in which one seems to refer to the same mental state despite changing that state’s transitivity. This case can be signified in the following way:
T1: S has a general undirected feeling of fear [the sense of fear seems to be intransitive-that is-without an object/condition of satisfaction]
T2: S has an experience which makes S say, ‘Now I know what I was afraid of.’ [now the sense of fear is described in a transitive way, but with the oddity that S (at T1) did not know the condition of satisfaction for that fear]
Now suppose S forms a conjunction between those two temporally distinct events. S might say: ‘well, earlier on I was afraid of something but I didn’t know what…later on I realized I had been afraid of performing poorly on my upcoming math test.’ That sort of description would seem to mean that the fear appeared intransitive initially but became transitive when S realized what S was afraid of (that is, acknowledge the condition of that that state’s satisfaction). That raises a question though, in what sense exactly does the description in T1 describe the same sensation of fear that the description in T2 describes?
S’s revision of the transitivity from one moment to another is problematic for theories of mental content in that it shows that the meaning of a psychological state like ‘to be in fear’ is not strictly determined by its ‘condition of satisfaction’. Since it is often the case that one’s description of an intentional state is revisable in this sense, it is possible to (in a sense) refer to the same intentional content despite fluctuations in what that content is intentionally related to! Fregean and post-Fregean accounts of intentional content require that the determination of a particular mental state (i.e. how that content is meant for the individual that experiences that state) is identified with what that state represents as being the case, that is, what the content of that state is intentionally related to. That sort of identification is a grammatical one, as Wittgenstein points out. However it does not accurately provide ‘what it makes sense to say’ about intentional states since it cannot block instances in which we refer to one and the same ‘mental state’ (after all, S is ‘talking about’ the meaning of to be in fear throughout) despite changing the transitivity of that state over time (what the content of that state represents/what that state is intentionally related to/what the satisfaction condition is).
The upshot of all of this is that the meaning of predicates we think of as somehow ‘mental’ is not logically determined. Some commentators refer to this idea as the ‘indeterminacy of psychological concepts’ but that sort of description is misleading. It is akin to the distinction between ‘what is thought’ and ‘what is said’ expressed in instances in which someone might say, “I know what I’m thinking of but I just can’t say”. The view that psychological concepts are indeterminate is misleading in the sense that it makes it seem as if ‘physical concepts’ for instance, are determinate. The sort of thinking that leads to a fundamental dualism between ‘mental’ predicates and ‘physical’ predicates is exactly the root of the problem, according to Wittgenstein.
 See Russell’s Analysis of Mind for his theory of mental-state predicates. Presumably it is from there that Wittgenstein is basing his interpretation and subsequent criticism.
 Russell’s definition of ‘wishing’ can be depicted in the following way: Russell’s view: if S wishes that p then there is some direct object R which is the condition of satisfaction for that state
 I have taken some liberty in presenting this case. For instance, Wittgenstein’s description involves the first-person, not the third-person. For the actual description see p. 22 of the Blue Book. Wittgenstein, Blue Book, 22