Download Fregeanism vs. Russellianism: Comparing Views on Content & Perception - Prof. Jeffrey Spea and more Study notes Introduction to Philosophy in PDF only on Docsity!
From transparency to Russellianism
phil 93507
Jeff Speaks
October 12, 2009
Recall our contrast between the following opposed views of content:
Russellianism: contents are structured objects the constituents of which are worldly items such as objects and properties. Fregeanism: contents are structured objects the constituents of which are ways of thinking about, or modes of presentation of, objects and properties.
A core difference between Russellian and Fregean views of content is that the Russellian thinks of contents as built up from worldly items like objects and properties, whereas the Fregean thinks of contents as built up from modes of presentation of, or concepts of, these items.
Unlike the Russellian, who constructs propositions out of objects and properties, the Fregean introduces a new class of items — senses — to play the role of contents. It is therefore reasonable to think that the Fregean should be able to provide some constraints on when two experiences have the same Fregean sense, and when they do not. When the bearers of content in question are sentences, such constraint is standardly provided by some version of Frege’s Criterion:
Two sentences S and S′^ have the same sense ≡ any rational agent who under- stood both would, on reflection, judge that S is true just in case he he would judge that S′^ is true.^1
But how are we to apply this to the case of perception? A straightforward generalization would yield something like this:
(^1) Obviously this ignores the need to relativize to contexts of utterance; this plays no role in what follows. An immediate problem for this statement of the Criterion is the possibility of two intuitively non- synonymous sentences which are such that anyone, on reflection, would judge them both to be true (or false). ‘1=1’ and ‘2=2’ do not have the same sense, even if we can’t imagine a situation in which a reflective agent would differ over their truth value. The most natural way around this problem is to allow the relevant truth-value judgements to be not only about S and S′, but also about any pair of complex sentences which differ only in the substitution of S for S′. The Criterion can naturally be extended in an analogous way to subsentential expressions, so that two expressions x and y have the same sense ≡ any rational agent who understood both would, on reflection, make the same truth value judgement about any two sentences which differ only in the substitution of x for y.
Two experiences e and e′^ have the same sense ≡ any rational agent who had both experiences would, on reflection, judge that e is veridical just in case he he would judge that e′^ is veridical.
But the latter criterion presents problems of interpretation which the former does not. Implicit in the criterion for sameness and difference of the senses of sentences is the re- quirement that the agent consider the two sentences with respect to the same circumstance of evaluation; we can’t demonstrate that ‘Bob is a bachelor’ differs in sense from ‘Bob is an unmarried man’ by noting that a friend of Bob’s might rationally judge the first to be true with respect to the morning of the wedding while judging the second to be false with respect to the evening of that day. The problem is that perceptual experience come with ‘built in’ circumstances of evaluation: the location, time, and world of the experience itself. For this reason, one can’t have two experiences ‘of’ the same time in the same way in which one can consider the truth of two sentences with respect to a single time, and so we can’t simply build the ‘same circumstance’ requirement into the second criterion. But, clearly, we need something like the ‘same circumstance’ requirement in the case of the second criterion just as much as in the case of the first. After all, suppose that an agent has experience e after being told by a reliable authority that his next experience will be veridical, and experience e′^ just after being told by the same authority that his experience will be illusory. The fact that such an agent might well be rational to judge that the first experience is veridical and that the second is illusory hardly suffices to show that the two experiences differ in sense — any more than the example above suffices to show that ‘bachelor’ and ‘unmarried man’ do.
It seems to me that we can capture the intuitive idea behind the second criterion like this: consider any agent, complete with a set of beliefs about his environment. Now consider two possible courses of that agent’s immediate future: that he has experience e, and that he has experience e′. If it could ever be the case that the agent would be rational to judge one of the experiences veridical and rational for him to judge the other not to be veridical, then the two experiences differ in sense; if it couldn’t be the case that any such agent’s rational judgements would differ in this way, then the two experiences have the same sense. This seems to do the work which the ‘same time’ requirement does in the case of the first version of Frege’s criterion without requiring, incoherently, that the agent simultaneously have two experiences of the scene before him.
If we adopt this interpretation of Frege’s criterion as applied to perceptual experiences, then it seems clear that any two experiences which have exactly the same phenomenal character must also have the same sense. For if two experiences have the same phenomenal character, then they would be indistinguishable to any subject. But then how could the pair of experiences supply a rational ground for any difference in veridicality judgements? They couldn’t; holding fixed all relevant aspects of an agent’s psychology, it can never be rational to judge one experience veridical if it would be rational, in just the same circumstances, to judge illusory an experience in principle indistinguishable from the first. So the Fregean about the contents of perceptual experience is committed to the following:
Phenomenology/Sense Principle If two experiences have the same phenomenology, then they have the same
supposition that f p 1 and f p 2 correspond to the same Russellian proposition because, if they did, they would not differ with respect to which objects and properties represented as being in the agent’s environment, but only with respect to the modes of presentation of the same objects and properties. So the hypothesis that the contents of perceptions are Fregean senses, along with the transparency of experience, entails a contradiction.^2
It seems to me that the only plausible reply for the Fregean who wants to respect the view of the transparency of experience sketched above is to try to block the derivation of the Phenomenology/Sense Principle. One initially plausible way to do this is to relax the criterion for sameness and difference of the senses of perceptual experiences to simply a criterion of difference. On this view, differences in the relevant kinds of veridicality judgements are sufficient for difference in sense, but sameness of the relevant kinds of veridicality judgements is not sufficient for sameness of sense. So the view in question would reject the principle above, namely
Two experiences e and e′^ have the same sense ≡ any rational agent who had both experiences would, on reflection, judge that e is veridical just in case he he would judge that e′^ is veridical.
in favor of
Two experiences e and e′^ have the same sense → any rational agent who had both experiences would, on reflection, judge that e is veridical just in case he he would judge that e′^ is veridical.
From the latter, there is no ready derivation of the Phenomenology/Sense Principle.
But this saves the Fregean view of experience only by making the notion of the Fregean sense of an experience unacceptably obscure. To see this, consider any two experiences which (according to the Fregean) have distinct senses as their contents, but correspond in the above sense to a single Russellian proposition. To borrow a phrase from Michael Nelson, call any pair of experiences related in this way a ‘puzzling pair.’^3 We know that, to block the conflict with the transparency of experience, the Fregean must claim that puzzling pairs never differ in phenomenology. (This was the point of the retreat from giving necessary and sufficient conditions for sameness of sense to giving necessary conditions.) We know from the definition of puzzling pairs that they do not differ with respect to which objects and properties they represent as being before the perceiver. But, despite the fact that puzzling pairs present just the same objects and properties in
(^2) One response to this problem that I have heard is that although in general the relationship between Fregean and Russellian propositions is many-one, this may not be so in the case of the contents of perception. So, for any Russellian proposition, there is at most one Fregean proposition which is such that it is a possible content of perception and is ‘about’ the same objects and properties as the Russellian proposition. For this to be a plausible reply, we’d need some motivation for making this restriction in the case of perception, but not thought or language. One might attempt to provide such a motivation via the claim that the senses which are the contents of perceptual experiences are always demonstrative senses (as in the view of Brewer (1999)). But this is not enough; the relationship of demonstrative senses to any given object or property is many-one. (^3) The phrase is due to Nelson (2002), though Nelson has in mind certain pairs of sentences which (allegedly) differ in Fregean sense but not Russellian content, rather than pairs of experiences.
phenomenally identical ways, the Fregean claims that there is some difference between the members of puzzling pairs in their mode of presentation of some object of property. But in what could such a difference of mode of presentation consist?
It is important to be clear that the worry here is not just the standard worry that Fregeans do not say enough about the nature of the senses invoked to explain linguistic and cogni- tive phenomena. The worry is that, even if we grant the distinctions between the senses of expressions which Fregeans typically accept and Russellians reject, this still gives us no purchase on the differences in sense claimed to obtain between puzzling pairs of expe- riences.
One way to show this is to compare the Fregean’s claim here with standard Fregean claims in the philosophy of language. There is no shortage of pairs of sentences which the Fregean claims to differ in sense, but which the Russellian regards as having the same content; the easiest and most well-known examples are sentences which differ only in the substitution of simple coreferential names. The difference which the Fregean claims to hold between puzzling pairs of experiences is analogous to the difference which the Fregean claims to hold between such a pair of sentences: they differ in sense, but correspond to the same Russellian proposition. The Russellian, of course, will deny that there typically are such differences in sense; but, for purposes of argument, let’s grant the Fregean’s claim about linguistic expressions. This gives us a test for the Fregean’s claim about the contents of experience.
Often, we can use language to report how a perceptual experience represents the world to us. If, as we are supposing, there are many sentences which differ in Fregean sense but not Russellian content then, if puzzling pairs of experiences really exist, one would expect that there is a puzzling pair of experiences e 1 and e 2 such that there is some sentence which correctly (even if partially) expresses the way that e 1 represents the world as being, but does not correctly express the way that e 2 represents the world as being.^4 But now consider a puzzling pair of experiences, which represent just the same objects and properties as before the perceiver, and are phenomenally identical. Is it ever the case that a sentence would be a correct report of the content of one member of that pair of experiences, but not of the other? It seems to me that the answer is ‘No’.
A similar argument can be run at the level of thought. The Fregean will hold that the contents of thoughts are often distinct, even though identical at the level of Russellian content. So, if there are puzzling pairs of experiences, we should expect that sometimes a thought would (even if partially) match in content one member of the pair, but not the other. But again consider a puzzling pair of experiences, which represent just the same objects and properties as before the perceiver, and are phenomenally identical. Is it ever the case that a thought or judgement would represent the world in the same way as one of those experiences, but not the other? If, as I think, the answer is again ‘No’, the Fregean is stuck with the following view about puzzling pairs of experiences:
There are indefinitely many pairs of experiences related in the following way: they present just the same objects and properties in phenomenally identical (^4) Otherwise, the Fregean would have to claim that, although there are differences in the senses of experiences which are not reflected as differences in Russellian content, and there are differences in the senses of linguistic expressions which are not reflected as differences in Russellian content, of necessity the two never coincide. But what could explain that?
