The Irreducibility of Intentionality

The Argument from Copies 

I have been reviewing Bill's postings on Intentionality recently and was struck by the argument here.   Bill writes,

The following argument is my interpretation of remarks made by Edward Feser in his Philosophy of Mind: A Short Introduction (One World, 2005), pp. 156-159)

1. Consider a representation such as a picture. You draw a picture of your mother. The picture represents her: it is of or about her, and it would remain about her even were she to cease to exist. The picture is a physical object with physical properties: the paper is of a certain size and shape and texture, the ink of a certain chemical composition, the lines have a definite thickness, etc. Now I would insist that these physical features cannot be that in virtue of which the picture represents your mother: they cannot be that in virtue of which the physical item is a representation. For it makes no sense to ascribe intrinsic semantic or intentional properties to merely physical items. But even if I am wrong about this, there remains a problem for a materialist theory of representation.

2. Suppose a 'copycat' comes along and makes an EXACT copy of your picture of your mother. The copycat's intention is not to represent your mother; his intention is merely to represent your representation of your mother. Now there are two pictorial representations, call them R (the original) and R' (the copy). The question arises: Is R' a representation of your mother, or is R' a representation of R? Suppose a second copycat comes along and produces a second copy R''. Does R'' represent R' or R or your mother? The situation is obviously iterable ad infinitum.

3. Clearly, there is a difference between saying that R' represents your mother, a human being, and saying that R' represents R, a nonhuman drawing of your mother. The reference is different in the two cases. But the reference is indeterminate if we go by the physical properties of the representations alone. Suppose I hand you two drawings of your mother, one an exact copy of the other, but you do not know which is the orignal and which is the copy. You cannot, by inspection of these drawings, tell which is which. Thus you cannot determine the reference from the physical properties.

4. The point is generalizable to other types of representations. Suppose I say 'cat' to refer to a cat and my copycat brother says 'cat' simply to copy me. If my brother mimics me perfectly, then it will be impossible from the physical properties of the two word-sounds to tell which refers to a cat and which does not.

Please do not say that we are both referring to a cat. For my copycat brother is a mere copycat: his intention is merely to reproduce the word-sound I made. To make it even clearer, replace my brother with a parrot who happens to be a perfect mimic. No one will say that the 'cat'-token produced by the parrot refers to a cat. The parrot is just an animate copy machine.

The same goes for any physical representation. Suppose a pattern of neural firings is taken to be a representation of X. An exact copy of that pattern needn't be a representation of X; it could be a representation of the original pattern. In general, no material representation of X is such that its physical properties suffice to make it a representation of X as opposed to a representation of a representation of X. 

5. Here is the argument:

P1. All thoughts have determinate objects.
P2. No purely material representation has a determinate object.
—–
C. No thought is a purely material representation.

I am not persuaded by this argument.   Bill says that we cannot, by inspection of the drawing R and its copy R', tell which is the drawing and which is the copy.   We cannot determine the reference (ie, that which is represented) from the physical properties.   I agree with this.   But the argument sets us off down a false path by treating representation as a two-place relation.  A (re)presentation is always a presentation to something, typically a mind.  Thus we should see representation as a three-place relation:  picture P represents object M to subject S.   If S has drawn a likeness P of his mother M  then seeing P will invoke a thought in S of M, and not necessarily so for another subject S'.   But since the copy P' is indistinguishable from the original P it will also invoke a thought of M in S.  Hence we should say both R(P, M, S) and R(P', M, S).   Likewise, the sound 'cat' invokes the thought Cat in any English-speaking subject S, regardless of how the sound was produced.   The representativeness of P or 'cat' is not determined by the picture or sound alone, but necessarily involves a suitably prepared mind S. 

In his summary (5) Bill is asking us to take his P1 premise for granted and to accept his P2 on the basis of the argument in (1) through (4).  I have shown above that the argument from exact copies that Bill makes does not render material representations indeterminate as to their object.  Bill is ignoring the essential role in representation of the prepared mind S.   Turning now to putative representations in the form of patterns of neural firings,  Bill reruns the argument from exact copies.   But there cannot be an exact but distinct copy of a pattern of neural firings.   For suppose some set N of neurons is stably firing away as a result of a pattern of inbound connections.  And suppose a distinct set N' of neurons is an exact copy of N.   Then N and N' have the same sets of input and output connections.  N and N' fire on the same inputs and produce the same outputs.  The firings of N' merely augment those of N.   They play no distinct role in the functioning of the neural system.  Rather like an original and a copy painted on glass and viewed in exact alignment.