Atheism and Theism 93channels to which the various parties have access. The present worry is that
any theory claiming that communication produces the channels – along with
everything else – faces the objection that without the channels there could be
no communication. They are part of what the organizational information
creates. The envisaged reply is that initially something arises which is less than
a power of transmission but enough to get the process of communication
started.
One way of framing the worry I have about this is that it seems to be trying
to account for a significant qualitative difference in terms of a merely quant-
itative one. Let me illustrate what I mean, and indicate why I think there
is a problem, by switching to a parallel case concerning the nature of mental
phenomena. I shall be saying more about the philosophy of mind and theism
shortly; at this point the feature to focus on is simply the structural analogy
between the case I am about to discuss and that of reproduction.
A few years ago I wrote an essay in the course of which I criticized the
efforts of Daniel Dennett to give an adequate reductionist account of mental
representation.^9 The problem is this. Thoughts are intentional in the tech-
nical sense that they are directed towards, or are about, something or other
(from the Latin ‘intendere’: to aim or direct). How is this possible? One much
discussed suggestion is that to think ‘There is a tree in the garden’ is for one’s
mental system to be in a computational state involving a representation – a
sentence in the mind and /or in the head – the content of which is that there
is a tree in the garden. Very crudely indeed, one might thus suppose that some-
one thinks that p when his or her information-processing system entertains
a mental sentence ‘S’ the meaning of which is that p.
Much could be said about this, but here simply note that it involves a
homuncular regress (‘homo’ (man), ‘-culus’ (little)). The problem of mental
representation has not gone away. It has just been moved from the personal
to the subpersonal level: I think that p because (in some sense or other) there
is something – a ‘processing module’ – in me that can interpret a symbol ‘S’
that means that p. To his credit, Dennett sees that this proposal is hopelessly
regressive if treated in realist terms, i.e. as maintaining that representational
power is derived from a representational subsystem, and so he offers an
alternative reductionist-cum-eliminativist version of it. He writes:
[You] replace the little man in the brain with a committee [whose members]
are stupider than the whole; they are less intelligent and ‘know’ less. The
subsystems don’t individually reproduce all of the talents of the whole. That
would lead you to an infinite regress. Instead you have each subsystem doing a
part, so that each homuncular subsystem is less intelligent, knows less, believes
less. The representations are themselves, as it were, less representational... a
whole system of these stupid elements can get to exhibit behaviour which looks
distinctly intelligent, distinctly human.^10