Re: Representation vs. Symbolization
- Edgar, you said "you have some very restricted view of
what 'representation' must mean". On the contrary: I've said there're
more than a dozen meanings of "representation", so we may agree or
disagree depending on which of the meaning we're talking about. This
is clear when you enumerate very different things (Memory, sensation,
abstraction, thoughts) and you say that all of them are
representations. The term "representation" is applied to very
different things. Another example is the answer of Jay: he
uses "representation" not as "symbol" but as "the use of certain
features of something to stand for the whole of something" (which I'd
call a kind of discriminative response).
Of course I don't deny that representations exist, and that some
functioning requires representation. For example, the words I'm using
in this message are symbols/representations of other things. But I
don't think that ALL complex behavior requires this kind of
symbols/representations. I'm just proposing to clarify the different
meanings of the term "representation", and evaluate each of the
meanings, case-by-case, instead of the indiferentiation and confusion
of very different cases under a single umbrella-term.
You said "there must be some kind of cognitive representation of that
real world for the organism to be able to function". But an adaptive
neural network is able to learn and behave, and it doesn't
have "representation" as you defined it ("a symbol of something"). It
just has nodes whose probability of activation changes according to
the input and the activation of other nodes. So, at least, some kinds
of complex behavior can exist without symbols/representations. Unless
you call this a "representation" too, but this would be a wider
conception than "symbol of something".
You said "it is quite clear that the actual physical world is not at
all like the world we imagine we live in". I agree, but this doesn't
imply that all activity is representational.
You said "There is either representation or we sense and live
directly in the real world in some magical spiritual god-like way
unmediated by our particular biology and brains". But it's a false
dichotomy, because I don't deny the mediation of biology and brains.
The issue is if all neural mediation is accurately described
as "representational", or there're some kinds of neural mediation
that are not accurately described as "representational".
I pretty much agree with your description of representation as 'the
use of certain features of something to stand for the whole', and how
this will be different for different organisms. But that is precisely
what a symbol is! The symbol is never the whole thing itself, it
couldn't be. It is a mental 'notation' of that whole thing which
represents it (symbolizes it) for the organism. In the dog it can be
a smell, in the human it can be a model of the salient features of a
face, or in the human it could be a word (the name of the person). In
the zebra it could be the cry that symbolizes 'hungry lion
attacking'. All of these are symbolic representations in terms of
salient characteristics, of a whole actual entity in the external world.
So far we are just quibbling over terminology, as you define symbol
in a much more restricted sense limited primarily to human activity.
Specifically you define it as something abstract and arbitrary which
has no inherent connection with the thing symbolized, in other words,
a word which arbitrarily denotes an object, but which has no innate,
only a conventional, association with that object.
I'm not sure it is worth arguing over this point, my real arguments
are mostly where you go with it in terms of the other conclusions you
draw. I will note though that as you are certainly aware, alphabetic
language is a direct outgrowth of earlier pictographic
(representational in your sense) languages.
On Jul 2, 2007, at 10:29 AM, Jay R. Feierman wrote:
> Edgar Owen: The question of whether there is 'representation' or
> not is the real pseudo-problem here, not the hard problem. Given a
> real external world and given the functional structure of the
> sensory organs and cognition, there must be some kind of cognitive
> representation of that real world for the organism to be able to
> function. Of course one can argue over the details, but if there
> were no representation of some kind, there could be no way to
> function in a real world of which we had no knowledge
> (representation) at all. There is either representation or we sense
> and live directly in the real world in some magical spiritual god-
> like way unmediated by our particular biology and brains. In
> reality, it is quite clear that the actual physical world is not at
> all like the world we imagine we live in. However enough of the
> logical organization of our mental world does map to that physical
> world, that we are able to function within it. Memory is
> representation, sensation (perception) is representation,
> abstraction (eg. thinking about something when it is not present)
> is representation . . .
> I think the problem here is that you have some very restricted view
> of what 'representation' must mean. Just use it in its original
> sense of being a symbol of something, and there should be no
> problem in understanding what I'm saying about the necessity of
> Jay R. Feierman: Representation is not the same as a symbol of
> something. Rather, representation is the use of certain features of
- Jay,The reason I prefer to use 'symbol' rather that 'representation' in discussing how organisms view the world is that one does not normally think of the logical manipulation of representations, as one does of symbols.The point is that for any animal to function in the world such logical manipulations in the mind are absolutely essential. Whatever one calls them, mental images such as prey, water, location, shelter, and the many more necessary, are in fact held in the mind as units and manipulated in logical ways that result in appropriate actions. Call them what you will, but it is much easier to understand the complexities of this process if they are called symbols rather than representations. We have excellent formal systems that describe the manipulation of symbols that can be adapted to this purpose. In the case of the actual mental representations, they will be more fuzzy and ill defined in a formal sense, but animals' mental processes obviously manage to manipulate them effectively and logically to produce intelligent and appropriate actions. In fact much or most of human mental activity also operates at this same non-verbal level.EdgarOn Jul 3, 2007, at 9:13 AM, Edgar Owen wrote: