From Aaron Thu Feb 22 01:24:19 GMT 1996 Newsgroups: sci.psychology.consciousness References: <9602201533.AA00611@medusa.oracorp.com> Subject: Re: Tucson II [Function and Experience] Daryl McCullough writes: > Date: Tue, 20 Feb 1996 10:33:02 EST > > Gregg Rosenberg (ghrosenb@phil.indiana.edu) writes: > > >Dualism is an *enlargement* of the world, not the end of it. As such, > >it opens up a new scientific frontier, it doesn't destroy science. Sit > >back. Relax. Enjoy the ride -- it should be one terrific show. [DMcC] > I think that there is a sense in which functionalism and behaviorism > are themselves dualistic theories, because they introduce nonmaterial > entities "function" and "behavior". These entities are patterns found > in the material world, but are not themselves material, just as > instances of the number 2 can be found in the material world (two > apples for instance) while the number 2 itself is nonmaterial. I think > that the enlargement of the world by considering abstract entities is > very fruitful. I agree with this statement. In particular, I include all the following within my ontology: powers, capabilities, strengths, propensities, inclinations, dispositions, constraints, positive and negative feedback loops, stable, meta-stable and unstable equilibrium, dangers, risks, opportunities, solubility, flexibility, rigidity, fragility, elasticity, hardness, permeability, etc. etc. and even causal connections between these things. I suspect most people include these things in their ontology during their ordinary lives, unless they are professional philosophers with an axe to grind. I think these things and others like them, exist in the world, are implemented in physical and chemical structures and are explainable by the underlying implementation. Reasoning about these "abstract" things and the possible states of affairs implicit in them, is an integral part of designing, planning, deliberating, and deciding. Moreover, as JJ Gibson saw, they are deeply involved in perceiving. (He called them "affordances".) Consciously or otherwise, every engineer deals with them all the time. A typical engineering design combines large numbers of these things in a new and useful way, thereby creating new useful powers, capabilities, constraints, opportunities etc. (And sometimes also some harmful ones.) HOW this is done is another matter. The abstractions are normally produced and combined by combining other things: the physical materials used in the *implementation*. Software engineering does the same but in a more subtle and complex fashion, which I shall try to analyse another time. It seems to me that many people who discuss functionalism lack an understanding of these issues, because they think of functionalism as somehow equivalent to a crude physicalist reductionism. This confuses the design with its implementation. Eventually I shall try to show what a huge difference this viewpoint makes to a functionalist analysis of mental concepts. But not now. [DMcC] > While properties of abstractions cannot be studied > empirically, they *can* be studied rigorously using the tools of > mathematics. I think we do study some properties of some of these abstractions empirically. E.g. we may discover the relationships between rigidity and fragility in a vase, or between fragility and thickness, or fragility and material used etc. We may discover empirically how to make a possibility more or less remote. Having your car's brakes checked regularly makes the possibility of a serious accident less remote. [I meant "more"] Others are conceptual connections. E.g. the connection between fragility and the possibility of shattering is not empirical, though it may be "news" to an individual who has never tried to analyse his notion of "fragility". A functionalist theory of mind will specify a mixture of such conceptual relationships and empirical ones. The conceptual relationships are what make the so called "hard problem" illusory, and undermine oft-repeated anti-functionalist arguments of the form: The discovery of the relationships between experience and functional states is a discovery of an empirical relationship between two things, i.e. our prior concept of experience and these functional states. So there are two separate things and the hard problem is to explain how they are connected. When the discovery is conceptual, like discovering that your concept of fragility includes the possibility of something shattering, then there are not two things to be connected. No bridging principles are needed. But that's something to spell out in more detail another time. [DMcC] > .... > We know that *we* have consciousness (because we purposely define the > word "consciousness" to mean the sort of thing that we have), but we > have no way of knowing whether anything else does. We can make > reasonable guesses---for example, other people are probably conscious, > and rocks are probably not---but we have absolutely no way to test > those guesses. When we have done an analysis of the sort I've described, checking whether other things have experiences, or consciousness becomes no more difficult *in principle* than checking which things are fragile or in unstable equilibrium or use negative feedback with damping. It may be very difficult *in practice* if these functional states exist in a very complex integrated network of functional states. That's like the difficulty of finding out which algorithms a complex piece of software is running if one did not design it. > Our beliefs about the consciousness of others are > incorrigible---if we have decided, for example, that a robot cannot be > conscious in spite of being functionally equivalent to a human, how > could we ever discover that we were wrong? By discovering that we were using an inadequate notion of functional equivalence, when we are presented with a new and better one. >... > Because of the incorrigible nature of beliefs about consciousness, I > don't consider it to be a scientific subject. The real reason is that people unwittingly use the word "consciousness" and related words in obscure, poorly defined, ambiguous fashion. Consequently what look like scientific questions are just muddles. (E.g. "What are the neural correlates of consciousness?") But it doesn't have to be just a lot of muddle. I'll try to show this in more detail one day. Aaron