From stapp@thsrv.lbl.gov Thu Sep 24 17:13:18 1998 Date: Thu, 24 Sep 1998 17:11:08 -0700 (PDT) From: Henry Stapp To: A.Sloman@cs.bham.ac.uk Cc: ghrosenb@ai.uga.edu, bdj10@cam.ac.uk, brings@rpi.edu, brucero@cats.ucsc.edu, chalmers@paradox.ucsc.edu, jmschwar@ucla.edu, keith@imprint.co.uk, klein@adage.berkeley.edu, patrickw@monash.edu.au, phayes@coginst.uwf.edu Subject: Re: consistency of descriptions and non-humean causal powers On Sat, 19 Sep 1998 A.Sloman@cs.bham.ac.uk wrote: >... > Moreover, Henry's argument, like the arguments of so many philosophers, > apparently depends mainly on his ability to *imagine* P¬-M. > > However, what people think they can or cannot imagine is of dubious > relevance to any of this, since (a) what people can imagine changes as > they learn things (e.g. you can learn to find something UNimaginable, as > the history of mathematics and physics shows) and (b) you can vary the > constraints on your imaginings so as to make almost anything imaginable. > E.g. I can easily imagine seeing a round square by looking at it end on. > > So someone's ability to imagine some very vaguely specified physical > configuration P occurring without any mental phenomena M occurring > proves nothing. > I consciously and intentionally formulated my argument in a way that did not rely upon what someone can or cannot imagine. I completely agree that "I can't imagine it" types of arguments are inadequate, and do not use them. To dump my argument into that easily refutable class does not properly address the issue. I shall return to this point. > It gets a little more interesting if we talk about some real cases (as I > think Henry recently proposed): e.g. a design for an engine which, > without relying on quantum phenomena, actually computes square roots, or > considers options and takes decisions, or detects intruders, or plays Go > better than most human beings, or evaluates its own problem-solving > strategies and improves them, or switches attention between different > sub-problems while doing its planning, or prefers one strategy to > another, or wants to win and is disappointed when it loses, etc. Are you claiming that there exists in association with the implementation of this computation "an actual feeling of disappointment" even when no one (no biological system) is around? What is the basis of that claim? Is it a consequence of the principles of classical physics that are---in this discussion---supposed to account for the causal structure of the computer? How does one get from these principles, together with (1) the complete classical description of what is going on in the computer, and (2) a complete understanding of the way [perhaps via many levels of programming] in which that classically describable process corresponds in some way to some conceptual ideas in the minds of the programmers and builders, to the conclusion that that some "experiential feeling of disappointment" is occurring in conjunction with the classically describable activity of the computer. > > We already know how to create relatively simple cases of such machines. > (When I say "we" I mean the scientific community as a whole. There are > many individuals who think they know how it is done but are basically > ignorant and get it wrong -- e.g. most philosophers who have never > designed, implemented, debugged, or documented a piece of software in > their life -- which is why I stopped reading philosophy written by > professional philosophers some time ago: I found the arguments mostly > irrelevant because if their ignorance, and I learnt more powerful and > relevant stuff by reading neuroscience, computer science, psychology, > studies of effects of brain damage, etc. etc.) > > Lots of people can use arguments similar to Henry's (and yours?) to show > that those software implementations which already exist don't "really" > suffice to produce the phenomena I know they do produce, You certainly have profound understanding to be able to KNOW that they do produce phenomena, if by phenomena you mean the existence of the experiential feelings under discussion here. > because *they* > can imagine all the physical part without the other capabilities being > present, or so they think. > MY argument is not based on what anyone can or cannot imagine. My claim is that there is no necessity from the principles of classical physical theory---given the complete classical-physics description of the implementing system---for there to exist any quality in nature beyond those that are completely representable in geometric terms (assignments of numerical values to locations in space-time). This is because that kind of quality is all that the classical description of the classically describable system directly describes: any other kind of quality must be reduced to it in order for it to be represented by the behavior of that system. All of the clear examples of implementation that you have mentioned are reducible to geometric properties in this sense. But our experiences do have qualities beyond the purely geometrical ones; uneasiness, dread, ecstasy > ..... > In discovering how to make mechanical clocks, calculators, various kinds > of control systems, vehicles, etc. engineers found ways of harnessing > networks of causal powers to implement new kinds of *emergent* causal > powers which are not definable in the concepts of physics (e.g. > numerical abilities, the ability to show the day of the week, comfort > and safety (in a house or carriage), protection from predators, easier > harvesting, etc. etc. > Yes, but the implementing of these concepts in a classically describable physical system require a translation or reduction to geometric terms. As a typical example take the "safety" of a carriage. If all human beings depart the scene the carriage retains all of its features that classical physical theory assigns to it, which includes how it will fare under all sorts of conditions that can be described in terms of the concepts of classical physics: collisions with other objects, overturnings, etc. We may use the English language word "safe" to denote some combination of objective classically describable tendencies for this carriage to respond to classically describable stresses. But within the framework provided by the principles of classical physical theory there is a distinction drawn between these properties that are conceptualized in purely geometrical terms, which are included in what the theory covers, and qualities such as `the feeling by some observer of the carriage, that "the carriage is beautiful." ' This beautifulness is `connected' to geometric features, but there is a residue of appreciation/valuation that is not conceptualized in geometric terms. Given the essentially geometric character of the principles of classical physical theory, and of the classical description of the implementing system (the carriage), there is no way in which those principles, and that classical description, can necessitate the existence of qualities---in conjunction with the existence solely of the implementing system---that are not conceptualized in geometric terms. I do not believe this argument is simply: I can "imagine" a world devoid of consciousness. It is rather a matter of the fact that geometric concepts are not co-extensive with concepts. ... > the attempts by Henry and others to > show that somehow the nature of consciousness inherently requires > something non-classical, or that quantum physics inherently involves > consciousness. > > > You could do me a great favor by critiquing that work. My claim is not that consciousness is incompatible with classical physical theory: one certainly can just add consciousness onto classical physics, as an ad hoc appendage, which however then appears as a logically disconnected superfluity. Nor do I claim that quantum physics inherently involves consciousness. I claim rather that the most rationally coherent and pragmatically useful way to understand quantum theory is to adopt the vonNeumann/Wigner pragmatic approach in which (1) the experiences that we are able to describe to ourselves and to our colleagues, regardless of whether they are couched in the language of classical physics---as Bohr would require---are taken to be the basic realities in our scientific description of nature, and (2) the representation of nature provided by quantum theoretical formalism be regarded as our representation of a very important component of the causal structure that connects these realities. The link between mind and matter then occurs in brains, in accordance with the formulas that I have set forth, which provide a satisfactory basis for a scientifically sound approach to the study of consciousness.