Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!usc!venera.isi.edu!smoliar From: smoliar@vaxa.isi.edu (Stephen Smoliar) Newsgroups: comp.ai Subject: Re: Another letter to the New York Review Summary: intelligence and things Message-ID: <12143@venera.isi.edu> Date: 2 Mar 90 16:11:43 GMT Sender: news@venera.isi.edu Reply-To: smoliar@vaxa.isi.edu.UUCP (Stephen Smoliar) Organization: USC-Information Sciences Institute Lines: 158 In article <2c7R02Sr8d8b01@amdahl.uts.amdahl.com> kp@amdahl.uts.amdahl.com (Ken Presting) writes: >In article <12085@venera.isi.edu> smoliar@vaxa.isi.edu.UUCP (Stephen Smoliar) >writes: >> Rather than agonizing >>over the "nature of intelligence"--as if it were a "thing" which had >>a "nature"--perhaps we should be reviewing our vocabulary as it pertains >>to talking about ANY form of behavior. > >Stephen, I agree wholeheartedly with your ultimate suggestion, but I'd >like to put in a few words in favor of hairsplitting. > >In the first place, human bodies (and computers) *are* objects. After >all, that is how Penrose gets into the debate at all. One way to think >about intelligence is to try to figure out what "object properties" the >body has which make possible intelligent behavior. The neurologists are >working on this from one angle. AI has been working on it from another. >So have philosophers. > I have no problem with this point; and I think Marvin Minsky (THE SOCIETY OF MIND) and Gerald Edelman (THE REMEMBERED PRESENT) have done the best jobs of articulating that THIS is the primary issue. >In the second place, the dynamical properties of a running computer are >neatly determined by its program. The program is of course a static >object (modulo the contents of its data structures, but you see what >I mean). The $64K question is: what properties does the program need >to have, so as to allow the emergence of intelligent behavior? > This is where I would like to move in an split a hair or two. It's that word "neatly" which is giving me trouble. After all, if there were a clean relationship between the static properties of a program and the dynamic properties of the device running that program, we wouldn't have all the software problems we have, would we? Even when we invoke all the motherhood of software engineering and enforce static properties which simplify that relationship, we STILL have problems! Furthermore, this is only the tip of the iceberg. If we consider cellular automata, instead of "clean" applicative LISP code (for example), we encounter devices with utterly trivial static properties exhibiting extraordinarily complex behavior with little clue as to the relationship between them. (At the Artificial Life Conference, Chris Langton presented a fascinating talk in which he tried to model this behavior in terms of the physics of phase transitions; and HE was still able to talk about his model at a relatively rough level.) I think we should be honest with ourselves and admit that we still do not have terribly good ways to describe the relationship between a program and the device which runs that program. I would argue that the reason for this is that we still lack good ways to describe and reason about the dynamic properties of processes. The best we have been able to do, thus far, is the abstract those processes into static objects. This is how software engineering SOMETIMES gives us a handle on reasoning about what our programs actually do, but the power of this approach is only as good as the abstraction we develop. Finding the right abstraction often remains the intractable problem in software engineering. Now I would like to push my argument one step further and conjecture that the reason we have so much trouble with the dynamic properties of processes is that we still have considerable trouble with time, itself. Many of the key issues of intentionality (at least as I have thus far been able to understand the story) emerged from Brentano's attempts to deal the ways in which time affects our abilities to perceive. These issues were further developed in a series of lectures of "internal time-consciousness" which Husserl delivered in 1905 and Heidegger subsequently edited and published under the title THE PHENOMENOLOGY OF INTERNAL TIME-CONSCIOUSNESS. (Probably the best place to read up on all this is in Izchak Miller's MIT Press book, HUSSERL, PERCEPTION, AND TEMPORAL AWARENESS.) The key questions seem to be how we perceive the passage of time and how we perceive what is happening as time passes. These remain very stick issues which tend to resist any attempts at reductive abstraction to static objects. I would like to insert an aside here to the effect that I feel my interest in music has been a major impulse to thinking about these questions. What makes music particularly interesting is that we have to deal with (i.e. reason about) the passage of time on several different scales simultaneously. This is nicely described in a passage I would like to quote from Stephen Handel's LISTENING, where, among other things, he tries to compare these time scales with respect to the perception of both music and speech: These variations often occur almost instantaneously (the changes in loudness, quality, and timbre of a note when a violin is initially bowed); they may also occur in short time periods (the changes in pitch, duration, and timbre for a syllable or a note) or occur in long time periods (the changes in loudness, duration, order, and rhythm among elements of a sentence or a musical phrase and the changes in position for a moving object). As Handel points out, what makes the situation particularly difficult is that these time scales cannot be segregated. In other words "what happens in the short time intervals affects what happens in the long time intervals and vice versa." Unfortunately, most people who "do" music theory have little to say about these dynamic processes (just as most software engineers try to abstract away the dynamic properties of running programs), although David Lewin has begun to ask some of these questions as a result of having read Miller's book on Husserl's time-consciousness. >In the third place, its easy to give lots of necessary conditions for >the intelligence of behavior. This is a worthwhile enterprise, but >until we can state a sufficient condition, those goal posts are gonna >keep on movin'. I doubt that Turing-style imitation conditions will >fill the bill, and in any case they are useless as design guides. > Returning to the matter at hand, Ken, you may agree wholeheartedly with my ultimate suggestion; but I think here you have fallen into the trap I have been trying to avoid. Necessary and sufficient conditions are devices we use when we are talking about OBJECTS. They carry an implicit assumption that we have an object which we can hold up and ask whether or not those conditions are satisfied. The point I want to hammer away at for a while is the skeptical question of whether or not processes, behaviors, and other "things that pass with time" can be viewed as such objects. The reason my intuition is putting up such a fight is because I worry about whether or not I can ever do more with a necessary or sufficient condition than apply it to a SNAPSHOT of a process (i.e. to "freeze time" and perform various forms of static analysis on that "frozen moment"). There is a nagging voice which keeps saying that no matter how many snapshots I accumulate, I may never be able to capture the process, itself. >Finally, words like "conscious", "intelligent", and "rational" have >very important roles in the conceptual framework of our moral and >legal system. You have no argument with me here. I would only assert that this makes it all the more important that we home in on better ways to talk about those words, reason with them, and use them. Consciousness may be far too important to be relegated to a universe of objects which are delimited by necessary and sufficient conditions. >A lot of philosophers have gotten badly burned by the concept of >emergent properties. I don't think Searle is relying on that concept, >but his arguments about programs not determining any "causal powers" >are similar to some arguments based on emergence. In my own thinking, >I have found that the concept of "normative property" does a much >better job of capturing the relation between (eg) rationality and >goal-seeking behavior. Dan Dennett's "Intentional Stance" is an OK >example of a normative approach, although he's not always as thorough >as someone like Davidson. If you find that the concept of emergence >is useful in formulating some ideas, you will probably find >normativity both more useful and more reliable. > I appreciate this observation and will try to follow though on it. Watch this space for further development. Meanwhile, I shall probably continue to use rec.music.classical as a forum for working out ideas specific to relating our understanding of the passage of time to what is roughly called "music appreciation" (otherwise known as "what the music critic's ear tells the music critic's brain . . . assuming he has either"). Perhaps at some point my struggles in these two arenas will come to a point of convergence. ========================================================================= USPS: Stephen Smoliar USC Information Sciences Institute 4676 Admiralty Way Suite 1001 Marina del Rey, California 90292-6695 Internet: smoliar@vaxa.isi.edu "Only a schoolteacher innocent of how literature is made could have written such a line."--Gore Vidal