krulwich-bruce@CS.YALE.EDU (Bruce Krulwich) (05/10/88)
In article <31024@linus.UUCP> bwk@mbunix (Barry Kort) writes: >It is not uncommon for a child to "spank" a machine which misbehaves. >But as adults, we know that when a machine fails to carry out its >function, it needs to be repaired or possibly redesigned. But we >do not punish the machine or incarcerate it. > >Why then, when a human engages in undesirable behavior, do we resort >to such unenlightened corrective measures as yelling, hitting, or >deprivation of life-affirming resources? This can be explained easily in light of AI theories of the roles of expectations in cognition and learning. Your example could be explained as follows: 1. Yelling at and hitting a person because of something he's done is irrational 2. People have the expectation that other people act rationally 3. When someone yells at you, it triggers a failure of this expectation 4. So, the person being yelled at or hit tries to explain this expectation failure, hopefully concluding that he did something that the other person feels strongly about 5. Thus he learns that the other person feels strongly about something, which most of the time is the goal that the yeller or hitter had in the first place Bruce Krulwich
bwk@mitre-bedford.ARPA (Barry W. Kort) (05/11/88)
I appreciated Bruce Krulwich's analysis of the cognitive chain initiated by a yelling/hitting episode. The fifth (and last) link in the chain of reasoning, is that the target of the verbal abuse draws a conclusion about the abuser: > 5. Thus he learns that the other person feels strongly about > something, which most of the time is the goal that the yeller > or hitter had in the first place Wouldn't it have been easier if the yeller had simply disclosed his/her value system in the first place? Or do I have an unrealistic expectation that the yeller is in fact able to articulate his/her value system to an inquiring mind? --Barry Kort
smoliar@vaxa.isi.edu (Stephen Smoliar) (05/12/88)
In article <31570@linus.UUCP> bwk@mbunix (Barry Kort) writes: >I appreciated Bruce Krulwich's analysis of the cognitive chain >initiated by a yelling/hitting episode. The fifth (and last) >link in the chain of reasoning, is that the target of the verbal >abuse draws a conclusion about the abuser: > >> 5. Thus he learns that the other person feels strongly about >> something, which most of the time is the goal that the yeller >> or hitter had in the first place > >Wouldn't it have been easier if the yeller had simply disclosed his/her >value system in the first place? Or do I have an unrealistic expectation >that the yeller is in fact able to articulate his/her value system to an >inquiring mind? > As long as the agents we are talking about are "all-too-human" (as Nietzsche put it), your expectation is quite unrealistic. I think you are overlooking how great an extent we rely on implict assumptions in any intercourse. If we had to articulate everything explicitly, we would probably never get around to discussing what we really wanted to discuss. The problem comes in deciding WHAT needs to be explicitly articulated and what can be left in the "implicit background." That is a problem which we, as humans, seem to deal with rather poorly, which is why there is so much yeeling and hitting in the world.
tomh@proxftl.UUCP (Tom Holroyd) (05/17/88)
In article <5499@venera.isi.edu>, smoliar@vaxa.isi.edu (Stephen Smoliar) writes: > I think you are overlooking how great an extent we rely on implict > assumptions in any intercourse. If we had to articulate everything > explicitly, we would probably never get around to discussing what we > really wanted to discuss. True communication can only occur when both parties understand what all the symbols used to communicate mean. This doesn't mean you have to explicitly define what you mean by "tree" every time you use the word tree, but it's a good idea to define it once, especially if it's something more complex than "tree" (with due respect to all sentient hardwood). >The problem comes in deciding WHAT needs to be explicitly articulated >and what can be left in the "implicit background." That is a problem >which we, as humans, seem to deal with rather poorly, which is why >there is so much yeeling and hitting in the world. Here's a simple rule: explicitly articulate everything, at least once. The problem, as I see it, is that there are a lot of people who, for one reason or another, keep some information secret (perhaps the information isn't known). A truly reasoning being doesn't hesitate to ask, either, if something hasn't been explicitly articulated, and it is necessary for continuing discussion. Tom Holroyd UUCP: {uunet,codas}!novavax!proxftl!tomh The white knight is talking backwards.
nau@frabjous (Dana Nau) (05/18/88)
In article <180@proxftl.UUCP> tomh@proxftl.UUCP (Tom Holroyd) writes: >True communication can only occur when both parties understand what all >the symbols used to communicate mean. This doesn't mean you have to >explicitly define what you mean by "tree" every time you use the word >tree, but it's a good idea to define it once, especially if it's something >more complex than "tree" (with due respect to all sentient hardwood). There's an obvious problem with this: since your definitions are themselves formulated symbolically, you can't define *everything* without running into circularities. This is why mathematical axiomatic systems always contain some terms that are explicitly left undefined. As an amusing example of the problems that arise with trying to define everything, I had a math prof about 10 years ago who pointed out that the dictionary definition of "didapper" is "a dabchick or other small grebe." Perhaps you all know what a dabchick or a grebe is, but none of us did. Dana S. Nau ARPA & CSNet: nau@mimsy.umd.edu Computer Sci. Dept., U. of Maryland UUCP: ...!{allegra,uunet}!mimsy!nau College Park, MD 20742 Telephone: (301) 454-7932
gilbert@cs.glasgow.ac.uk (Gilbert Cockton) (05/19/88)
In article <180@proxftl.UUCP> tomh@proxftl.UUCP (Tom Holroyd) writes: >Here's a simple rule: explicitly articulate everything, at least once. Sorry, I didn't quite get that :-) > >A truly reasoning being doesn't hesitate to ask, either, if something >hasn't been explicitly articulated, and it is necessary for continuing >discussion. Read Irvine Goffman's "The Presentation of Self in Everyday Life" and you'll find that we do not correspond to your "truly reasoning being". We let all sorts of ambiguities and incompleteness drop, indeed it's rude not to, as well as displaying a lack of empathy, insight, intuition and considerateness. Sometimes, you should ask, but certainly not always, unless your a natural language front-end, then I insist :-) This idealisation is riddled with assumptions about meaning which I leave your AI programs to divine :-) Needless to say, this approach to meaning results in infinite regresses and closures imposed by contingency rather than a mathematical closure n+1 n information = information where n is the number of clarifying exchanges between the tedious pedant (TP) and the unwilling lexicographer (UL). i.e there exists an n such that UL abuses TP, TP senses annoyance in UL, TP gives up, UL gives up, TP agrees to leave it until tomorrow, or ... TP and UL have a wee spot of social negotiation and agree on the meaning (i.e. UL hits TP really hard) -- Gilbert Cockton, Department of Computing Science, The University, Glasgow gilbert@uk.ac.glasgow.cs <europe>!ukc!glasgow!gilbert The proper object of the study of Mankind is Man, not machines
bwk@mitre-bedford.ARPA (Barry W. Kort) (05/19/88)
The discussion seems to be turning to the shortcomings of human-human communication. It might by helpful to review some estimates of the human information processing capacity. It is estimated that the human mind accumulates and retains over a lifetime enough information to fill 50,000 volumes. That's quite a library. The human input/output channel operates at about 300 bits per second (30 characters per second). Exchanging personal knowledge bases is a time-consuming operation. We are destined to remain unaware of vast portions of our civilization's collective information base. Much of what we know is not easily reduced to language. That which cannot be described in words may have to be demonstrated in action. Some people speak of secret knowledge or private language. I don't think there is anything nefarious here. If I attempted to express every notion that crossed my mind, I would surely drive people nuts. --Barry Kort
bryan@trsvax.UUCP (05/20/88)
/* ---------- "Re: Acting irrationally (was Re: Fr" ---------- */ In article <5499@venera.isi.edu>, smoliar@vaxa.isi.edu (Stephen Smoliar) writes: >> I think you are overlooking how great an extent we rely on implict >> assumptions in any intercourse. If we had to articulate everything >> explicitly, we would probably never get around to discussing what we >> really wanted to discuss. /* Written 9:44 am May 17, 1988 by proxftl.UUCP!tomh (Tom Holroyd) > Here's a simple rule: explicitly articulate everything, at least once. ^^^^^^^^^^!! That's the rub, "everything" includes every association you've ever had with hearing or using every word, including all the events you've forgotten about but which influence the "meaning" any particular word has for you, especially the early ones while you were acquiring a vocabulary. You seem to have some rather naive ideals about the meanings of words. > A truly reasoning being doesn't hesitate to ask, either, if something > hasn't been explicitly articulated, and it is necessary for continuing > discussion. A truly reasoning being often thinks that things WERE explicitly articulated to a sufficient degree, given that both parties are using the same language. ------------------ Convictions are more dangerous enemies of truth than lies. - Nietzsche ...ihnp4!convex!ctvax!trsvax!bryan (Bryan Helm)
smoliar@vaxa.isi.edu (Stephen Smoliar) (05/23/88)
In article <180@proxftl.UUCP> tomh@proxftl.UUCP (Tom Holroyd) writes: >In article <5499@venera.isi.edu>, smoliar@vaxa.isi.edu (Stephen Smoliar) >writes: > >> I think you are overlooking how great an extent we rely on implict >> assumptions in any intercourse. If we had to articulate everything >> explicitly, we would probably never get around to discussing what we >> really wanted to discuss. > >>The problem comes in deciding WHAT needs to be explicitly articulated >>and what can be left in the "implicit background." That is a problem >>which we, as humans, seem to deal with rather poorly, which is why >>there is so much yeeling and hitting in the world. > >Here's a simple rule: explicitly articulate everything, at least once. > >The problem, as I see it, is that there are a lot of people who, for >one reason or another, keep some information secret (perhaps the >information isn't known). > No, the problem is that there is always TOO MUCH information to be explicitly articulated over any real-time channel of human communication. If you don't believe me, try explicitly articulating the entire content of your last message.
gilbert@cs.glasgow.ac.uk (Gilbert Cockton) (05/23/88)
In article <32403@linus.UUCP> bwk@mbunix (Kort) writes: >We are destined to remain unaware of vast portions of our civilization's >collective information base. And so therefore are computers! Still want to get mind on silicon? Only if it's information free? Anyone remember GPS? Can you get anything on without requiring some knowledge (remember expert systems?) A computer will be beyond programming whereever its input bandwidth is substantially narrower than ours. It will just take far too long to get the knowledge base in, except for small well-defined areas of technical knowledge with a high pay off (e.g. Prospector, genetic engineering advisors). Thus the potential of AI is obviously limited even ignoring formalisation problems related to the large area of non-technical knowledge which underlies most social interaction outside of high-tech. work. -- Gilbert Cockton, Department of Computing Science, The University, Glasgow gilbert@uk.ac.glasgow.cs <europe>!ukc!glasgow!gilbert The proper object of the study of humanity is humans, not machines