From newshub.ccs.yorku.ca!torn!cs.utexas.edu!sun-barr!olivea!uunet!trwacs!erwin Wed Sep 16 21:22:08 EDT 1992
Article 6817 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca sci.cognitive:400 comp.ai.philosophy:6817
Path: newshub.ccs.yorku.ca!torn!cs.utexas.edu!sun-barr!olivea!uunet!trwacs!erwin
>From: erwin@trwacs.fp.trw.com (Harry Erwin)
Newsgroups: sci.cognitive,comp.ai.philosophy
Subject: More Mind Reading
Keywords: cognition
Message-ID: <716@trwacs.fp.trw.com>
Date: 8 Sep 92 13:13:33 GMT
Followup-To: sci.cognitive
Organization: TRW Systems Division, Fairfax VA
Lines: 82


Some comments on my "mind reading" posting. (There are other notes, too,
but I'm in the middle of a proposal and won't be able to get to them for
3-4 days.)

>From: Gordon Joly <G.Joly@cs.ucl.ac.uk>

>How would artificial mind reading (c.f. AI) be distinct from real mind
>reading? The sort of thing we experience from time to time 

>``Let's go for a walk to buy bread'' ... ``I was just thinking about
>exactly that ...''

>is too subtle to be measured objectively.

These are probably the same phenomenom.

>But how does that feedback into the (model of the) individual?

>Gordon.

See the next item...

>From: scarter@almaak.usc.edu (Steve Carter)

>Harry -
  
>     I'm intrigued by your posting regarding mind reading.  PUtting aside
>for the moment the fact that I think that there are enormous problems
>with the methodology you describe (essentially a sort of triple-dissoci-
>ation!), and thus the conclusions you draw, I find your hypothesis
>intuitive appealing.
  
>     Could you do the following?:
>     1) Send me a description of how you "operationalize" personality

I thought about this quite a bit. The conscious mind appears to me to be a
quasiperiodic process that seems to be able to monitor around 7 "chunks"
at a time by cycling through images (for me, sound images) in turn. The
ego seems to be able to control this process in some manner. For example,
if I'm driving, I seem to spend most of my time in the polysensory and
visual areas monitoring for novelties. Interestingly, memory seems to work
by my leaving a request for data at one of those stages. When I get back,
the retrieval has usually been done. There is a tight connection (for me)
between this cognitive cycle and speech generation. I suspect that if we
were to study other species, we would find that the 7+/-2 rule is no
longer true. I know that when I try not to use verbal images, I can't
handle as many unconnected "chunks." I also suspect that other people use
other mechanisms for cognition.

One point is that if this process is chaotic, it should be inexpensive to
control. Hence, the ego can be an emergent process with relatively low
cost to maintain. (Relatively low may still be quite high!)

I tried to see what conditions allowed me to "read minds." One thing I
have to do is ignore my own body sensations. Another thing is that a good
many of the cues I use for calibrating a personality model seem to involve
frequency, phase, and texture data. If I scratch myself, I know how it
feels, sounds, and appears. If someone else scratches themself, there's
overlap, and I can imagine how it feels from the sound and appearence. I
suspect a lot of the processing occurs in one of the polysensory areas,
the difference being that I've redirected the inputs. The effort seems to
be involved in doing the input redirection.
 
It's very hard to run two personalities at once, although I can do it,
using some sort of switching between input sources. Apparently if the time
spent in each personality model is short enough, the other doesn't fade to
non-existence while it is cut off from sensory inputs. It's a strange
experience, though, to be knowingly running two models simultaneously.


>     2) Send me citations on the two authors to which you refer.

I owe this to a number of people. I have the Pecora and Stafford papers
somewhere in my office... Please give me 3-4 days. I don't have Paul
Raup's papers--I have listened to him speak at the Washington Evolutionary
Systems Society, so I'll try to get the citations from a colleague.
 
Cheers, 
-- 
Harry Erwin
Internet: erwin@trwacs.fp.trw.com


