From newshub.ccs.yorku.ca!ists!torn!utcsri!rpi!zaphod.mps.ohio-state.edu!cs.utexas.edu!uunet!trwacs!erwin Tue Jun 23 13:21:09 EDT 1992
Article 6305 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!torn!utcsri!rpi!zaphod.mps.ohio-state.edu!cs.utexas.edu!uunet!trwacs!erwin
>From: erwin@trwacs.fp.trw.com (Harry Erwin)
Newsgroups: comp.ai.philosophy
Subject: Re: Transducers
Message-ID: <639@trwacs.fp.trw.com>
Date: 18 Jun 92 16:27:34 GMT
References: <1992Jun17.132117.9273@Princeton.EDU> <60837@aurs01.UUCP> <1992Jun17.182829.18441@mp.cs.niu.edu>
Organization: TRW Systems Division, Fairfax VA
Lines: 23

On analog transducers...
If the way my current train of thought is leading me holds water, there
are at least four layers involved. The innermost layer is action, the next
layer out is consideration, the third layer is scrutiny, and the outermost
layer is scanning. I act in, consider, and examine a realistic representation
of the world that is maintained by scanning. The world scan consists of
objects, keyed by texture and narrow-band sound spectrum, localized only
generally until I scrutinize them. Small objects are represented by Gabor
functions (intrinsically analog). Harnad's transduction is a way of
describing the process by which this world scan is maintained. Once I
become interested in a specific object in the world scan and orient to it,
I shift to a different representation.

Some comments:
1. this is speculative, and
2. this is deliberately laid out at a specific level of abstraction,
so please don't bite my finger, instead look where I'm pointing.

Cheers,
-- 
Harry Erwin
Internet: erwin@trwacs.fp.trw.com



