From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!sdd.hp.com!think.com!ames!sun-barr!news2me.ebay.sun.com!exodus.Eng.Sun.COM!orfeo.Eng.Sun.COM!silber Sun May 31 19:04:54 EDT 1992
Article 5989 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rpi!usc!sdd.hp.com!think.com!ames!sun-barr!news2me.ebay.sun.com!exodus.Eng.Sun.COM!orfeo.Eng.Sun.COM!silber
>From: silber@orfeo.Eng.Sun.COM (Eric Silber)
Newsgroups: comp.ai.philosophy
Subject: Homunculus and the witch's brew
Date: 31 May 1992 20:38:33 GMT
Organization: Sun Microsystems, Mt. View, Ca.
Lines: 49
Message-ID: <l2iea9INN44p@exodus.Eng.Sun.COM>
References: <1992May31.145204.16357@Princeton.EDU>
NNTP-Posting-Host: orfeo
Keywords: computation, transduction, homunculus, sensorimotor physiology

In article <1992May31.145204.16357@Princeton.EDU> harnad@phoenix.Princeton.EDU (Stevan Harnad) writes:

e.s: Let us reduce the sauce on high heat stirring constantly ::
s.h.::
>         COMPUTATIONALISM = HOMUNCULARISM
>             (Or, I AM A TRANSDUCER)
>
>Many readers still do not seem to have understood ....
>transduction,  ....
>...People keep reverting to
>the rival computational hypothesis  ...
>... in which you are a computational
>core, with the transducers simply carrying information TO it ("you"),
>...But this computational view, besides all the other points against it,
>is homuncular: ...
>...[W]e know that we are
>still "us" if we are blind, or deaf or paralyzed. But none of that
>refutes the hypothesis that we are transducers; nor does it make any
>sense of what would be left if every bit of transduction (and analog
>extesnions of it) were ablated from our brains (leaving, I assure you,
                                                          ^^^^^^^^
>very little in place). One thing I'm sure would not be left over would
                                  ^^^^^^^^
>be the little homunculus that normally sees what we see, hears what we
>hear, and thinks what we think. That is just the wrong view of
>cognition, and I am beginning to think that that perseverative
>computationalism is just homuncularism (an unhelpful form of dualism).
>
e.s:
 On the one hand, 
 S.H.'s transduction hypothesis becomes unconvincing
 exactly at that point where he ''assures'' me !
 I AM convinced that for humans,
 real world experience is neccessary for normal cognitive development, 
 and I AM convinvced that real-world-'grounding' may be essential for
 the instantiation of certain classes of symbols, but I am not
 convinced that the 'grounding' cannot be encoded. And what is a
 'transducer' anyway? Is it not ultimately equivalent to a 
 mathematical 'transformation' ? That is , a 'transducer' which
 converts a physical-event-stream-of-type-A into a 
 physical-event-stream-of-type-B  is equivalent to a mathematical
 transformation of a sequence. 
 A program, that is, a running-program executing on a von-neumann-computer,
 might 'compute' appropriate transformations of appropriate time-series,
 producing appropriate 'grounding'.  But why couldn't this satisfy
 S.H.'s grounding criteria anyway, it is a RUNNING-INSTANCE-of-a-program
 executing on a REAL-computer!  It is the HALT-ed-program which can
 have no grounding for any of its 'symbols' !!!!!!!



