From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!mips!spool.mu.edu!uunet!dg-rtp!sheol!throopw Tue Apr  7 23:24:14 EDT 1992
Article 4929 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca sci.philosophy.tech:2510 comp.ai.philosophy:4929
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!mips!spool.mu.edu!uunet!dg-rtp!sheol!throopw
>From: throopw@sheol.UUCP (Wayne Throop)
Newsgroups: sci.philosophy.tech,comp.ai.philosophy
Subject: Re: A rock implements every FSA
Summary: seems not to be Putnam's argument anymore
Message-ID: <4401@sheol.UUCP>
Date: 6 Apr 92 01:13:18 GMT
Article-I.D.: sheol.4401
References: <1992Mar31.070425.1633@csustan.csustan.edu> <1992Mar31.104855.10515@husc3.harvard.edu> <60507@aurs01.UUCP> <1992Apr2.012749.10596@husc3.harvard.edu>
Lines: 67

WT: throop@aurs01.UUCP (Wayne Throop)
MZ: zeleny@zariski.harvard.edu (Mikhail Zeleny)

MZ:
> if you can formulate a theoretical distinction
> between a material object "responding" to input by "producing" output, and
> another material object "just sitting there", please let the world know
> about it.  

WT:
> An object can be said to produce output in a
> nontrivial sense if the map used to interpret output states does not
> depend upon time.  

MZ:
> Looks good.  How about an existence proof?  Start at the beginning.

Pardon.  Existence of what?

WT:
> [...] the requirement not to depend upon time is very modest
> indeed, yet seems to kill Putnam's line of reasoning quite dead.

MZ:
> You don't seem to have thought this through.  

Agreed.  I was only concerned with the formulation of "to implement"
in the "all objects implement every FSA" reasoning.  The formulation
of "objects" I was taken as given.

Now looking at the formal notion of objects instead of that of
implementation, I suppose it possible that Mikhail was asking for proof
of the existance of objects at all, within some formal framework.  Fair
enough, I suppose, but but whether the notion of "objects" can be
formalized adequately or not, it still seems quite clear that in any
reasonable formalization of the notion, time co-ordinates would not be
physical attributes of objects, and hence can't legitimately be used as
the basis of attributing states to objects.  This is enough to rule out
Putnam's construction, whether or not there are other problems with the
notion of objects implementing FSAs. 

MZ:
> Start by individuating a
> material object that persists through time.  Keep in mind the "ship of
> Theseus" problem of loss and replacement of its material constituents.
> Proceed to individuate inputs and outputs in a strict materialistic
> framework.  Pay attention to the problems of context-dependence and
> type-identity.

Quick sketch.

Objects are abstractions that organize observables.  The bounds on
what can "be an object" are only the constraints of observation.  The
object "persists through time" as long as the observables that are
modeled by the object persist.

The replacement of components is not a problem, as long as they
maintain the observable invariants used to abstract the object.

The notion of "object" IS context dependent, and only has a
type-identity in so far as it convenient in forming the abstraction. 

Presumably the above was considered and rejected by western philosophy
many years ago, and I'm a crude philistine to bring it up despite its
obvious and manifold faults. 
--
Wayne Throop  ...!mcnc!dg-rtp!sheol!throopw


