From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!cs.utexas.edu!swrinde!mips!spool.mu.edu!cs.umn.edu!uc.msc.edu!shamash!map Tue Nov 19 11:09:19 EST 1991
Article 1232 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.ecf!utgpu!cs.utexas.edu!swrinde!mips!spool.mu.edu!cs.umn.edu!uc.msc.edu!shamash!map
>From: map@u02.svl.cdc.com (Mark Peters)
Newsgroups: comp.ai.philosophy
Subject: Re: Animal Vs. Human Intelligence
Message-ID: <37801@shamash.cdc.com>
Date: 7 Nov 91 21:32:07 GMT
References: <191674@tiger.oxy.edu>
Sender: usenet@shamash.cdc.com
Reply-To: map@svl.cdc.com
Distribution: na
Organization: Control Data Corporation, Silicon Valley Operations
Lines: 143

In <191674@tiger.oxy.edu> khan@oxy.edu (Onnie Lynn Winebarger) writes:

>>map@u02.svl.cdc.com (Mark Peters) writes:

>>What I said is exactly correct.  Nobody can tell a lie (and know he is)
>>unless he already has some grasp of what honesty is, simply because
>>lying *is* one form of NOT being honest.

>  It seems to me this argument is invalid.  This isn't about Washoe being able
>to conceptualize that she's not telling the truth.  It's about the fact that
>she conceptualized an event which didn't happen.  What's more, I doubt either
>of the assistants she claimed to be responsible for the "doodoo" had ever 
>decided to break toilet training, so she had never even seen the act which 
>she described.  So she couldn't have based this story on something she had 
>seen happen.  So she must have imagined it.	Correct me if I'm wrong, but 
>I've always thought the ability to imagine entails an ability to 
>conceptualize.

As I said before, all Washoe needed to do this is a functioning sensory/
perceptual mechanism, memory, the ability to associate one perceptual
level entity with another, and an aversion to pain/unpleasantness.        
Washoe had signs for herself, the humans, the doo-doo, and she knew
(from memory) that she would be punished for soiling the carpet.  She
saw a human associate the sign for "Washoe" with the doo-doo, and
knowing that punishment was imminent, repeated what the human signed,
replacing the sign for "Washoe" with the sign for a human - all in 
an attempt to avoid punishment.  

Even if Washoe did conjure up an image of a human doing this, that does not  
qualify as conceptualization - it is merely a slightly more sophisticated 
use of memory, i.e., of her sensory/perceptual mechanism.  The question
of a link between imagination and a conceptual faculty depends entirely
on *what* is being imagined, not merely *that* imagination happens.  To
conjure up images of nothing but actually existing things that have been
observed, and then merely rearranging them, doesn't require conceptualization.
To do the same with things that are not, in principle, directly observable, 
and to do so not only with things, but with classifications of things,     
actions, attributes, and relationships - that does require conceptualization.
No evidence of this exists with respect to apes.

>>                                   A human child who does know what
>>lying is also knows:
>>
>> 1) That he has the power to choose (volition)
>> 2) That thinking is what ought to guide his choices, not whim (reason)
>> 3) That thinking requires identifying facts (logic)
>> 4) That there are things in the world that are important to him,
>>    some more than others, which he therefore acts to obtain (values)
>> 5) That obtaining values demands a certain course of action (virtue)
>> 6) That accepting the facts of reality and acting accordingly is
>>    an example of such a course of action. (honesty)

>  This part seems fairly unnessecary.  When did one have to have a concept of
>something to be able to do it?	Did man have to have a concept of how gunpowder
>actually worked before he started making explosives out of it?	No, he just
>needed to know how the explosion worked.  Why does one need to have a 
>concept of lying to actually lie?

>   Also, why does the child need to grasp any of (1)-(6) to actually be able 
>to "do" them.  For example, (3) suggests you have to know that thinking 
>requires identifying facts.  You don't.  You just start identifying facts.  
>Having these concepts sorted out is not how one begins.

Having a concept of something and having a scientific explanation of
that something are not the same thing.  You're right - we don't need
to know *why* gunpowder works in order to make and use it.  The maker  
of gunpowder needs to know that it is made of certain proportions of 
certain compounds; the user of gunpowder needs to know that certain
conditions must be met for it to explode.  Both people have the concept
"gunpowder," with the maker having a deeper grasp of it than the user,
but both have a less detailed concept than does the chemist who tries
to make better gunpowder.

I tried to indicate an analogous situation with my description of 
the knowledge a child would need to grasp "lying."  The child doesn't
require an adult understanding of the concepts I mentioned, just as
the user of gunpowder doesn't require the understanding of a chemist.
I tried to pare down highly abstract concepts like "logic" and "virtue"
to the essentials a child would have to grasp in order to lie.  This
was not a claim that a child would be able to define "logic," even   
to the level I indicated - only a claim that a child has to grasp,
in some terms, these essentials before he could tell a lie.

To qualify as a lie, what is said must not only contradict the facts, but 
the speaker must *know* that it does before he says it, and he must   
have said it for the express purpose of getting something (or avoiding
something).  This does, I think, imply at least a primitive grasp of
all the concepts I mentioned above.  

In reference to my reduction of "logic" to the grasp that "Thinking requires 
identifying facts," you denied that this grasp is needed, saying that 
"You just start identifying facts."  I would like to amend my
reduction to "Thinking *is* the identification of facts (by a particular
method)."  We aren't born with the ability to automatically distinguish
fact from non-fact.  We have to learn that this is necessary, that not
just any old method will do (e.g., wishing doesn't make it so), and that
we have to do it by choice.  We can't "just start identifying facts", 
we have to struggle mightily, making alot of mistakes on the way before
we can do it reliably.  A child that has no grasp of this cannot tell
a lie.

>>I think "apparent lying behavior" is just that - apparent behavior that
>>does not require conceptualization to perform.  Washoe doesn't have *any*
>>grasp (even a primitive one) of the concepts mentioned above, so she
>>can't know what "lying" is.  What looks to us like "lying behavior" is
>>explainable in perceptual terms, so I think that positing a conceptual
>>faculty in apes on the basis of that behavior is not justified.

>  Behaviorists said the same thing about humans.  However, I'm not sure 
>how much of this argument is about Washoe knowing that she's lying or if 
>it's about Washoe having a conceptual faculty.

The possession of a conceptual faculty is a prerequisite of being able to 
lie and to know that you are, so the argument is about both.  It is very 
easy for us to mistakenly project human attributes and abilities into other 
animals on the basis of superficial similarities, and I think this is a
classic example of exactly that.

>    IMHO, anyone that uses natural language must have the ability to
>conceptualize, or what you say with your words has no meaning.	This fact is
>probably why it's so difficult (read impossible) to have computers learn 
>natural language.  Until they are able to conceptualize what they're saying, 
>they'll be unable to give meaning to their words.  At any rate, the fact 
>that Washoe even uses a language in an apparently meaningful way is enough 
>to grant that she has a conceptual faculty.  To tell me this is not so 
>would take out at least half (probably more) of the reasoning I use to 
>assume other humans have minds, including you.

But implicit in what you say is the claim that what apes do with signs
is language - I deny this.  Communication is not language; signalling
is not language; language is much more than the ability to associate
one percept with another.  What the apes do with signs (or colored disks,
or whatever) is no different, in principle, from what a dog does with a
trick, i.e., it is nothing more than a perceptual-level mechanism for
getting what it wants.  The animal research doesn't support the opposite
view, and even it did, a massive amount of other evidence would have to be 
explained away/refuted before such a claim could even be thought possible.

(Sorry about the length, but I couldn't see any way to cut more.)
--
Mark A. Peters                              ****** ======================
Control Data Corporation                    ****** == "What a save!!!" ==
Internet: map@svl.cdc.com                   ****** == "What an idea!!" ==


