Article 2 of 212

Subject:      Re: McKinstry's 400 terabyte genome
From:         gall@umanitoba.ca (Norman R. Gall)
Date:         1997/04/07
Message-Id:   <gall-0704971823320001@usrts1m15.uwinnipeg.ca>
Newsgroups:   comp.ai.philosophy
[More Headers]

In article <5ibs3l$4u4@now.cirl.uoregon.edu>, dej@now.cirl.uoregon.edu
(David Joslin) wrote:

>In article <gall-0704971120340001@usrts1m8.uwinnipeg.ca>,
>Norman R. Gall <gall@umanitoba.ca> wrote:

>>Not at all.  Why respond to nonsense in a refereed journal and make
>>yourself look at assinine as the drivel to which you are responding.
>
>On the other hand, if it did make sense to publish responses
>to to this kind of stuff, just look at all of the publications
>you could add to your CV!  After publishing something explaining
>what is wrong with the MISTIC project, you could tackle the "AI
>Heaven" claims posted every few weeks in comp.ai.  Then move on
>to the Mentifex model of the mind.  And remember the guy who
>had "discovered" the exclusive-or, and claimed that it was the
>prevailing use of inclusive-or logic that was preventing AI from
>achieving major breakthroughs?

Quite true.  I hadn't looked at things in just that way.

One's CV could be serveral pages longer in a weekend or two.

>People complain all the time about the kookery that shows up in
>comp.ai, but this would put an end to those complaints.  People
>would welcome the appearance of yet another buzzword-inspired
>crackpot, because it would mean an easy publication to put them
>one step closer to tenure.

And the way tenure has been handed out in some institutions these last few
years, it wouldn't be much of a change from present practices.

>On the other hand, it is so easy
>to imitate a crank, that people might try to pad their resume
>by posing as one.  They could prepare the rebuttal first, and
>then post their imitation-crackpot article ("Holographic quantum
>tomography for neural nets," or "Olfactory Turing Test Proposal,"
>or "The Turin Test shows Italy is the key to AI"), followed in
>short order by submitting the rebuttal for publication.

Well, isn't this the other side of the coin of Sokal's 'hoax' in _Social
Text_? If only doing so could have the political/pedagogical/philosophical
point as Sokal's endeavour.

>Serously, the problem in most of these cases is that it is very
>hard to prove that the person making the claims is wrong.

In fact, it might be _logically_ impossible to prove many (if not all) of
them wrong.  My philosophical approach to these issues in AI has been to
show that the assumptions underlying these AI 'views' is just that they
employ _metaphysical_ theses and as such _cannot_ be shown to be false.

If there are strange mathematical interpretations involved (which I
suspect is the immediate problem with McKinstry's thesis), then we have a
compounded mess since getting someone to renounce such a mathematical
mistake is almost as difficult as seeing a metaphysical thesis as such AND
mathematics is a powerful club with which to hit someone over the head.

>But what they uniformly fail to understand is that the burden of
>proof is on them.

This problem is rife not only in AI but in UFOlogy, parapsychology, etc.
The psychology isn't exactly the same, but very allied in form and
tenacity against counter argument - and provision of *good* arguments for
their case.

>McKinstry is free to *demonstrate* that his
>claims are true, as is the AI Heaven guy, and the mentifex guy,
>etc. etc.  Wild unsubstantiated claims will generally be laughed
>at.

Interestingly, those of us who would laugh at such wild unsubstantiated
claims are usually slammed as 'unscientific,' 'obstructionist,' and 'close
minded.'

As well, even if some of these people *can* provide some evidence for
their claims, we still have to carefully examine the assumptions against
which this 'evidence' is interpreted.  If the assumptions are flawed, the
interpretatios cannot be trusted.

>(To the credit of the Mentifex guy, he's always courteous,
>unlike McKinstry and some of the others.  The right-justified
>text -- without adding exta spaces -- is unique, and he certainly
>seems to be intelligent.  If we have to have people making
>wild unsubstantiated claims, I wish more of them were like him.)

I couldn't agree more.

Norm Gall

--
"All philosophy can do is to destroy idols. And that means not
creating a new one - for instance as in the 'absence of an idol.'
        - L. Wittgenstein