3.274 evidence, cont.; the Logical Detectives (128)

Willard McCarty (MCCARTY@VM.EPAS.UTORONTO.CA)
Thu, 20 Jul 89 17:49:16 EDT


Humanist Discussion Group, Vol. 3, No. 274. Thursday, 20 Jul 1989.


(1) Date: Thu, 20 Jul 89 09:20:32 BST (28 lines)
From: Donald Spaeth (0532) 33 3573 <ECL6DAS@CMS1.UCS.LEEDS.AC.UK>
Subject: Evidence (and CHEd)

(2) Date: Wed, 19 Jul 89 17:30:00 EDT (24 lines)
From: P.Burnhill@edinburgh.ac.uk
Subject: Re: 3.265 computers in the humanities, cont. (54)

(3) Date: Thu, 20 Jul 89 09:17:49 -0400 (EDT) (51 lines)
From: Leslie Burkholder <lb0q+@andrew.cmu.edu>
Subject: Re: Logical detectives, issue 3.268

(1) --------------------------------------------------------------------
Date: Thu, 20 Jul 89 09:20:32 BST
From: Donald Spaeth (0532) 33 3573 <ECL6DAS@CMS1.UCS.LEEDS.AC.UK>
Subject: Evidence (and CHEd)

I agree wholeheartedly with Steven DeRose's adjustments to my
comments that the the attempt to count cases is often invalidated
by the poor nature of the evidence. I did not wish to suggest
that counting is never valid, nor that the computer does not
make it possible to improve the quality of research. There are
many areas of my specialisation, history, where quantitative
methods are very useful; that's why I stressed the importance
of knowing one's sources (a point Joseph Rudman has made well in
the context of Authorship Attribution studies, a highly quantitative
field).

I was reacting against two commonly-stated (but, I hope, rarely
believed!) arguments: that the citing of examples ("cut and paste
history", it's been called) is in some way less valid than the
citing of numbers--each is appropriate to particular types of
sources and fields; and that computers will make for more
scientific, by which is (wrongly) meant more objective,
results. In the golden age of humanities computing, I expect
conflicts over interpretations to continue to rage!

Cheers,
Don (Spaeth)
Arts Computing Development Officer
University of Leeds
(2) --------------------------------------------------------------35----
Date: Wed, 19 Jul 89 17:30:00 EDT
From: P.Burnhill@edinburgh.ac.uk
Subject: Re: 3.265 computers in the humanities, cont. (54)

The term 'sample' implies notions of representativeness, which may be why
John W Tukey, a Princeton statistician of some note, when confronted
with some observations for the first time liked
to refer to a 'batch' of data, rather than grace it with
undeserved status. If we know something of the 'sampling method',
that is, how the data arose (survived) then we may risk the term.

Big batches in themselves should not inspire confidence, except to
let us know that an occurrence was more than an isolated incident.
For us to estimate (or judge) whether
such occurrences were widespread or major happenings
requires access to a 'sample' which we reckon to have come from some
'population' in some predictable way.

Surely the fact that computers are high speed idiots capable of doing sums
across very large batches of data shouldn't mean that we should
believe what these high speed idiots conclude.

Peter Burnhill
.
(3) --------------------------------------------------------------54----
Date: Thu, 20 Jul 89 09:17:49 -0400 (EDT)
From: Leslie Burkholder <lb0q+@andrew.cmu.edu>
Subject: Re: Logical detectives, issue 3.268

Matthew Gilmore refers us to an article in the 19 July 1989 issue of the
Chronicle of Higher Education on some programs written largely by
philosophy undergraduates at Trinity University. The group of students
writing the programs is called the Logical Detectives. There are several
programs with various topics. One is about Jack the Ripper, one about a
wierd treasure pit in Nova Scotia, one about AIDS, and one about the
drug problem.

I attended a session at a recent IBM ACIS conference during which Peter
French, the organizer of the project and a professor of philosophy at
Trinity, gave a talk about the programs and demonstrated parts of some
of them. Anyone interested in the programs should read the article. I'll
not summarize it here. Professor Gilmore asks for comments from someone
in philosophy. I'll give some.

In general the programs look exciting and interesting. What is exciting
and interesting about them is that they (at least the Jack the Ripper
and the Money Pit Madness programs) are more involving and more detailed
exercises than those usually provided in critical thinking or logic
classes. I wonder about some claims made about them:
(1) That in writing the programs, the students exercised skills they
were taught in philosophy classes. Perhaps. A strong version of this
would be: If they had not taken some or other philosophy classes, they
would not have been able to write the programs at all or would at least
have produced worse programs or would have had a harder time producing
the programs. A weaker version of this would be: If the students writing
the programs had not certain skills and knowledge (wherever they
obtained the skills or knowledge, in a philosophy class or elsewhere),
they would not have been able to write the programs at all or would at
least have produced worse programs or would have had a harder time
producing the programs. Are either of these true? What are the skills
that are supposed to be employed?
(2) That in using the programs, students exercise skills they are taught
in philosophy classes (especially in critical thinking or logic
classes). This remark is relevant only to Jack the Ripper and Money Pit
Madness. Again, a strong version of this would be: If a student does not
understand some of things taught in these classes or have the skills
that are supposed to be acquired from these classes (whether or not the
knowledge and skills were acquired in the philosophy classes or
somewhere else), then they won't be able to solve the problems the
programs pose. Is this true? Another possibility would be that: Solving
the problems the programs pose, like doing any other exercise, helps
students to understand or acquire the knowledge that they are supposed
to be acquiring in certain philosophy classes. What knowledge or skills
is this?

LB