13.0399 decision-making; experimenting

From: Humanist Discussion Group (willard@lists.village.virginia.edu)
Date: Sun Feb 13 2000 - 11:02:39 CUT

  • Next message: Humanist Discussion Group: "13.0398 conferences: Virtuality in Europe; Design; Electronic Art"

                   Humanist Discussion Group, Vol. 13, No. 399.
           Centre for Computing in the Humanities, King's College London

       [1] From: lachance@chass.utoronto.ca (Francois Lachance) (21)
             Subject: Re: 13.0391 decision-making alternatives

       [2] From: "Osher Doctorow" <osher@ix.netcom.com> (42)
             Subject: Subject: Experimenting?

             Date: Sun, 13 Feb 2000 10:53:33 +0000
             From: lachance@chass.utoronto.ca (Francois Lachance)
             Subject: Re: 13.0391 decision-making alternatives

    Willard and Osher,

    The recent posting on decision-making and probability theory is for me
    fascinating, exciting and thrilling. Osher's explanation invities one to
    extend the very physical examples of rain and snow to textual elements.
    Given the presence or absence of an element how are we to assess the
    chances for another element to occur (before, after, concurrently) or not?
    I use "occur" here loosely since the occurance of an element can also
    stand it for the interpretation given an ambiguous element. What I do like
    about this type of calculation or computing is that it allows, for
    example, researchers working on the stylistic dimensions of oral delivery
    of poetic matter to compare their modelling with that of researchers
    tracing out the occurence of distinct letter forms in a set of manuscripts

    Willard, I was wondering if you would venture a few comments as to how the
    territory brought to our attention by Osher might shape your own research
    and experimentation into the problem of "weighting" and markup as related
    to the Onomasticon project...

    Francois Lachance
    Post-doctoral Fellow
    projet HYPERLISTES project

    --[2]------------------------------------------------------------------ Date: Sun, 13 Feb 2000 10:54:13 +0000 From: "Osher Doctorow" <osher@ix.netcom.com> Subject: Subject: Experimenting?

    Date: Sat. 12 Feb 2000 07:09:05 From: Osher Doctorow, <mailto:osher@ix.netcom.com>osher@ix.netcom.com

    Dear Colleagues:

    I am replying to Willard McCarty's question as to whether, when we do humanities computing, we are conducting an analogue of a laboratory experiment, and if we are, what are we experimenting on, and where is the reality we are trying to determine, and what is uncertain?

    The question is deep, and only a few points will be mentioned here. The advantages of a scientific laboratory experiment approach are perhaps obvious, although even in astronomy it needs to be generalized to the sense of many repeated controlled observations by trained persons. However, there are also dangers arising from a lack of philosophical training by scientists. Bayesian statisticians, for example, have presented a number of proposals formulated as axioms to outlaw competing types of statistics and probability which have already permeated the field of Artificial Intelligence in the USA and some of its friends under the name of Non-Monotonic Reasoning. Their favorite arguments involve half-thought-out arguments about birds flying in the air and the supposed ability of Bayesian approaches to update or correct knowledge based on new experimental or related data via Bayes Rule or versions of it. However, its competitor Logic-Based Probability (LBP) has an analogous updating equation involving addition and subtraction rather than multiplication/division as in the Bayesian approach. If an LBP person tries to publish a paper in mainstream Bayesian journals or in statistics journals sympathizing with this mainstream approach, it simply is rejected with such obscure comments as: It does not seem right, or even no comment, or 3 out of our 4 reviewers recommended rejecting it and the Editor concurs.

    What is happening here is largely bureaucratic defense of the accepted mainstream approach having nothing to do with laboratory or other experiments and much to do with protection of reputations, publishing versus perishing, promotions based on numbers of publications, and simply in-group over-identification versus the outsiders. Philosophy would cure the problem if carefully administered in large doses, but psychology itself is only at the initial phases of beginning to understand character and personality disorders which seem to go along with bureaucracy whether in higher education or politics. There is as yet absolutely no treatment for it. Perhaps a philosophy-psychology dialogue would be a starting point.

    Yours truly and sincerely, OD --------------------------------------------------------------------------- Dr. Osher Doctorow, Ph.D., Doctorow Consultants, Culver City, California, voice: 310-398-0693, fax 310-398-7954, <mailto:osher@ix.netcom.com>osher@ix.netcom.com

    This archive was generated by hypermail 2b29 : Sun Feb 13 2000 - 11:07:10 CUT