Re: evolution and diet

From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Wed Apr 23 2003 - 20:13:32 MDT

  • Next message: Mitchell Porter: "Doomsday vs Diaspora"

    gts wrote:
    > Eliezer S. Yudkowsky:
    >
    > You did not answer my question, Eliezer.
    >
    > Would Mr. Hugh M. Species in 12,000 BC have been acting unreasonably if he
    > asked the new dairy and grain farmers to prove their case that their
    > new-fangled foods were healthy additions to the diet? Please answer this
    > important question.

    Mr. Hugh M. Species is acting unwisely if he does not investigate the
    question of how new dairy and grain products correlate with health, just
    as he is acting unwisely if he does not investigate the question of
    whether good-tasting and bad-tasting foods correlate with health, or how a
    proposed strategy of "eating ancestral foods" correlates with health.
    Investigation is generally a good idea. It is burden-of-proof to which I
    object.

    In matters such as these, the wise Bayesian attempts to accumulate enough
    evidence to make all reasonable differences of priors moot.
    Burden-of-proof amounts to relying on your priors and that, I think, is
    not wise. If you don't have enough evidence to settle the issue
    regardless of where any reasonable burden of proof rests, you don't know.
      Moreover, mentioning burden-of-proof is a common tactic used by arguers
    who simply increase the burden of proof whenever their opponents come up
    with new evidence, or who try to establish burdens of proof so great that
    no one can possibly meet them.

    Furthermore, allocating the burden of proof to one side or the other, in
    advance, restricts your spirit of creative inquiry. In the beginning Hugh
    M. Species should simply examine these new dairy and grain products to see
    what they are, gathering information, not immediately launching a war
    between one position and the other. Perhaps neither position is correct.

    For these reasons, arguing about burdens of proof should be considered
    harmful. 

    -- 
Eliezer S. Yudkowsky                          http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence

    This archive was generated by hypermail 2.1.5 : Wed Apr 23 2003 - 20:25:03 MDT