Re: evolution and diet

From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Wed Apr 23 2003 - 13:55:22 MDT

  • Next message: Robert J. Bradbury: "HEALTH: weight control"

    gts wrote:
    >
    > I'm glad to see another paleo-guy in this thread. I think that makes about
    > 2.5 of us, where Eliezer seems to be the .5 :) (I'm still not clear about
    > what Eliezer really thinks, given that he keeps disagreeing with me even
    > while I keep trying to agree with him :-)

    Don't mind me. Just being precise. I count the right conclusion for the
    wrong reasons as a loss. Specifically, I object to the way you are
    attempting to claim the paleodiet is an a priori null hypothesis, rather
    than an experimentally supported *new* null hypothesis. A priori, there
    is nothing in particular to distinguish "eat what your ancestors ate" and
    "eat what tastes good" as evolutionarily motivated hypotheses. However,
    extensive research has shown that violation of ancestral invariants is bad
    for you, such as, e.g., eating ice cream, which mixes sugar and fat and
    messes up your insulin appetite regulation mechanism. You can't eat a
    whole stick of margerine. Your body will scream "Enough!" But mix lots
    of fat with lots of sugar - which you almost never find in ancestral foods
    - and you can eat a whole quart of ice cream. It tastes good, but it
    breaks the engineering assumptions. Is it bad for you? Yes. Does it
    taste good? Yes. Is it in the ancestral dietary range? No. So if you
    accumulate enough examples like these, you eventually conclude from
    *investigation*, and not a priori argument, that the paleodiet is a good
    null hypothesis.

    I've also noted a disturbing tendency for responders in this thread to
    attack the weakest arguments for the paleodiet - that means *yours*, gts -
    rather than the strongest arguments, or those arguments which appear in
    the literature. In other words, gts, you're using an argument which is
    much weaker than necessary - I'm not sure why - and people who, for some
    odd reason, want to argue about this, are responding to you, instead of
    reading the literature. Note that nobody is bothering to respond to my
    posts, only yours. This is because your posts are easy and fun to attack,
    because they use the weakest possible argument for a correct conclusion.

    > In my view the paleodiet is, however, still the base diet upon which we
    > should formulate an optimal diet. Again, it's a question of burden-of-proof.
    > As I see it, any deviation from the paleodiet needs to be justified by solid
    > scientific evidence.

    Because of the empirical regularity with which past deviations have turned
    out to be negative. *Not* because of a priori considerations. The
    paleodiet took its fair turn in the burden-of-proof barrel *and now it is*
    a good null hypothesis.

    Trying to argue the a priori plausibility of paleodiet theory, without
    citing the supporting evidence, is, in fact, a worthless argument. If you
    go on providing a worthless argument for paleodiet theory, the skeptics in
    this thread will go on arguing it to the exclusion of all other points raised.

    Accept that the burden of proof rests on paleo, then answer it.

    -- 
    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 - 14:07:14 MDT