From: Eliezer S. Yudkowsky (sentience@pobox.com)
Date: Wed Jan 15 2003 - 09:26:07 MST
You know, you meet people who grew up in the age of newspapers and
magazines, and they'll read just one article and get their information
from that. Sad, really.
My child... have you let Google into your heart?
I've visited Lomborg.com, Anti-Lomborg.com, read the Scientific American
critique, and have simultaneously read Lomborg's point-by-point reply to
the critique along with Scientific American's counterreplies to Lomborg's
critique, split over two monitors.
http://www.lomborg.com/critique.htm
http://www.sciam.com/article.cfm?articleID=00000B96-9517-1CDA-B4A8809EC588EEDF
My advice is to ignore nontechnical articles, as of the popular news
media, as they are rich on accusations (on both sides) but short on facts.
Overall, my impression would be that while Lomborg takes hits, he scores
two to three times as many serious points against his opponents as they do
against him. In only three cases does a Scientific American critic make
what seems to me to be a serious point; in one case, that Lomborg quotes
nuclear power as supplying 20% of power used in nuclear-equipped
countries, when it actually generates 20% of grid electricity and only 10%
of power. Lomborg correctly says "Sorry, you've caught a genuine error"
and promises to post an erratum on his website. In the second case,
Lomborg's analysis of species extinction involves a projection of forested
area which fails to distinguish between old-growth habitats and new
forests. In the third case, Lomborg quotes a researcher as saying species
extinction rates were "incalculable" when examining the surrounding
context reveals that the researcher probably meant "very large", although
the critic fails to dent Lomborg's point that the species extinction rate
*is* incalculable.
These points scored against Lomborg stand in contrast to the severer
errors made by his critics, most noticeably including Schneider's
misrepresentations of a study on the adaptive iris effect and of a study
on the interaction of cosmic rays with global cloud cover, both of which
were corrected by letters to Scientific American from the researchers
involved.
But then counter-errors like this one aren't really the issue; we're
interested in whether Lomborg has made severe errors, not whether his
critics have made worse. In general, it looks to me like, with the three
exceptions noted above, Lomborg replies successfully to most points made,
even after the counterreply is taken into consideration. Note that
Scientific American receives the last word in this exchange; and as
Scientific American sued Lomborg to prevent him from posting his original
point-by-point reply on his own site, they have control over who gets the
last word.
I am not a specialist in environmental science, and am not qualified to
make a statement about the facts of the matter being disputed. I am not
even qualified to know whether Lomborg is, as his critics accuse him,
selectively quoting data.
However, if Scientific American is claiming that they, in their critique
or their counterreply, have not just asserted but *shown* Lomborg to be
guilty of bias - to say nothing of a systematic pattern of deliberate
scientific dishonesty - then they are simply wrong. Having read the
critique, the reply, and the counterreply, SciAm simply has not publicly
presented any evidence which warrants this conclusion.
Had SciAm managed to present many errors in the same class of the three
above, they would have at least managed to keep the evidence *compatible*
with an accusation of blatant mistakes and consistently poor scholarship
by Lomborg. As it is, if this is the strongest critique they could make,
that upper bound on strength exonerates Lomborg. The SciAm presentation
was characterized by selective quoting and outright errors, and when these
are pointed out, the points against Lomborg turn out to be either trivial
or outright wrong. Given the overall pattern of scholarship by Lomborg's
critics I expect that they chose Lomborg's seemingly weakest points to
criticize, rather than sampling Lomborg's book and performing deeper
investigations at randomly selected points. Given that even after this
selection bias most attempted criticisms still fail, with only three
points still standing, it seems unlikely to me that Lomborg's book is
really of below-average scholarship.
SciAm's criticism and counterreply are laced with invective, ad hominems,
and points that are simply irrelevant to the subject at hand. This would
perhaps be excusable had they succeeded in making their accusations stick.
They did not.
For example, SciAm points out that Lomborg is not himself an
environmentalist researcher but a social scientist. So what? Lomborg may
be a kindergarten student for all I care, if you cannot refute his facts.
If you expect Lomborg as a nonspecialist to make mistakes, then it
should be easy for you to find specific cases where Lomborg messed up; so
easy, in fact, that there is absolutely no need for you to argue that
Lomborg is a priori likely to make mistakes. Actually, given the enormous
range of different fields covered by the debate, even had Lomborg been a
specialist in one field he could not have been a specialist in all.
Let us assume that Lomborg's basic thesis is correct; when you consider
the academic reaction to Lomborg, it is no wonder that this thesis was not
presented by an insider to the field. An environmental researcher who
attempted to present Lomborg's thesis would have been permanently throwing
away their career. Even as an outsider Lomborg's scientific career has
been damaged by the accusations leveled against him (regardless of their
truth). Given this, to argue that Lomborg is not an insider is
disingenuous (cannot be evidence against Lomborg's main point), even if
the speaker was unfamiliar with the basic rules of science and had not
read Kuhn's "The Structure of Scientific Revolutions".
The evidence does not definitely disprove any of the following scenarios:
1) Lomborg has engaged in a subtly biased presentation of a healthy
scientific field, based on careful selection of evidence; the experts are
capable of detecting this, but not capable of explaining what it is they
see or why. In frustration they have leveled massive, angry, accurate
denunciations of Lomborg, but without being able to back up their claims
to nonspecialists.
2) The environmentalist field has been engaging in a systematically
biased presentation of data, which Lomborg has reviewed and then presented
neutrally. The environmentalist field is invested in the belief that
environmental *activism* is critical to the survival of Earth, tends to
selectively ignore or underreport evidence which would interfere with the
dramatic claims needed to promote that activism, and has built up a
fundamentally flawed scientific consensus based on the selective
presentation of evidence at conferences, selective remembering of evidence
by researchers to each other, and so on. Lomborg's expose of this massive
bias has been met with incredible anger by the activist-scientists whose
work to promote specific political agendas has now been undermined;
believing in their own scientific consensus, they issue drastic
condemnations of Lomborg as a means of political damage control, but are
then unable to back up these accusations.
3) The environmentalist field has arrived at an academically dominant
consensus which is the result of systematically biased focus on data that
supports the political agenda of environmental activists. But, Lomborg's
presentation of this data also admits of some bias in the opposite
direction - with the note that this bias is (a) not deliberate and (b)
considerably less than that of the environmental activist-scientists.
None of these hypotheses are definitely ruled out by the evidence, but I
would say that the evidence seems to support (3).
Analyzing the pattern of critique and countercritique, Lomborg has a
clearer focus on the facts, confronts issues more squarely, and usually
ends up in the lead after the exchange of arguments, even though his
critics have had the opportunity to select his weakest points out of an
entire book.
What most strongly gave me the impression of empty righteous indignation
was the repeated accusations that Lomborg has contradicted the scientific
consensus. Duh. Lomborg knows that already. He's trying to show the
scientific consensus is wrong.
Accusations that Lomborg fails in scholarship, references, or other forms
of scientific professionalism are simply not supported by his critics.
Why (3) instead of (2)? At one point, a critic takes Lomborg to task for
presenting the species extinction rate as 0.7% over 50 years, instead of
1,000 times the natural rate of species extinction, since the public might
become less worked up over a rate of 0.7% over 50 years. (Jesus Haploid
Christ on a smoking birch bark canoe! Changing the presentation of a
statistic to manipulate public opinion? Why not just hack off your own
foot with an apple slicer and be done with it?) Lomborg retorts that it
is the 0.7% over 50 years figure that is relevant to any estimate of
ecological wealth or its supposed decline. Perhaps the most depressing
part of the entire readthrough, from my perspective, was that neither the
critic nor Lomborg thought to suggest providing BOTH figures to the public
and letting us MAKE UP OUR OWN DARNED MINDS about which one was important.
Given this, it seems likely that Lomborg has an ordinary human
rationalization mechanism operating normally (not all that suprising a
conclusion, really) and there is probably an ordinary amount of bias in
his reports. The attempt by his critics to make this out to be a
systematic pattern of dishonesty fails; they end up picking nits and often
even the nits fail, and at any rate demonstrate a noticeably larger amount
of rationalization and bias in their replies than they manage to accuse
Lomborg of in his original.
Final verdict: Lomborg's not perfect, but he wins the exchange.
Scientific American gets another black eye and absolutely does not have
enough ammunition on hand to justify a title like "Science defends itself
against _The Skeptical Environmentalist_".
-- Eliezer S. Yudkowsky http://singinst.org/ Research Fellow, Singularity Institute for Artificial Intelligence
This archive was generated by hypermail 2.1.5 : Wed Jan 15 2003 - 17:35:51 MST