[PREV - CSICOP] [TOP]
CRAZYIDEAS
January 24, 2003
A review of
Nine Crazy Ideas in Science
A Few Might Even Be True
by Robert Ehrlich
2001, Princeton University Press
The concept of this book
is absolutely superb:
someone with a technical background
and an open mind investigates in
detail a number of "crazy" ideas, to
see if there's anything to them.
The execution of the idea is not
perfect, but Robert Ehrlich has done
better at this difficult job than
anyone else I know of. This book is
highly recommend as a good review of
the evidence on some scientific
controversies.
Here's the deck of nine ideas under consideration:
More Guns Mean Less Crime
AIDS is Not Caused by HIV
Sun Exposure is Beneficial
Low Doses of Nuclear Radiation Are Beneficial
The Solar System Has Two Suns
Oil, Coal, and Gas Have Abiogenic Origins
Time Travel is Possible
Faster-than-Light Particles Exist
There Was No Big Bang
Part of the idea here is that
Ehrlich is not telling you in
advance what his conclusions were
on any of these issues, though
there's an "answer key" summary in
the back if you'd like to peek
ahead of time. He says he's tried
to keep an open mind, and claims
that his investigations changed
his mind about some things (though
he never says about what exactly).
So in this review I'm going to
give you generalities first, and
bury "the butler did it" type
information after a SPOILER
warning.
One of the problems with the
execution of this work is that you
can pretty often tell when Ehrlich is
enthusiastic about an idea just from
his general tone as he writes about
it... and conversely, in retrospect I
think I should've been able to spot
the chapters about issues he
disagreed with, because there the
writing was a little unclear.
Part of his schtick is that at the
end of each chapter he rates the
idea on a scale of 0 to 4
"cuckoos", and oddly enough I
often find that I strongly
disagree with his cuckoo ratings
even just based on the evidence
that he presents, though, the
absolute magnitude of my
disagreements are typically no
more than a single "cuckoo".
I was worried about some of his
evaluation criteria, because he
includes several points that
strike me as fairly dicey:
"Who proposed the idea?";
"How attached is the proposer to the idea?"
"Does the proposer have an agenda?"
These all relate to judging
the *person* rather than the (Consider that
idea itself. "consider the
source" and "ad
But he does clearly understand hominem argument"
that these are just rules of are the same as far
thumb, and I note with some as logical
amusement that he doesn't resort validity goes.)
to these particular rules anywhere
in the following chapters.
He's more interested in the logic
of the arguments, which is as it
should be.
I could bring up lots of quibbles
(and will after the spoiler warning),
but overall I found this to be a great
breezy read.
I learned quite a bit from
it. While nothing here made me do
a reversal of my beliefs, often I
was surprised that the evidence
for something was stronger or
weaker than I'd supposed.
Here we have an educated, astute,
person doing a relatively
independent review of some
controversial, interesting
technical subjects. Why aren't
there more books like this?
Now, into the nitty gritty. Here's your SPOILER WARNING.
Skip the following if you want to play the "guess where he's
going" game with this book.
Chapter by chapter:
More Guns Mean Less Crime
I was surprised that the data
doesn't support private ownership
of guns as a crime deterrent.
Ehrlich argues persuasively that
the statistical evidence for this
is very weak. I appreciate the
fact that Ehrlich concludes that
both the pro and anti gun sides 3 = "almost certainly not true"
are nutty: he rates them 3 and 2 2 = "very likely not true".
"cuckoos" respectively.
But here, we come to my first
strong disagreement with him. If
the effects aren't strong enough
to measure, why the asymmetry in
the "cuckoo" rating for the pro Is it worth mentioning that
and anti side? I might rate them this whole "do guns discourage
both at a 2 myself. crime?" question has nothing
to do with the actual purpose
(But then, when Ehrlich wrote this, of the second ammendment?
he still believed in the claims
of that historian who said The actual question is,
he'd demonstrated that the do private guns discourage
gun-toting American pioneer tyrrany.
doesn't exist. This historian
has been discredited since Does distributed power
then...) help you maintain
distributed power.
AIDS is Not Caused by HIV
I've had the impression that the
the Duesberg hypothesis was
pretty screwy, but I was willing For example, what about the possibility
to tentatively consider it might that multiple diseases have been
have something of value. incorrectly lumped into the one
category of "AIDS"?
But Ehrlich's analysis satisfies
me that there's not much of
scientific value in Duesberg's
ideas at all. I don't argue with
his 3 cuckoo rating (but I
wouldn't blame you if you thought
it deserved the full 4).
Sun Exposure is Beneficial
Ehrlich concludes that this looks
fairly plausible, and gives it a 0
cuckoo rating, pretty much as I would
have expected. Many people might find
this surprising though: the popular
impression these days seems to be that
sunlight is deadly.
Low Doses of Nuclear Radiation Are Beneficial
Here lays out the case for
"radiation hormesis":
Myself, I don't think this
notion is that fantastic (the
difference between a poison and
a medicine is often a matter of
dosage, why wouldn't this be
true of radiation?).
But radiation is so demonized
in the popular imagination, the
idea that "radiation is good
for you" comes off an insane
joke.
Ehrlich takes this notion
seriously, and essentially
concludes that while there are
reasons for suspecting that
this effect exists, it hasn't
been established.
And here we have one of my quibbles:
he awards it 1 cuckoo, which
translates to "probably not true, but Some people seem to think
who knows". But there is *no* reason I'm confused about burden
for saying it's probably not true. of proof here, but I think
If something is not crazy, just not they're not counting their
established, I would be inclined to cuckoos correctly:
award it "0 cuckoos", aka "Why not?".
A zero cuckoo rating does
not mean "I will believe this
unless proven otherwise",
The Solar System Has Two Suns it just means "this isn't
crazy".
This is the "Nemesis"
hypothesis and it should
come as no surpris ehat this
is rated at 2 cuckoos.
The short version of the story:
originally they looked at part
of the extinction record, and
it looked like there was a
definite cycle. But if you look
at the whole record it doesn't
seem to be there.
Oil, Coal, and Gas Have Abiogenic Origins
I've been interested in this
ever since I heard Thomas Gold
give a talk on this idea about a GOLD
decade ago. It turns out that
this is now looking much less
like "an intriguing possibility"
and much more like a truth
awaiting a few funerals before
it will be declared established.
GOLD
The odds are good that "fossil
fuels" don't actually come from
fossils, rather they're from
hydrocarbons that pre-existed
the formation of the earth, So that means we can ignore
which means we're probably not those environmental wackos,
going to run out of them. right? Nope: imagine what
happens to the atmosphere if we
Ehrlich rates this at 0 keep ramping up the rate at
cuckoos, but maybe he should which we burn this stuff.
have invented a "-1 cuckoo"
for this one. The "peak oil" gang
are a bunch of optimists.
Time Travel is Possible
2 cuckoos: no surprises.
Faster-than-Light Particles Exist
Ehrlich mentions in his introduction
in the interests of "full disclosure"
that he's actually strongly attached
to one of the ideas discussed here
(the existence of tachyons), but by
the time I'd gotten to that chapter
I'd entirely forgotten about this.
I was disappointed to realize that he
was being an advocate in this one
chapter, not an independent reviewer
(it includes a picture of him wearing
a "no tardy-centrism" T-shirt).
Ehrlich rates this at 0 cuckoos,
but come on. Even just based on
the write-up he presents, it's a
clear 1 cuckoo.
There Was No Big Bang
Clocks in at 3 cuckoos, as you might expect.
--------
[NEXT - MIND_PENROSE]