> I wonder if anyone else sees these research findings through Kuhnian
> eyes? When do the anomalies trigger the response in us that these
> data are not coherent and that the consensus picture of the origin of
> phyla is in urgent need of a revolution?
These data seem to me (and to the authors of the study you quoted, and
to other biologist I know) to fall quite short of being "paradigm-
shifting" data, for the following reasons:
1) It's not at all clear that molecular clock techniques are calibrated
well enough to support the small error bars given by Wang et al. Their
study is impressive, but molecular clock techniques extended over
hundreds of millions of years are much better at giving relative dates
than absolute ones. Their error bars take into account statistical
errors, and perhaps some known systematic errors. However, there could
well be (and probably are) some other, unaccounted, systematic errors,
because obtaining absolute dates from such techniques relies on several
assumptions which are accurate (in my opinion at least) to only about
"plus or minus a factor of 2."
2) I would not at all be surprised if many of the phyla split from each
other several hundred million years before the Cambrian explosion, with
their representatives being microscopic and soft-bodied -- living "in
the shadow," as it were, of the Ediacaran critters. (My opinion on
this point my be very much a minority one. I don't know that field
well enough to say.)
3) The paradigm of common ancestry is judged to derive strong support
from the pattern of nested homologies which we see in gene sequences,
genome organization, and developmental programs -- patterns well-
matched to the fossil record. These patterns don't stop at the level
of family, order, or class. They continue right up to the phylum
level.
Recent developments in cosmology -- regarding the age of the universe -
- offer an interesting story with many parallel features. (Did you
happen to follow this story?) Throughout the 80's and early 90's,
cosmologists had been estimating the age of the universe as 12-15
billion years (errors bars reflecting the error estimates in measuring
Hubble's constant). Meanwhile, stellar physicists had been using
independent techniques to measure the age of the oldest stars at 12-15
billions years. The oldest stars are *supposed* to be 1-3 billion
years younger than the universe. So far, no problem. But in the mid-
90's, as we got better measurements of Hubble's constant, the best
estimates of the age of the universe kept creeping towards the lower
end of that range, while the age of the oldest stars stubbornly stayed
just as high or slightly higher than the best estimates of the age of
the universe. Problem? Well, cosmologists weren't TOO worried about a
"Kuhnian crisis." Why? First, comfort was taken in the fact that
these two independent techniques gave numbers which really were pretty
close to each other. Second, each technique relied upon several
assumptions which could -- upon better examination -- potentially shift
their predictions by perhaps 10 or 20 percent. And that is exactly
what happened in the last few years! As it turns out, one big
assumption in previous estimates of the age of the universe was that
the "cosmological constant" was exactly zero. But in the last few
years, astronomers have found pretty good evidence that the
cosmological constant is not zero. If we plug the new non-zero value
for the cosmological constant into the equations for the age of the
universe, and the age of the universe is once again comfortably longer
than the age of the oldest stars. Crisis resolved (for now).
This story which played out so recently in cosmology looks -- to me --
just like what's happening right now regarding the different techniques
(geological and molecular) for dating the origin of phyla.
Loren Haarsma