Review of The Extended Phenotype

From: Michael Ragland (ragland37_at_webtv.net)
Date: 09/03/04 

Date: Fri, 3 Sep 2004 20:27:22 +0000 (UTC)

http://www.ttaxus.com/

Dawkins, Richard -- The Extended Phenotype
The Extended Phenotype: The Long Reach of the Gene
Richard Dawkins 1982, rereleased 1999
Oxford University Press

By Jonathan Badger

Why am I reviewing this 20 year old book? Well, despite getting a
doctorate in microbiology and publishing several papers on the topic of
molecular evolution, I had never actually read any of Dawkins' works
dealing with the gene as the unit of selection, although I could hardly
avoid being at least superficially exposed to it via the popular media.
I chose to read "The Extended Phenotype" rather than "The Selfish Gene"
because I was under the impression that the former was a more serious
book; Dawkins agrees with my assessment: he states in the preface (p.
viii) of "The Extended Phenotype" that even if one never reads anything
else by him, one should at least read this book. Fair enough, Richard.

The first thing one notices about the book is that the font is extremely
small -- a colleague of mine even claims that is why his copy sits
unread on his shelf. Ignoring the font issue, the first chapter of "The
Extended Phenotype" opens quite well, with Dawkins arguing that the
gene-centric view of evolution is not meant to be taken as a testable
scientific hypothesis in itself but instead as a framework that can
suggest testable hypotheses in specific cases. Not only is this part
well argued, it is extremely well written as well.

Another interesting part about this chapter is it discusses Dawkins'
background. Although the popular media often refer to Dawkins as "a
British geneticist" (presumably because of "The Selfish Gene"), there is
a good explanation as to why he was never mentioned in any of my
genetics classes; he isn't a geneticist at all but rather an ethologist
(a student of animal behavior), having studied under the Nobel
prize-winning ethologist Niko Tinbergen. Understanding Dawkins'
background makes many of his later arguments more understandable (if not
always more convincing); while assuming the genetic nature of a trait
without experimental information may strike a geneticist as being
ungrounded, it follows naturally from the traditional ethological
technique of envisioning evolutionary explanations for observed
behaviors.

Chapter two, "Genetic Determinism and Gene Selectionism", is by far the
weakest chapter in the book. Essentially Dawkins argues that he is not a
genetic determinist (he has often been so accused) because he admits
that the environment interacts with genes to shape behavior. However, in
the sense that Dawkins likes to postulate genes that incline (if not
force) animals, including humans, towards specific behaviors, Dawkins
obviously is a genetic determinist. It is true that the phrase "genetic
determinist" often had (and has) connotations of racism and reactionary
politics, and it is understandable that the non-racist, non-reactionary
Dawkins would want to clear himself of unpleasant characterizations.
However, there must be some descriptive term for Dawkins' belief that
behaviors can be explained in terms of specific (albeit always
conveniently unknown) genes that were selected for on the basis of
influencing specific behaviors. If not "genetic determinism", then what?
(Just for the record, I don't deny that any given behavior could in
principle be influenced by specific genes -- I just think that such
claims should only be made on the basis of experimental evidence
concerning the genes and behaviors in question. Simply assuming that
genes for behaviors must exist is as intellectually bankrupt as assuming
that life on other planets must exist.)

The next chapter, "Constraints on Perfection", finds Dawkins yet again
on the defensive, although it is a better written and more interesting
chapter than the previous one. Dawkins uses the chapter to argue that
his views on adaptionism aren't really much different from those of
critics like Lewontin and Gould -- he agrees with them that there have
been many naive adaptionist explanations, and that unless a population
contains the needed genetic variation, natural selection is powerless.
Still, he can't help wondering if Lewontin and Gould aren't purposely
overstating their dissatisfaction with adaptionism simply in order to
score some debating points or to sound more revolutionary than they
really are. After all, despite all the fuss, both Gould and Lewontin
have used typical adaptionist explanations in their own work. Dawkins
may have a point there.

In chapter 4, "Arms Races and Manipulation", Dawkins finally gets down
to science. In this chapter Dawkins argues that organisms have evolved
mechanisms to influence ("manipulate") the behavior of other organisms
and that these organisms in turn have evolved mechanisms to avoid being
so influenced. The "arms races" refers to the coevolutionary history of
the manipulator and manipulated. The examples range from the obvious (an
angler fish's lure manipulates prey into coming nearer) to the more
surprising and less obviously antagonistic (male song birds use song to
manipulate the females' hormonal levels). All of this is very
interesting but the reader may wonder what these descriptions of
traditional organism level selection has to do with Dawkins' own gene
selectionism. Dawkins ties it together near the end of the chapter with
the example of the cuckoo manipulating other birds to work for the
benefit of the cuckoo's egg, and thereby its genes.

In chapter 5, "The Active Gene-line Replicator", Dawkins presents his
arguments for the basis of gene selectionism. The chapter starts out,
bizarrely enough, with the description of various forgotten definitions
of the gene from the early days of genetics (the "recon", the minimum
unit of recombination; the "muton" the minimum unit of mutation; the
"cistron", a contiguous stretch of chromosome encoding a protein).
Dawkins isn't bringing up these archaic terms for nostalgia or humor; he
thinks it's completely appropriate to have multiple definitions of the
gene -- and in fact he wants to add a fourth definition, the "optimon"
(the minimum unit of selection).

What is so odd about Dawkins' love of definitions is that even in 1982
it was obvious that genes were no more or less than nucleotide
sequences, obsoleting all previous speculations; while one may argue
about various properties of these nucleotide sequences (such as whether
or not they are the units of selection), it makes no sense to introduce
a definition like the "optimon" as an alternative to the idea of the
gene as a nucleotide sequence. But I think I understand where Dawkins is
coming from -- his training in genetics was in the classical tradition
of Mendelian genetics when genes were abstractions encoding traits
rather than real concrete entities encoding proteins. He has not yet
internalized the reality of molecular genetics; genes are still
abstractions in his heart of hearts.

However, as Dawkins himself says (p. 86), "...let us not become worked
up over terminology". What is Dawkins actually getting at here?
Basically Dawkins argues for the existence of replicating units
("replicators", stretches of a chromosome which need not be identical to
a gene). Replicators that increase the chances of the host reproducing
tend to survive over time; others are lost. It is easy to see that this
is essentially the same model as in classical population biology in
which nicely atomic alleles independently increase or decrease the
"fitness" of their host and thus either increase or decrease their
frequency in the population. The problem is that molecular biology has
taught us that genes work in pathways and so their effects cannot be
independent. While the loose definition of "replicator" could encompass
multiple genes, with the exception of bacterial operons, these genes
tend not to be adjacent and thus could not be in general part of the
same replicator. I do not see how this could be resolved in Dawkins'
framework.

I have come to this point in the review without mentioning what is the
most interesting idea presented in the book -- the "extended phenotype"
itself. In essence Dawkins argues that the external products of
genetically determined animal behavior (say a bee hive) can be treated
exactly in the same way as body phenotypes. Indeed, as behavior modifies
the environment, in some sense the distinction between organisms and
their environment (and even other interacting organisms) is artificial.
This idea is fascinating and I can see its possible utility easier than
I can for his "replicator" theory.

The 1999 rerelease concludes with an afterword by the philosopher
Dennett. Whether this a good addition depends on your point of view.
Personally, while respecting ancient philosophers like Plato, I don't
find modern philosophers very insightful. As with theology, almost all
of the interesting questions of philosophy have long since been
appropriated by science. In regard to Dennett's afterword, my respect
for philosophy is not particularly raised by Dennett's constant
confusing of the results of the research program of molecular genetics
with those of traditional evolutionary biology. Despite Dennett's
assertions, Dawkins' ideas have had essentially no influence on such
molecular fields as HIV research. I'm not saying that HIV research (for
example) couldn't benefit from Dawkins' ideas, just that they are not
being used -- count the few citations to Dawkins in molecular biology
papers if you disagree.

Although I have been quite critical of the book, I have to admit that it
was probably the most thought-provoking book I've read in quite some
time. A book well worth reading.

"It's uncertain whether intelligence has any long term survival value.
Bacteria do quite well without it."
 Stephen Hawking

Relevant Pages

  • Re: Review of The Extended Phenotype
    ... > that the environment interacts with genes to shape behavior. ... > the sense that Dawkins likes to postulate genes that incline (if not ... I think that the current academic climate is very hostile to philosophy. ...
    (sci.bio.evolution)
  • Re: Darwins Dangerous Idea-No Middle Ground
    ... >> John Wilkins wrote: ... Dennett certainly likes Dawkins' view ... > process of information generation according to algorithms, and treats genes as ...
    (talk.origins)
  • Re: Richard Dawkins
    ... "When we die we can leave behind us only two things, genes and memes. ... And Dawkins offers these examples - ... lines "lives of great men all remind us, we can make our lives sublime, ... Many cnturies before either Dawkins or Longfellow, Quintillian taught ...
    (talk.origins)
  • Re: Richard Dawkins
    ... "When we die we can leave behind us only two things, genes and memes. ... And Dawkins offers these examples - ... lines "lives of great men all remind us, we can make our lives sublime, ... Many cnturies before either Dawkins or Longfellow, Quintillian taught ...
    (talk.origins)
  • Re: Exactly what are "recessive genes"?
    ... Dawkins is the best populariser of the utter misuse of ... population genetics models within evolutionary. ... That theory remains classic Darwinism which to this ... very day is NOT understood as a REFUTABLE theory of nature. ...
    (sci.bio.evolution)