citation of Sokal's work

Joe Felsenstein joe at
Mon Oct 13 12:42:57 EST 1997

Mark Siddall <msiddall at> complained that
>Joe Felsenstein wrote:
>> >  [And why aren't any of these papers cited by systematic
>> >biologists who use the technique?]
>> Partly it's just become so well-known that no one cites an origin for
>> it.  But partly it's the Not Invented Here syndrome.  The phylogenetic
>> systematics school has let all work before 1969 fall into the memory hole,
>> preferring to cite Hennig as the originator of numerical work on phylogenies.
>This is a wholesale misrepresentation intended, in would seem, to merely
>slander the school of thought with which Felsenstein retains so much
>disdain.  In the first place, none in the phylogenetic systematics
>school would have much reason to reference UPGMA at all for obvious
>reasons, and so Felsenstein's comment is irrelevant.  Secondly, no one
>ever has attributed UPGMA to Hennig.  Third, a mere cursory reading of
>the literature will demonsatrate that cladists were more aware of the
>origins for numerical methods than many others (e.g., Nelson's
>revisiting Jardine's work).  That Hennig did not happen to use the word
>"parsimony" hardly means that Camin and Sokal originated the idea. 

There is some justice in Siddall's complaint that my comment is irrelevant.
Byron Adams had asked about why the papers that were the origin of
UPGMA were not cited by later authors using this method.  This set me off
into a tirade (above) about lack of citation of pre-1969 work on parsimony
by the phylogenetic systematics school.  Of course they were not using UPGMA

UPGMA's originators mostly don't get cited because the method has become so
widely used that everyone can't remember when it wasn't around.

However the phylogenetic systematics school has not given enough credit to
Sokal and Michener (1958) as being the first one to outline a numerical method
for inferring phylogenies (I have to say I too have not given them enough
credit for that either).  They also under-credit Camin and Sokal for their 1965
paper, which led directly to almost all late-1960's work on parsimony methods,
and instead cite Hennig or Herb Wagner as the inventors of parsimony.
Neither was (Herb does not claim to be).

Where all this hangs together is that there was a late-1950's to late-1960's
literature on numerical methods for classification and for phylogenies, and
it developed independently of Hennig's work.  It included papers by
Sokal, Michener, Sneath, Edwards, Cavalli-Sforza, Wilson, LeQuesne, Estabrook,
Farris, and Kluge.  The merger with the phylogenetic systematics school
comes later, with a tentative start by Farris, Kluge, and Eckhardt in 1970.

In the retelling of the history by that school, numerical work on phylogenies
springs directly from Hennig, like Minerva springing from the head of Zeus.
For an example, see the Farris and Kluge 1997 paper mentioned below.

>Rather than taking Felsenstein's ipse dixit, the scholarly student would
>be wise to read Farris, J. S. and Kluge, A. G. 1997. Parsimony and
>History. Syst Biol 46: 215, and references therein.

Being ipse, I still dixit.  In my view the arguments of Farris and Kluge in
that article are particularly unconvincing, so I would suggest a close read
of the "references therein".  But then, I'm biased.

Joe Felsenstein         joe at
 Dept. of Genetics, Univ. of Washington, Box 357360, Seattle, WA 98195-7360 USA

More information about the Mol-evol mailing list