527 
Rainbows in retrospect; L.A.S. Johnson's 
contributions to taxononnic philosophy 
David L. Hull 
Hull, David L. (Department of Philosophy, Nortinoestern University, Evanston, IL 60208) 1996. Rainbows 
in Retrospect: L.A.S. Johnson's Contributions to Taxonomic Philosophy Telopea 6(4): 527-539. Thirty 
years ago, L.A.S. Johnson wrote an extensive evaluation of the principles and techniques of 
numerical (or phenetic) taxonomy. Ten years ago he returned to these topics and this time included 
the principles of cladistic taxonomy as well. In this paper 1 re-examine Johnson's criticisms of 
phenetic and cladistic taxonomy to see how well they stand up to the test of time. On the main 
they stand up very well indeed. 
Introduction 
In 1968 in a Presidential Address to the Linnean Society of New South Wales, Johnson 
pubhshed a blockbuster of a paper criticizing numerical taxonomy. At the time I wrote 
at the top of my reprint of this paper 'The most stimulating paper that I've ever read.' 
Two years later this paper was reprinted in Systematic Zoology with an Addendum. 
Thereafter, Johnson remained all but silent on taxonomic methodology and philosophy 
until 1987 when he participated in a symposium at the 14th International Botanical 
Congress in Berlin (Johnson 1989). In his 'Rainbow's End,' Johnson presented a highly 
sophisticated view of systematics, not to return to these topics again for almost two 
decades. Why the long hiatus? Johnson (1989: 95) himself explains: 
Discussions of methodology in science tend to be much more prolix and less 
profitable than those of what we might loosely call fact, or of theory as it 
relates to the linking of facts and extracting generalizations about their relations 
to each other, whether in a causal framework or simply a correlative one. Far 
too much time, I believe, has already been spent on the methodology of 
taxonomy and indeed of phylogeny. 
Although my own contributions to systematics have been entirely within the context 
of taxonomic philosophy and methodology, I am forced to agree with Johnson about 
the relative value of methodological discussions to matters of fact and theory (Hull 
1979: 419). As in the case of political systems, many scientific systems may look 
good in principle, but when it comes to applications, they are disasters. Putting a 
particular method to work in order to see what the results are is the ultimate test of 
any methodology, and Johnson was in a position to do just that. 
Roughly thirty years have passed since Johnson and I started writing on systematics. 
Throughout this period we have found ourselves in basic agreement. We both agreed 
with the numerical taxonomists that systematics practice should be made as explicit, 
objective, quantitative, and repeatable as possible but were skeptical of what we 
took to be the overly empirical philosophy underlying numerical taxonomy. Johnson 
also presented specific criticisms of some of the mathematical techniques suggested 
by the numerical taxonomists. Finally, when cladistics arose, we had pretty much 
the same reaction to it. It seemed too 'rigid.' We also were equally put off by the 
later development of 'pattern' or 'transformed' cladistics, once again because we 
had our doubts about the extremely empirical philosophy that seemed to underlie it. 
