[Taxacom] Reproducibility of phylogenetic analysis

Richard Jensen rjensen at saintmarys.edu
Wed Jan 27 11:18:59 CST 2010

I couldn't agree more. We have to assume that there are patterns (e.g., 
regularities that can be detected, whether with respect to morphological 
features, nucleotide sequences, atomic structure, response to light 
photons, etc.) that represent the consequences of real phenomena (my 
words are fraught with loaded meanings - no apology for that). After 
all, if every event is unique, and every event requires an independent 
explanation, then we can never have hypotheses, laws and theories, 
except for the conclusion that we can never know how or why anything 
happens!). So, if we want to explain why things are the way they are, we 
start with observations, notice patterns (regularities) and then, given 
whatever assumptions we use to organize our thinking, try to find out if 
we can explain these patterns and if our explanation is useful for 
interpreting other patterns.


Dick J

Richard Jensen, Professor
Department of Biology
Saint Mary’s College
Notre Dame, IN 46556
Tel: 574-284-4674

Richard Zander wrote:
> Well, Dick, I think we need to go beyond patterns. The particular
> processes and the particular historical events that lead to patterns and
> clustering in nature should be a focus. Doubtless patterns lead to and
> guide study that might reveal processes and historical events. Patterns
> also are convenient for classification, but surely knowledge (when such
> can be had) of a process and/or an historical event should also inform
> classification. Also, patterns of what? With morphological data we get
> inferred clusters of shared homologous traits fixed at nodes in a
> gradualist model of evolution, with molecular data we get inferred
> genetic continuity and isolation events without relevance to taxa. Of
> course, compared to these bad interpretations in phylogenetic analysis,
> phenetics is refreshingly clean in what it is doing.
> *****************************
> Richard H. Zander 
> Voice: 314-577-0276
> Missouri Botanical Garden
> PO Box 299
> St. Louis, MO 63166-0299 USA
> richard.zander at mobot.org
> Web sites: http://www.mobot.org/plantscience/resbot/
> and http://www.mobot.org/plantscience/bfna/bfnamenu.htm
> Modern Evolutionary Systematics Web site:
> http://www.mobot.org/plantscience/resbot/21EvSy.htm
> *****************************
> -----Original Message-----
> From: Richard Jensen [mailto:rjensen at saintmarys.edu] 
> Sent: Wednesday, January 27, 2010 9:54 AM
> To: Richard Zander
> Cc: Stephen Thorpe; Bob Mesibov; TAXACOM
> Subject: Re: [Taxacom] Reproducibility of phylogenetic analysis
> I think the key is that analytic methods produce results that we wish to
> interpret according to some set of a priori 
> assumptions/hypotheses/guesses. What we get, as Stephen noted, is a 
> function of what those assumptions etc. are. Are they a function of some
> explicit process, or are they just one possible reflection of a 
> particular set of assumptions? There are patterns in nature - the 
> questions are: Can we detect them? If so, how do we detect them? How do 
> we know when we have detected them? Can we explain them?
> Dick J
> Richard Jensen, Professor
> Department of Biology
> Saint Mary's College
> Notre Dame, IN 46556
> Tel: 574-284-4674

More information about the Taxacom mailing list