Three-taxon analysis

[SKÁLA Zdenek]

-----Original Message-----
From: P.Hovenkamp [mailto:Hovenkamp at NHN.LEIDENUNIV.NL]

... Apart from that, my point stands: if any of these proposed 
mechanisms results in a particular pattern of sister-group relationships, 
we may reasonably hope that we can find that pattern by studying the 
characters of the taxa. Of course, many of these mechanisms will result in 
very similar-looking patterns of relationship, and that is why there still 
may be controversy and uncertainty.
============

Not the only problem, IMO. Think about evolutionary scenarios that would not yield any recoverable phylogenetic signal. One such example can be heavily parallel evolution (some parallelisms are known; what we seem to know about its 'real' amount rely on the hypothesis that parallelisms are rather rare). Another such model is quick character evolution; provided the character states change frequently enough, any information available can be noise, etc.
Indeed, you will get some "cladogram" anyway but it will not include information about phylogeny pattern. More importantly, we seem to be unable to distinguish if the phylogenetic analysis is or is not informative just due to missing knowledge about the true evolutionary scenario... 

Best!
Zdenek Skala