# 3TA (correction)

P.Hovenkamp Hovenkamp at NHN.LEIDENUNIV.NL
Mon Apr 14 13:48:19 CDT 2003

At 12:13 PM 4/14/03 +0100, David Williams wrote:

>Peter Hovercamp suggests that the A(BCD) answer requires justification, and
>asks "why do you prefer the outcome (BCD) over the outcome (ABCD)." Peter
>neglects to add that ABCD is not a solution at all; it represents no
>solution. But, of course, standard parsimony analysis of AB(CD) + AC(BD)
>finds two solutions, identical to the original characters: AB(CD) and
>AC(BD), which is also not an answer. One might imagine some poor taxonomist
>who has diligently studied his organisms and finds one thousand AB(CD)
>characters and one thousand AC(BD) characters, and still nothing may be
>said of those data.
OK, this is a new argument. (BCD) is not preferred because it is the
solution directly perceived by untrained students, but simply because it is
a solution, and (ABCD) is not. I might ask "why is it then to be preferred
over (ABC), which is also a solution?", but I won't.
Instead, I'll draw attention to the possibility that if the solution to a
problem is not closer after making a number of observations (1 or 1000, I
don't care), the observations may have been irrelevant to the problem. If
you're interested in the position of A, and all you can come up with is
contradictory information about the position of B, C, and D, you indeed
have a problem. David seems to deny the existence of irrelevant
observations, insisting instead that if we have collected data
"diligently", we may be assured that they will be relevant. I commiserate
with the poor taxonomist, but practice shows this to be not true generally.

Peter Hovenkamp