|
<Your
question is a little bit like asking "will
a calculator produce a correct sum without the intervention of the analyst"? The calculator can't press its own buttons. If you enter in the wrong numbers, it isn't the calculator's fault if it spits out the wrong answer.> Given the correct numbers, the calculator will produce the correct answer. There is only one possible mechanism for addition. However, evolution works in a number of different ways, and my question is whether the evolutionary model built into the program algorithm does not in itself put a limit on the number of possible solutions compared to the larger number of outcomes available in nature. If I and someone else use the program, we are likely to come up with similar solutions, not because the answers we are tending to find are more true necessarily, but because the algorithm is guiding us in a particular direction. The characters I put in will produce a variation in the outcome, but this is a variation within the limited range of outcomes the program has available. <I think cladistic analysis is a basically a way of ordering information that helps us to see particular sorts of patterns that we are looking for. I think you are (somewhat disingenuously) trying to make cladistics sound invalid by creating the impression that cladistic analysis is has to omnipotently be able to program its own data without the subjective limited input of a human analysts and produce perfect results in order to tell us anything useful. This is balony. Its like saying that radiometric dating is useless because it has a margin of error, or (perhaps a better analogy) that all the eyewitness accounts of a crime should be completely ignored because we know that eyewitnesses can make errors. There is still interesting infromation to be gleaned out of an analysis, regardless of whether or not it gives perfectly flawless results.> Agreed. If the purpose of the analysis is to provide 'interesting information' to the analyst, who then uses this information along with other information to draw a conclusion, this is an entirely valid use. I can imagine the possibility that a conclusion would be entirely impossible to coax out of a cladistic algorithm, but the analysis would still have been helpful in creating a hypothesis the analyst rejects. I'm not being disingenuous, honest, I'm trying to distinguish between a tool belonging to the analyst and an objective means of finding 'truth' (in this case defined as a useful hypothesis). <You don't get a Borg implant when you buy PAUP that forces you to think only what the analysis gives you.> Definitely agreed. And if you were critiquing an analysis documented by a cladistic printout, you would be able to argue against it without producing a different printout yourself. < Again, cladistic analysis isn't "defeated" just because PAUP isn't an omnipotant and can't identify every case of reversal and convergence that could skew the analysis Tweaking the data matrix is a way of trying to get around the limitations of the data that the analyst programs in.> Could I paraphrase this to say that because the cladistic diagram can be used to explicate the results of your analysis, it can be adapted to show your conclusion in a manner immediately apprehended by the people you are explaining it to? |