Molecular taxonomy, etc.

[John Grehan]

> Behalf Of Don.Colless at CSIRO.AU

> The current thread on this topic is, like many before it, a perfect
> example of the confusion caused by not distinguishing between
"character"
> and "character state". I hate to point the finger, John Grehan, but
you
> are the prime offender - and your confusion renders a lot of your
> statements nonsense. 

So far I have not seen anything really problematic with respect to this
issue in TAXACOM postings.

For instance, how does one set up a data matrix that
> contains only derived states (NOT characters). Do you leave all the
cells
> blank for what you believe to be primitive states, or enter some dummy
> symbol, or what? 

The usual way - outgroup coded zero for primitive state, ingroup coded 1
derived state for absence/presence. Multistate character treated as
ordered from zero in outgroup to higher numbers in the ingroup.

Indeed, John, have you EVER made a data matrix? 

Yes.

And have
> you never grasped that most cladistic algorithms do not - and cannot -
> consider polarities. Without further input they provide an UNROOTED
tree.

Agreed. Although I see nothing cladistic about that. What is the
critiorn by which you see unrooted trees as being cladistic?

> If a designated outgroup was input, its point of attachment might then
> provide a respectable estimate for the root; 

Right. The difference in my viewpioint appears to be in restricting
character states for the ingroup to derived states before the analysis.
 
> I am, by the way, not available for debate on the currently accepted
> terminology for characters, states, etc.

Looks like you are admitting that yours is not the only view on these
terms. However, if I slip up from your standards of definition you need
not read any further.

John Grehan