[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Cladism is not an inkblot



HP Pharris made a request:
One of my profs has been using a technique known as "clique analysis" to
determine the phylogeny of Chinese dialects.  In this method, as he's
explained it, the favored tree is the one congruent with the distributions
of
the greatest number of characters.  Is anyone on the list familiar with this
method, and can anyone give me a critique?

Have to admit, all I remembered about cliques is that they're a form of
cluster analysis.  Here's a reference with a lot more detail:
http://www.analytictech.com/borgatti/analyzing_clique_overlap.htm

This doesn't sound too promising:
However, once these cohesive subgraphs have been identified researchers are
often left with a long list of overlapping subgroups and have no means of
assessing the structure or importance of these groups.

Apparently, what the researcher already knows is important.  In your example
of the Chinese characters, how many other ways would there be for a dialect
to be central other than historical origination?  Do characters have
linkages other than origination?
Sounds to me like cladistics, an expression of expertise with fulsome
documentation rather than a replacement for expertise...
Hope this helps a bit.