Geography Reference
In-Depth Information
to component coding, applied inBPA, inthree-item codingeachnodeiscon-
sidered a relationship between branches, where some branches are related
more closely than other branches. Each separate relationship is expressed
minimally as a three-item statement (Williams and Humphries 2003). All
area statements are input in a data matrix and analyzed with a parsimony or
The three-item statement analysis has received some criticism (Farris
2000; Farris and Kluge 1998; Harvey 1992; Kluge 1993), and a few authors
have defended it (de Pinna 1996; Marques 2005; Scotland 2000; Siebert
and Williams 1998). The main reasons for controversy have to do with the
relationship of this approach to cladistics, observations and homology, data
transformation, parsimony, and synapomorphies (Marques 2005). Discus-
sions are of a highly technical nature, but I concur with de Pinna (1996:11):
Regardless of the fate of three-item analysis as a method of char-
acter analysis, I predict there will be an increase awareness of
the problem of ancestor-descendant relationships among character
states. Even if three-item analysis turns out not to be an appropriate
solution, I think it will have its place in history as the first concrete
attempt at terminating ancestor-descendant thinking at the level of
character coding.
