Linguistics and cladistics

[Andrew K. Rindsberg]

Ken, thank you for resurrecting some linguistic memories. I have kept up
with linguistics only in a desultory fashion, but a few quick memories may
be useful as a loose analogy to modern cladistic thought. Barry, thank you
for keeping us on track. Allow me to "think out loud"...

Linguists use the terms "family" and "phyla" for groups of related
languages, but do not have anything comparable to the formal Linnaean
hierarchy of taxonomic ranks.

They tend to name all the nodes (Balto-Slavic, Thraco-Illyrian,
Celto-Italo-Ligurian...) and they use them all, too. However, most of the
nodal names are little used except in specialized papers.

If someone feels the need to rename a node, he does so. Usually no one pays
much attention; there is no rule of priority.

The number of extant languages is only a few thousand, even according to
splitters.

They do not ignore paleographic data, their equivalent of paleontologic
data. Instead, they seek it out as particularly valuable information. There
is no equivalent to the crown-group; a linguist would be astonished if
someone told him to ignore Gothic or Hittite just because they had no modern
descendants.

The raw material of linguistic reconstruction consists mainly of words built
of sounds; the raw material for cladistic analysis has a broader range, but
includes molecular and morphologic data. Both work under the assumption that
analysis of a large database will reveal historical patterns even from
incomplete samples.

Linguists commonly deal with very incomplete data. If only a few words are
known from an ancient language, they are not discarded as useless. It is
hard not to see this as an advantage of the linguistic approach, compared to
which the standard cladistic methods seem brittle.

+++

In sum, linguists have been operating with an intuitively based cladistic
system for decades, albeit without a hard statistical basis. The differences
seem to be based on the more disorderly nature of linguistic data,
especially compared to molecular data; and to the much larger number of
species than languages. I don't think that taxonomy should emulate
linguistics, rather the reverse; but the parallels are intriguing -- at
least to one with a background in linguistics!

I am particularly impressed by the fact that linguists make relatively full
use of relatively chaotic databases. I wonder if they have developed any
software that might be applicable to biological taxonomy?

Andrew K. Rindsberg
Geological Survey of Alabama