Linguistics, Please
Dr Elizabeth Pearce writes to The Listener.
Dear Editor,
Marc Wilson (Psychology February 8) says that Russell Gray has developed an algorithm that "is used to test theories about language evolution". No, it is the other way around. The theories that linguists have developed about language evolution provide us with the means to test the validity of Gray¹s algorithm.
The broad schema of the Pacific migrations that is shown in the map accompanying Wilson¹s article has long been known to linguists as a central hypothesis reflecting the linguistic family tree reconstructions for Austronesian languages. However, the linguists' trees themselves do not come with dates and, for this, we need to look principally at evidence coming from work by archaeologists.
If it is the case that Russell's algorithm can be shown to mirror the findings of scientific work from other fields, it would indeed be interesting if it means that we could use computer power to take over, for example, the massive task of linguistic reconstruction.
In the meantime, where there are unknowns for linguists (and there are many of these), whatever results come out of Russell's algorithm can provide hypotheses to be further probed by linguists.
Elizabeth Pearce
Linguistics Programme
Victoria University of Wellington