Re: Unknown Language Identifier!

John Cowan wrote:

> J Matthew Pearson wrote:
>
> > Wearing my linguist hat, I have to say that I'm extremely dubious that
> > this would be useful tool in determining the relatedness of languages.
>
> In all fairness, it's not meant for that.  It's meant when you have
> a sample of an unknown (to you) language, and want to know where to
> go with it.  That being so, strong superficial resemblance is usually
> what you want.
The author of the website says it's based on a computer program for
calculating the relatedness of languages (echoes of lexicostatistics).  I
was commenting on its usefulness in that capacity.  In particular, I was
criticising the passage where the author explains the scoring system, in
which he says that if the sample text you enter scores in the medium range
against one of the comparison languages, it will likely be a member of the
same family as that comparison language (the example he gives is that if you
enter a text in Shona, a Bantu language, it will score highest against the
Bantu comparison languages, namely Swahili and Sotho).  I don't see how he
can seriously claim that, given that orthographic similarity is a bad test
for genetic relatedness (Malagasy being a case in point).  Of course, I
would need to know more about what exactly the program was doing when it
compared texts.  Maybe he compensates for this somehow...

Matt.