minor quibble: the value of precision

[Larry Trask]

----------------------------Original message----------------------------
Alexis writes:
 
> I obviously agree with Tom Cravens.  Indeed why would
> anyone WANT to say that Basque is "unrelated" to any
> other language rather than "not known to be related"?
> In no other science or branch of mathematics that I
> am familiar with do people go out of their way to
> make this kind of a jump.  The proposition that P equals
> NP is not known to be true or false, and no computer
> scientist feels compelled to say that P does NOT equal
> NP (though I think most of us (a clarification for those
> who do not know: for the last decade I have been
> teaching theoretical etc. computer science)) suspect
> that they are unequal).  It is one thing to note that
> once upon a time most linguists said things like this.
> It is quite another for anyone TODAY, when we know
> better, to say it.
 
To be honest, I do not see this as a problem.  When I say that two
languages are "unrelated", I expect everybody in the business to
understand that I am asserting that there exists no evidence of a
relationship, and not that I am asserting that the languages go back
to independent inventions of human speech in our dim prehistory.  Life
is just easier if we use one word instead of five words.
 
Such usages are in no way a foible of linguists.  When I say that I am
unrelated to you, I expect everyone to understand only that I know of
no link between you and me, and not that I am claiming we belong to
separate species.
 
When we are writing textbooks for beginners, we must be careful to
explain that "unrelated" means "not discoverably related" or "not
known to be related", sure.  But, among ourselves, why can't we just
use the shorter form?
 
Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK
 
larryt at cogs.susx.ac.uk