isolates

[Larry Trask]

Guy Deutscher writes:
 
> Behind all the merriment that Gonzalo Rubio's posting has helped
> create, I think there lies a serious point concerning isolates .
 
> It is unquestionable that the romantic attraction of finding
> linguistic/cultural relatives among exotic peoples is one of the
> main reasons for the obsession of amateurs with isolate
> languages. But wouldn't it be fair to say that this attraction is
> strongly encouraged by the relative rarity of such languages, and so
> by the conception that an isolate language is abnormal, something
> strange which needs to be explained away?  (And what better
> explanation than: it is not an isolate after all...)
 
> But why are isolate languages actually so rare? Why e.g. are there
> only a handful of 'Basque's 'Sumerian's or 'Etruscan's in
> Europe/Western Asia, rather than dozens of unrelated languages? Of
> course the question can be explained ad hoc each time by the
> peculiar historical circumstances of any given language, if we know
> what they are.  But wouldn t it be helpful if linguists discussed in
> general terms what the conditions are for the survival of isolates ,
> and why these conditions are (relatively) so rarely met?
 
These questions have been at least briefly addressed by Johanna
Nichols in some of her papers.  If I remember correctly, she too
expresses some surprise at the comparative rarity of isolates, on the
ground that the typical number of daughters produced by a mother
language is only around 1.6, which ought to give us more isolates than
we see.  (Hope I'm recalling this correctly.)
 
In fact, there are perhaps more isolates around than we sometimes
suppose, especially if we are cautious about accepting some of the
remote proposals involving isolates.  Among living languages, we have
only Basque in Europe, but Asia gives us (at least) Burushaski,
Nihali, Gilyak, Ainu, Korean, Japanese, Yukaghir, and Ket, the last
two being languages which had known relatives that died out only
recently.  (Here I ignore several remote proposals which have not won
general acceptance, as well as the Uralic-Yukaghir hypothesis, which
seems better supported but perhaps not yet generally accepted.)  In
Africa, all of Songhai, Sandawe, and Hadza are assigned to various
families perhaps more for bookkeeping reasons than because of hard
evidence.  There are several apparent isolates in New Guinea,
including Porome, and perhaps quite a few more in North America,
including Yuchi, Kutenai, and Haida, if the Na-Dene hypothesis for
this last is not accepted.  South America is messy, but may have more.
 
Still, this is not an overly impressive total.  But, rather than ask
"What circumstances favor isolates?", I'd prefer to ask "Why are there
so few isolates?"  Or, perhaps I should put it "Why do so many
isolates get blatted?"
 
Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK
 
larryt at cogs.susx.ac.uk