--- In
cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@...> wrote:
> Actually what I criticize is this:
>
> Typical approach: Researcher NN finds a root in PGerm. that is
similar
> to a root in PFinn. (or vice versa). Why are they similar? If it's a
> loan, there are three logical possibilities
>
> 1) PGerm. -> PFinn.
>
> 2) PGerm. <- PFinn.
>
> 3) Some unknown language X -> PGerm., PFinn.
>
> Invariably, NN chooses 1).
>
> Why?
>
>
> Torsten
>
OK now we are talking business!
In so many cases 2) must be excluded because for phonological reasons
i.e. it would not work for the words. Substitution rules are not
reversible of course. Finnic drops clusters: that's not reversible;
Finnic drops voicing: that's not reversible. Finnic makes lots of
fricatives plosives: that may be reversible given the case is right,
and pre-grim's law, but usually it is not. As there are usually more
than one phoneme in a word :) the likelyhood that there are
irreversible rules will be high.
The article you posted of Burrows is perfectly right. You should
always be aware of the possibilty of a 'reverse' borrowing, and a
word with narrow distribution is alsways more suspect. But mind then
to be clear about the substitition rules for Uralic => IE. The
material is so scarce that parallell cases for any particular
supposed substitution rule may be impossible to find.
There has for long been a discussion on the direction of borrowings
between Baltic and Finnic. Here we at least have the case Gmc. *flauj-
=> Finnic laiva 'ship' (with metatesis) => Baltic
There is the complicated case for the word 'kaima' on my homepage.
Here I admit I have added some thought of my own, which are not
direct quotes from the litterature. And then there is Finno-Samic
salo 'island (> backwoods)', which does not have a further etymology
in Baltic either, so which way did it go?
I posted one link for a possible borrowing of the word knife (Finnish
veitsi) PU > PI-A. The article is in Finnish though.
A clear case is the word 'moth' where the substitution could have
worked Saami > Gmc but not the other way. Phonological criteria
leaves no mercy. But you are right, it is controversal because of the
lack of parallell borrowings. That's why I threw it out to you. The
fact that the language were neighbours at that time is not
controversial though, because borrowings in the other direction are
plenty.
As for why not choose 3) (separate borrowings from substrate): Given
a situation where the explanation 1) PGerm. -> PFinn. works well
without irregularities it is impossible to construct an etymology
under 3), which would benefit from 'economy of presumptions'. It will
always be a more complicated solution.
The scope for explanations under 3) is therefore limited, unless
A) it would be possible to come up with extradisciplinary reasons,
e.g. making the presumption of contact impossible.
B) A completely different situation would occur if it would be
possible to find enough cases, for which neither 1) or 2) works, but
the similarity is striking and distribution on both sides narrow.
Strangely enough not very many has occured to my knowledge. The so
calle "substrate words" are different on both sides. The word for
salo 'island' fullfills the criteria of narrow distribution on both
sides, but here unfortunatey substitution rules are reversible so 1)
and 2) works. Can you think of any better examples?
Jouppe