Re: Genitives NPs as Relative Clauses

--- John Cowan wrote:
> Hebrew?!?! Ok, that's out of the blue!
Not at all, it's just one of the large collection of Semitic/Celtic
symmetries, of which VSO is the most obvious.

Personally, I suspect a very old Sprachbund effect either in Spain
or just possibly Southwest Britain.
--- end of quote ---

I'm suspicious of this kind of explanation. After all, having inflected
prepositions is a common property of verb-initial languages (or languages that
pattern as if they were verb-initial, as in the case of Hebrew). Malagasy has
inflected prepositions:

  amiko  "with me"
  aminao  "with you"
  aminy  "with him/her/them"

Malagasy shares other properties with Celtic and Hebrew, such as determiners and the
use of a "construct state" type construction for marking possession, among
others.

I doubt if anybody would want to suggest that Malagasy (which originated in
Southern Borneo, after all) is part of the Hebrew-Celtic Sprachbund. And if it
isn't, then we must assume that this cluster of structural properties can arise
independently in verb-initial languages in different parts of the world. But if
you allow that assumption, then positing a Sprachbund effect to explain the
Semitic-Celtic connection doesn't buy you much--especially since there doesn't
seem to be any evidence of shared vocabulary between Semitic and Celtic to
suggest prolonged contact. (This isn't to say that there wasn't *any* early
contact between Semitic- and Celtic-speaking peoples. Given their geographic
proximity, it's quite likely that there was.)

To me a structural explanation seems much more likely than a historical one. That
seemingly disparate morpho-syntactic properties often cluster around a
particular word order type has been known since Greenberg's work in the early
60s. If you compare, say, Turkish and Quechua (both SOV languages), you find
that they share an extraordinarily large number of features: Postpositions,
case-marking suffixes on nouns (except in the case of the subject, which
triggers agreement on the verb), morphological causatives, pro-drop, non-finite
verb forms used in embedded clauses, prenominal adjectives, prenominal
possessors (which trigger agreement on the possessed noun), relatively free
order of subject and object, topic- and focus-fronting, tense/aspect and number
marked by suffixes, etc. etc. etc.. It would be strange to argue that these
parallels are due to some sort of historical contact between speakers of
(proto-)Turkish and (proto-)Quechua. For some as yet unknown reason, our men!
tal grammars seem to be set up to statistically favour certain constructions over
others, depending on the basic word order of our language.

Matt.

Matthew Pearson
Department of Linguistics
Reed College
3203 SE Woodstock Blvd
Portland, OR 97202
503 771 1112 x 7618

Reply