Theiling Online    Sitemap    Conlang Mailing List HQ   

Re: Linguistic Universals?

From:Ed Heil <edheil@...>
Date:Friday, November 12, 1999, 1:33
Charles wrote:

> "Grandsire, C.A." wrote: > > > OV languages are more likely > > to use postpositions (here again, Arabic and Japanese are perfect > > examples). SVO languages can have both, and exceptions exist
(Latin is
> > mostly SOV -at least for the unmarked order- but uses
prepositions).
> > Latin is the one that puzzles me most ... > Why have both morphological cases (inflections) > and also have prepositions? Were the prepositions > "originally" (there was no true origin I suppose) > adverbs? Apparently they also glued them onto verbs > to make productive series like in-* and pro-* etc. > Odd, because this pattern does not extend back > into PIE. I guess some pre-Latin conlangers > did it around 1000 BC?
I think that PIE is supposed to have been OV with the expected postpositions, which became the case endings; it shifted towards VO in most of its dialects, resulting in a mixed system. You're likely right about the originally "adverbial" nature of the prepositions; or else one could cop out and call them "particles," which i think is Linguistese for "small word that's hard to classify." :) The historical linguist Winifred Lehman thinks that the mixed VO and OV patterns of languages like Latin and Old English mean that they're "in transition" from one pure pattern to the other. However, some typologists, like William Croft, would take a more complex view: particular syntactic patterns can individually be dominant or non-dominant (for example, noun-adjective seems to be dominant, and demonstrative-noun is also dominant), and pairs of patterns may be harmonic or disharmonic (For example, Noun Adjective is harmonic with Noun Determiner, and disharmonic with Determiner Noun). They would say that languages tend to maximize the dominance and the harmony of their syntactic patterns, but these two goals are often incompatible: for example, DemN and NA are both dominant, but they are disharmonic with each other. DemN AN are harmonic, as are NDem NA, but each contains a non-dominant element. What this predicts is that you will not find a non-dominant element unsupported by harmonic elements: within the restricted scope of the discussion above, you would not expect to find Adjective-Noun-Demonstrative, as it is disharmonic *AND* non-dominant. So under this system, Latin's prepositions would be harmonic with it's noun-adjective order (which is dominant), and verb-object sentences are not at all unknown in classical Latin, so that would be another element of harmony (note that there's no reason a language can't allow two different possible orders for a syntactic combination!). So it would not be *necessary* to explain it as a "transitional state" between "pure OV" and "pure VO." There are lots of stable configurations in between the idealized extremes. (BTW, a handy shortcut for dominance: as a rule of thumb, the dominant order puts the shorter element before the longer element: demonstrative-noun, noun-adjective, subject-verb, and so on.) ------------------------------------------------- edheil@postmark.net -------------------