Object-initial languages

[Östen Dahl]

Joining Spike in being nervous at "the suggestion that the subordinate
clause word order be considered a default for typing an entire
language", here are some considerations that have worried me for quite
some time:

In all Germanic languages that I know of, a simple sentence such as "I
love you" has SVO word order. Still, the majority opinion seems to be
that whereas English and Scandinavian do have basic SVO word order, the
continental Germanic languages are really SOV, the order in subordinate
clauses being decisive. Similarly, in Standard Mainland Scandinavian,
negation and various other elements follow the finite verb in main
clauses and precede it in subordinate clauses. So syntacticians tend to
declare the latter order basic. But subordinate clauses are both less
frequent and acquired much later than simple main clauses. One wonders
if there isn't something fundamentally wrong with linguists' assumptions
about "basicness" and "markedness". Clearly, subordinate clauses are
more rigid with respect to word order, and this has something to do with
"degree of grammaticalization", which in turn has to do with historical
processes that fix the word order in constructions. But the result of
such processes may not necessarily be "basic" in the language in any
sense.

Östen Dahl