pseudo-generalizations

[Matthew Dryer]

Private communications from a number of people during the past week have
made me realize that there was one thing in my message a week ago about the
relative value of crosslinguistic generalizations and explanations that may
have been misunderstood.  Some people took my statement that I value
crosslinguistic generalizations more than I value explanations to imply
that I am more interested in crosslinguistic generalizations than I am in
explanations.  But this is definitely not the case.  I am much more
interested in questions of explanation than I am in the generalizations.
The amount of interest I have in the questions is very different from the
value I place on hypothesized answers.

Let me make the point more concrete by illustrating it with some of my own
work.  I published a paper in 1992 called "The Greenbergian word order
correlations".  There were two goals of the paper.  One was to document
which pairs of elements correlate in order with the order of object and
verb; this involved documenting crosslinguistic generalizations, based on
data from over 600 languages.  The other was to offer an explanation for
why the correlations exist.  The details are not important here.  When I
had given talks based on this paper prior to its being published, I had
used as a title "The Branching Direction Theory of word order
correlations", after the explanation I gave for the correlations.  But I
eventually chose the title "The Greenbergian word order correlations"
because I considered the detailed documentation of the correlations
themselves (and, more importantly, what does NOT correlate) a more valuable
contribution to the field than the explanation I offered.  Lots of people
don't believe the idea that the correlations are due, even partly, to the
nature of sentence processing, and that's fine, precisely because it's so
hard in general to give convincing explanations.  On the other hand, if
someone were to question the empirical results I presented, for example, if
someone were to insist that OV languages tend to use Adjective-Noun order
more often than VO languages do, despite the evidence I presented, I would
consider that rather odd.  I think we can be much more confident of the
crosslinguistic patterns I document than we can be of any explanations,
mine or others', and that's why I consider the evidence for the
crosslinguistic patterns more valuable.  But I find the question of WHY
these patterns exist more interesting than the question of what these
patterns are.

Matthew Dryer