central embedding
Geoffrey Sampson (geoffs@cogs.susx.ac.uk)
Wed, 13 Mar 96 12:46 GMT
It seems to me that Martin Wynne's claim that sentences of the precise
type "the rat the cat the dog tossed bit chased" don't occur is very much
more limited than the usual claim that central embedding of the same
grammatical construction to more than two levels don't occur. I have
been quoting to various individual participants in this discussion, and
will now quote to the whole list since the discussion seems to be drawing
more people in by the day, a recent article of my own ("From central
embedding to corpus linguistics", in Jenny Thomas & M.H. Short, eds.,
_Using Corpora for Language Research_, Longman, Jan. 1996) which gives lots
of evidence that multiple central embedding _does_ occur in real-life
usage, fairly frequently. None of my examples are quite like the example
quoted above from Martin Wynne, and that seems not surprising. Austere
clauses like "the dog tossed the cat" are characteristic of punchy,
informal genres of language; heavy use of subordinate clauses, in any
grammatical position, is characteristic of intellectualized, formal genres
-- so Martin Wynne's example is strange because it exemplifies contradictory
genres. But there are plenty of examples of relative clause within relative
clause within main clause, though the clauses are mostly more complicated.
There is at least one in the million-word LOB Corpus.
Geoffrey Sampson
University of Sussex