Uses of the genitive in an English corpus
- On Tue, Jan 8, 2013 at 12:29 PM, Jim Henry <jimhenry1973@...> wrote:
> I shouldn't be surprised if a corpus analysis showed the ownershipBehold! I did a quick and dirty corpus analysis of the Project
> sense to be less common than some of the others, or at least less
> common than all the others put together.
Gutenberg etexts from 1990 to 2001. I searched for apostrophe-ess
(using Emacs' grep-find feature) and then randomly scrolled down a few
thousand lines at a time through the results, mostly though not always
getting no more than one instance from a given book. Shakespeare, who
is overrepresented in the corpus, appears more than once, but I don't
think other authors do (in a lot of cases I didn't bother to check the
author of the book). The texts included a mix of novels, plays,
poetry, and nonfiction ranging from the sixteenth to the early
twentieth centuries. Some search results were were borderline and
hard to classify, even after going to look at the context.
contraction of "is" 24
contraction of "has" 3
agentive genitive 13
kinship (etc) 15
Some of the uses I classified as "kinship (etc.)" refer to relation by
marriage, or to other personal relationships, as between empoyer and
employee, leader and follower, or vice versa. Some of the "ownership"
uses could have referred to rental, borrowing or other temporary
possession; I would have had to read a lot more context to be sure
Ownership is not only somewhat less common in this sample than some of
the other genitive senses, like part-whole and kinship, but far less
common than the other genitive senses put together. The former effect
is small enough that it might disappear or reverse in a different
sample of a different corpus, but the latter effect is big enough I
suspect it will hold with other English corpora.
Ideally one would check for genitive pronouns as well, but I'm not
sure how to combine those samples with the apostophe-ess sample to get
a balanced sample.