(some) linguists and the PAP

[Arnold Zwicky]

in their new book The Language Organ, anderson and lightfoot
list what they see as some everyday but remarkable facts about
english.  their first example, in chapter 2 (p. 19), has to do
with pronoun-antecedent linkages:

  Pronouns like _she_, _her_, _he_, _him_, _his_ sometimes may
  refer back to a noun previously mentioned in a sentence (2.1a-c).
  However, one can only understand (2.1d) as referring to two men,
  Jay and somebody else; here the pronoun may not refer to Jay,
  unlike (2.1a-c).

  (2.1)  a.  Jay hurt his nose.
         b.  Jay's brother hurt him.
         c.  Jay said he hurt Ray.
         d.  Jay hurt him.

notice (2.1b), a straightforward violation of the Possessive
Antecedent Proscription.  (yes, the pronoun is potentially ambiguous,
referring either to jay or to some other man, but that's true of
(2.1a) and (2.1c) as well, so there's no special problem with
possessive NPs here.)

it never even occurred to these particular linguists that there might
be something odd about (2.1b).  for them, it's just a routinely
grammatical sentence, usable as a textbook illustration.

my point is not that linguists (in general, shudder, or these
particular ones) should serve as models of grace and style in writing
- i'd edit a&l's sentence above in several ways, and sticklers would
object to where they place _only_ in it - but that as observers of
english (their native language) a&l see no constraint on possessives
as antecedents as pronouns.

arnold (zwicky at csli.stanford.edu), flouting the PAP himself