It is generally agreed that some anaphoric pronouns with (what appear to be) quantifier antecedents occur outside the syntactic scope (i.e., the c-command domain) of their antecedents. First, there is “donkey anaphora,” of both the conditional and relative clause varieties:

1. (1) If Sarah owns a donkey, she beats it.

2. (2) Every woman who owns a donkey beats it.

Without going through the details, let me just assert that there is good reason to think that the pronouns in (1) and (2) do not occur in the syntactic scope of the quantifier’ a donkey’ . A second sort of case in which a pronoun with a quantified antecedent occurs outside the syntactic scope of its quantifier antecedent is one in which the pronoun and its antecedent occur in different sentences.

A second sort of case in which a pronoun with a quantified antecedent occurs outside the syntactic scope of its quantifier antecedent is one in which the pronoun and its antecedent occur in different sentences. Examples of such “discourse anaphora,” from the very simple to the slightly complex, include:

1. (3) A man is following Sarah. He is from the Internal Revenue Service.

2. (4) A man is following Sarah. Melanie believes he is from the Internal Revenue Service.

3. (5) It is possible that several students flunked at most five exams. Melanie believes they didn’t study for them.

4. (6) Suzi ought to apologize to most of Ann’s dinner guests. It is certain that she insulted them. But it is unclear whether they noticed.