Please read the following passage from a linguistics article and answer the questions that follow.

Linguistic theories and psycholinguistic models of anaphora. Many linguistic theories of anaphora posit that different anaphoric forms are subject to different syntactic constraints. Consider (1)-(5):

(1) John likes himself.
(2) *John likes himself.
(3) *John likes him.
(4) John likes him.
(5) *John likes John.

In (1)-(2), the sentence is grammatical when the reflexive, himself, is co-referential (i.e. co-indexed, indicated by matching subscripts) with an antecedent in the same clause, John, and ungrammatical when it is not; yet, the reverse is true when the reflexive is replaced with a pronoun, as in (3)-(4). In this case the sentence is ungrammatical when the pronoun, him, is co- referential with an antecedent in the same clause and grammatical when it is not. The proper name in (5) patterns with the pronoun and is only grammatical under a reading where each proper name, John, represents a different person. Thus, sentences (1)-(5) illustrate that reflexives are in complementary distribution with pronouns and proper names.

The most famous linguistic theory of anaphora, Chomsky's (1981) Government and Binding Theory, was created to explain these facts (among others). Under this theory, a reflexive (or reciprocal) must be bound in its governing category, a pronoun must be free in its governing category, and a referential expression (R-expression, e.g. a proper name or an NP) must be free everywhere. To oversimplify, these constraints basically translate into a requirement that the antecedent of a reflexive must occur in a very local syntactic domain (often the same clause), while the antecedent of a pronoun or a noun phrase may not occur in this domain. Predicate-Based Binding Theories (e.g. Reinhart & Reuland, 1993 ) explain the same set of facts in a slightly different way, focusing on the relationship between a predicate and its arguments. While pronouns and referential expressions must not be dominated by a co-argument in the same predicate, reflexives may. If a reflexive is dominated by a co-argument, then its reference is taken from that argument; however, if it is not dominated by a co-argument, then its reference is determined pragmatically. This essentially reduces to a distinction between reflexives in argument positions (positions required by the verb that receive theta-roles) and those in non-argument positions (i.e. logophors).

The linguistic binding theories of anaphora discussed so far attempt to explain why, for a certain syntactic position in a given construction, one anaphoric form is grammatical and another is not; in a different vein, linguistic theories of accessibility (Ariel (1990) and Ariel (1991) ; Givon, 1983 ; Murphy, 1985 ) attempt to explain patterns of anaphoric usage in situations where multiple anaphoric forms are grammatical. Accessibility theories propose that under optimal conditions, speakers produce anaphors that help the hearer identify the appropriate antecedent. Two factors interact to determine the anaphoric form used: (a) characteristics of the anaphor that determine how effective it is in uniquely identifying its antecedent and (b) characteristics of the antecedent that determine its accessibility to the hearer (pp. 236-237).

Passage from Callahan, S. (2008). Processing anaphoric constructions: Insights from electrophysiological studies. Journal of Neurolinguistics, 21 (3), 231-266.