Consciousness, Backward Coreference, and Logophoricity

Contrasts like those between the (a) and (b) sentences of (1)-(7) suggest that Self-anaphors are subject to a coreference requirement that operates in addition to binding theory.¹

(1)

1. a. {That book about herself_i}_A {hit Sara_i}_B.

2. b. *{That book about itself_i}_A {hit the Hope diamond_i}_B.

(2)

1. a. {That picture of herself_i}_A {pushed Sara_i off of the log}_B.

2. b. *{That picture of itself_i}_A {pushed the Hope diamond_i off of the display case}_B.

(3)

1. a. {That story about herself_i}_A {caused Sara_i to become famous}_B.

2. b. *{That story about itself_i}_A {caused War and Peace_i to become famous}_B.

(4)

1. a. {Lightning striking herself_i}_A {caused Sara_i to get hot}_B.

2. b. *{Lightning striking itself_i}_A {caused the cari to get hot}_B.

(5)

1. a. {That picture of himself_i}_A {caused folks to think they should avoid Bill_i}_B.

2. b. *{That picture of itself_i}_A {caused folks to think they should avoid the cave_i}_B.

(6)

1. a. I put {that picture of herself_i}_A {next to Sara_i}_B.

2. b. *I put {that picture of itself_i}_A {next to the Hope diamondi}_B.

(7)

1. a. They saw {that picture of herself_i}_A {next to Sara_i}_B.

2. b. *They saw {that picture of itself_i}_A {next to the rock_i}_B.

It seems clear that these contrasts cannot be due to binding principles, since in no case does one coreferring element c-command the other. Moreover, on standard assumptions the (a) and (b) sentences in each of (1)-(7) are structurally identical to each other, so that any binding-theoretic treatment would yield identical judgments, not contrasting ones, for the members of each pair. This squib proposes a principle to account for the noted contrasts, comments on ways this principle differs from principles of binding, and discusses implications it may have for theories of logophoricity.

1 Consciousness and Coreference

I believe the contrasts in (1)-(7) are due to a coreference requirement on Self-anaphors, whose first, overly broad formulation I give as Principle E in (8).

(8) Principle E (nonfinal)

A Self-anaphor must corefer with an expression whose referent typically possesses consciousness.²

Principle E correctly accounts for the contrasts in (1)-(7).On the one hand, in each of the (a) examples, the referent typically possesses consciousness, so that Principle E is satisfied, and the result is acceptable. On the other hand, in each of the (b) examples, the referent lacks consciousness, so that Principle E is violated, and the result is unacceptable.

2 Consciousness, Coreference, and Binding

Of course, Principle E as formulated in (8) cannot possibly hold in all circumstances, since there exist sentences such as those in (9)-(10), in which Self-anaphors happily corefer with expressions regardless of whether their referents typically possess consciousness. In its current formulation, Principle E would wrongly predict the same contrast for the sentences in (9)-(10) as it does for those in (1)-(7), deeming (9a) and (10a) acceptable, but (9b) and (10b) not so.

(9)

1. a. Joshua_i destroyed himself_i.

2. b. The machine_i destroyed itself_i.

(10)

1. a. Joshua_i took a picture of himself_i.

2. b. The automatic camera_i took a picture of itself_i.

Therefore, I believe Principle E would be more accurately formulated as in (11).

(11) Principle E (nonfinal)

A free Self-anaphor must corefer with an expression whose referent typically possesses consciousness.

Now Principle E correctly accounts for the contrasts in (1)-(7) without wrongly extending those contrasts to (9)-(10). On the one hand, the Self-anaphors in (1)-(7) are free, so Principle E applies, making these sentences acceptable only when the referent typically possesses consciousness. On the other hand, the Self-anaphors in (9)-(10) are bound, so here Principle E does not apply, leaving these sentences acceptable regardless of whether or not the referent possesses consciousness.

3 The Backward Coreference Domain

Next, it seems clear that Principle...