Comments on Graff Fara's Paper

I have uploaded a rough draft of my comments on Graff Fara's super-interesting paper for the Arizona Ontology Conference to my website. In the comments I briefly outline an epistemic solution to the problem of contingent identity. But, as Chalmers has just pointed out to me, it is not obvious that sentences involving subjunctive conditionals such as 'Goliath is identical to Lumpl but if I broke off his finger he wouldn't be' can be handled in this way. The puzzle of contingent identity: (1) and (2) seem true.

(1) Goliath is identical to Lumpl.

(2) But Goliath might not have been identical to Lumpl.

But they can't be if identity is absolute, and names are rigid. I suggest that the modality in (2) is (deeply) epistemic.

UPDATE: Formula (4) didn't go through. It should be: (x)(y)(x = y --> <>(x is not identical to y) iff <>(y is not identical to y)).