I meant to not do late-night philosophy here, but I might make an exception this once. I just flipped through the Karen Bennett paper on coincidence that I joked about below, basically browsing until I got to the bits that were about my preferred view, and I was sort of struck by this argument. The background is that we’re discussing the views of a theorist who believes, roughly, every transworld fusion is an object, and our ordinary referring terms are indeterminate between many such fusions, but such indeterminacy is normally harmless because we can supervaluate away. (This summary is a cross between just how Karen describes the position and just how I’d describe my position, but it gets the essentials right I think.) The argument is going to be that for such a theorist, all de re modal claims are in a sense analytic. I probably believe this too, but if I didn’t I wouldn’t be convinced by Karen’s argument for it.

A claim like ‘Goliath would not survive being squashed’ is true because 1) ‘Goliath’ picks out the statue there, and 2) all admissible precisifications of ‘statue’ refer only to things that cannot survive squashings. Both of these are semantic claims.

Maybe, but let’s try a slighly different example.

A claim like ‘Bill Clinton essentially has DNA with a double-helix structure’ is true because 1) ‘Bill Clinton’ picks out the human there (the male one) and 2) all admissible precisifications of ‘human’ refer only to things that essentially have DNA with a double-helix structure. Both of these are semantic claims.

I think in this case, (2) is arguably the combination of a semantic claim and a non-semantic claim. The semantic claim is that, as Sidelle might put it, all admissible precisifications of ‘human’ refer only to things that are like these ones (picking out some paradigm-case humans) in their most fundamental features. The non-semantic claim is that having those features entails having DNA with a double-helix structure. Put another way, the semantic facts do pick out the property of having DNA with a characteristic shape as being one of the characteristic properties of humans. But they do not pick it out as such, rather only under the description being a fundamental feature of these here creatures. And the only analytic facts are those that are semantic facts as such.