Superlatives

[dotdv]

If the head is a concrete entity I guess it should be:
[This hotel]_A is_S [the_E friendliest_E place_C [I've ever stayed at]_E]_A
also in a toplevel scene: "Friendliest_S [place_C [I've ever stayed at]_E]_A

But what if the head is a P/S:
One_Q of the best_D lodging_A experiences_P [I've ever had]_?
Best_D photographer_P+A [I've ever worked with] ?
(there is the option of considering marking these two heads C when they are inside an A scene, but that's not possible if they are inside an H, right?)

[omriabnd]

This is again the problem of relative clauses over P/S. How about the two Hs solution for this case as well? [One_Q of the best_D lodging_A experiences_P]_H [I_A 've_F ever_D had_F (experiences)_P]_H Note: I'm aware the foundational layer doesn't handle superlatives well. This will be dealt with in a refinement layer.

…

On Mon, Nov 19, 2018 at 10:34 AM dotdv ***@***.***> wrote: If the head is a concrete entity I guess it should be: [This hotel]_A is_S [the_E friendliest_E place_C [I've ever stayed at]_E]_A also in a toplevel scene: "Friendliest_S [place_C [I've ever stayed at]_E]_A But what if the head is a P/S: One_Q of the best_D lodging_A experiences_P [I've ever had] *? Best_D photographer*{P+A} [I've ever worked with]_? (there is the option of considering marking these two heads C when they are inside an A scene, but that's not possible if they are inside an H, right?) — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#48>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AIG865h4dU6a-K6IGxp60pEA6CZduPv1ks5uwm0WgaJpZM4YooEF> .

[nschneid]

Why not

[[One_Q of the best_D lodging_A experiences_P]_C [I_A 've_F ever_D had_F (experiences)_P]_E]_H

?

[omriabnd]

We could, but that would mean that we allow Hs that are C+E, rather than Scenes. Isn't that a high price to pay for a cxn we don't handle well anyway.

…

On Mon, Nov 19, 2018 at 2:01 PM Nathan Schneider ***@***.***> wrote: Why not [[One_Q of the best_D lodging_A experiences_P]_C [I_A 've_F ever_D had_F (experiences)_P]_E]_H ? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#48 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AIG86_cUI2GW9wdpdqMHrk2K_c2W_LUcks5uwp2sgaJpZM4YooEF> .

[nschneid]

It's not specific to superlatives, though: e.g. "an experience that broke my heart".

I think the issue is that elaboration is an information structural notion, and it can happen between scenes just as it can happen between non-scenes, or a scene and a non-scene.

[omriabnd]

I generally agree (though what about the issue of restrictiveness? where an E serves to restrict the denotation, it's not equivalent to a flat structure). We could decide to add a restrictive/non-restrictive distinction, and decide that non-restrictive elaboration is simply "syntactic sugar" for flat structures. In this sense, it doesn't matter if you analyze it like you suggested, or like I did, and we can normalize it in the end. What do you think?

…

On Mon, Nov 19, 2018 at 2:08 PM Nathan Schneider ***@***.***> wrote: It's not specific to superlatives, though: e.g. "one of the experiences that broke my heart". I think the issue is that elaboration is an information structural notion, and it can happen between scenes just as it can happen between non-scenes, or a scene and a non-scene. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#48 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AIG86yj6E2skm8z6JaoyBacLiF0aEsFzks5uwp9RgaJpZM4YooEF> .

[nschneid]

If we were doing model-theoretic semantics, then we'd want to make a restrictiveness distinction. But I'm not sure it's something we need in the foundational layer, and annotators might find it difficult. Both restrictive and nonrestrictive RCs seem to elaborate on the head; whether that is narrowing the set of things denoted by the head or not is orthogonal, right?

[omriabnd]

But so is information structure (ideally). I think the two Hs solution is the least of evils

…

On Tue, 20 Nov 2018, 04:24 Nathan Schneider ***@***.*** wrote: If we were doing model-theoretic semantics, then we'd want to make a restrictiveness distinction. But I'm not sure it's something we need in the foundational layer, and annotators might find it difficult. Both restrictive and nonrestrictive RCs seem to elaborate on the head; whether that is narrowing the set of things denoted by the head or not is orthogonal, right? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#48 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AIG864VXIAPac-sN92adFYawIOv2Kj6dks5uw2f0gaJpZM4YooEF> .