Woodley: Another question for Ann. What are the ramifications of the class of scopal predicates being unbounded?

Ann: Nasty. We've been aware of a few places that that comes up, and we're sticking our heads in the sand.

Emily: What are those examples?

Woodley: One on the board: I texted her that the plane landed.

oe: At any given point in time the language has a finite set, but that can change over time.

Ann: That one I wasn't so worried about.

oe: Nearly but not quite every; -?

Ann: No there's some case I can't quite remember, where you get a that …

Emily: so surprising that everyone gasped

Woodley: so difficult to believe that everyone laughed

Dan: so nearly surprising that everyone laughed

Woodley: Isn't so just a two-place predicate?

Ann: Might be able to handle it.

Dan: so, like comparatives (biggest book that I could find) takes two arguments, rather than building an open-ended class of things like so difficult.

Woodley: How does it take them?

Dan: Like degree modifiers in general: ARG0 of adjective is ARG1 of so, or superlative. I don't think that so is a problematic case, just one more in the inventory of rather complicated degree modifying thingies.

Ann: There's some case that I remember which doesn't have a clear lexical item.

Dan: I can believe that. But so far the great majority of these have morphological or lexical thingies that you can attach them too.

Dan: If there were constructions like That John went to the store, I really smiled. as a productive process of English, could be problematic. Not clear where you hang the LTOP for the clause.

Emily: Looks like you could just put in another EP contributed by the hypothetical construction that licenses that type of phrase.

Emily: [recap question about whether generalized quantifiers have to be a finite set.

Ann: No, GQs don't have to be a finite set. There are analyses of numbers like that. There is a thesis on this (can track down the name of person who wrote it). Within the last 10 years or so a formal semanticist has been finally working on the problem of complex quantifiers like almost but not quite every.

Emily: Since Woodley's sticking his tongue out at me, if you can have a infinite set of those, but don't have the interpretive rules for them, have you really solved the problem?

Ann: Yeah… I don't think there's a problem with semantics as such, but there is one with compositionally constructing the semantics.

Emily: So I'd be interested in seeing that thesis if you can track it down, to see if the author talks about that problem.

Woodley: If we wanted to attack that problem, we'd have to work out some way of describing the shape of a non-atomic quantifier, per se.

Ann: Semantically, the non-atomic quantifiers are not a problem, so long as you want work out what they mean in terms of two sets, but actually constructing that semantics compositionally is something I've got no clue about how to do.

Woodley: If you're willing to suspend the notion that we don't know what almost means, and we don't know what every means, but we might be able to do something about almost but not every in terms of what they mean. … or at least 3

Ann: I'm not sure I can do at least 3 compositionally, but I know what it should mean.

Woodley: Attacking at least 3 would be easier to attack almost but not every, as a starting point.

Ann: But we don't treat 3 (or at least 3) as a generalized quantifier.

Dan: And we don't have the GQ treatment because we're aiming for uniformity for 3 in the case of these 3 books and at least 3 books.

Ann: Assuming you all plural entities, it's fine to have an intersective modifier analysis.

Woodley: These blue three books?

Dan: Adjective order is a separate and vexed problem. My three books and four records --- don't need another quantifier. We're stuck in English with at least a usage of 3 and 4 as adjectives. That still leaves the question of whether there is an ambiguity.

Ann: There's some issues about that, but it's nothing to do with GQs.

TheAbbey/Chrysalis2014OpenEndedPredicates (last edited 2014-02-19 22:25:48 by EmilyBender)

(The DELPH-IN infrastructure is hosted at the University of Oslo)