-
Notifications
You must be signed in to change notification settings - Fork 245
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
as X as
expressions as fixed
with ExtPos - what qualifies?
#1043
Comments
https://universaldependencies.org/en/dep/fixed.html has:
But regular comparatives like "I am as tall as you" or "I have as much as you" are not |
The line gets a bit blurry on the quantity approximators for
dev set:
test set:
In some cases it seems to be easier to separate and therefore perhaps doesn't warrant
|
Going by the example "as much as fifty percent", I think the |
Given that it is syntactically regular, is there a good reason to treat as much as fifty percent as |
I think that the test for "as X as Y" could be to see if it can replaced by X:
For the other examples, I let you native guys decide, but I don't see why "as much as fifty percent" should be necessary fixed:
|
This gets into semantics but "as much as one percent" does not entail "much". :) It means "up to one percent". With the caveat that the |
The fact that "as much as fifty percent" could be non compositional doesn't qualify it as a "fixed" expression because we are not annotating semantics. If the construction is syntactically regular it must be annotated with regular syntactic relations.
|
A canonical use of the as-as comparative for quantities would be like (1): (1) I have as many cookies as you (have). This describes two distinct sets of cookies and says their sizes are equal. There is a variant (2) I have as many cookies as you have bagels. where the two sets are of different kinds of things, but again, their sizes are the same. Whereas (3) I have as many as five cookies. likely describes only one set of objects, and is imprecise about its size. It reflects bounded uncertainty: the speaker thinks they might have five cookies, or possibly less, but not more than 5. One could read (3) as having two nominals with ellipsis in the first, i.e. (4) I have as many cookies as five cookies. which is oddly redundant, but if forced to interpret it I get not the imprecision meaning, but the same-quantity meaning. That is why one could argue "as many as" is morphosyntactically fixed with an idiomatic meaning. What does this mean for UD? The guidelines say: The guidelines also list similar uses of "more than", "less than", "up to", and "all of" as If the test is semantic, it would also apply to (5) In this neighborhood, houses are on the market for as much as a million dollars. But I wonder if the construction is actually more productive, and could include "as high as", "as long as", etc.: (6) In this neighborhood, house prices are as high as a million dollars. I can get (6) to mean the same thing as (5), i.e. the houses have a range of prices and the most expensive ones are a million dollars. (Though "as high as a million dollars" is clearly not a nominal. "As much/*high as a million dollars was spent on the project.") So this makes me wonder if we are bringing too much semantics into the syntactic analysis. Maybe the UD structure should be the same for standard comparatives where the head noun is omitted, and noncompositional or noncanonical meaning belongs in a different layer. (P.S. I take back the bracketing I gave earlier—I was confusing the CGEL analysis with EWT's.) |
Where is the line to draw for
as X as
expressions? There are some marked in EWT, such asbut then many others are not marked, such as
The text was updated successfully, but these errors were encountered: