Skip to content
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

WHAT-initial clauses (free relative clauses) #523

Open
nschneid opened this issue Jul 26, 2017 · 7 comments
Open

WHAT-initial clauses (free relative clauses) #523

nschneid opened this issue Jul 26, 2017 · 7 comments
Labels

Comments

@nschneid
Copy link
Contributor

nschneid commented Jul 26, 2017

As part of UniversalDependencies/docs#454, it is difficult to tell from the guidelines what the conventions are/should be for initial clauses like these (including examples from the English treebank):

  1. What was left in place was a...

image

  1. What your morning paper did not tell you was...

image

  1. What I heard was...

image

In (1) the WH-word occupies a missing core role (the passive subject). The WH-clause is treated as a clausal subject of the item left in place (the answer to the WH-question). (2) and (3) have the main verb of the sentence as the root, even though it is a copula. In (2), the WH-word corresponds to the clausal complement of the embedded verb, and is analyzed as the head of a relative clause. In (3), the analysis is the same, though "heard" normally licenses either a direct object ("I heard the announcement") or a complement clause ("I heard that...").

Is the clausal complement vs. subject/object distinction the defining criterion for deciding between these analyses?

@sylvainkahane
Copy link

Your examples are free/fused "relative" clauses, that is "relative" clause where the antecedent is fused with relative pronoun. They must be definetely analyzed as phrases headed by WHAT. When they are subjects they must be analyzed as nsubj and when they are objects as obj:

I bought what I needed.
obj(bought,what)
acl:relcl(what,needed)

I wouldn't have any sense to analyze this latter case as a ccomp, BUY can never have a ccomp.

@nschneid
Copy link
Contributor Author

@sylvainkahane are you saying that the tree for (1) is incorrect?

@sylvainkahane
Copy link

Yes

@dan-zeman
Copy link
Member

We need documentation of free relative clauses in the guidelines.

@dan-zeman dan-zeman changed the title WHAT-initial clauses WHAT-initial clauses (free relative clauses) May 14, 2020
@nschneid
Copy link
Contributor Author

nschneid commented Feb 6, 2021

We need documentation of free relative clauses in the guidelines.

Yes! See also #23

@nschneid
Copy link
Contributor Author

nschneid commented May 5, 2024

These are now documented in the English guidelines as pseudoclefts, a construction that utilizes free relatives.

I was able to find false negatives in

Need to also check for PronType=Rel (#278)

@nschneid nschneid transferred this issue from UniversalDependencies/docs May 5, 2024
@amir-zeldes
Copy link
Contributor

Agreed, some of the GUM ones are errors, will fix those!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

4 participants