Hopper and Thompson (1980) defined a multi-axis theory of transitivity that goes beyond simple syntactic transitivity and captures how much "action" takes place in a sentence. Detecting these features requires a deep understanding of lexical semantics and real-world pragmatics. We propose two general approaches for creating a corpus of sentences labeled with respect to the Hopper-Thompson transitivity schema using Amazon Mechanical Turk. Both approaches assume no existing resources and incorporate all necessary annotation into a single system; this is done to allow for future generalization to other languages. The first task attempts to use language-neutral videos to elicit human-composed sentences with specified transitivity attributes. The second task uses an iterative process to first label the actors and objects in sentences and then annotate the sentences' transitivity. We examine the success of these techniques and perform a preliminary classification of the transitivity of held-out data.
[1]
Philip Resnik,et al.
More than Words: Syntactic Packaging and Implicit Sentiment
,
2009,
NAACL.
[2]
G. Miller,et al.
A Semantic Network of English Verbs
,
1998
.
[3]
Perry R. Cook,et al.
How well do visual verbs work in daily communication for young and old adults?
,
2009,
CHI.
[4]
Dan Klein,et al.
Accurate Unlexicalized Parsing
,
2003,
ACL.
[5]
S. Thompson,et al.
Transitivity in Grammar and Discourse
,
1980
.
[6]
Martin F. Porter,et al.
An algorithm for suffix stripping
,
1997,
Program.
[7]
George A. Miller,et al.
Nouns in WordNet: A Lexical Inheritance System
,
1990
.
[8]
Jason Eisner,et al.
Machine Learning with Annotator Rationales to Reduce Annotation Cost
,
2008
.
[9]
M. F. Porter,et al.
An algorithm for suffix stripping
,
1997
.
[10]
Ian H. Witten,et al.
The WEKA data mining software: an update
,
2009,
SKDD.