[Paper Review] Automatic Identification of Support Verbs: A Step Towards a Definition of Semantic Weight

Mark Dras|ArXiv.org|Oct 25, 1995

Natural Language Processing Techniques7 references3 citations

TL;DR

This paper proposes a computationally tractable definition of semantic weight based on word frequency in syntactic constructions, focusing on semantically light verbs like 'make' and 'take' when used as support verbs. By applying a statistical approach inspired by Grefenstette and Teufel (1995), the method successfully identifies light verb instances in a corpus, demonstrating that frequency-based syntactic patterns can serve as a reliable proxy for semantic lightness, with potential extension to other word classes.

ABSTRACT

Current definitions of notions of lexical density and semantic weight are based on the division of words into closed and open classes, and on intuition. This paper develops a computationally tractable definition of semantic weight, concentrating on what it means for a word to be semantically light; the definition involves looking at the frequency of a word in particular syntactic constructions which are indicative of lightness. Verbs such as "make" and "take", when they function as support verbs, are often considered to be semantically light. To test our definition, we carried out an experiment based on that of Grefenstette and Teufel (1995), where we automatically identify light instances of these words in a corpus; this was done by incorporating our frequency-related definition of semantic weight into a statistical approach similar to that of Grefenstette and Teufel. The results show that this is a plausible definition of semantic lightness for verbs, which can possibly be extended to defining semantic lightness for other classes of words.

Motivation & Objective

• To develop a computationally tractable definition of semantic weight that moves beyond intuitive, class-based divisions of closed and open classes.
• To investigate whether syntactic frequency patterns can serve as indicators of semantic lightness in verbs.
• To test the feasibility of automatically identifying support verbs—semantically light verbs like 'make' and 'take'—using statistical methods.
• To evaluate whether this frequency-based approach can be extended to define semantic lightness across other parts of speech.

Proposed method

• The method defines semantic lightness through the frequency of a word in specific syntactic constructions known to signal low semantic content.
• It applies a statistical classification model similar to that of Grefenstette and Teufel (1995) to detect instances of 'make' and 'take' used as support verbs.
• The approach uses corpus-based frequency data to identify patterns where these verbs appear in constructions like 'make a decision' or 'take a walk', which are semantically underdetermined.
• The model is trained and evaluated on a corpus, using syntactic frames as features to distinguish light from heavy verb uses.
• The definition is operationalized by measuring how often a verb appears in these low-semantic-content constructions across a large text collection.

Experimental results

Research questions

• RQ1 Can semantic lightness be defined computationally using syntactic frequency patterns rather than intuition?
• RQ2 To what extent can frequency of occurrence in specific syntactic constructions predict whether a verb is semantically light?
• RQ3 Is the proposed frequency-based definition effective in identifying support verbs like 'make' and 'take' in context?
• RQ4 Can this method be generalized to define semantic lightness for other parts of speech beyond verbs?

Key findings

• The frequency-based definition of semantic lightness successfully identifies instances of 'make' and 'take' used as support verbs in a corpus.
• The method achieves a plausible level of accuracy in distinguishing light from heavy uses of these verbs, supporting the validity of the approach.
• The results indicate that syntactic constructions with high frequency of light verbs are strong indicators of semantic lightness.
• The approach provides a data-driven, computationally feasible alternative to traditional, intuition-based lexical class divisions.