MetaNet: Repository, Identification System, and Applications
Miriam R. L. Petruck and Ellen K. Dodge
The ubiquity of metaphor in language (Lakoff and Johnson 1980) has served as impetus for cognitive linguistic approaches to the study of language, mind, and the study of mind (e.g. Thibodeau & Boroditsky 2011). While native speakers use metaphor naturally and easily, the treatment and interpretation of metaphor in computational systems remains challenging because such systems have not succeeded in developing ways to recognize the semantic elements that define metaphor. This tutorial demonstrates MetaNet’s frame-based semantic analyses, and their informing of MetaNet’s automatic metaphor identification system. Participants will gain a complete understanding of the theoretical basis and the practical workings of MetaNet, and acquire relevant information about the Frame Semantics basis of that knowledge base and the way that FrameNet handles the widespread phenomenon of metaphor in language. The tutorial is geared to researchers and practitioners of language technology, not necessarily experts in metaphor analysis or knowledgeable about either FrameNet or MetaNet, but who are interested in natural language processing tasks that involve automatic metaphor processing, or could benefit from exposure to tools and resources that support frame-based deep semantic, analyses of language, including metaphor as a widespread phenomenon in human language.
Part I: Background to FrameNet and the FrameNet Constructicon
a. Frame Semantics and FrameNet
b. Construction Grammar and the FrameNet Constructicon
c. FrameNet treatment of metaphor
Part II: Overview of MetaNet
a. Conceptual Metaphor Theory
b. The MetaNet Repository
c. The Metaphor Identification System
Part III: Applications
a. Metaphor Identification and Analysis
c. Information Extraction
Part IV: Challenges and Opportunities
a. Integrating FrameNet and MetaNet
b. Further system development of MetaNet
About the presenters
Miriam R. L. Petruck (firstname.lastname@example.org) received her PhD in Linguistics from the University of California, Berkeley. A key member of the team developing FrameNet almost since the project’s founding, her research interests include semantics, knowledge base development, grammar and lexis, lexical semantics, Frame Semantics and Construction Grammar.
Ellen K. Dodge (email@example.com) received her PhD in Linguistics from the University of California, Berkeley. Since 2000, she has worked in ICSI's AI Group, first as part of the Neural Theory of Language project, and also FrameNet. She is the primary linguist continuing to develop Embodied Construction Grammar. Since 2012, she has worked on MetaNet, developing formal representations of frame and metaphor networks, as well as automatic methods to identify and analyze metaphors in text.