Session	Introductory
Duration	1 hour
Date/Time	November 6 2019 17:00 GMT
	9:00am PST/12:00pm EST
	5:00pm GMT/6:00pm CET
Convener	Ken Baclawski

Ontology Summit 2020 MetaNet Metaphor Repository

Agenda

Title: MetaNet: Deep semantic automatic metaphor analysis Slides in pdf format Slides in pptx format Video Recording YouTube Video
Speaker: Professor Elise Stickles
Abstract: As analyzed in Conceptual Metaphor Theory, metaphors allow language users to exploit their rich and complex knowledge of one domain, such as the physical world, to understand and reason about another, often less structured and/or more abstract domain. Building on the foundation of Conceptual Metaphor Theory, the MetaNet project has developed formal representations of metaphors as mappings from one domain (the Source domain) to another (the Target domain) , and has built a structured repository containing a systematic network of searchable and interrelated metaphors, as well as a network of semantic frames that act as source and target domains of metaphors. The MetaNet metaphor repository consists of a very large compendium of attested metaphors, including time metaphors, mind metaphors, event structure metaphors, emotion metaphors, and morality metaphors. It also includes novel metaphors pertaining to target domains of interest to the project, which have centered on social problems such as issues of poverty, taxation, bureaucracy, governance, gun violence, and cancer. In addition to the searchable encyclopedia of systematically-linked metaphors and the semantic frames that constitute the source and target domains, the MetaNet project has been exploring metaphors in corpus data across four different languages, namely American English, Mexican Spanish, Iranian Persian, and Russian as spoken in Russia. The American English MetaNet repository is available as a Semantic MediaWiki at [1]. The goal of the MetaNet project has been to build a system that will extract linguistic manifestations of metaphor and automatically interpret them.
Bio: Elise Stickles is an Assistant Professor at the University of British Columbia. She is a cognitive semanticist, meaning that she studies the relationship between linguistic meaning and form from a cognitive-functional perspective. Her research lies at the intersection of conceptual metaphor theory and embodied construction grammar approaches to syntax and lexical semantics. She focuses particularly on multimodal metaphoric constructions, comprising both linguistic and gestural content. Most of her research is on American English, but she also studies the metaphor and argument structure of American Sign Language.

Conference Call Information

Date: Wednesday, 06-November-2019
Start Time: 9:00am PST / 12:00pm EST / 6:00pm CET / 5:00pm GMT / 1700 UTC
- ref: World Clock
Expected Call Duration: 1 hour

The Video Conference URL is https://zoom.us/j/689971575
- iPhone one-tap :
  - US: +16699006833,,689971575# or +16465588665,,689971575#
- Telephone:
  - Dial(for higher quality, dial a number based on your current location): US: +1 669 900 6833 or +1 646 558 8665
  - Meeting ID: 689 971 575
  - International numbers available: https://zoom.us/u/Iuuiouo
Chat Room

Participants

Discussion

[12:12] David Eddy: So this is WordNet Enhanced?

[12:14] David Eddy: https://www.kurzweilai.net/ambiguous-words

[12:17] David Whitten: Hello everyone.

[12:17] David Whitten: Thank you for whoever invited Elise Stickles to talk today.

[12:18] Ken Baclawski: I invited Elise to speak.

[12:19] David Whitten: I don't know if this uses WordNet or just FrameNet.

[12:20] David Whitten: Thank you Ken. She is very well informed.

[12:22] David Whitten: I wonder what inference methods exist that use these metaphor models.

[12:25] David Whitten: If you can draw a conclusion, when you have a physical frame, it may involve spatial reasoning.

[12:25] David Whitten: She said Entailment relation is part of the inferencing structure.

[12:27] ToddSchneider: Were the instances of relations among frames derived from the corpora used?

[12:27] David Whitten: They have about 600 metaphors and 500 frames.

[12:28] David Whitten: Is a "wiki form" part of the Semantic Wiki enhancement to MediaWiki ?

[12:29] Ken Baclawski: Yes, they are called "Semantic Forms" in Semantic MediaWiki. Plain MediaWiki has a notion of a Page Form. This generates an HTML form with text input and dropdown lists that are very commonly used on the internet. Semantic Forms allows the dropdown lists to be generated dynamically via a query rather than being statically specified. The queries can even depend on the answers to other parts of the form, so that the form will change dynamically. Again, this is commonly used on the internet, but usually requires complex programming. Semantic Forms allows one to specify the form declaratively.

[12:29] John Sowa: The metaphors and the hierarchy is a good example of an ontology for knowledge graphs.

[12:29] Mark Underwood @knowlengr: some work I see as related is to identify metaphor/analogy in model-based systems engineering, e.g., https://apps.dtic.mil/dtic/tr/fulltext/u2/1058354.pdf

Transforming Systems Engineering through Model-Centric Engineering
A013 Interim Technical Report SERC-2017-TR-111

[12:29] BobbinTeegarden: Is there a relationship around 'contains' e.g. aggregation (for a fractal-like structure)?

[12:30] David Whitten: I would assume "contains" is a relation tied to the metaphor of a "container"

[12:30] John Sowa: Each metaphor can be stated in a subset of logic (RDF or knowledge graphs).

[12:31] David Whitten: I wonder if Object Type is contrasted with Stuff Type or something else?

[12:32] David Whitten: "Achieving a purpose" is a frame. relations exist between frames.

[12:32] John Sowa: A type is a type is a monadic relation.

[12:33] John Sowa: A frame uses the same subset of logic as RDF or knowledge graphs.

[12:33] ToddSchneider: What was/is the corpus used?

[12:33] David Whitten: John Sowa, that is consistent with your ideas about type from your previous writing.

[12:34] John Sowa: The inheritance can be stated in Aristotle's syllogisms.

[12:34] David Whitten: I assume the corpus is some text that they are analyzing as part of fulfilling some grant.

[12:34] John Sowa: The corpus is any collection of documents.

[12:35] David Whitten: John, didn't you have some analogical reasoning system to draw conclusions when VivoMind was around?

[12:36] John Sowa: A non-metaphor pattern is a knowledge graph derived from another sentence about the subject.

[12:37] David Whitten: She seems to be talking about some kind of "transfersThrough" relation we discussed several years ago to direct inference along a non-generalization relation augmenting the generalization hierarchy.

[12:38] ToddSchneider: Was the corpus constrained to a particular domain or source (e.g. New York Times)?

[12:38] John Sowa: David, yes. The Cognitive Memory system was designed to handle arbitrary analogies.

[12:39] David Whitten: I thought there was some funding from the US Government.

[12:39] John Sowa: Formal logic uses a precisely defined special case of analogy called unification.

[12:39] David Whitten: cool term : Metaphoricity.

[12:40] John Sowa: Cool terms are good for sales. But when you translate them to logic, you get same old, same old.

[12:41] David Whitten: I think this relation is a hierarchy inducing relation: mo:isSubCaseOfFrame* Is the * a Kleene closure?

[12:41] David Eddy: @DaveW ==> IARPA for funding

[12:41] David Whitten: Metaphoricity seems to be a measure like probability that ranges from 0 to 1.

[12:42] John Sowa: Re funding for cogmem: Most of the funding for VivoMind and Kyndi came from some branch of the US govt.

[12:42] John Sowa: But most of it was for specific deliverables -- i.e., working applications, not research.

[12:42] Ram D. Sriram: @John: Any thoughts on the methodology based on your experience with Vivomind. I remember you had something similar in terms of matching and extracting information

[12:43] Ram D. Sriram: @john: saw your message after I typed mine.

[12:43] David Whitten: The iterative developmental process seems to be an example of a Virtuous Cycle (Beneficial Network Effect)

[12:44] John Sowa: Most of the information is in the cogmem.pdf slides or the references i the slides and the final page.

[12:44] John Sowa: See http://jfsowa.com/talks/cogmem.pdf

[12:45] David Whitten: mostly Non-Fiction text based on the news.

[12:45] John Sowa: All the metaphors and frames can be represented in conceptual graphs and cogmem can find analogies, informal or formal.

[12:46] David Eddy: excellent

[12:46] David Whitten: A well grounded definition for each relation is essential if you are having a semi-automated approach.

[12:46] David Eddy: @Elise... thx re WordNet

[12:48] David Whitten: Inference model is grounded in human psychological patterns. Does this entailment match logical entailment?

[12:48] John Sowa: But I don't want to minimize anybody's implementations.

[12:48] John Sowa: If they have good applications, that is important.

[12:49] John Sowa: But when you get to the theory, there is nothing new.

[12:50] David Whitten: John Sowa, could you expand on what "cogmem" means to you? Is it just a short form for "Cognitive Memory" or do you mean more?

[12:50] TerryLongstreth: Does linguistic analysis include establishing or measuring or judging the appropriateness of particular metaphors? "Why is a raven like a writing desk?"

[12:51] John Sowa: Terry, there are huge problems with NLs.

[12:51] David Whitten: They advocate using Statistics in the discovery process.

[12:51] John Sowa: See http://jfsowa.com/talks/natlog.pdf

[12:52] David Whitten: Doesn't Cyc have 45 or more meanings for "in" ? Each would potentially be a specialized Container metaphor, I guess.

[12:52] John Sowa: Statistics are very important.

[12:53] John Sowa: See http://jfsowa.com/pubs/fuzzy.pdf

[12:53] David Eddy: @Elise... is "repository" just an inflated term for "database?"

[12:54] David Whitten: David Eddy. If a database has inference, isn't that a necessary precursor to be a knowledge base?

[12:55] ToddSchneider: How stable are metaphors from a historical perspective?

[12:55] David Eddy: @DaveW... big bafflegab words I do not grok

[12:55] MikeBennett: Do you treat synecdochal statements as metaphor? Or have similar relations to these? (part/whole?)

[12:55] John Sowa: David, see the slides in natlog.pdf for the many issues about determining word sense.

[12:55] David Whitten: I really like MetaNet's focus on evidence to decide what is a relation.

[12:57] David Whitten: Eve on her team is involved in historic linguistics.

[12:58] David Whitten: strong dependency on bottom up development.

[13:01] David Whitten: reminds me of the manual steps done when creating Cyc's KB.

[13:02] David Whitten: Has examining causality between target and source domains induced some particular analogies ?

[13:03] janet singer: Is identification of a metaphor as more basic from observation of use or from theory? Do you have a observations/theory re what makes a given frame more likely to be used as a source or target?

[13:04] David Whitten: invoking a frame may use metonymy. "Have you read Margaret Atwood?" She isn't a book, but writes books.

[13:06] David Whitten: MetaNet uses construction grammar. Langacker's construction grammar?

[13:07] David Whitten: FrameNet blurs inferential structures needed by MetaNet. Up vs Down. direction affects inferencing.

[13:07] David Whitten: Construction grammar allows for bottom-up development.

[13:10] David Whitten: MetaNet seems to have a lot of influence from linguists and their vocabulary.

[13:11] David Whitten: Why am I not surprised John Sowa is skeptical re buzzwords and Knowledge Graphs ?

[13:12] MikeBennett: By the way I will be spending time in Vancouver, we might know some people in common?

[13:13] John Sowa: For more info about metaphors by Lakoff & Johnson, see http://jfsowa.com/pubs/lakoff.pdf

[13:13] David Whitten: Janet Singer asked re whether certain types of frames are used either as sources or targets ?

[13:14] David Whitten: Source Domains are more objective and shared but more cognitive theory uses a different term.

[13:14] Ram D. Sriram: @john: I think the problem with KGs is standardization. If we can agree on a representation standard, we don't need to reinvent work done earlier.

[13:15] David Whitten: Emotions are not seen, but Running activities are easily seen.

[13:16] David Whitten: more is up and less is down is a cross-cultural physiological basic metaphor.

[13:16] John Sowa: That's http://jfsowa.com/pubs/lakoff.pdf

[13:17] David Whitten: Metaphor: Intimacy is Closeness has a grounding in Physical Closeness.

[13:18] David Whitten: Zoltan Kovecses is working on cross-cultural metaphors.

[13:19] David Whitten: There is a Croatian Meta-Net with language specific examples.

[13:19] Elise Stickles: Zoltan Kovecses

[13:19] Elise Stickles: Anna Wierzbicka

[13:20] janet singer: https://www.researchgate.net/publication/291103119_Metaphor_and_Culture

[13:20] David Whitten: John Sowa mentions a Polish researcher

[13:20] David Whitten: Ah. Anna Wierzbicka is the Polish researcher John Sowa mentioned.

[13:22] ToddSchneider: Have to go. Thank you.

[13:22] David Whitten: Daniel Everett is a linguistic researcher in Amazonian Language - Pirahã

[13:23] janet singer: https://en.wikipedia.org/wiki/Natural_semantic_metalanguage

[13:23] David Whitten: Anna works in "Semantic Metalanguage"

[13:23] Elise Stickles: The language is Pirahã

[13:23] Ken Baclawski: Pirahã (also spelled Pirahá, Pirahán), or Múra-Pirahá

[13:26] TerryLongstreth: Is 'grok' a metaphor?

[13:26] David Whitten: MetaNet core premise is that metaphors are present in common language, even in dictionaries

[13:26] David Whitten: Thank you Ken.

[13:26] David Whitten: Thank you Elise.

[13:27] David Whitten: Maybe we can get Elise next week.

[13:28] janet singer: @Terry: grok may be a new frame that can be used as a source or target in a metaphor

[13:30] Elise Stickles: thanks everyone!

[13:31] Mark Underwood @knowlengr: Elise, thanks, really interesting and relevant

[13:36] Elise Stickles: I've got another meeting to get to, but I'm happy to answer further questions, or if you have suggestions/ideas I'd love to hear them - my email address is elise.stickles@ubc.ca

[13:36] Elise Stickles: Thanks again

Resources

Previous Meetings

... further results

Next Meetings