Abstract

In this paper, we show how a rich lexico-semantic network which has been built using serious games, JeuxDeMots, can help us in grounding our semantic ontologies as well as different sorts of information in doing formal semantics using rich or modern type theories (type theories within the tradition of Martin Löf). We discuss the domain of base types, adjectival and verbal types, hyperonymy/hyponymy relations as well as more advanced issues like homophony and polysemy. We show how one can take advantage of this wealth in a formal compositional semantics framework. This is a way to sidestep the problem of deciding how your type ontology should look like once you have made a move to a many sorted type system. Furthermore, we show how this kind of information can be extracted from JeuxdeMots and inserted into a proof-assistant like Coq in order to perform reasoning tasks using modern type theoretic semantics.

Highlights

  • Modern Type Theories (MTTs), i.e. Type Theories within the tradition of Martin-Löf (1975); Martin-Löf (1984), have become a major alternative to Montague Semantics (MS) in the last twenty years

  • A number of influential approaches using MTTs have been proposed throughout this period (Ranta 1994; Luo 2011; Retoré 2014; Cooper et al 2014), showing that the rich typing system offered by these approaches has considerable advantages over simple typed systems predominantly used in mainstream formal semantics

  • How does one decide on the base types to be represented? On the one hand, we do have a way to get a more fine-grained type system unlike the monolithic domain of entities found in MS, but on the other hand, constructing such a type ontology is not at all a straightforward and easy task

Read more

Summary

Direct crowdsourcing

Playing games in order to fill the lexical network is a kind of indirect crowdsourcing, where people (players) do not negotiate their contri-. Direct crowdsourcing (with negotiation between contributors) is desirable. Some lexical relation might be complicated enough to be playable without some linguistic knowledge. This is for example the case for telic role, which is the goal/purpose of an object (or action). A butcher knife has the telic role of cutting meat. It is to be differentiated from the instrument of a predicate, which indicates what can be done with the object. In some other cases (depending on each term), a given relation might not be productive enough to be playable. A voting mechanism is at the core of the validation (or invalidation) of proposed relations between terms

Inside the JDM Lexical Network
Negative relations
Aggregate nodes
Some figures
Deductive scheme
Induction scheme
Relation annotations
Annotation values
Common nouns as types and subtyping
Base types and instances of base types
Predicates and world knowledge information
Findings
Unfold the definition for some and use intro
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call