Actions
Actions (or more fully, Action types) represent individual semantically disambiguated verbs.
Three valencies: Entity type valency, grammatical valency, and semantic valency
Actions acquire three kinds of valencies per any actant slot (subject, object 1, object 2 –2; the data model is potentially extensible further, beyond trivalent verbs):
entity type valency,valency, which defines which entity type is allowed in the given actant slot;
grammaticalmorphosyntactic valency,valency, which is a free text field defining the
prepositions,prepositions and grammatical cases, etc.but –uses ita formalized notation (grammatical cases are noted with numbers 1-6, prepositions are in quote marks "", alternative is
usefulmarked forwith thea usepipe of our Actions for machine understanding;"|"); and
semantic valency,valency, i.e. what kind of role
any the
holderentity ofoccupying the given actant slot
canhas acquireby implication (e.g.,
in our example, the subject of
the Action “to travel” would have the semantic valency C “traveller”
, and thus, in a research-oriented data projection, Alice could be tagged as traveller, and we could find all travellers throughout the dataset)).
ValenciesThe main benefits from valencies are that they:
guide coders in their choice of the correct Action (or towards creating a new one if none
among the existing yet fits the syntactic and semantic definition)
.;
They also allow us to implement
data validation features in a data collection
interface.interface;
Finally,facilitate theymachine areunderstanding usefulof fortext, NLPallowing semantic disambiguation of
different verbs based on their
grammaticalmorphosyntactic valency,valency (recognized
throughby dependency
parsing,parsing), and
optionally theiroptionally, entity type
valency,valency (recognized
e.g. through
some other procedure (such asthrough named entity recognition).
