Verb Edge encodes various semantic qualifiers in addition to the verb body. Together with Semantic Role, they form the complete meaning of a predication.

Qualifier List

CategoryEnglish nameData typeValue mapping
Core verbCore VerbIdentifierSemantic-aligned absolute identifier
Participant listParticipant ListComposite type list{entity, grammatical role, semantic role}
SpeakerSpeakerReferencePredication subject (required)
ListenerListenerReferencePredication target (nullable)
EvidentialityEvidentialityFloat [-1.0~1.0]-1=inference, 0=direct experience, 1=hearsay
MoodMoodFloat [-1.0~1.0]-1=hypothetical, 0=declarative, 1=imperative
ModalityModalityFloat [0.0~1.0]Degree of volition
TenseTenseFloat [-1.0~1.0]-1=past, 0=present, 1=future
AspectAspectBitmask1:progressive, 2:perfective, 4:resultative
PolitenessPolitenessFloat [-1.0~1.0]-1=informal, 0=neutral, 1=honorific
PolarityPolarityFloat [-1.0~1.0]-1=negative, 0=indeterminate, 1=affirmative
VolitionalityVolitionalityFloat [-1.0~1.0]-1=unintentional, 0=indeterminate, 1=intentional
ConfidenceConfidenceFloat [-1.0~1.0]-1=speculative, 0=indeterminate, 1=certain
IterativityIterativityInteger0=indeterminate, 1=once, MAX=infinite

Evidentiality

Expresses the source of information.

ValueMeaningExample
-1.0Inference“It seems like ~”
0.0Direct experience“I saw that ~”
1.0Hearsay“They say that ~”

Mood

Expresses the function of the utterance.

ValueMeaningExample
-1.0Hypothetical/counterfactual“If only ~ had happened”
0.0Declarative/factual“~ is”
1.0Imperative/request“Do ~”

Tense

Expresses the temporal location of the event.

ValueMeaningExample
-1.0Past“~ did”
0.0Present“~ does”
1.0Future“~ will do”

Aspect

Expresses the internal temporal structure of an event as a bitmask.

BitsMeaningExample
001Progressive“is doing ~”
010Perfective“has done ~”
100Resultative“has ~ ready”
011Progressive + Perfective“has been doing ~”

Politeness

Expresses the social relationship between speaker and listener.

ValueMeaningExample
-1.0Informal/casual“Do it”
0.0Neutral“Please do it”
1.0Honorific/humble“Would you kindly ~”

Design Principles

  • Continuous values: Expressed as Float rather than discrete classifications to allow gradient representation
  • Bipolar: Most use a [-1.0, 1.0] range to represent extremes
  • Indeterminate: 0.0 can also mean “indeterminate” rather than “neutral” (for Polarity, Volitionality, Confidence)
  • Combination: Hybrid of bitmask (Aspect) + Float for representing compound meanings