Scenify is a natural-language processing tool that translates natural-language narratives into ScenIC scenarios.
Narrative: Unconstrained natural-language paragraphs that describe user-system interactions.
Dictionary: Princeton Wordnet dictionary.
(Optional) Interactive presentation of ScenIC scenario: A screen-based tabular or graphical presentation of a ScenIC scenario for editing.
SCML-0 document: A plain-text representation of the scenario in SCML-0 (Scenario Markup Language 0).
...
A hostile country mobilizes some advanced missiles and threatens its neighbor, an ally of the US.
Hostile country Achieve: Ally threatened
Hostile country Mobilize missiles
Ally threatened
A US surveillance satellite reports the nature of the mobilization
US Know: nature of mobilization US Maintain: Assistance of ally
Satellite Detect mobilization Satellite Report to US nature of mobilization
US knows nature of mobilization.
<Scenario Threat to ally> <Background> ... </Background> <Episode hostile mobilization> <Point> <Actor hostile country> <Goal ally threatened> </Point> <Actor hostile country> <Task mobilize missiles> <Outcome> Ally threatened </Outcome> </Episode> <Episode reporting on mobilization> <Point> <Actor US> <Goal know nature of mobilization> <Actor US> <Goal maintain alliance with ally> </Point> <Actor satellite> <Task detect mobilization> <Actor satellite> <Task report on mobilization> <Outcome> US knows nature of mobilization </Outcome> </Episode> ... </Scenario>
A ScenIC actor is usually a the agent of a verb, which is usually the subject of a verb.
A ScenIC actor is usually a noun phrase dominated by a noun denoting a political entity, person, occupation, machinery.
A passive sentence without an explicit subject probably denotes a task done by the default actor ("the system").
Inanimate anaphoric reference ("it", "they") of a verb subject probably refers to the last inanimate actor referred to.
Animate anaphoric reference ("he", "she", "they") of a verb subject probably refers to the last animate actor referred to.
Verbs that denote knowledge actions probably are tasks (a.k.a. achievement or knowledge goals) in ScenIC, for example: reports, analyze, identify, request (information), assess, diagnose, select.
Verbs that denote coordination or organization actions are probably tasks in ScenIC, for example: task to, request (an action), interact, assign.
Transitive non-stative verbs (e.g. not "become", of "have") probably denote non-information tasks, for example: mobilize, put on alert, etc.
The pattern "verb1 (in order) to verb2" probably indicates that verb2 is the point of verb1.
The pattern "verb2 by (means of) verb1-ing" probably indicates that verb2 is the point of verb1.
The pattern "try to verb" probably indicates that verb is both a task in an episode and the point (achievement goal) of it.
The pattern "try not to verb" probably indicates an avoidance goal.
A stative verb may indicate a maintenance goal, for example: "the US tries to keep its ally informed."
The pattern "want to verb" probably indicates that verb is the point but may or may not be achieved.
The pattern "S1 contrastive S2" (where contrastives include "but" and sentence-opening "however," etc.) probably means that S1's goal is thwarted by an obstacle described in S2 (in which case S2 may not represent an action), for example: "The satellite was tasked to observe but the signal never reached it."
Negation of an action verb ("not", "never"...) may indicate an obstacle.
Marked negative verbs ("un-", "dis-") may indicate a mitigation task to overcome an obstacle.
Repetition of actions ("again", "continues to",
"a second time") or marked repetition verbs ("re-")
may indicate a mitigation task to overcome an obstacle.