computational-pragmatics

Speech act classification identifies illocutionary force like assertions and directives using transformer-based zero-shot NLP. The skill applies candidate labels to tag utterances and provides rule-based fallbacks for scenarios lacking trained models.

Computational pragmatics analyzes implied meaning by detecting Gricean maxim violations like quantity and manner. The skill identifies hedging patterns and implicature signals to infer indirect requests and unstated speaker intentions in conversation.

Yes, politeness and hedging scoring extracts social interaction cues from real-world text. The skill computes feature-level outputs for downstream analytics, quantifying face-management strategies used in interviews and social media dialogue.

Discourse coherence analysis maps conversational structure through adjacency pairs and turn transitions. The skill identifies how utterances connect sequentially, revealing the underlying organization of dialogue in interview transcripts and chat logs.

Yes, the skill implements zero-shot classification with candidate labels, eliminating the need for a pre-trained dialog act model. It also provides rule-based fallback heuristics to ensure baseline speech act tagging functionality.

Zero-shot classification for indirect request detection depends on candidate label quality and may miss nuanced implicature. Rule-based fallbacks provide baseline coverage but lack the deep contextual understanding needed for highly ambiguous discourse.