COMPLEXITY — Syntactic Complexity Ratio

Status: Current Last updated: 2026-05-12 10:07 EDT

Computes syntactic complexity by counting subordinating dependency relations in the %gra tier.

Usage

chatter clan complexity file.cha
chatter clan complexity corpus/ --speaker PAR
chatter clan complexity file.cha --format json

COMPLEXITY reads the %gra dependency tier and counts clause-embedding relations that indicate syntactic subordination. The complexity ratio is the number of subordinating relations divided by total tokens (non-punctuation).

UD Relations (auto-detected)

Per crates/clan/clan-core/src/commands/complexity.rs:10:

Relation	Description
CSUBJ	Clausal subject
CCOMP	Clausal complement
XCOMP	Open clausal complement
ACL	Adnominal clause
ADVCL	Adverbial clause modifier
APPOS	Apposition
EXPL	Expletive

Legacy CLAN Relations (auto-detected)

Relation	Description
CSUBJ	Clausal subject
COMP	Complement
CPRED	Clausal predicate
CPOBJ	Clausal object of preposition
COBJ	Clausal object
CJCT	Clausal adjunct
XJCT	Non-finite clausal adjunct
NJCT	Nominal adjunct
CMOD	Clausal modifier
XMOD	Non-finite clausal modifier

Output

Per speaker:

Count of each subordinating relation
Tokens: total matched complexity relations
TotalTokens: all non-PUNCT tokens
Ratio: Tokens / TotalTokens

Requirements

Requires a %gra dependent tier. Utterances without %gra are skipped.

Differences from CLAN

Auto-detects UD vs legacy relation names (CLAN requires compile-time config)
Uses typed AST GraTier with GrammaticalRelation entries
Supports JSON and CSV output
Sub-relations (e.g., CSUBJ:pass) are properly matched to their base relation

Batchalign