code: IN
biber_number: I39
xiao_number: F77
bohmann_number: 158
mfte_code:
- IN
- INother
name: Prepositions
definition: >-
Prepositions (residual count after removing specific subcategories).
normalization: words
detection:
- requires:
- pos
cql: '[cat="IN"]'
combine: "_ & !CONC"
description: >-
Uses cat refinement for THSC, COND, CUZ, LIKE (all have refines: IN).
CONC needs explicit exclusion (crosses POS boundaries). SO default rule
only refines RB, not IN, so it doesn't affect this default rule.
- source: mfte
requires:
- pos
cql: '[pos="IN|TO"]'
combine: "_ & !THSC & !COND & !CONC & !CUZ & !LIKE & !SO"
description: >-
MFTE converts all remaining TO → IN (line 1234) after to-clause
features have been tagged. So IN = original IN + remaining TO,
minus features that REPLACE IN-tagged tokens: subordinators
(THSC, COND, CONC, CUZ), plus LIKE and SO "bin" features.
Note: MFTE removed Biber's OSUB/OTHADVSUB category entirely
(Le Foll comment, line 1082); while→CC is handled separately;
most "as" subordinators stay as IN. CAUSE and ELAB do not
overlap with IN-tagged tokens in practice.
- source: pybiber
requires:
- pos
- dep
semgrex: '{dep:prep}'
sources:
- biber_1988
- mfte
- pybiber
- xiao_2009
- bohmann_2019