NN YAML source

code: NN
biber_number: E16
xiao_number: A2
mfte_code:
  - NN
  - NNother
name: Total other nouns
definition: >-
  Residual noun count: all nouns (including proper nouns) minus nominalizations
  and gerunds. A marker of informational density.
normalization: words
detection:
- requires:
  - pos
  cql: '[cat="NN|NNS|NNP|NNPS"]'
  combine: "_ & !GER"
  description: >-
    Uses cat refinement for NOMZ (refines: NN|NNS). GER is explicit exclusion
    because it matches VBG tokens (not NN) — conceptual subtraction, not index overlap.
- source: pybiber
  requires:
  - word
  - pos
  - upos
  - dep
  parts:
    ger_noun:
      cql: '[word=".*ings?$" & upos="NOUN" & dep="nsubj|dobj|pobj|nsubjpass" & word!={words_ger_stop}]'
  words:
    ger_stop:
    - according
    - anything
    - beijing
    - bing
    - bings
    - boeing
    - bring
    - ceiling
    - ceilings
    - cling
    - clings
    - darling
    - ding
    - dings
    - during
    - evening
    - evenings
    - everything
    - fling
    - flings
    - inning
    - innings
    - irving
    - king
    - kings
    - morning
    - mornings
    - nothing
    - notwithstanding
    - offspring
    - offsprings
    - outstanding
    - ping
    - pings
    - ring
    - rings
    - sing
    - sings
    - something
    - spring
    - springs
    - sterling
    - sting
    - stings
    - string
    - strings
    - thanksgiving
    - thanksgivings
    - thing
    - things
    - wedding
    - wing
    - wings
    - wrongdoing
    - wyoming
  cql: '[upos="NOUN|PROPN" & word!=".*-.*"]'
  combine: "_ & !NOMZ & !ger_noun"
  description: >-
    All nouns (NOUN/PROPN coarse POS, no hyphens). pybiber uses coarse POS
    which excludes pronouns tagged NN (nothing, something, anyone, anything).
    Only NOUN-tagged gerunds are subtracted (not PROPN like "Building").
sources:
- biber_1988
- mfte
- pybiber
- xiao_2009
notes: >-
  Strongest negative loading on D1. Biber's f_16 "other nouns" is the residual
  after subtracting nominalizations (f_14) and gerunds (f_15). pybiber also
  excludes hyphenated tokens.