This page provides documentation for the parts-of-speech.
POS | Name | Extra |
---|---|---|
n | noun | |
v | verb | |
a | adjective | |
r | adverb | |
s | adjective satellite | |
c | conjunction | |
p | adposition | preposition, postposition, etc. |
x | non-referential | particle, classifier, exclamative, etc. |
u | unknown |
Princeton WordNet defines the original four open class parts of speech: either noun, verb, adjective, or adverb. It later added satellite adjectives, which are in clusters representing a concept that is similar in meaning to the concept represented by its head synset. OMW maps them internally to adjectives.
Conjunctions and adpositions are reserved for extending to new semi-open classes.
x is used for non-referential words such as particles, classifiers, exclamatives, greetings, determiners and so forth.
z is used for phrases, for example to
represent phrasets, introduced by Luisa Bentivogli and
Emanuele Pianta (2003).
DEPRECATED better to
use lexicalized instead.
u Is for automatically produced wordnets that want to add words for which they are not sure of the part-of-speech.
We have an optional attribute lexicalized
on the synset and sense types, with a boolean value, defaulting to
true. This was inspired by the Basque and MultiWordNet projects.
If a synset is marked as lexicalized False
then it means it has no
lemmas, and this is a deliberate decision on the part of the wordnet
builders (but the synset may be included to keep the hierarchy in sync
with other projects). For example the synset dedos "fingers and toes"
in Spanish would be lexicalized=false
in English.
If a sense has lexicalized True
then it has been validated in some
standard lexicon for the language. If it has lexicalized False
, then
it is believed to be compositional and only added as an aid to
multilingual users (similar to phrasets in multiwordnet). For example
harimau anak "young tiger" in the Indonesian synset for tiger cub is
lexicalized=false
, or dedos pedas "foot
finger-and-toes" in the synset for toe in Spanish.
The original documentation for the parts of speech in the Princeton Wordnet is here. We have added some new possibilities.
Source code hosted at https://github.com/omwn/omwn.github.io.