What are all possible pos tags of NLTK?

Question

How do I find a list with all possible pos tags used by the Natural Language Toolkit (nltk)?

Answer 1

The tag set depends on the corpus that was used to train the tagger. The default tagger of nltk.pos_tag() uses the Penn Treebank Tag Set.

In NLTK 2, you could check which tagger is the default tagger as follows:

import nltk
nltk.tag._POS_TAGGER
>>> 'taggers/maxent_treebank_pos_tagger/english.pickle'

That means that it's a Maximum Entropy tagger trained on the Treebank corpus.

nltk.tag._POS_TAGGER does not exist anymore in NLTK 3 but the documentation states that the off-the-shelf tagger still uses the Penn Treebank tagset.

Answer 2

You can download the list here: ftp://ftp.cis.upenn.edu/pub/treebank/doc/tagguide.ps.gz. It includes confusing parts of speech, capitalization, and other conventions. Also, wikipedia has an interesting section similar to this. Section: Part-of-speech tags used.