Part of Speech Tags 3

A frequently asked question is “What do the Part of Speech tags (VB, JJ, etc) mean?” The bottom line is that these tags mean whatever they meant in your original training data. You are free to invent your own tags in your training data, as long as you are consistent in their usage.

Training data generally takes a lot of work to create, so a pre-existing corpus is typically used. These usually use the Penn Treebank or Brown Corpus tags.

The most common part of speech (POS) tag schemes are those developed for the Penn Treebank and Brown Corpus. Penn Treebank is probably the most common, but both corpora are available with NLTK.

Penn Treebank POS Tags

Here are the POS tags used in the Penn Treebank:

POS Tag Description Example
CC coordinating conjunction and
CD cardinal number 1, third
DT determiner the
EX existential there there is
FW foreign word d’hoevre
IN preposition/subordinating conjunction in, of, like
JJ adjective big
JJR adjective, comparative bigger
JJS adjective, superlative biggest
LS list marker 1)
MD modal could, will
NN noun, singular or mass door
NNS noun plural doors
NNP proper noun, singular John
NNPS proper noun, plural Vikings
PDT predeterminer both the boys
POS possessive ending friend‘s
PRP personal pronoun I, he, it
PRP$ possessive pronoun my, his
RB adverb however, usually, naturally, here, good
RBR adverb, comparative better
RBS adverb, superlative best
RP particle give up
TO to to go, to him
UH interjection uhhuhhuhh
VB verb, base form take
VBD verb, past tense took
VBG verb, gerund/present participle taking
VBN verb, past participle taken
VBP verb, sing. present, non-3d take
VBZ verb, 3rd person sing. present takes
WDT wh-determiner which
WP wh-pronoun who, what
WP$ possessive wh-pronoun whose
WRB wh-abverb where, when

The official annotation guidelines including full descriptions can be found here (GZip-compressed Postscript file). This includes confusing parts of speech, capitalization, and other conventions.


Brown Corpus POS Tags

The Brown Corpus POS tags are very similar, and there is the potential for some confusion. However, there are differences. For example, the Penn Treebank has three types of adjective (JJ, JJR, JJS) but the Brown Corpus divides JJS into JJS and JJT.

The Brown Corpus also has rules for combining tags. For example, the colloquial “wanna” means “want to” and is tagged “VB+TO” (“want/VB to/TO”). Similarly, a suffix asterisk indicates a negative, so that “aren’t” becomes “BER*”.

The Brown Corpus manual is available here,and useful summaries can be found at the University of Leeds and at Wikipedia.

3 thoughts on “Part of Speech Tags

  1. Reply Subhan Ullah Nov 2,2013 4:52 am

    How i can use the NLP for the POS in PHP

    • Reply Richard Marsden Nov 2,2013 7:58 am

      You would need to find an NLP library for PHP. PHP isn’t really designed for things like computer learning and NLP type processing – I would use a different library, and call that from PHP. E.g. NLTK in Python or Stanford NLP for Java.

  2. Reply grokorg Feb 22,2016 6:22 pm

    PHP is a POS when it comes to NLP.

    I would advise you use Python to do the processing and possibly ZeroMQ to actually connect the two if you must. But that’s just how I like to roll things.

Leave a Reply