1. Background
Grammar writers tend to just make up nonterminal symbols and
part-of-speech tags (or preterminal symbols) as they go along, relying on
native speaker intuition, context of use and tradition to provide their meaning.
I've attempted to pull together the symbols and tags I've used in this
course and given brief definitions and examples below.
Note that I have not attempted to cover the whole Penn
Treebank or Brown Corpus tagsets. Jurafsky & Martin have a
brief listing of version 1 of the Penn tags inside the front cover. For
complete lists, see
2. Grammar symbol glossary
2.1. Lexical categories
Sometimes called pre-terminals. The ones in
bold below are closed-class, that is, there is a
(small) fixed list of these in the language at any given time, and that changes rarely
if at all. Everything else is open-class.
- Adj
- Adjective: big, serious, yellow
- Adv
- Adverb: quickly, early
- Art
- Article: the, a, an
- Aux
- Auxiliary verb: did, have, will
- Mod
- Modal verb: must,
should, might
- N
- Noun: pet, democracy
- P, Prep
- Preposition: in, among, to
- Pro
- Pronoun: they, her, him
- PropN
- Proper Noun: Paris, Robin,
Appleton Tower
- V
- Verb: eat, embrace, give
- Wh, WhAdv
- Question word: who, why, how
2.2. Nonterminals
Some of these (marked with a *) are (sometimes) used as
terminals or nonterminals or a mixture of the two.
- AdjP
- Adjective phrase: very big, nearly new
- AdvP
- Adverbial phrase: very quickly, more easily
- D, Det
- Determiner*: the>, this,
those, the child's
- Nom
- Nominal: big house, snow storm
- NP
- Noun phrase*: Edinburgh,
the sleepy children on the bus, sheep
- PP
- Prepositional phrase: to the river, in
a ship
- S
- Sentence: the farmer killed the duckling
- VP
- Verb phrase: easily closed the window
3. Subscripts
In the absence of a grammar formalism which supports features, some
attempt to address issues of agreement and sub-categorisation can be made by
expanding the category space. In this course I've used subscripted categories
for this purpose, as well as to distinguish sub-types within a category.
- person-number agreement
- Occurring on e.g. D,
N, NP, V, VP:
sg (singular), pl (plural), possibly in combination with 1,
2, 3 (1st, 2nd, 3rd person), e.g.
NPpl →
Dpl Npl - verb sub-categorisation
- Occurring on e.g. VP,
V: i (intransitive), t (transitive),
dt (ditransitive), e.g.
VPdt →
Vdt NP NP - Sentence subtypes
- Occurring on S:
decl (declarative), imp (imperative), q
(question), qyn (yes-no question), whadv (why/how question)
- Noun subtypes
- Occurring on e.g. NP,
N: sc (singular count), mp (mass
or plural), e.g.
Nsc → 'child',
Nmp → 'rice', Nmp → 'children' - Case marking
- Occurring on e.g. NP,
Pro: nom (nominative), acc (accusative),
e.g.
VPt → Vt NPacc, Proacc
→ 'them'