English Grammar Specification

This document provides a comprehensive formal specification of English grammar as implemented in Nasty. It serves as the authoritative reference for the English language parser implementation.

Table of Contents

1. Part-of-Speech Tags
2. Phrase Structure Rules
3. Dependency Relations
4. Morphological Features
5. Sentence Types
6. Lexical Categories

Part-of-Speech Tags

Nasty uses the Universal Dependencies (UD) part-of-speech tagset.

Open Class Words

Closed Class Words

POS Tag Mapping in Code

In lib/language/english/pos_tagger.ex, tags are represented as atoms:

:adj, :adv, :noun, :propn, :verb,
:adp, :aux, :cconj, :det, :intj,
:num, :part, :pron, :punct, :sconj, :x

Phrase Structure Rules

Nasty uses a simplified Context-Free Grammar (CFG) for phrase parsing.

Formal CFG Rules

# Sentence Level
S    → CLAUSE+ PUNCT
SENT → MAIN_CLAUSE COORD_CLAUSE* SUBORD_CLAUSE*

# Clause Level
MAIN_CLAUSE   → NP VP
COORD_CLAUSE  → CCONJ NP VP
SUBORD_CLAUSE → SCONJ NP? VP

# Phrase Level
NP   → DET? ADJ* (NOUN | PROPN | PRON) PP* RC*
VP   → AUX* VERB NP? PP* ADVP*
PP   → ADP NP
ADJP → ADV? ADJ
ADVP → ADV+
RC   → REL_PRON_ADV CLAUSE

# Terminal Symbols
DET       → the | a | an | this | that | some | my | ...
ADJ       → big | small | happy | fast | ...
NOUN      → cat | dog | tree | idea | ...
PROPN     → London | Mary | Monday | ...
PRON      → I | you | he | she | it | we | they | ...
VERB      → run | eat | think | walk | ...
AUX       → be | have | do | will | can | should | ...
ADP       → in | on | at | by | for | with | ...
CCONJ     → and | or | but | ...
SCONJ     → because | if | when | while | ...
ADV       → very | quickly | often | well | ...
REL_PRON_ADV → who | whom | whose | which | that | where | when | why

Phrase Structure Detailed Specifications

Noun Phrase (NP)

Structure: DET? ADJ* HEAD POST_MOD*

Components:

1. Determiner (optional): DET

   • Articles: the, a, an
   • Demonstratives: this, that, these, those
   • Possessives: my, your, his, her, its, our, their
   • Quantifiers: some, any, every, each, all, both, many, much, few, several

2. Pre-modifiers (0 or more): ADJ*

   • Adjectives: big, old, happy
   • PROPN (for multi-word names): New in "New York"

3. Head (required): NOUN | PROPN | PRON

   • Common noun: cat, tree, happiness
   • Proper noun: London, Mary
   • Pronoun: I, you, he, she, it, we, they

4. Post-modifiers (0 or more): PP | RC

   • Prepositional phrases: on the mat, in the house
   • Relative clauses: that sits, who I know

Examples:

"the cat"                    → [DET, NOUN]
"the big cat"                → [DET, ADJ, NOUN]
"the big black cat"          → [DET, ADJ, ADJ, NOUN]
"the cat on the mat"         → [DET, NOUN, PP]
"the cat that sits"          → [DET, NOUN, RC]
"New York"                   → [PROPN, PROPN]
"I"                          → [PRON]

Verb Phrase (VP)

Structure: AUX* MAIN_VERB COMPLEMENT*

Components:

1. Auxiliaries (0 or more): AUX*

   • Be: am, is, are, was, were, be, been, being
   • Have: have, has, had, having
   • Do: do, does, did
   • Modals: will, would, shall, should, can, could, may, might, must, ought

2. Main Verb (required): VERB

   • Action: run, eat, write, think
   • State: be, seem, appear, know

3. Complements (0 or more):

   • Direct object (NP): the cat
   • Prepositional phrase (PP): on the mat
   • Adverbial phrase (ADVP): quickly, very fast

Examples:

"runs"                       → [VERB]
"is running"                 → [AUX, VERB]
"will have been running"     → [AUX, AUX, AUX, VERB]
"eats the food"              → [VERB, NP]
"sits on the mat"            → [VERB, PP]
"runs quickly"               → [VERB, ADVP]
"gave the book to Mary"      → [VERB, NP, PP]

Copula Constructions:

When no main verb is present, the last auxiliary serves as the main verb:

"is happy"        → [AUX-as-VERB, ADJP]
"are engineers"   → [AUX-as-VERB, NP]
"was in the house" → [AUX-as-VERB, PP]

Prepositional Phrase (PP)

Structure: PREPOSITION NP

Prepositions:

• Location: in, on, at, inside, above, below, behind, beside, between
• Direction: to, from, toward, into, through, across, along
• Time: at, on, in, during, before, after, since, until
• Other: of, by, with, for, about, without

Examples:

"on the mat"      → [ADP, NP("the mat")]
"in the house"    → [ADP, NP("the house")]
"to New York"     → [ADP, NP("New York")]
"with a smile"    → [ADP, NP("a smile")]

Adjectival Phrase (ADJP)

Structure: INTENSIFIER? ADJ

Intensifiers (optional): ADV

• Degree: very, quite, rather, too, so, extremely, incredibly

Examples:

"happy"           → [ADJ]
"very happy"      → [ADV, ADJ]
"quite sad"       → [ADV, ADJ]
"extremely fast"  → [ADV, ADJ]

Adverbial Phrase (ADVP)

Structure: ADV+

Currently implemented as single adverbs. Future versions may support:

• very quickly (degree + manner)
• right here (directional + locative)

Examples:

"quickly"     → [ADV]
"very well"   → [ADV, ADV] (not yet supported)

Relative Clause (RC)

Structure: RELATIVIZER CLAUSE

Relativizers:

• Relative pronouns: who, whom, whose, which, that
• Relative adverbs: where, when, why

Clause Patterns:

1. Relativizer as subject: REL_PRON VP

   • "that sits" → [that, [VP: sits]]

2. Relativizer as object: REL_PRON NP VP

   • "that I see" → [that, [NP: I] [VP: see]]

Examples:

"that sits"              → [REL_PRON, VP("sits")]
"who I know"             → [REL_PRON, NP("I"), VP("know")]
"which is on the table"  → [REL_PRON, VP("is on the table")]
"where we met"           → [REL_ADV, NP("we"), VP("met")]

Dependency Relations

Nasty uses Universal Dependencies relation taxonomy.

Core Arguments

Non-Core Dependents

Nominal Dependents

Coordination

MWE and Other

Special

Dependency Extraction Rules

From phrase structures to dependencies:

From NP

NP(determiner=D, head=H, modifiers=[M1, M2], post_modifiers=[PP])
→ det(H, D)
  amod(H, M1)
  amod(H, M2)
  [dependencies from PP with H as governor]

From VP

VP(auxiliaries=[A1, A2], head=V, complements=[NP, PP, ADVP])
→ aux(V, A1)
  aux(V, A2)
  obj(V, NP.head)
  [dependencies from NP]
  [dependencies from PP with V as governor]
  advmod(V, ADVP.head)

From PP

PP(head=P, object=NP)
→ case(NP.head, P)
  [with governor G:]
    obl(G, NP.head)    # if G is verb
    nmod(G, NP.head)   # if G is noun
  [dependencies from NP]

From Clause

Clause(subject=NP_subj, predicate=VP, subordinator=S)
→ nsubj(VP.head, NP_subj.head)
  [dependencies from NP_subj]
  [dependencies from VP]
  mark(VP.head, S)  # if subordinator present

From Relative Clause

RelativeClause(relativizer=R, clause=C, attached_to=N)
→ mark(C.predicate.head, R)
  acl(N, C.predicate.head)
  [dependencies from C]

Morphological Features

Nasty tracks morphological features for each token based on Universal Features.

Verb Features

Implementation (in code):

%{
  tense: :past | :present | :future,
  aspect: :progressive | :perfect,
  person: 1 | 2 | 3,
  number: :singular | :plural
}

Noun Features

Implementation:

%{number: :singular | :plural}

Adjective Features

Implementation:

%{degree: :positive | :comparative | :superlative}

Morphological Rules

Verb Inflection

Regular Verbs:

Base form:     walk
3rd sg present: walk + s → walks
Past:          walk + ed → walked
Progressive:   walk + ing → walking
Past participle: walk + ed → walked

Irregular Verbs (dictionary lookup):

go → went (past), gone (past participle), going (progressive)
be → am/is/are (present), was/were (past), been (past participle)
have → has (3sg present), had (past)

Noun Inflection

Regular Plurals:

cat → cats
box → boxes (after s/x/z/ch/sh)
fly → flies (y → ies after consonant)

Irregular Plurals (dictionary lookup):

child → children
man → men
woman → women
tooth → teeth
mouse → mice

Adjective Inflection

Regular Comparison:

fast → faster → fastest
big → bigger → biggest (consonant doubling)
happy → happier → happiest (y → i)

Irregular Comparison (dictionary lookup):

good → better → best
bad → worse → worst
far → farther/further → farthest/furthest

Sentence Types

By Function

Function Inference:

• . → Declarative
• ? → Interrogative
• ! → Exclamative (or Imperative if no subject)

By Structure

Structure Determination:

• Simple: 1 independent clause, 0 subordinate clauses
• Compound: 2+ independent clauses, 0 subordinate clauses
• Complex: 1 independent clause, 1+ subordinate clauses
• Compound-Complex: 2+ independent clauses, 1+ subordinate clauses
• Fragment: 0 independent clauses (only subordinate)

Clause Types

Lexical Categories

Closed-Class Word Lists

These are finite sets of words that rarely change.

Determiners (60+ words)

Articles:

the, a, an

Demonstratives:

this, that, these, those

Possessives:

my, your, his, her, its, our, their, whose

Quantifiers:

some, any, no, every, each, either, neither
much, many, more, most, less, least
few, several, all, both, half

Pronouns (50+ words)

Personal (Subject):

I, you, he, she, it, we, they

Personal (Object):

me, you, him, her, it, us, them

Possessive:

mine, yours, his, hers, its, ours, theirs

Reflexive:

myself, yourself, himself, herself, itself,
ourselves, yourselves, themselves

Interrogative:

who, whom, whose, which, what

Demonstrative:

this, that, these, those

Indefinite:

someone, somebody, something
anyone, anybody, anything
everyone, everybody, everything
no one, nobody, nothing

Prepositions (50+ words)

Common Prepositions:

in, on, at, by, for, with, from, to, of, about
above, across, after, against, along, among, around
before, behind, below, beneath, beside, between, beyond
down, during, except, inside, into, like, near
off, over, past, since, through, throughout, till
toward, under, underneath, until, up, upon, within, without

Conjunctions

Coordinating Conjunctions (7 words):

and, or, but, nor, yet, so, for

Subordinating Conjunctions (30+ words):

after, although, as, because, before, if, once, since
than, that, though, till, unless, until, when, whenever
where, wherever, whether, while

Auxiliaries (20+ words)

Be:

am, is, are, was, were, be, been, being

Have:

have, has, had, having

Do:

do, does, did, doing

Modals:

will, would, shall, should
can, could, may, might, must, ought

Adverbs (100+ words)

Manner:

well, badly, carefully, quickly, slowly, easily

Frequency:

always, never, often, sometimes, usually, rarely, seldom

Time:

now, then, soon, later, already, yet, still, just

Place:

here, there, everywhere, nowhere, anywhere, somewhere

Degree:

very, really, quite, rather, too, so, enough

Interrogative:

how, why, when, where

Conjunctive:

however, therefore, moreover, furthermore,
nevertheless, nonetheless, besides, otherwise

Particles (10+ words)

Common Particles:

to (infinitive marker)
up, down, out, off, in, on, away, back

Interjections

Common Interjections:

ah, oh, wow, hey, hi, hello, goodbye, bye
yes, no, yeah, nope
thanks, please, sorry
ouch, oops, ugh, hmm, huh

Common Verbs (Top 150)

High-Frequency Verbs:

be, have, do, say, go, get, make, know, think, take
see, come, want, use, find, give, tell, work, call, try
ask, need, feel, become, leave, put, mean, keep, let, begin
seem, help, show, hear, play, run, move, like, live, believe
bring, happen, write, sit, stand, lose, pay, meet, include, continue
set, learn, change, lead, understand, watch, follow, stop, create, speak
read, spend, grow, open, walk, win, teach, offer, remember, consider
appear, buy, serve, die, send, build, stay, fall, cut, reach
kill, raise, pass, sell, decide, return, explain, hope, develop, carry
break, receive, agree, support, hit, produce, eat, cover, catch, draw

Common Adjectives (Top 100)

High-Frequency Adjectives:

good, bad, big, small, large, little, new, old, young, long
short, high, low, great, right, left, different, same, next, last
early, late, public, important, able, free, real, sure, certain, wrong
ready, clear, white, black, red, blue, green, hot, cold, open
happy, sad, easy, hard, strong, weak, full, empty, rich, poor
heavy, light, fast, slow, clean, dirty, safe, dangerous, cheap, expensive
quiet, loud, wide, narrow, deep, shallow, thick, thin, bright, dark
soft, hard, smooth, rough, wet, dry, simple, complex, common, rare
perfect, terrible, beautiful, ugly, wonderful, awful, excellent, fine, nice, special

Common Nouns (Top 100)

High-Frequency Nouns:

time, person, year, way, day, thing, man, world, life, hand
part, child, eye, woman, place, work, week, case, point, government
company, number, group, problem, fact, people, water, room, money, story
book, word, question, school, state, family, student, system, program, teacher
house, home, office, door, car, street, city, country, name, area
idea, body, face, food, job, night, power, end, side, week
mother, father, friend, girl, boy, business, service, health, law, level
hour, game, line, member, mind, minute, music, party, result, death

Ambiguity and Disambiguation

Common Ambiguities

1. POS Ambiguity:

   • book → NOUN ("I read a book") or VERB ("Book a flight")
   • fast → ADJ ("He is fast") or ADV ("He runs fast")
   • that → DET ("that book"), PRON ("I see that"), or SCONJ ("I know that he left")

2. PP Attachment:

   • "I saw the man with a telescope"
     • Attach to VP: I used a telescope to see the man
     • Attach to NP: The man had a telescope

3. Coordination Scope:

   • "old men and women"
     • [old men] and [women]
     • [old [men and women]]

Disambiguation Strategies

1. Lexical Lookup Priority: Check closed-class lists first
2. Morphological Cues: Use suffixes to infer POS
3. Contextual Rules: Use local context (e.g., word after DET is likely NOUN)
4. Statistical Models: Use HMM or neural models for better accuracy
5. Selectional Preferences: Verbs prefer certain argument types
6. Semantic Plausibility: More plausible interpretations preferred

Grammar Extensions (Future)

Planned extensions to the grammar:

1. Comparative Constructions: "John is taller than Mary"
2. Passive Voice: "The cat was seen by Mary"
3. Wh-Questions: "What did you see?"
4. Ellipsis: "John likes apples and Mary [likes] oranges"
5. Coordination: Better handling of coordinated phrases
6. Negation: Explicit negation marking
7. Modality: Modal auxiliary semantics
8. Aspect: Progressive, perfect aspect marking

References

• Universal Dependencies: https://universaldependencies.org/
• Penn Treebank: Marcus et al. (1993)
• Cambridge Grammar of English: Huddleston & Pullum (2002)
• English Grammar: Quirk et al. (1985)
• Universal Features: https://universaldependencies.org/u/feat/
• UD English-EWT Treebank: https://github.com/UniversalDependencies/UD_English-EWT