Group: grammar

topics > communication > Group: natural language

attribute grammar
BNF grammar
case frame grammar
grammar modification
lexical analysis
pattern matching
syntax analysis
van Wijngaarden grammar
program representation

formal methods and languages
natural language as a system
semantic grammar
string transformation languages
structure transformation languages
Turing machine
words in natural languages


Grammars show the structure underlying sentences in a language. A grammar is unambiguous if each sentence has only one interpretation. Grammars may be formalized by a set of production rules turning a 'sentence' non-terminal into valid sentences of the language. Valid sentences are formed of terminal symbols. Formal grammars can be categorized as phase-structured, context sensitive, context free, and regular. Natural languages have not been described by formal grammars. Generative grammars are used which define deep structures underlying each surface sentence structure. During the design of Thesa, locally-determined grammars were developed. They are phases-structured grammars with ordered production rules. Sentences are unambiguously analyzed by a recursive descent over possible rules. (cbb 5/80)
Subtopic: grammar for natural language up

Quote: the linguist seeks a simple and revealing grammar that explains the ability to produce and understand new sentences and recognize ungrammatical ones; hopefully leads to a general theory [»chomN9_1956]
Quote: statistical approximation is irrelevant to grammar; e.g., 'colorless green ideas sleep furiously' is rare but grammatical [»chomN9_1956]
Quote: the class of grammatical sequences must be predetermined, but can't simply list all of the morphemes as done with phonemes [»quinWV8_1951]
Quote: a language always has a regular and abstract way of conveying distinctions of number, time, manner, space, hierarchy, difference [»fishS5_2005]
Quote: can use any formal devise to signal grammatical distinctions -- word order, word endings, prefixes, suffixes, numbers, brackets, fonts, colors [»fishS5_2005]

Subtopic: Astadhyayi, self-contained grammar up

Quote: Panini's mathematical grammar described the language as spoken, concisely defined by enumerations and rules; transmitted orally; transform from syntactic relationship to phonemic realization [»misrVN_1966]
Quote: Panini wanted a self-contained science that perfected grammatical description, without recourse to extraneous definition [»misrVN_1966]
Quote: Panini stood at the culmination point of a rich grammatical tradition; 7th-4th century B.C.; replaced the old systems [»misrVN_1966]
Quote: Panini's grammar Astadhyayi consists of general definition and rules, substitution and declension, primary suffixes, addable suffixes, morphoponemics, miscellaneous [»misrVN_1966]
Quote: Panini provides a full description of organized data from a living speech; algebraic without justification or extralinguistic definition; e.g., definition by ommission [»misrVN_1966]
Quote: to produce an utterance one has to refer back to general rules and forward to specific rules while reconstructing the rules from continuity, context, and exceptions [»misrVN_1966]
Quote: Panini's grammar, Astadhyayi, is richly interdependent; must understand the complete grammar; well preserved

Subtopic: generative grammar up

Quote: a sentence is words placed into a relationship with one another; generate 20 sentences from 5 random words [»fishS5_2005]
Quote: mimetic sign language is like spoken language: complex forms from a small number of discrete components in a shell-like structure [»newpEL_1982]

Subtopic: poetry as grammar up

Quote: oral poetry is like a second grammar that operates with the primary grammar of a language [»sowaJF_1984]

Subtopic: pidgin grammar up

Quote: pidgin grammars lack surface and morphological complexity. Pidgins use semantic transparency, a limited vocabulary, and limited function words [»sebbM_1997]

Subtopic: grammar as rules and symbols up

Quote: a language is a set of strings; a production is an ordered pair, a character and a nonempty string [»floyRW2_1964]
Quote: Frege does not use subject/predicate because it is part of the interaction between speaker and listener [»fregG_1879]
Quote: a grammar specifies a set of strings and a structure; different grammars can specify the same set of strings with different structures
Quote: a transformational grammar consists of a kernel of basic sentences and optional transformations to derived sentences; e.g., active to passive [»chomN9_1956]
Quote: limit kernel to a small set of simple, declarative sentences; defines content since transformations preserve meaning [»chomN9_1956]
QuoteRef: grieD_1971 ;;18 grammar is a set of rules using non-terminals and terminals where each non-terminal is defined in terms of non-terminals and terminals
QuoteRef: grieD_1971 ;;20 a language is a set of sentences which are sentential forms of only terminal symbols which are derivable from the grammar
QuoteRef: grieD_1971 ;;46 types are phrase structured, context sensitive, context free, regular
QuoteRef: grieD_1971 ;;46 grammars have a distinguished symbol from which language is produced

Subtopic: regular set up

Quote: any finite automaton is a regular set -- closed under disjunction, product, and repetition [»kleeSC_1956]

Subtopic: phrase-structure up

Quote: a phrase-structure grammar is a finite set of rewriting rules on a finite alphabet and initial strings; uses subset for terminals [»chomN9_1956]
Quote: simplify a phrase-structure grammar by ordering the rewriting rules and identifying obligatory rules [»chomN9_1956]

Subtopic: context sensitive up

Quote: a context-sensitive grammar for verb phrases is too complex [»chomN9_1956]

Subtopic: tree grammar up

Quote: formalization of XML Schema using tree grammars; named types and structural types; matching and validation [»simeJ1_2003]

Subtopic: LR(k) grammar up

Quote: an LR(k) grammar needs k characters of look-ahead [»knutDE6_1965]
Quote: LR(k) grammars are unambiguous; nearly every unambiguous grammar is LR(k) from left, right, or both; powerful test [»knutDE6_1965]
Quote: for many practical grammars, one character of context is sufficient; construct a table of character pairs and tables of first and last characters of each derivative for each character [»floyRW2_1964]
Quote: example of LR(1) grammar for reduced, fully parenthesized, '+' expressions [»knutDE6_1965]
Quote: ALGOL 60 is almost an LR(1) grammar; would make it unambiguous [»knutDE6_1965]
Quote: deterministic languages have an LR(k) and an LR(1) grammar, and vice versa [»knutDE6_1965]

Subtopic: unambiguous grammar up

Quote: every sentence of a programming language should have one and only one meaning; unambiguous grammar; each syntactic rule has a semantic rule
Quote: a bounded context grammar is unambiguous with syntactic analysis in linear time; appropriate for most programming languages [»floyRW2_1964]
Quote: a precedence grammar is unambiguous [»wirtN1_1966]
Quote: associative language descriptions (ALD) use permissible contexts instead of nonterminals; subset of context-free languages; unique ALD for a set of trees [»reghSC1_2000]
Quote: Backus used Post's productions to formalize IAL's grammar; simplified by Naur for ALGOL 60 [»backJ_1980]
QuoteRef: grieD_1971 ;;26 unambiguous grammars if all sentences have only one syntax tree

Subtopic: algorithms up

Quote: efficient computation of LALR(1) look-ahead sets; used Tarjan's algorithm for strongly-connected components [»dereF8_1979]
Quote: A->x|xAx is a pathological language; not bounded context; syntactic analysis is ad hoc, breaks down, or goes factorial [»floyRW2_1964]

Subtopic: deep vs. shallow structure -- semantic up

Quote: similar sentences may have different deep structures; e.g., 'I persuaded/expected John to leave' [»chomN_1965]
QuoteRef: goodJB_1968 ;;767 analyze expressions as a surface string (literal input), b) surface structure (isomorphism), c) deep structure (syntactic paraphrase), d) semantic interpretation (process). e) effective interpretation(result)
QuoteRef: goodJB_1968 ;;768 generative grammar-- a deep structure base with a surface and a semantic superstructure

Subtopic: finite-state grammar up

Quote: finite-state grammars are the simplest that generate an infinite number of sentences from a finite apparatus [»chomN9_1956]
Quote: English is not a finite-state language because 'If S then S' has an unbounded dependency set [»chomN9_1956]

Subtopic: data as grammar up

Quote: a data stream is described by a producing grammar and a consuming grammar which both describe the same result [»jackMA5_1976]
Quote: define symbol table with Abstract Syntax Description Language (ASDL); auto generate code for data structures [»hansDR8_1999]

Subtopic: locally determined grammar up

QuoteRef: cbb_1973 ;;9/10/75 locally determined grammars

Subtopic: problems with grammar up

Quote: formal systems have a high threshold for using them; e.g., define a grammar to express sequencing in dialogs [»myerB3_2000]
Quote: Panini assumed a thorough knowledge of Sanskrit and metalinguistic jargon

Group: grammar up

Topic: attribute grammar (9 items)
Topic: BNF grammar (9 items)
Topic: case frame grammar (5 items)
Topic: grammar modification (10 items)
Topic: lexical analysis (14 items)
Topic: pattern matching (42 items)
Topic: syntax analysis (29 items)
Topic: van Wijngaarden grammar
(9 items)

Related Topics up

Group: formalism   (9 topics, 478 quotes)
Group: patterns   (8 topics, 179 quotes)
Group: program representation   (25 topics, 659 quotes)

Topic: formal methods and languages (53 items)
Topic: natural language as a system (43 items)
Topic: semantic grammar (23 items)
Topic: string transformation languages (17 items)
Topic: structure transformation languages (7 items)
Topic: Turing machine (30 items)
Topic: words in natural languages
(40 items)

Updated barberCB 11/05
Copyright © 2002-2008 by C. Bradford Barber. All rights reserved.
Thesa is a trademark of C. Bradford Barber.