Difference between revisions of "Libraries and tools/Linguistics/Applicative universal grammar"

From HaskellWiki
Jump to navigation Jump to search
(→‎Implementing these idea: Giving another location for link to Mark P. Jones, Paul Hudak and Sebastian Shaumyan:``Using Types to Parse Natural Language'', because the first link may be sometimes busy)
((1): Splitting section `Introduction' to →‎Introduction: and →‎General resources: . (2): Writing introductory text. (3): Reference to Combinatory logic)
Line 3: Line 3:
 
== Introduction ==
 
== Introduction ==
   
  +
Applicative universal grammar (AUG) is a far-reaching approach in linguisics.
Bernard Paul Sypniewski's article [http://elvis.rowan.edu/~bps/ling/introAUG.pdf An Introduction to Applicative Universal Grammar].
 
  +
Seeing its main goals (and comparing it to other approaches), it could be characterised as
  +
;semiotic
  +
:because it does ''not'' follow a “grammar vs semantics” view, and has a relational approach to understand sign and meaning
  +
;explanatory
  +
:wanting not only to enable us to write better natural language processing projects, but also to understand better how language works. Also ultimate questions (like explaining why the languages of the word show both deep resemblences and difffereces, or why children can lern such a difficult thing like language) are tried to solve
  +
;universal
  +
:because it makes a difference -- it separates two levels in understanding how language works: a common ''genotype grammar'' is underlying in each language, enabling us to discover invariants and general resemblences, leaving the explanations of specific differences to the level called ''phenotype grammar''
  +
  +
These levels in understanding language are the followings:
  +
* a genotype grammar is common in all human languages -- AUG aims to find invariants, explain concepts abstracting away from conrete languages.
  +
* phenotype grammars are responsible special features of languages
  +
  +
Attribute universal grammar uses many thoughts from [[combinatory logic]].
  +
  +
References for this section:
 
* Bernard Paul Sypniewski's book [http://elvis.rowan.edu/~bps/ling/introAUG.pdf An Introduction to Applicative Universal Grammar]
  +
* Shaumyan's [http://citeseer.ist.psu.edu/shaumyan98two.html Two Paradigms of Linguistics: The Semiotic versus Non-Semiotic Paradigm]
  +
(see them also below).
  +
  +
== General resources ==
  +
  +
Bernard Paul Sypniewski's book [http://elvis.rowan.edu/~bps/ling/introAUG.pdf An Introduction to Applicative Universal Grammar]. This book does ''not'' require the reader to know [[combinatory logic]]. Mainly this book
   
 
The same author has also an [http://elvis.rowan.edu/~bps/ling/ling.htm entire homepage for the topic].
 
The same author has also an [http://elvis.rowan.edu/~bps/ling/ling.htm entire homepage for the topic].

Revision as of 20:09, 19 July 2006

Introduction

Applicative universal grammar (AUG) is a far-reaching approach in linguisics. Seeing its main goals (and comparing it to other approaches), it could be characterised as

semiotic
because it does not follow a “grammar vs semantics” view, and has a relational approach to understand sign and meaning
explanatory
wanting not only to enable us to write better natural language processing projects, but also to understand better how language works. Also ultimate questions (like explaining why the languages of the word show both deep resemblences and difffereces, or why children can lern such a difficult thing like language) are tried to solve
universal
because it makes a difference -- it separates two levels in understanding how language works: a common genotype grammar is underlying in each language, enabling us to discover invariants and general resemblences, leaving the explanations of specific differences to the level called phenotype grammar

These levels in understanding language are the followings:

  • a genotype grammar is common in all human languages -- AUG aims to find invariants, explain concepts abstracting away from conrete languages.
  • phenotype grammars are responsible special features of languages

Attribute universal grammar uses many thoughts from combinatory logic.

References for this section:

(see them also below).

General resources

Bernard Paul Sypniewski's book An Introduction to Applicative Universal Grammar. This book does not require the reader to know combinatory logic. Mainly this book

The same author has also an entire homepage for the topic.

As an article describing what AUG is, see also Shaumyan's Two Paradigms of Linguistics: The Semiotic versus Non-Semiotic Paradigm.

Fragments

Sebastian Shaumyan's long answers to comments in Disc: Ungrammatical Sentences gives examples, thus can help at understanding the above mentioned materials -- by discussing the notions like meaning, grammatical vs syntactic etc.

Some other citings and fragments from Sebastian Shaumyan can be read here.

Details

Long-Distance Dependencies and Applicative Universal Grammar (written by Sebastian Shaumyan and Frédérique Segond) compares (presenting some advantages of the latter)

  • combinatory categorial grammar
  • applicative universal grammar

Linguistic Types and the Valence of Operators in Applicative Universal Grammar written by Bernard Paul Sypniewski. In case of problems, here is an alternative link.

Implementing these ideas

A Haskell application for natural language parsing, based on Applicative Universal Grammar (AUG) is described in Mark P. Jones', Paul Hudak's and Sebastian Shaumyan's Using Types to Parse Natural Language (if this link is busy, here is another location). The Haskell source code given by the article is full, it can be run by Gofer, and after a few modification, by GHC too (transpose must be explictly imported from standard library module Data.List, and class Text renamed to Show).

A more detailed description of the topic of this previous article described in Sebastian Shaumyan and Paul Hudak's Linguistic, Philosophical, and Pragmatic Aspects of Type-Directed Natural Language Parsing