Top Banner
23

Taddhita Generation

Apr 12, 2017

Download

Engineering

Amrith Krishna
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Taddhita Generation

Page 2: Taddhita Generation

SANSKRIT

PĀṆINI

AUTOMATION

Nullam eu tempor purus. Nunc a leo magna, sit amet consequat risus.

Page 3: Taddhita Generation

Evaluation

Conflict Resolution

Implementation

Objective

Related Work

Linguistic aspects of Taddhita

Page 4: Taddhita Generation

4 4

• Automation of

Taddhita section

• simulate the affixation

process

• Handling affix

polysemy,

homonymy and

synonymy.

• methods for rule selection, conflict

resolution and blocking of rules

• Effects in Taddhita section and in other

general cases.

• Features obtained by the entities

participating in derivation

• How it can be used for later derivations and

analysis.

Objective

Page 5: Taddhita Generation

Motivation

praises from domain experts in the fields of linguistics, mathematics and computer science, for the deep

computational insights it carries

Preserve etymological information of nouns so formed

Supplementary information to lexical databases like IndoWordnet

Pedagogy tool for learning

Page 6: Taddhita Generation

• •

Page 7: Taddhita Generation

Linguistic Aspects of Taddhita

माथुरा - मथुर औपगव - उपग ुआश्वऱायन - अश्वऱ नैपुण - ननपुण

m:n

Page 8: Taddhita Generation

Linguistic Aspects of Taddhita

Rules from sūtra A.4.1.76 to A.5.4.160

Semantic relations : arthādhikāra rules

Types of affixation Prātipadika + Taddhita-affix Prātipadika + Taddhita-affix + strī-affix

Default affix rules : pratyayādhikāra rules

Exception rules

An instance of inheritance hierarchy in Taddhita section

Page 9: Taddhita Generation

1-3-2 उपदेशे अच ्अनुनाससक् इत ्1-3-3 हऱ ्अन्त्यम ् 1-3-4 न ववभक्तौ तुसमा् 1-3-5 आदद् ञिटुडव् 1-3-6 ष् प्र्ययसय आदद् 1-3-7 चुटू 1-3-8 ऱशकु अतविते

3.3.1 प्र्यय् (Head) 4.1.76 तविता् (Head) 4.1.83 प्राक् दीव्यत् अण ्(Head, Default affix Rule) 4.1.92 तसय अप्यम ्(Default semantic rule) 4.1.123 शुभ्राददभ्य् च (Exception Rule)

.

Anuvṛtti and Adhikāra

Page 10: Taddhita Generation

उपग ुङस ्अप्यम ्

notify

notify

Conflicting RUles

Winner rule

Overview

Page 11: Taddhita Generation

sūtra

IVone76

IVone92

IVone123

+ sūtra text : String

+ registerd rules : List

+ check()

+ register()

+ notify()

+ anuvṛtti : String

+ execute()

+ taddhita : String

+ default checker()

Multilevel Inheritance & Observer Design Pattern

Page 12: Taddhita Generation

SHR – Same Head

Rules

TLR – Top Level Rules

R – Rule

RG – Rule Group

Rule Triggering Hierarchy

Rule triggering schema

Page 13: Taddhita Generation

Rule Triggering Input - चटका ङस ्अप्यम ्

Affixation under patronymic relation for चटका

Output - चटका ङस ्ऐरक्

Page 14: Taddhita Generation

Attain technical terms during derivation ->

Operational rules trigger

Helps in forming relation between nouns

Rich linguistic features get stored. -> Useful for new

derivations

Individual entity property remains with itself like the

‘it’ marker

Śabdarūpa Representation

Page 15: Taddhita Generation

Conflict Resolution and Rule Triggering

Utsarga-Apavada

General – Exception case

The inheritance network captures this notion

A.4.1.128 wins over A.4.1.92

Antaraṅga - Bahiraṅga

Bracketing

Internal is preferred

A.3.4.86 wins over A.6.1.77

Nitya – Anitya

Obligatory rule wins

A.6.1.77 wins over A.6.1.8

Para – Pūrva

Later rule emerges as the winner

The triggering by default enables pūrva

Internalizes the concept of A.1.4.2

Aṣṭādhyāyi provides limited information about conflict resolution

No general consensus among scholars

Implemented as a pluggable entity in the system

Page 16: Taddhita Generation

Specificity Hierarchy

Specificity Hierarchy

Priority wise ordering from the most concrete to most

abstract

Page 17: Taddhita Generation
Page 18: Taddhita Generation
Page 19: Taddhita Generation

गगग ङस ्गोत्र, कवप ङस ्गोत्र

Rules applied

A.4.1.151 and A.4.1.122

Desired rules

A.4.1.86 and A.4.1.85

Resolved

by specificity hierarchy

(refinement inside semantic

class)

वपतषृ्वस ृङस ्अप्यम ्

Multiple rules to be applied

A.4.1.133

Talks of elision with an affix

as condition

उ्स ङस ्अप्यम,् ददनत ङस ्अप्यम ्

Rules applied

A.4.1.95 and A.4.1.122

Desired rules

A.4.1.86 and A.4.1.85

Resolved

by specificity hierarchy

Rule A.4.1.122,

Need to know of stems with ‘इि’् pratyaya

beforehand

Another application of

our representation,

which stores this

information

Page 20: Taddhita Generation

The Schema as a General Schema for Modelling Aṣṭādhyāyī

Principles used in the system like Anuvṛtti, Adhikāra, Conflict resolution techniques etc. are Applicable to entire Aṣṭādhyāyī

Examples based on Scharf’s paper

A.6.1.87 आद्गुण्

A.6.1.88 ववृिरेञच ।

A.6.1.77 इको

यणञच ।

A.6.1.101 अक् सवणे दीर्ग् ।

A.7.3.111 रे्र्ङगनत ।

A.7.1.73 इकोऽञच ववभक्तौ ।

Internalises

Vipradhishedha

Case 3, Passes to Conflict

Resolution module

Resolved with

Specificity Hierarchy

Page 21: Taddhita Generation

h

Page 22: Taddhita Generation

Bottlenecks

Linguistic Features

Specificity Hierarchy

Metarules

• Panini uses rich linguistic features (Phonetic to Semantic)

• Intention of the speaker is one of them

• A.4-1-147 गो ि या् कु नेण च

• Multiple entities with different specificity

• Specificity for entitles in anuvritti

• Semantic entity in A.4.1.86

• Like Vipratishedha, other metarules

• System Internalized them

Page 23: Taddhita Generation

Generation of Correct forms, Sequence of rules

Preserve Etymological Information.

Pedagogy Tool.

Supplementary Information to

Lexical Databases

Handling Affix Synonymy, Homonymy and other features

Rule Selection, Conflict Resolution

Preserve rich linguistic features along with the noun object

Conclusion