Contextualising concordances for corpusCALL - PowerPoint PPT Presentation

Loading...

PPT – Contextualising concordances for corpusCALL PowerPoint presentation | free to view - id: 13e02b-MmQzM



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Contextualising concordances for corpusCALL

Description:

... drank several glasses of whiskey, brandy, champagne, and at least one high ... Yet, consumerism lived, even if it didn't fill too many champagne glasses. ... – PowerPoint PPT presentation

Number of Views:35
Avg rating:3.0/5.0
Slides: 35
Provided by: edme4
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Contextualising concordances for corpusCALL


1
Contextualising concordances for corpusCALL
  • Hans Paulussen Piet Desmet
  • K.U.Leuven / KULAKALT Research Center on CALL

2
Overview
  • Corpora for CALL samples
  • Types of sample rendition
  • XML new opportunities

3
Corpora for CALL /1
  • Corpora for learning activities
  • before preparation of exercises
  • during corpus material as part of the learning
    activity
  • after corpus material for feedback

4
Corpora for CALL /2
  • Corpora as reference material
  • learner dictionaries
  • learner grammars

5
Corpora during learning activities
  • corpus is part of learning activity
  • Mariana (Vordingburg Gymnasium, Denmark)
  • http//www.vordingbg-gym.dk/km/ict4lt/
  • corpus supports learning activity
  • NEDERLEX (FUNDP Namur)
  • http//obelix.droit.fundp.ac.be/droit1/index.php

6
(No Transcript)
7
REBECA
  • Ressources Electroniques Bilingues Extraites de
    Corpus Alignés (bilingual electronic ressources
    extracted from aligned corpora)
  • Parallel corpus5,000,000 Dutch ? 5,000,000
    French
  • automatic corpus selection
  • sentence alignment

8
(No Transcript)
9
Resource links
10
http//corpora.informatik.uni-leipzig.de/
11
Drinking glasses /1
  • At the winery, Evxinograd's director, Ivan
    Penkov, is pouring out glasses of his 20-year-old
    brandy. (source Wall Street Journal 1991)
  • Drinking glasses are plastic and are weighted on
    the bottom so that they are tough to knock over.
    (source Wall Street Journal 1991)
  • Noting that Winston Churchill regularly drank
    several glasses of whiskey, brandy, champagne,
    and at least one high-ball during the working
    day, the Economist observed on March 4, "he could
    never have been trusted to run the Pentagon."
    (source Wall Street Journal 1989)
  • The two drain their glasses in one gulp. (source
    Wall Street Journal 1991)

12
Drinking glasses /2
  • Yet, consumerism lived, even if it didn't fill
    too many champagne glasses. (source Wall Street
    Journal 1988)
  • After Investcorp took over, Tiffany played down
    its 10 wine glasses to concentrate on the
    high-priced diamonds and gold jewelry that had
    made it famous. (source Wall Street Journal
    1990)
  • At a black-tie benefit hosted by ARA Services
    Inc. a few weeks ago, Chairman Joseph Neubauer
    and members of his management team exuded
    confidence as they moved from one dinner table to
    the next, shaking hands, patting backs and
    clinking glasses. (source Wall Street Journal
    1988)

13
Spectacles
  • Mr. Brown wears tinted aviator glasses, combat
    boots, and a Soldier of Fortune cap on his
    closely shaved head. (source Wall Street Journal
    1991)
  • She must wear prism glasses to correct double
    vision caused by the accident. (source Wall
    Street Journal 1991)
  • In meetings, he often can be seen chewing on the
    end of his reading glasses sometimes, he speaks
    so softly that he can't be heard. (source Wall
    Street Journal 1990)
  • His horn-rimmed glasses and rakish beret were
    irresistibly photogenic. (source Wall Street
    Journal 1987)

14
Problematic glasses
  • And its replaceable filters are good for only
    about 100 glasses. (source Wall Street Journal
    1991)
  • I still have to find his glasses and keys for
    him. (source Wall Street Journal 1988)
  • The police confiscated her watch and glasses.
    (source Wall Street Journal 1989)
  • He plans to hand out 100 glasses when he performs
    in Washington, D.C., in December at the Kennedy
    Center's Mozart Festival. (source Wall Street
    Journal 1991)
  • He often hands out glasses to his audience and
    has them play chords. (source Wall Street
    Journal 1991)
  • The glasses were my idea. (source Wall Street
    Journal 1988)

15
Rendering authentic text samples
  • Linked samples
  • Extracted samples
  • Embedded samples

16
Linked samples
  • The sample is linked to the original document
    (e.g. pdf document)
  • Original context layout
  • Full context
  • Problem sample skimming

17
Extracted samples
  • The example is extracted from the original
    document (e.g. KWIC concordance)
  • Sample shown in immediate context
  • Layout not authentic
  • Context limited

18
Embedded samples
  • The example is embedded in the original document
  • Sample shown in full context
  • Layout authentic
  • Problem recreating and indexing the document

19
XML -gt XHTML
  • XML extensible markup language
  • Stylesheets
  • CSS cascading style sheet
  • XSLT XML style sheet transformations
  • XHTML

20
Web reinvents standardisation
  • SGML standard generalized markup language (1968
    ISO in 1986)
  • HTML hypertext markup language (1993)
  • XML extensible markup language (1998)
  • XHTML extensible HTML

21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
lt?xml version "1.0" encoding"ISO-8859-1"?gt lt!DOC
TYPE poème SYSTEM "poemfr.dtd"gt ltpoèmegt ltpréambule
gt lttitregtChanson d'automnelt/titregt ltrecueilgtPoèmes
saturnienslt/recueilgt ltdategt1866lt/dategt ltauteurgtPa
ul Verainelt/auteurgt lt/préambulegt ltcorpsgt ltstancegt
ltlignegtLes sanglots longslt/lignegt ltlignegtDes
violonslt/lignegt ltlignegtltr/gtDe l'automnelt/lignegt ltl
ignegtBlessent mon coeurlt/lignegt ltlignegtD'une
langueurlt/lignegt ltlignegtltr/gtMonotone.lt/lignegt lt/st
ancegt
26
poem.dtd
lt?xml version"1.0" encoding"ISO-8859-1"?gt lt!--
poemfr.dtd DTD pour poésie M. Goossens
--gt lt!ELEMENT poème (préambule,
corps)gt lt!ELEMENT préambule (titre, recueil?,
date?, auteur)gt lt!ELEMENT titre
(PCDATA)gt lt!ELEMENT recueil
(PCDATA)gt lt!ELEMENT date
(PCDATA)gt lt!ELEMENT auteur
(PCDATA)gt lt!ELEMENT corps
(stanceligne)gt lt!ELEMENT stance
(ligne)gt lt!ELEMENT ligne (PCDATAr)gt lt!ELE
MENT r EMPTYgt
27
xpath
xpath -e '///stancecontains(., "langueur")'
Verlaine1.xml Found 1 nodes in
Verlaine1.xml -- NODE -- ltstancegt ltlignegtLes
sanglots longslt/lignegt ltlignegtDes
violonslt/lignegt ltlignegtltr /gtDe l'automnelt/lignegt lt
lignegtBlessent mon coeurlt/lignegt ltlignegtD'une
langueurlt/lignegt ltlignegtltr /gtMonotone.lt/lignegt lt/s
tancegt
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
Conclusion
  • Recreating an authentic document containing
    indexed samples is feasible
  • At what cost?
  • Full control of production cycle
  • Text and images?
  • Optimisation of on-the-fly rendition
About PowerShow.com