Language learning from an audio description corpus - Eva Schaeffer-Lacroix, Sorbonne Université/Espé de Paris - Faculty of Education
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Language learning from an
audio description corpus
Eva Schaeffer-Lacroix, Sorbonne Université/Espé de Paris
TaLC13 (Teaching and Language Corpora Conference), Faculty of Education, University of Cambridge
AD produced by future German teachers 2
Web soap for German
learners: Jojo sucht das
Glück [Jojo’s pursuit of
happiness]
Episode 12: It's coffee time
AD tool: Youdescribe.org
2018-07-20
Audio description, a highly normed text genre 3
Characteristics of ADs:
• Description of visual events
• Speaking indications
You can see… No, I can't!
Recommendations for AD authors:
• Be short
• Be precise
• Be objective
• Put yourself in the shoes of the visually impaired 2018-07-20
Expert AD corpus 4
be short compound words AD manuscripts from Neues
aus Buettenwarder, a series
present participles telling rural stories located
in northern Germany.
be precise prepositions
Corpus uploaded to Sketch
verb particles Engine and TXM.
wide range of verbs 336.723 tokens and 69 text
files
Part of speech annotations
with TreeTagger & RFtagger.
2018-07-20
Expert AD: Buettenwarder manuscript 5
10:01:46
"Ich bin beim Essen!" ["I am eating!"]
ss Adsche nimmt Griem den Teller weg. [Adsche takes the plate away from Griem.]
Reference points for recordings:
- time indications;
- film script prompts surrounded by quotation marks;
- speech indications highlighted in bold (e.g. "ss" : speak very quickly).
2018-07-20
Learner AD: Jojo manuscript 6
* = error
1.19 Aufnahme * auf dem 1.19 * on the
Tisch: (…) table: (…)
1.36 Jojo ist verlegen, Lena senkt den 1.36 Jojo is embarrassed, Lena
Blick. lowers her eyes.
1.50 Jojo zeigt * einen Teller. 1.50 Jojo * a plate.
2018-07-20
« Be short with compound words. » 7
Aufnahme auf dem Essen auf dem Tisch [Shot of the meal on the table]
Too long; preposition error
> Großaufnahme des Essens [> Close-up view of the meal]
= compound noun & genitive object
2018-07-20Compound words in the Buettenwarder corpus 8
2018-07-20« Be precise with prepositions. » 9
Jojo zeigt xxx einen Teller >
Jojo zeigt auf einen Teller
[to show a plat vs.
to point at a plate]
2018-07-20Zeigen [to show] in the Buettenwarder corpus 10
You can show… You can point…
a beer coaster, a "bird", your at/behind/in direction of sth or sb
thumb, a photograph, … and with sth.
2018-07-20« Be short with present participles. » ü 11
Vier lachende Jugendliche Four smiling young people
kopf schüt teln d [shaking t heir head]
grü belnd [ brood ing]
W o rd s p e r m illio n
sch munzelnd [ grinni ng]
lächelnd [smiling]
all p resent part iciples
0 50 100 150 200 250 300 350 400
Weissensee Buet te nwar der
2018-07-20« Vary your verbs. » ü
« Don’t use ‘see’. » ü 12
• erscheinen [to appear]
• bringen [to bring]
• servieren [to serve]
Jojo sucht das Glück
• trinken [to drink]
• 3x essen [to eat]
• 2x sein [to be]
• (den Blick) ab·wenden [to look away] sp a c e
37% 37%
• antworten [to answer]
• (den Blick) senken [to lower the eyes] m im ic a n d
• (auf etwas) zeigen [point at sth] p e rc e pt ion
• nehmen [to take]
o th e r
• etw auf etwas legen [to put sth on sth]
• sich zu jm wenden [to turn to sb] 26%
• lachen [to laugh]
• zu·blinzeln [to wink at sb]
• lächeln [to smile] 2018-07-20Next steps to do 13
• XML annotations for the expert corpus
è time indications
è speech indications
è compound words
è …
• Storing the Buettenwarder corpus on
Ortolang – CLARIN (rights granted by the
owners of the data)
2018-07-20Action-research with a research group and a
control group 14
Independent
RG task CG task variable Dependent variables
Design an audio Design an audio Type of tools - Type of language
description scenario description scenario with descriptions;
with the help of the help of other tools - type of created
corpus tools (e.g. online grammars, language learning
text books) activities.
2018-07-20References 15
• Eberlein, N. (1997-2017). Neues aus Büttenwarder. Television series.
• Heiden, S. (2010). The TXM Platform: Building Open-Source Textual Analysis Software Compatible with the TEI
Encoding Scheme. In R. Otoguro, K. Ishikawa, H. Umemoto, K. Yoshimoto & Y. Harada (eds), 24th Pacific Asia
Conference on Language, Information and Computation - PACLIC24 (pp. 389-398). Institute for Digital Enhancement of
Cognitive Development, Waseda University.
• Jojo sucht das Glück (n.d). Web soap for learners of German, produced by Deutsche Welle.
http://www.dw.com/de/deutsch-lernen/telenovela/s-13121
• Kilgarriff, A., Rychly, P. & Pomikalek, J. (nd). Sketch Engine. Corpus management system.
http://www.sketchengine.co.uk/
• Schiller, A., Thielen, C., Teufel, S. & Stöckert, C. (1995/1999). STTS (Stuttgart-Tübingen Tagset). http://www.ims.uni-
stuttgart.de/projekte/corplex/TagSets/stts-table.html
• Schmid, H. & Laws, F. (2008). Estimation of Conditional Probabilities with Decision Trees and an Application to Fine-
Grained POS Tagging. COLING 2008, Manchester, England.
• YouDescribe. (2017). Free online tool which can be used to add description to YouTube videos. Developed by The
Smith-Kettlewell Eye Research Institute.
2018-07-20You can also read
