Testing natural language use insights from naturalistic experimental - - PowerPoint PPT Presentation

1/ 13

Experimental traditions The current approach Comparison of results Conclusion

Testing natural language use

insights from naturalistic experimental paradigms Katerina Danae Kandylaki

University of Marburg, Germany

• 26. June 2015

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

2/ 13

Experimental traditions The current approach Comparison of results Conclusion

Outline

1 Experimental traditions 1

“controlled stimuli”

2

“ecological validity”

2 The “naturalistic” approach 3 Comparison of results 4 Conclusion & Take home message ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

3/ 13

Experimental traditions The current approach Comparison of results Conclusion

Controlled stimuli

Carefully created/selected stimuli Example experiment: Domahs et al. (2013)

stress violation, “Vitamin” Carrier sentence: Er soll nun *(’vi.ta)(min) vs. *vi(’ta.min) vs. (vi.ta)(’min) sagen

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

3/ 13

Experimental traditions The current approach Comparison of results Conclusion

Controlled stimuli

Carefully created/selected stimuli Example experiment: Domahs et al. (2013)

stress violation, “Vitamin” Carrier sentence: Er soll nun *(’vi.ta)(min) vs. *vi(’ta.min) vs. (vi.ta)(’min) sagen

Unnatural setup, decontextualised language

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

4/ 13

Experimental traditions The current approach Comparison of results Conclusion

Why is decontextualisation a problem?

Isolating language from its natural environment: Theoretical problem of scaling the results Disorders’ symptoms often occur only in context Applications in a noisy environment: maybe studying language use improves applications

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

5/ 13

Experimental traditions The current approach Comparison of results Conclusion

Ecological validity

inspired from recent advances in cognitive neuroscience Example experiment: Bartels & Zeki (2004)

natural viewing of movies no a priori hypotheses “events” automatically extracted from the movie features

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

5/ 13

Experimental traditions The current approach Comparison of results Conclusion

Ecological validity

inspired from recent advances in cognitive neuroscience Example experiment: Bartels & Zeki (2004)

natural viewing of movies no a priori hypotheses “events” automatically extracted from the movie features

paradigms using language in context / as natural as possible

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

6/ 13

Experimental traditions The current approach Comparison of results Conclusion

The “naturalistic” approach

hybrid between completely controlled designs and completely natural tightly controlled stimuli embedded into natural context Benefits: design and statistical procedures as in “controlled stimuli” no decontextualisation, higher ecological validity no explicit task

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

7/ 13

Experimental traditions The current approach Comparison of results Conclusion

Naturalistic paradigm: story listening

20 two-minute long stories

Story 105 – Zirkus Der Clown Valentino genießt das Leben im Zirkus 'Fiforello', da sie viel herumreisen. Jedoch will der Zirkus bald auch ein festes Zirkuszelt für die Winterzeit bauen. Jetzt benötigen sie nur noch ein passendes BaugrundstückCLASH und schon kann der Bau

• beginnen. Neulich trugACT-HIGH der Clown bei einem Auftritt die Prinzessin auf Stelzen durch

die Manege, was die Zuschauer sehr amüsiert hat. Der Clown machte dabei lustige Fratzen, griff in sein rotes Stoffsäckchen und warf der Dame in der ersten Reihe ein kleines HolzspielzeugSHIFT zu. Im improvisierten Sketch nahm der Clown einen Kuschelbär, eine Zahnbürste und ein Kissen aus seinem Säckchen. Als er an den Rand der Manege kam und vorgab, sich die Zähne zu putzen, stand er mit dem Rücken zur Prinzessin. In dem Moment klaute die Prinzessin den Kuschelbär und versteckte ihn unter ihrem Kleid. Als der Clown sich ins Bett legen wollte, konnte er sein Kuscheltier nicht finden. Und weil er so verwirrt aussah,TOM mussten alle Zuschauer herzlich lachen. Dann wurde der Clown von der Prinzessin gestoßenPASS-HIGH und unter dem Kleid kam der Kuschelbär wieder

• hervor. Der Clown freute sich, sein Kuscheltier wiedergefunden zu haben! Dann nahm er

einen KohlemalstiftLAPSE aus seiner Tasche und fing an, etwas zu zeichnen. Als er seine Zeichnung dem Publikum zeigte, schien ein Zuschauer sehr fasziniert davon. Der Clown ging näher zu ihm und unterhielt sich mit ihm über das Kunstwerk.NONTOM Der Zuschauer mochteACT-LOW den Clown, und weil die Zeichnung ein ganz skurriles Porträt der Prinzessin war, wurde die Diskussion sehr lustig. Schließlich stand der Zuschauer auf und brachte das ganze Publikum zum Klatschen. Zum Dank schenkte der Clown dem Zuschauer die Zeichnung.Der Zuschauer wurde von dem Clown für seine freundliche, mitreißende Art geschätztPASS-LOW. Das Publikum klatschte und johlte begeistert. Der Zuschauer war nämlich vielen aus den Medien als der LandesvolkswirtNOSHIFT von Hessen bekannt.

Reference tracking Discourse Semantics Syntax Phonology

false-beliefs verb causality grammatical voice rhythm ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

8/ 13

Experimental traditions The current approach Comparison of results Conclusion

Phonological rhythm manipulation

time beat strength strong weak strong weak time beat strength strong weak weak well-formed rhythm ill-formed rhythm ill-formed rhythm time beat strength strong weak strong weak strong strong strong strong

no phonological task

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

9/ 13

Experimental traditions The current approach Comparison of results Conclusion

Analysis

In order to achieve a “clean” baseline, we modelled:

1 Contrasts of interest (full factorial design) 1

well-formed (SHIFT, NOSHIFT)

2

ill-formed (CLASH, LAPSE)

2 Regressors of no interest 1

the rest of the story (speech excluding manipulation)

2

question reading

3

answer reading

4

responses

5

jitters

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

10/ 13

Experimental traditions The current approach Comparison of results Conclusion

Results

Domahs et al. (2013) Kandylaki et al. (in prep.) task decide whether auditory probe stressed correctly listen carefully to stories (for answering questions) results

correct vs. violation mild vs. severe

billateral SMA premotor cortex rIFG

LAPSE vs. CLASH

Post-central gyrus STG Insula

LAPSE vs. NOSHIFT ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

11/ 13

Experimental traditions The current approach Comparison of results Conclusion

Conclusion

Our results

replicated previous findings established their ecological validity added new brain regions to the findings

Contribution of naturalistic designs: free environment brain “in action”, how it categorises language language in interaction with other cognitive systems (neurobiological basis of language)

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

12/ 13

Experimental traditions The current approach Comparison of results Conclusion

Take home message

Investigating brain processes of language in action can inform linguistic theory

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece

Thanks to my advisors, colleagues, collaborators

Fiona Weiß Alexander Dröge Miriam Burk

• Dr. Jona Sassenhagen
• Dr. Jens

Sommer

• Dr. Marie

Josephine Rocholl

• Dr. Phillip Alday
• Prof. Dr. Ina Bornkessel-

Schlesewsky

• Dr. Johannes Kanus
• Dr. Karen

Henrich

• Prof. Dr. Tilo

Kircher

• Prof. Dr.

Richard Wiese

• Prof. Dr. Ulrike

Domahs

• Dr. Arne Nagels

Thank you all! Ευχαριστώ! Dankeschön!

13/ 13

13/ 13

Experimental traditions The current approach Comparison of results Conclusion

Bartels, A., & Zeki, S. (2004). Functional brain mapping during free viewing of natural scenes. Human brain mapping, 21(2), 75-85. Domahs, U., Klein, E., Huber, W., & Domahs, F. (2013). Good, bad and ugly word stress–fMRI evidence for foot structure driven processing of prosodic violations. Brain and language, 125(3), 272-282. Willems, R. M. (Ed.). (2015). Cognitive neuroscience of natural language use. Cambridge University Press.

ExLing 2015 - 6th International Conference on Experimental Linguistics, 26-27 June, Athens, Greece