Bilingual Processing of English Aspectual Properties

1
Bilingual Processing of English Aspectual
Properties

• S. Seegmiller, D. Townsend, M. Call, S. Mancini,
  N. Ilia
• Montclair State University

2
Introduction Telicity and Aspect

• Comprehension requires understanding of the
  aspectual properties of events
• Three morphological, syntactic, and/or semantic
  properties seem to play a role in denoting events
  as ongoing, recurring, and bounded

3
Introduction Telicity and Aspect

• Three morphological, syntactic, and/or semantic
  properties seem to play a role in denoting events
  as ongoing, recurring, and bounded
• Morphological aspect (also known as Viewpoint
  aspect)

4
Introduction Telicity and Aspect

• Three morphological, syntactic, and/or semantic
  properties seem to play a role in denoting events
  as ongoing, recurring, and bounded
• Morphological aspect (also known as Viewpoint
  aspect)
• Object number and definiteness

5
Introduction Telicity and Aspect

• Three morphological, syntactic, and/or semantic
  properties seem to play a role in denoting events
  as ongoing, recurring, and bounded
• Morphological aspect (also known as Viewpoint
  aspect)
• Object number and definiteness
• Telicity

6
Morphological Aspect

• Morphological aspect (also known as viewpoint
  aspect) occurs when there is some morphological
  marker indicating whether an event is completed
  (perfective) or ongoing (imperfective)
• Mario fixed the tire
• Mario was fixing the tire

7
Object Number and Definiteness

• Singular objects, either definite or indefinite,
  may denote either bounded or unbounded events
• Fred pushed a/the cart (unbounded)
• Fred baked a/the cake (bounded)
• Typically, plural objects are associated with
  ongoing or repeated events
• Fred baked cakes

8
Telicity

• Telicity is a semantic property of verbs and/or
  predicates that refers to events that have (or
  don't have) and inherent end point
• Susan arrived (telic)
• Susan drew well (atelic)
• Susan drew a circle (telic)
• Susan drew a circle every day (atelic)
• Susan pushed the cart (atelic)
• Susan pushed the cart to the store (telic)

9
Previous Research and Findings with Native
English Sepakers

• We are exploring a series of interrelated
  questions about on-line comprehension
• What information do comprehenders (readers and
  listeners) use to determine aspectual
  information about an event?
• What is the source of this information?
• At what point in the process of comprehending
  does this information become available?

10
Previous Research and Results with Native
Speakers of English

• Research methods
• Self-paced reading experiments
• Speaker detection experiments
• Probe recognition experiments
• Forced-choice tests
• Analysis of corpus data

11
Previous Research with Native Speakers of English

• Brief overview of our methods and main findings
• We used garden-path (GP) sentences like those
  below in self-paced reading experiments
• The actor tripped on the stage amused the
  audience
• ... rested ...
• ... captured ...
• ... admired ...
• Reading times are faster for on after tripped
  and rested

12
Previous Research and Results with Native
Speakers of English

• Brief overview of our main findings
• Comprehenders use subcategorization information
  (e.g. transitive/intransitive) immediately for
  parsing decisions, but they do not use
  information on telicity for parsing
• In self-paced reading tasks, comprehenders read
  the words following atelic verbs faster than
  words following telic verbs, but sentence-final
  words are read faster after telic verbs

13
Story Comprehension

• In a story comprehension task, participants read
  a story one sentence at a time, pressing a key
  when ready for the next sentence.
• The firemen rescued a survivor.
• The firemen were rescuing a survivor.
• The firemen rescued survivors.
• The firemen were rescuing survivors.
• Probe rescue survivor

14
Story Comprehension Results

• The telicity of the verb and the object type had
  an effect on response times
• Response times were faster after telic verbs than
  after atelic verbs
• Response times were faster with singular than
  with plural objects for immediate probes
• The effect of predicate telicity (VerbObject)
  appears only for delayed probes, not immediate
  probes

15
Other Results

• Atelic verbs are more flexible aspectually than
  telic verbs
• In an atelic context, the reading time on a
  sentence-final word is longer when the
  sentence-final word is singular than when its
  plural
• Tony worked in a bakery. He made a cake.
• Tony worked in a bakery. He made cakes.

16
General Summary

• Comprehenders make immediate use of verb
  subcategorization for parsing but not verb
  telicity
• Probe recognition times are faster after telic
  verbs
• Reading times are faster after atelic verbs
• The telicity of a predicate depends on a variety
  of factors, including the semantics of the verb,
  the number of the object, adverbials, etc.

17
Extending the Research

• Are our results valid only for speakers of
  English, or are these characteristics true
  cross-linguistically?
• To answer these questions, we are beginning to
  explore two areas
• Experiments in other languages
• Experiments with non-native speakers of English
  to investigate the influence of L1 on the
  processing of English

18
Extending the ResearchPilot Study with ESL
Students

• Saving the Children
• The tornado struck the school with little
  warning,
• It was destroyed.
• The firemen didn't think
• there would be many survivors.
• Suddenly, one of the men heard voices in the
  wreckage.
• Did the tornado strike a school?

19
Extending the ResearchPilot Study with ESL
Students

• Participants read one sentence at a time on the
  screen and press a key when they are ready for
  the next sentence
• A question appears on the screen either
  immediately after the sentence or three sentences
  later
• Reading times (and errors) are recorded

20
Participants

• 63 native speakers of English
• 24 ESL students, native speakers of
• 7 speakers of aspectually explicit languages
• Polish (1), Russian (4), Turkish (2)
• 4 speakers of an aspectually non-explicit
  language Mandarin
• 13 speakers of other languages

21
Results for Native Speakers of English
22
Results for Non-native Speakers of English
23
Comparison of Native and Non-native Speakers of
English

• Overall target sentence reading times
• Native speakers of English 2112 ms
• Non-native speakers of English 3274 ms
• For native speakers, morphological aspect and
  verb telicity had an effect on reading times
• For non-native speakers, object number had an
  effect on reading times

24
Examining Morphological Aspect

• Although the effect of morphological aspect on
  reading times seems large for non-native
  speakers, the difference is not significant.
• To investigate this further, we selected two
  subsets from our sample of non-native speakers
• Speakers of L1s in which aspect is marked
  morphologically (Russian, Polish, Turkish)
• Speakers of a language in which aspect is not
  marked morphologically (Mandarin)

25
Results Native and Non-native Speakers of
English
26
Tentative Conclusions

• Reading times are generally faster in contexts
  that allow fewer options (telic verbs, singular
  objects)
• The extent to which the L1 explicitly marks
  morphological aspect influences the way
  English-language learners attend to explicit cues
  in processing English

27
Future Directions

• Increase the sample size of non-native speakers
  of different L1's, especially those with
  different verbal and nominal morphology
• Prepare materials in other languages (e.g.
  Mandarin, Russian, Turkish) to compare processing
  strategies across languages

28
Acknowledgements

• We would like to thank
• J. Magliano and M.C. Schleich for sharing their
  story comprehension materials with us. See J.P.
  Magliano and M.C. Schleich (2000), Verb aspect
  and situation models. Discourse Processes, 29,
  83-112.
• Tina Bollettieri for scoring the CELT test of
  English grammar
• Joanna Musial for providing experimental results
  on the role of context for native speakers of
  English