Foundations of Verb Learning: Infants Categorize Path and Manner in Motion Events

1
Foundations of Verb Learning Infants Categorize
Path and Manner in Motion Events

• Shannon M. Pruden, Kathy Hirsh-Pasek
• Temple University
• Mandy J. Maguire Meredith A. Meyer
• University of Louisville University of
  Oregon

2
Not just verbs

• Relational terms
• In English, relations are encoded in, not only
  verbs, but also in prepositions
  

3
What we know about verbs

• Verbs are hard to learn (Gentner, 1988)

• Actions are ephemeral

• Verbs are polysemous
• Run - 42 entries vs. ball - 9 entries

• Verbs can encode diverse components
• Path, manner, result, and instrument

4
The Paradox

• Verbs appear in childrens earliest vocabularies
  
  
• Choi, 1998
• Choi and Bowerman, 1991
• Fenson, et al., 1994
• Nelson, 1989
• Tardiff, 1996

5
Demonstration Verbs are hard
Watch, Merediths blicking? What does blicking
refer to?
6
Possible meanings of blicking

• Path the trajectory of agent
• e.g. enter, come, approach
• Manner the way in which the agent moves
• e.g. walk, dance, swagger, sway, stroll
• Result
• e.g. open, close
• Instrument
• e.g. hammer, shovel

7
Path and Manner

• Focus on path and manner
• (1) Appear in most languages.
• (2) They are treated differently across
  languages.
  
• English - Manner encoded in verb path encoded in
  preposition.
  
• Spanish - Path encoded in verb manner encoded in
  adverb (optionally).

8
Most of what has been done on verbs

• Early production of relational terms
• Choi Bowerman, 1991
• Tardif, 1996
• Gopnik Choi, 1995
• Mapping relational terms onto actions and events
• Choi, et al., 1999
• Maguire, et al., 2003
• Naigles, 1996

9
ButBuilding verbs requires three steps

• A) Attention to non-linguistic components of
  action
  
• B) Where action meets words
• C) Productive use of verbs in grammar.
• Little work has been done on attention to
  non-linguistic components of action.

10
This Talk is in Four Parts

• Part 1 Path manner in non-linguistic motion
  events
  
• Part 2 Two Studies- Can infants form
  categories of path and manner?
  
• Part 3 Interpreting these results
• Part 4 Future Directions

11
Part 1 Path and manner in non-linguistic motion
events

• Pulverman and colleagues (2002 2003)
• 7 month olds discriminate path and manner
• 14-17 month olds discriminate path and manner.
• Casasola, Hohenstein, Naigles (2003)
• 10 month olds discriminate path and manner.
• To date, this is of what is known about path and
  manner in non-linguistic motion events.

12
So Whats Missing

• Oakes Rakison (2003)
• wordsrefer to categories of objects and events,
  or properties of these things.
  
• Therefore, verbs label categories of actions and
  events rather than single events.

13
For example, running

• Running is considered the same action whether
  performed by Carl Lewis or Grandma.

14
Part 2 Two Studies

• Study 1 Can infants form categories of path
  across multiple exemplars of manner?
  
• Study 2 Can infants form categories of manner
  across multiple exemplars of path?

15
How to address these questions

• Use a proven paradigm
• Use novel, easily manipulated and controlled
  stimuli
  
• Several exemplars of path and manner
• A consistent design across both studies

16
Paradigm

• Preferential Looking Paradigm forced-choice
  split-screen
  
• (Hirsh-Pasek Golinkoff, 1996)
• Non-linguistic task
• Dependent Variable Looking Time

17
Novel, easily manipulated and controlled stimuli
18
Stimuli across studies

• 6 Paths
• Over
• Under
• Past
• Around
• Behind
• In Front

• 6 Manners
• Flap
• Spin
• Twist
• Side Bend
• Bend Forward
• Toe-Touch

19
Design across studies

• Introduction
• Salience Trials
• Four Familiarization Trials
• Test Trials
• Trials are 12 seconds

20
Introduction Trial

• Purpose To ensure infants look to both sides

21
Salience Trial

• Purpose
• To show that infants do not have any a priori
  preferences for test events.
  
• What they see
• Two clips simultaneously.
• Same clips they see at test.
• Assumption
• Infants will not have a preference for either
  clip.

22
Familiarization Trials

• Four exemplars of the category are shown.
• Trials are separated by attention-getter
• Picture of a baby
• Accompanied by music

23
Test Trials

• Test trials
• Two clips shown simultaneously
• In-category event (familiar exemplar)
• Out-of-category event (novel exemplar)
• Predictions
• Infants who categorize will show a preference for
  one of these clips.

24
Study 1 Path Categorization

• Subjects
• 24 7-9 month olds
• (M 8.72, SD 1.01)
• 24 10-12 month olds
• (M 11.29, SD 0.87)
• 15 13-15 month olds
• (M 14.80, SD 1.07)

• Mono-lingual English-speaking homes.
• Equal numbers of males and females.

25
Salience Trial
Flap Around
Flap Past
26
Familiarization Trials for Path

• Four familiarization trials
• Same path across multiple exemplars of manner
• Vary manner across same path
• Example, around

27
Familiarization Trial 1
Side Bend Around
28
Familiarization Trial 2
Twist Around
29
Familiarization Trial 3
Spin Around
30
Familiarization Trial 4
Toe Touch Around
31
Test Trials
Flap Around Novel Manner, Familiar Path In-cat
egory event
Flap Past Novel Manner, Novel Path Out-of-cate
gory event
32
Was there a salience preference?
33
Results- Path Categorization
34
Study 1 Conclusions

• No a priori preferences for the test clips
• 7-9 month olds were not able to categorize path
• 10-12 and 13-15 month olds categorized path
• Familiarity preference

35
Study 2 Manner Categorization

• Subjects
• 24 7-9 month olds
• (M 8.47, SD 0.96)
• 24 10-12 month olds
• (M 11.49, SD 0.80)
• 23 13-15 month olds
• (M 14.75, SD 0.94)

• Mono-lingual English-speaking homes.
• Equal numbers of males and females.

36
Salience Trial
Toe Touch Under
Twist Under
37
Familiarization Trials for Manner

• Four familiarization trials
• Same manner across multiple exemplars of path
• Vary path across same manner
• Example, twist

38
Familiarization Trial 1
Twist Over
39
Familiarization Trial 2
Twist Around
40
Familiarization Trial 3
Twist In Front
41
Familiarization Trial 4
Twist Past
42
Test Trials
Toe Touch Under Novel Manner, Novel Path Out-o
f-category event
Twist Under Familiar Manner, Novel Path In-cat
egory event
43
Was there a salience preference?
44
Results- Manner Categorization
45
Study 2 Conclusions

• No a priori preferences for the test clips
• 7-9 and 10-12 month olds were not able to
  categorize manner
  
• 13-15 month olds categorized manner
• Novelty Preference

46
Novelty/Familiarity Preference

• Why do infants prefer to look at novelty in
  manner categorization study, but familiarity in
  path categorization study?
  
• Infants prefer familiar stimuli when stimuli are
  complex and need time to process (Hunter, et al.,
  1983)
  
• Maybe the infants need more time to process these
  stimuli
  
• Independent Samples t-test with average
  familiarization time for path study vs. manner
  study t (132) 2.472, p
  
• Infants look longer at familiarization clips for
  path study.

47
Part 3 Summary- Our interpretation

• 7-9 months
• Path no preference
• Manner no preference
• No categorization

10-12 months Path familiar Manner no prefe
rence
Categorize path
13-15 months Path familiar Manner novel
Categorize path and manner
What do these results mean?
48
What does all of this mean?

• First study to investigate whether infants can
  categorize path and manner
  
• Developmental Progression
• Path first, then manner
• Preverbal infants can abstract and categorize
  relations
  
• Learning verbs is hard, but conceptual
  foundations are present

49
Part 4 Future Directions

• Does labeling facilitate categorization?
• What other types of event categories can infants
  form?
  
• Would we see similar results with other stimuli?
• Would we see same trends for infants learning
  other languages?
  

50
Acknowledgements

• Natalie Hansell
• Beate Müller
• Heike Herrmann
• Dr. Nora Newcombe
• Carolyn Fenter

• Dr. Roberta Golinkoff
• Rachel Pulverman
• Anthony Dick
• NSF

Thanks to all the parents and children who
participated in these studies at the Temple
University Infant Lab.
51
QUESTIONS???
Correspondence Shannon Pruden (email
spruden_at_temple.edu)