The Conceptual VS' The PerceptualAssociationist View In Infant Word Acquisition

1
The Conceptual VS. The Perceptual-Associationist
View In Infant Word Acquisition

• Slides are available at
• www.cs.gonzaga.edu/depalma/links/links.html

2

• Booth, A.E., Waxman, S.R., Huang, Y.T (2005).
  Conceptual Information Permeates Word
  Learning. Developmental Psychology , 41,
  491-505.
• Booth and Waxman are in the psychology department
  at Northwestern
• Huang is a graduate student in psychology at
  Harvard
• You can find the article at
• www.psych.northwestern.edu/waxman/

3
The Dispute

• Booth, Waxman and Huang argue that conceptual
  knowledge influences lexical acquisition in
  infancy (p. 491)
• Linda Smith (and colleagues) at Indiana
  University maintain that early word learning can
  be accounted for by the childs associations
  among perceptual objects, syntactic cues, and
  words in his or her own productive vocabulary.
  (p. 492)

4
Conceptual Knowledge?

• Infants form categories.
• They make inferences based on category
  membership.
• Gelman and Markman Gelman, S. , Markman, E.
  (1986). Categories and induction in young
  children. Cognition 23, 183-209.
• Showed children a picture of a dolphin and told
  them it was a mammal that breathed by jumping out
  of the water.
• Showed children a picture of a shark and said it
  was a fish.
• Children had to decide whether the shark was more
  likely to breath under water or by jumping out of
  the water.
• Children overrode perceptual similarity and made
  judgments on the basis of category membership
• Called inductive inference and has been
  demonstrated in infants as young as 13 months
  Graham, S. et. al. (2004). Thirteen-Month-Olds
  Rely on Shared Labels and Shape Similarity for
  Inductive Inferences. Child Development 75 (2),
  409427.

5
The Claim

• Conceptual knowledge influences infant word
  acquisition
• Infants extend NOUNS names of objects on the
  basis of shape by 13 months and they extend on
  the basis of texture by 21 months. This cannot be
  accounted for solely by correlations between
  perceptual object features and words.

6
The Experiment

• 24 2-year olds were randomly assigned to one of
  two groups
• The artifact-kind group
• The animate-kind group
• Each group was shown a target object labeled with
  a novel word Dax. This object had no features
  that made it plausibly animate (in Smiths study
  the target object had eyes. Waxman argues that
  this feature is perceptual, but also signals
  animacy.
• Children were then shown three other objects that
  varied along a single dimension (shape, texture,
  size).
• Children in both conditions say exactly the same
  objects. The only thing that varied was the
  story the experimenters told them.

7
The Objects
8
The Vignettes

• Both groups were told this story
• Wow, look at this dax! I have something very
  special to tell you about this dax. Do you want
  to hear it? Listen carefully now because Im
  going to ask you some questions about what I say.

9
But

• The animate-kind group were told this story
• This dax has a mommy and a daddy who love it
  very much. They love it so much that when this
  dax goes to sleep at night, they give it lots of
  hugs and kisses.

10
And

• The artifact-kind group heard this story
• This dax was made by an astronaut to do a very
  special job on her spaceship. The astronaut
  always takes her dax with her when she flies to
  the moon.

11
Test Phase I

• With the target object visible, the experimenter
  said to both groups
• Now I am going to show you some other things.
  Each one might be a dax or it might not be a dax.
  I need you tell me if you think each one is a
  dax or is not a dax, ok?

12
Test Phase II

• She then showed the objects one by one, asking
  after each
• Is this another dax?

13
Prediction

• Infants in the artifact condition would extend
  words primarily on the basis of shape
• Infants in the animate condition would reject an
  object with the same shape as the target but with
  a different texture
• Infants in both groups would accept the size
  change

14
Results

• Performance in the artifact and animate
  conditions did not differ on shape change
• Infants in the artifact group accepted the
  texture changed object as a dax (presumably
  because it was the same shape as the original.)
• Infants in the animate group were more likely to
  reject the texture changed object (creatures that
  look like dogs but without fur probably arent
  dogs)

15
Conclusion 1

• The difference in performance must be attributed
  to the conceptual information provided in the
  vignettes because infants in both conditions saw
  precisely the same objects. (p. 495)

16
Conclusion 2

• These results are inconsistent with the claim
  that word learning is initially impervious to
  conceptual information.

17
Grand Conclusion

• Both perceptual and conceptual information are
  end points in a continuum of abstraction from
  sensory input
• Pereptual __________________Conceptual
• Much conceptual knowledge can be derived from
  information closer to the perceptual end (dogs
  have four legs and tails)
• But much cannot (how organisms function, what
  drives their behavior)

18
Yet More

• The authors conducted two more experiments
• A forced-choice experiment to counter the fact
  that infants reveal a strong yes bias.
  Conditions were very similar, except children
  were asked to pick out the Daxs from the three
  objects arrayed before them.
• A forced-choice experiment with children of 18
  months.

19
More Reading About The Dispute

• Smith, L.B., Samuelson, L. (2006). An
  attentional learning account of the shape bias
  Reply to Cimplian Markman (2005) and Booth,
  Waxman, Huang (2005). Developmental
  Psychology.
• Booth, A.E., Waxman, S.R. (2006). Re-re-
  revisiting the conceptual status of early word
  learning. Developmental Science 42(6), 1344-
  1346.