Back to Basics: Whats Actually Wrong with Good OldFashioned Cognitive Development

1
Back to BasicsWhats Actually Wrong with Good
Old-Fashioned Cognitive Development?

• Nora S. Newcombe
• Temple University

2
Criticisms of GOFCD

• Purely descriptive, age-fixated, static
• Little if any attention to mechanism and process
• Bogged down in dichotomies
• Nature vs nurture
• Abrupt vs continuous change

3
But I would argue that

• GOFCD is not that bad
• New alternatives suffer from most of the same
  issues as GOFCD
• Overall, researchers in cognitive development are
  doing pretty well
• In part due to the new analytic tools offered by
  connectionism and dynamic systems analyses

4
GOFCD is not that bad

• After all, it defined the phenomena dealt with
  today
• More important, it has evolved
• Age is no longer (really) treated as an
  independent variable, but rather process accounts
  are formulated that predict at what age certain
  achievements should be observed ? age as DV

5
Example of the Geometric Module
6
Age Differences in Use of Featural Cues
Note Perfect Performance 100
7
Room Within Room Studies

• Small waist-high enclosure (Hermer-Spelke size)
  centered within large room
• Large room had one colored wall
• Children stay within small enclosure

8
An Adaptive Combination Modelof Use of
Geometric and Feature Information
Large Action Target Proximal Age
at Success Feature? Possible? to Feature?
Hermer-Spelke No No
Yes 6 years Learmonth
Yes Yes
Yes 18 months et al.
(earliest tested) Study 1
Yes No No
6 years? (not 5
years) Studies 2 3 Yes No
Yes 5
years (not 3 years, 4
years not tested)
9
New Alternatives, Old IssuesVan der Meer

• Van der Meer appeals to increases in capacity
• An old construct that is not well defined or
  easily quantifiable
• Deus ex machina?
• One wonderful aspect of his analysis shows the
  need for caution in diagnosing mechanism
• When people see older children taking longer to
  decide, they might think delay is causalhis data
  suggest that is not true

10
Unsolved Issues with a Classic Flavor

• Why do children start to use distance?
• Role of input?
• Why do they ever learn torque rule?
• Schooling?
• Specificity versus generality
• Note that abrupt change is NOT Piagetian stage
  transition
• Any generalization to other reasoning situations?
• Any applicability to cognitive development in
  domains that involve less conscious
  problem-solving?

11
New Alternatives, Old IssuesMareschal

• Mareschals model and other connectionist models
  frequently rely on assumptions regarding input
• its fundamental nature, and also sequences in
  what is presented when in development
• But these assumptions have typically not been
  independently studied, even though they are
  crucial elements of the model

12
Consider Teeter Totter Experience
13
What Counts as Explanation?

• Why not equilibration?
• If fully specified
• Why not variation and selection?
• But note that both Siegler, and two presenters
  today, still need to say from whence their
  starting points derive..
• Can description ever equal explanation ?

14
In Sum, Dynamic Systems and Connectionism Are New
Ways of Thinking About GOFCD

• But it is not clear that they are immune from
  GOFCD issues
• What they have done is to bring exciting new
  analytic tools and a high degree of specificity
  to GOFCD