Title: Beyond generalization: Continuous repertoires ABA
1Beyond generalizationContinuous
repertoiresABA 3175/26/03900 -1020
2Teaching aims to establish more than what is
directly taught
- Making generalization effective
- Programming generalization
- Beyond generalization
3Misuse of the term generalization
- As an explanation
- As a redundant qualifier
- in conjunction with imitation and identity
matching to sample - As a designation of many different phenomena
- continuous repertoires
4Continuous repertoires
- Skinner (1953)
- . . . any unit of operant behavior is to a
certain extent artificial. Behavior is the
coherent, continuous activity of an integral
organism. Although it may be analyzed into parts
for theoretical or practical purposes, we need to
recognize its continuous nature in order to solve
certain common problems. (p. 116)
5Continuous repertoiresWildeman D. G. Holland
J. G. (1972).
- Control of a continuous response dimension by a
continuous stimulus dimension. Journal of the
Experimental Analysis of Behavior, 18, 419-434.
6Examples
- Imitation
- Drawing from copy
- Singing on key
- Visual joint attention skills
- (gaze following)
7Continuous fields Train 1
8Continuous fields Train 2
9Continuous fields Test 1
10Continuous fields Test 2
11The A-B-C of behavior analysis
12Identifying instances of SD and R
- Skinner (1931) The reflex as a relation (S ?R)
- Skinner (1935) Classes of Ss and classes of Rs
- Catania (1973) Operant class -- correlation
between descriptive and functional classes - Discriminated operant class Descriptive vs.
functional (cf. Wildemann/Holland results)
13When the basic SD?R relation breaks down
- a. SD ? R time delay
- b. SD ? R complexity
14Complex phenomena
- Largely left to cognitive psychology (memory,
problem solving) - Squeezed into the model (tacts, rule following)
- given a new name (relational frames)
- A third option Analysis in terms of the basic
formula, filling the gaps by incorporating
precurrent behavior
15RFT (S. Hayes, et al., 2001)
- In contrast to accounts in terms of naming or
joint control, RFT can account for derived
relations more directly. Naming, joint control, .
. . and the like, can be viewed as aspects of
specific kinds of relational responses given
specific histories, but they are not necessary to
the relational operant as such.
16RFT (S. Hayes, et al., 2001)
- Postulating such a mediating behavioral
- process seems not to be required in order to
- mount a behavior analytic research program
- into derived relational responding. . . . There
- is no immediate need to posit any additional
- behavioral processes behind that of a learned
- functional response class, based on multiple
- exemplars.
17Behavioral mediationStokes Baer, 2003
- Some problem classes require mediation for
generalized correct solutions (p. 134)
18Mediating or precurrent responses(Palmer, 2002)
- What is the tenth letter after F? . . . If we
- restrict our consideration to just those
- responses that can be observed, the behavior
- of announcing the answer is puzzling.
19SD ? R distortions
- at the level of our observations, behavior
- does not seem to be related in an orderly
- way to environmental antecedents or
- consequences but seems to emerge
- from within the individual
20Precurrent behavior
N
I
M
J
H
O
K
L
P
G
(E), F. . .
21SD ? R complexity
22SD ? Precurrent behavior(e.g., counting)
23Degrees of SD ? R complexity
- Same dimension
- Rotated dimension
- Compound dimensional relations
- Arbitrarily related dimensions
24Same dimension
25Rotated dimension
26Compound dimensional relations
27Arbitrarily related dimensions
28Math problems
- 65 x 65
- 85 x 85
- 35 x 35
- 75 x 55
- 35 x 55
- 65 x 75
- 85 x 45
- 73 x 77
4225
7225
1225
4125
1925
4875
3825
5621
29Math problems 1
30Math problems 2
31Math problems 3
32Math problems 4
33Stokes Baer (2003)
- Mediated generalization appears to be a promising
but as yet underdeveloped forefront for
facilitating desired behavior change. As such, it
represents an unfinished portrait in the
analysis of generalization strategies
34Continuous repertoires- mosaics of increasing
texture
- lack of order inevitably arises when there are
gaps in our data - for cognitive behavior such gaps are common
- order is restored to the data by referring to
plausible unobserved variables that fill in the
gaps - continuous repertoires of precurrent behavior
should be considered to fill in important gaps