tions: free wielding and restricted exploration. This comparison is of interest
first because the latter case precludes a mechanical variable found to be of
significance in previous research (e.g., Solomon &
Turvey, 1988), and second
because of the general issue of active versus passive touch. The restricted
situation qualifies as a reduction experiment, however, and for students of the
ecological approach to perception, relatively unrestricted exploration is critical
for the generalizability of the results. Thus, we begin by justifying the impor
tance of this reduction experiment and, building on analyses of the general
nature of experimentation, consider the limits of the applicability of the results.
ABSTRACTION AND THE INHERENT AMBIGUITY OF ABSTRACTION Rosen ( 1978 ), in a rigorous analysis of experimentation, emphasized the integral
and unavoidable role that abstraction plays in this process: Experimental science, just as much as theoretical science, rests upon the genera
tion of abstractions . . . For instance, when we measure the values of one or more
observables of a system, we thereby obtain a description of the system in terms of
a set of reduced states; these states are characterized only by the observables we have
measured, and the remainder have been abstracted away. If we observe a system by
utilizing the states of another system to impose dynamics on it, we can see only
those dynamics generated by the observables of the second system; all other
dynamics have been abstracted away. (p. 98)
Thus, any controlled experiment must include some reduction of the situa
tion of interest -- even if the only goal is careful observation, the investigator
must observe some variables and decline to observe others. On the other hand,
the general goal of most experiments is to apply the results of the abstraction to
the original system abstracted. Methodologically, proponents of the ecological
approach to the study of perception prefer to be cautious about the act of
abstraction and the scope of the conclusions drawn. It is recognized that any
abstraction (which is to say, any experiment and any act of science) is inherently
ambiguous and leaves open two possibilities: It is either true that the results
indeed apply to the system abstracted, or that a crucial variable has been left out
of the subsystem, reflecting on the particular variables included in the abstrac
tion but not on the system from which it was abstracted. This ambuguity does
not diminish the value of abstractions but implies that conclusions can only be
drawn confidently with respect to the variables actually included.1 ____________________
| 1 | Consult Rosen ( 1978 ) for descriptions of even more complications inherent in the act of
abstraction. For example, two ... |
| 
To continue reading this publication, you must have a Questia Subscription.
