p.110 Least Effort We have seen that rats prefer
spatially shorter and temporally shorter routes; and we have suggested that probably they also can differentiate and prefer
gravitationally (i.e., force-resisting) shorter routes. The possibility now suggests itself, however, that perhaps all three
of these preferences may, in the last analysis, be reducible to a single, more fundamental, one - to be called the preference
for "Least Effort."
p.110 It appears possible, in other words, that further experiments,
in which spatial and temporal and force-resisting delays were paired in all possible permutations and combinations, one against
another, would uncover the fact of a single dimension of "Least Effort" running through them all.
p.111 We cannot, that is, predict how in concrete types of means-end
dimension, this Principle of Least Effort is going to express itself. In each instance, what actually has to be done
is to discover separately, for each type of organism, and for each type of permutation and combination of
dimensional extents - spatial, temporal, force-resisting, and the like - in how far differences between or within such dimensions
will or will not give rise to preferences.
p.134 a memory in the most general sense, we shall define as a means-end-expectation
in which the expectation of the specific character of the more distant object... arise in part out of past stimuli no longer
sensorially present at the moment in question. A perception, on the other hand, we shall define as
a means-end-expectation in which the expectation of the character of the more distant object obviously depends primarily upon
present stimuli, i.e., stimuli coming then and there from this more distant object.
p.137 The definitive feature of a perception is that, in it, stimuli corresponding
to all the parts of the total complex (the remote parts as well as the more immediate) are directly present.
p.179 if... the three fundamental types of dimension - space, time, force-resistance
- can be shown to be ultimately intercommensurable - all to reduce to one single one of physiological "Least Action"... then
means-end-distance will be ultimately synonymous with, and can be equated to, physiological "Least Action."
p.179-180 we should like to introduce one more term - that of means-end-capacity.
This will be the capacity of the given animal relative to types and extents of means-end-fields. The more complicated
and refined the given animal's success in expecting degrees and types of distance, direction, succession,
common final pathness, multiple trackness, alternativeness between paths, reverse ends of one and the same
detour, etc., the greater, by definition, will be that animal's means-end-capacity. We coin the term means-end-capacity
to designate the underlying capacities which allow successful sign-gestalt-readinesses and sign-gestalt-expectations relative
to complicated types of means-end-field.
p.183 The degree of an animal's capacity to build up correct sign-gestalt-expectations
and hence perceptions, mnemonizations and inferences relative to complex means-end-fields defines the degree of his means-end-capacity.
p.183-184 When representative, i.e., typical, individuals of different species
are presented with one and the same complex means-end situation, they will be found to exhibit relatively different adequacies
of means-end-readiness and means-end-expectation. These different adequacies may be said to define and result from their differences
of means-end-capacity. Let us illustrate by some concrete examples from Kohler.
Kohler presented a chick, a dog, an ape, and a child with
practically one and the same problem - that of being barred from food by a three-sided barrier (transparent on the
side towards the food). He found that the chick, when put in this situation, will, for the most part, merely run back and
forth opposite the food. A chick will not, or will at the most only after much trial and error, learn to go round
one of the flanges of the barrier in order to get to the food. A dog in this same situation, on
the other hand, will relatively easily go around one of the flanges or sides, though he is decidedly less ready to do this,
if the food be placed extremely near the center of the barrier... Finally, the ape and the child will have no difficulty
at all. After only a second or two of trying to get directly at the food, they will turn, see the unblocked
flange, and run around it.
p.185 The higher the species, the more extensive and complicated and "highbrow"
the types of field-relationship to which the individuals are sensitive.
p.186 We shall... introduce the more specific term dimensional
means-end-capacities for the abilities for the special "material" embodiments in space, in time, in gravitation,
in the number system, in society, in words, in rhetoric, and the like, of such formal means-end-principles.
Such embodiments constitute specific "dimensions." And the capacity for means-end-relation as specifically
embodied in the material of some one of these dimensions, we shall call therefore a dimensional capacity.
p.187 it follows that the possession of a given type of dimensional capacity
will be bound up with the possession of certain types of especially good discriminanda- and manipulanda-capacities.
p.198 Means-end-capacity as we use it, is the capacity behind successful
expectations, irrespective of whether these expectations are "consciously experienced" or not.
p.200 we may, perhaps, say that, in general, they seem to have conceived
of insight as some sort of a "conscious grasping" of the field relationships or field "rules."
p.200 what is the relation of such a definition of "insight" to our concept
of means-end-capacity? Our answer is that a capacity for grasping field-relationships is what we mean by
means-end-capacity. All learning, therefore, will involve "insight" in this mere capacity sense.
p.209-210 It appears, in other words, that it is
only by virtue of the possession of at least this degree of means-end-capacity that consciousness will be possible.
The means-end-capacity for responding at least to differential and predictive relations, will be the necessary pre-condition
of a conscious awareness
p.323 the conditioned reflex is the unit out of which the whole
habit is formed.
p.329 we may point out in general that most behaviors,
that is all save simple physiological reflexes... necessarily require... complementing supports from
the environment. They require, that is, not merely stimuli to release them, but also appropriate discriminanda and
manipulanda to support them.
p.339 We turn now to the trial and error doctrine of learning. It, like
the conditioned reflex doctrine, has grown out of a special set of the facts. More specifically, it has grown out of the sorts
of facts exemplified by the rat in the maze.
p.342-343 Our own assertion now, however, will be that, whatever the special form of trial
and error theory - whatever its special statement concerning the relative importances of exercise and effect, and whatever
its conception as to the character of the "effects" - no trial and error theory is really tenable.
[JLJ - I disagree - an effective trial and error theory can be formed from an understanding of the cues in the environment
and a physical understanding of the environmental dynamics itself. Consider an American football coach calling plays for his
offense at the start of a highly contested game. He has years of experience in similar positions and understands the
players and their capabilities. He will proceed meticulously to probe-test each structure of the defense to look
for weakness that he can exploit with the strength of his players. "Trial and error" can be thought of as "read the cues,
guess by experienced inference, explore, perceive, alter model of situation, repeat". The challenge of every living
being is to make sense of the partial vagueness which constantly confronts - it must strive to change "I don't know my way
about" to "now I know how to go on" [Wittgenstein].
p.358 In general, the results were that the guided trials, simply by themselves,
were never enough to cause complete learning, although they did usually help to a greater of less extent. When the guidance
was removed both types of subject seemed always impelled to do at least some exploring of the newly opened up blinds.
p.365 Learning consists in the building up of appropriate sign-gestalts
both as regards the correct responses and as regards the incorrect responses... for the development of each individual
sign-gestalt-expectation involved in trial and error the favoring conditions are undoubtedly the same as those we discovered
in the last chapter for the case of conditioned reflexes.
p.374-375 The capacity laws are to be conceived in
general as listing certain fundamental traits, capacities, aptitudes, the possession of which in the given individual or species
will favor learning.
(a) Formal means-end-capacities (i.e., capacities for simple differentiation
and prediction) are obviously essential even to such simple sign-gestalt formation as is involved in conditioned
reflex learning. An animal must, then and there, be capable of differentiating the given sign, significate, and means-end-relation
from others and of forming the connecting predictive relations. And the quicker and more facile he is at such differentiation
and such prediction, the better he will be at such conditioned reflex learning.
(b) Dimensional means-end-capacities and discriminanda- and manipulanda-capacities.
The animal must be capable also of achieving correct expectations with respect to the specific sorts of dimensional
material - space, time, gravitations, social relations, or what not - constituting the stuffs of the given sign,
the given significate and the given means-end-relations. And the greater his capacities for such materials, again the better
he will be at the given conditioned reflex learning.
(c) Retentivity. Finally, the animal must be capable of
"retaining" these differentiations and predictions from trial to trial. And the better he is at retaining, the sooner
he will thus learn.
p.375 Consider now the new capacities which come to the fore in trial and
error learning. They are two: (d) more complicated means-end-capacities, such as those of alternativeness, roundaboutness,
final common-pathness, etc., and (e) the capacities for consciousness and for ideation (consciousness-ability and ideationability).
p.376 Consider now, inventive learning. All the capacities listed under
trial and error learning will again play a part, and in addition we note the role of one new capacity, viz., creativity.
(f) Creativity (creative instability). It seems obvious that
a capacity to break out into new lines of behavior, ideationally or actually, will be fundamental for inventive learning.
p.449 Means-end-capacity (formal, dimensional). The term is used to designate
the innate (and acquired) capacities whereby a given organism or species is capable of having commerce-with and expecting
means-end-relations
