Perception And Action Research Paper

1. Reflexes

In most species, behavior is triggered by the present situation and, thus, directly reflects the animal’s immediate environmental conditions. Such reflexes can also be observed in humans, especially in infants, but here they constitute a negligible part of the behavioral repertoire. Interestingly, however, even reflexes already show the close mutual relationship between perception and action. Perhaps the best demonstration of this relationship is provided by the orientation reflex, which we experience when encountering a novel and unexpected event. On the one hand, this reflex inhibits ongoing actions and tends to freeze the body—a prime example of a stimulus– triggered response. At the same time, however, it draws attention towards the stimulus source by increasing arousal and facilitating stimulus-directed body movements. That is, the novel stimulus triggers actions that lead to a better perception of itself, thus producing a full S–R–S cycle.

Even though reflexes themselves represent a relatively inflexible way to coordinate perception and action, some researchers have suspected them to provide the basis for voluntary action (Easton 1972). For instance, the tonic neck reflex, an asymmetric pose observed in newborns with head and arm extended to one side and arm and leg flexed on the other, might facilitate the development of eye–hand coordination. Likewise, the stepping reflex, in which babies move their feet in succession when coming in contact with a solid surface, might underlie our ability to walk.

2. Associations

Human behavior (and that of other higher species) is surely much more flexible than exclusive control by reflexes would allow. Not only can we learn to react to particular environmental conditions and situations in a certain way, we also can unlearn what we have acquired and learn new relationships between situations and actions. Our ability to associate actions with stimulus conditions was the major topic of American behaviorism around 1890–1940, when most of the basic principles of S–R learning were empirically established. In particular, stimuli and responses become associated only if they co-occur in time, if there is at least some degree of contingency between them, and if the response is judged to be appropriate (Thorndike 1927). Although the empirical approach and the theoretical language of behaviorism has long been abandoned, its results still have a major impact on modern connectionism, the attempt to model psychological phenomena by means of artificial neural or neurally inspired networks on computers or in robots.

A major objection against the behavioristic approach to S–R learning relates to the assumed role of action outcomes. In behavioristic theories, the outcome of a given action is only judged regarding its hedonic value: a positive evaluation results in strengthening the association between the action and its antecedents whereas a negative evaluation weakens the association. However, whether it feels good or bad, the outcome of an action also informs the actor about its consequences, that is, about what he or she can achieve by performing it (Tolman 1932). And as actions aim at producing intended outcomes, perception–action learning should not be restricted to forming stimulus–response associations but comprise response–effect associations also. Indeed, studies reveal that humans (and also rats and pigeons) do acquire very stable associations between their actions and the consequences that these actions produce. Moreover, there is evidence that these associations play a major role in the control of voluntary action. That is, people control their overt behavior by forming or reactivating perceptual representations of intended goal events, which through learning have become associated with the motor patterns that have been- —and must be—carried out to reach them.

3. Rules

The ability to learn new relations between environmental conditions and appropriate behavior provides an enormous gain in flexibility for an individual in allowing it to adapt to environmental change. Yet, learning and relearning take time and require at least some experience with a given new condition—it is thus necessarily reactive and, in a way, conservative in reflecting what one has become used to. Indeed, many forms of human behavior show these characteristics, as witnessed by the difficulties of introducing new behavioral patterns regarding, say, women, ethnic minorities, or the use of environmental resources. Nevertheless, there are many instances where people can switch between different reactions to the same stimulus conditions on the spot, and hence more or less independently of the amount of experience with the situation. For example, even though one may have used one route to go to work 1,000 times already, it is easy to take an alternative route from one day to another, without having to unlearn the old habit or to learn the new one, and although the fact that in the case of absent-mindedness one might again go by the old route shows that the overlearned associations still exist. Therefore, people must be able to choose deliberately among alternative ways to relate environmental situations to their actions, that is, to select one out of many possible S–R rules, and behave accordingly.

The interplay between overlearned S–R associations and the voluntary implementation of intentionally selected S–R rules (‘habit’ vs. ‘will’) was a major issue in the psychological literature between 1870 and 1935, and Narziss Ach (1935) was the first to study this interplay empirically in a systematic way. In his ‘combined method’ he first had subjects build up new, strong associations between nonsense syllables and one type of action, and then asked them to perform another than the practiced action to the same stimuli. The resulting increase in reaction time and errors as compared with actions with neutral stimuli was taken to represent the individual ‘will power’ needed to overcome the previously acquired habits. After many years of neglect, the issue of how people implement and switch between alternative S–R rules received renewed interest during the 1980s and 1990s, especially in experimental and neuropsychology (Monsell and Driver 2000). The (still preliminary) results of this research provide new insights into the relationship between, and the control of, perception and action. As one would expect, the implementation of S–R rules takes effort and time. If people are to apply several sets of S–R rules concurrently, or in short succession, their performance is impaired and more error prone. In particular, they will sometimes apply the wrong rule or apply the right rule at the wrong time, especially if effort and attention deteriorate. However, even under ideal conditions, intentional control is not absolute. That is, voluntarily implementing a particular rule does not exclude or prevent contributions from overlearned S–R associations (habits). Therefore, although any S–R rule can be implemented in principle, those rules are easier to implement and to apply if they are in accordance with natural S–R relations, acquired S–R associations, and previously implemented S–R rules than if not. For instance, performance is better if stimuli and responses have features in common than with arbitrary S–R relations, so that pressing a left or right key, or turning to the left or right, is easier if signaled by the location of a stimulus that also appears on the left or right side than by the color, shape, or meaning of a stimulus.

The ability to implement and switch between the most arbitrary S–R rules allows for the highest degrees of behavioral flexibility but it is at the same time costly in terms of attentional resources. Therefore, it makes sense to see intentional rule implementation and the more automatic guidance by overlearned S–R associations as mechanisms that complement each other: implementing and acting out the same rules repeatedly and with sufficient success provides the basis for forming increasingly stronger associations between the corresponding environmental conditions and actions, in this way transforming rules into habits.

4. Plans

In one sense, perception and action are interdependent to a degree that makes it difficult to say where action ends and perception begins—just think of the eyemovement example. In another sense, however, evolution has provided humans with the important ability to temporally decouple our actions from perception. That is, actions can be planned, prepared, and scheduled long before their environmental trigger conditions occur. Although we are only beginning to understand how action planning works, there is evidence that it can be likened to self-automatization.

At least three phases of planning can be distinguished. First, the features of the intended action effect (i.e., the goal) need to be specified, such as the direction or end location of a reach, and the effector to be used. Next, the features belonging to one plan are integrated (i.e., their cognitive representations are temporarily coupled), to avoid confusing features belonging to different, concurrently maintained action plans. This integration process will often involve linking the plan to its anticipated trigger conditions, and hence to representations of the event that is intended to trigger the planned action. That is, planning comprises anticipating both perceptual and action events. Accordingly, planning an action changes our perception. For instance, after having planned a particular action, action-related objects and events are more salient and more easily processed, action-related features of objects are more conspicuous than other features, and objects appearing at action-related locations catch more attention than others.

The final step of planning consists of initiating the planned action. Initiation may be triggered internally, such as when a plan is carried out immediately after construction, or externally, such as when the anticipated trigger event is perceived. Interestingly, this last step does not seem to underlie overly strict cognitive control. For instance, internal initiation is more or less indifferent to the content of the plan or the progress of planning, and externally triggered plans can be seen to be called up even under inappropriate conditions or at the wrong point in time. Thus, planning is like automatizing oneself by means of delegating control to future, often environmental, events.

Bibliography:

1. Ach N 1935 Analyse des Willens. Urban and Schwarzenberg, Berlin
2. Dewey J 1896 The reflex arc concept in psychology. Psycho-logical Review 3: 357–70
3. Easton T A 1972 On the normal use of reflexes. American Scientist 60: 591–9
4. Gibson J J 1979 The Ecological Approach to Visual Perception. Houghton Mifflin, Boston
5. Jordan J S Systems Theory and A Priori Aspects of Perception. North-Holland, Amsterdam
6. Monsell S, Driver J 2000 Control of Cognitive Processes. MIT Press, Cambridge, MA
7. Piaget J 1946 La Formation du Symbole chez l’Enfant. Delachaux and Niestle, Paris
8. Prinz W 1990 A common coding approach to perception and action. In: Neumann O, Prinz W (eds.) Relationships Between Perception and Action. Springer, Berlin, pp. 167–201
9. Thorndike E L 1927 The law of eff American Journal of Psychology 39: 212–22
10. Tolman E C 1932 Purposive Behavior in Animals and Man. Century, New York
11. Turvey M T 1977 Preliminaries to a theory of action with reference to vision. In: Shaw R, Bransford J (eds.) Perceiving, Acting, and Knowing. Erlbaum, Hillsdale, NJ, pp. 211–65