3.5 DISEÑO DE RED PASIVA
3.5.3 DISEÑO DEL SUBSISTEMA DE CABLEADO HORIZONTAL
Primatologists have described two types of emulation learning in which an observer reproduces the end result of a demonstrator’s behaviour instead of copying the behaviour that brought about the end result. According to Tomasello (1990, 1996), the term “emulation learning” refers to situations where an individual during observation learns something about the changes of state in the environment (e.g.,
objects) as a result of the behaviour of the model, but not about the model’s
behaviour or behavioural strategy. For example, having seen a model use a hammer to crack open a nut, the observer might seek to reproduce the same result through his or her own behavioural strategy, such as biting into the nut to open it. Under such circumstances, the observer could learn little about the specific actions performed by the model, but produce the effect of these observed acts by learning about the
dynamic properties of the nut (e.g., its solidity). In emulation learning, an observer’s behaviour might involve reproducing the end result of the object that the model intended to achieve. However, this does not always mean that he has understood the goal or intention of the model.
Whereas Tomasello (1990, 1996) emphasised that the observer’s reproduction of the end result in emulation learning involves learning neither the model’s
behaviour nor even the model’s goal. Whiten and Ham (1992) suggested that the observer might reproduce the end result via his or her own method with an understanding that the model held a goal towards such an outcome. Thus, Whiten and Ham distinguished a special type of emulation learning called goal emulation from Tomasello’s original definition. In goal emulation, an observer learns that a modelled action sequence was designed to produce a particular goal of the model, and devises his or her own method to achieve that goal by producing the end result of the model’s behaviour. For example, having seen a model use a hammer to open a nut but without success, the observer might learn that the final act in the modelled action sequence is to open the nut, and reproduce the end result using his or her own method (e.g., biting into the nut or using a stone to crack open it).
In addition to these two types of emulation learning, a potential effect of emulation learning might take place when an observer is exposed to the particular end result of a model’s actions but without seeing the model’s actions. It is possible that the observer could derive action information required for reproducing the end result by himself or herself from seeing only the end result of the model’s acts. Under such circumstances, exposure to the end result of the model’s actions might elicit an effect of emulation learning similar to that due to exposure to the model’s actions. For example, the individual might learn to crack open nuts with a stone from
observing two halves of the cracked ones without seeing a model do so beforehand. That is, in any of the above mentioned types of emulation learning, the individual devises his or her own behavioural strategy to achieve the end result that he or she has learnt during observation of the demonstration.
Even if, in emulation learning, an individual overlooks the behaviour of the model and is attempting to replicate the outcomes involved in the modelling with his or her own devised method, this does not imply that emulation involves a less
intelligent learning process than imitation. Instead, under some circumstances, emulation is a more adaptive and creative strategy than imitation, which by contrast is a more social one (Tomasello, 1999). For example, knowing the affordances or typical uses of objects allows children to incorporate this knowledge into their own attempts to find out how novel objects could be used to fulfil the same role. Similar to Tomasello’s characterisation of emulation. Whiten and Custance (1996) referred to a cognitive ability to reproduce the outcome by intelligently selecting and
extracting useful information from the demonstration. Whereas Tomasello (1996) has attempted to exclude emulation from phenomenon of imitation. Whiten suggested that emulation can be specified as a cruder level of imitation on a continuum of copying fidelity and it may not be necessary to draw a clear-cut distinction between imitation and emulation (Whiten & Ham, 1992; Whiten & Custance, 1996). At some point and in some circumstances, emulation learning presents an advanced cognitive strategy compared to imitation. For example, a model attempted to achieve a particular goal in the task but used an invalid or inefficient method. If the individual learns from his or her observation the affordances or functional relations in the task as result of emulation learning, the individual might adopt a valid or more efficient strategy devised by himself or herself without copying the model’s actual actions. In contrast, in imitative learning, the individual is
expected to precisely copy the model’s inefficient or invalid method in the task completion.
The notion of emulation learning raises several questions about the
interpretation of the child’s production of target acts in Meltzoff’s failed-attempt tasks. Could the infants understand nothing about the demonstrator’s behaviour, but
only the dynamic affordances of the objects that they might not have discovered in the control conditions? Or could the infants fail to copy the demonstrator’s failed attempts, because the results they themselves found out during observation were more attractive? Or perhaps the affordances of the objects that the infants learned as a result of the failed actions coincided with the target acts that the demonstrator was taking to be intended in the failed-attempt tasks?
Meltzoff (1988 b) used an adult-manipulation control to provide a basis for emulation learning in an early study of infant deferred imitation. In this control, the experimenter presented the same consequences of the objects as in the imitation group but without demonstrating the target acts. For example, one of the objects was a box with a recessed button on its top surface, as well as a hidden beeper underneath the button. Whereas the experimenter pushed the button to activate the beeper inside the box in the imitation group, he secretly triggered the beeping via a hidden switch when handling the box in the adult-manipulation control. The results showed that the 9-month-old infants exposed to the end results of the target acts did not produce the target acts as frequently as they did when observing the experimenter demonstrate these acts. However, no control was used to elicit emulation learning in Meltzoff’s (1995) study of 18-month-olds’ imitation in failed-attempt tasks. As emulation has been thought to be a creative and intelligent behavioural strategy under appropriate circumstances (Tomasello, 1996; Whiten & Custance, 1996; Whiten & Ham, 1992), it is quite probable that 18-month-olds may be more capable than 9-month-olds of using what they learn about the consequences of the target acts to recreate the entire actions. Therefore, the demonstration (intention or failed attempt) condition used in Meltzoff’s (1995) study was more likely to involve emulation learning than the adult-manipulation control, for the experimenter in the failed-attempt display was seen to attempt to bring about the target acts but not to succeed and that seemed to be sufficient to instigate infants to explore what dynamic affordances of the objects are likely to elicit.
On the other hand, in Meltzoff’s (1995) study, the pattern of the object movements characterising the failed attempts was similar and relevant to that presented in the target display as compared with the acts included in the adult-
manipulation control. Take, for example, the plastic square and the dowel (see appendix): the square was moved along the sides of the wooden base in a vertical position in the adult-manipulation control. In contrast, in the failed-attempt demonstration, the hole of the square overshot the dowel as a result of the square sloping down on the base almost horizontally. Presumably that movement resembled and had a relevance to the intended target act—the hole was aligned over the dowel as a result of the square lying on the base horizontally. Likewise, for example, the dumbbell (see appendix): the experimenter pushed its two ends inward in the adult- manipulation control, while he exhibited the action of moving his hands outward in both the target and failed-attempt displays. Thus, even if the infants did not learn to produce the target acts through stimulus enhancement by observing the demonstrated failed attempts, it is quite probable that they may have learned the dynamic
affordances of the objects from seeing the movements featuring the failed attempts and that may have led to their performance on the target acts. Similarly, it is possible that the infants learnt simply the final configurations of the objects by observing the modelled failed attempts. In other words, observation of the demonstrated failed attempts might have evoked the infants’ knowledge of the causal structure of the test materials—what end points the sequence of movements could lead to— and thereby resulted in the infants’ producing the target acts more frequently than they did by watching the adult-manipulation control acts that were dissimilar and irrelevant to the target acts. If this were the case, Meltzoff’s (1995) Failed-Attempt paradigm would seem to provide a basis for emulation learning rather than imitation or re enactment of the intentional actions of others.