Clarifying the structure of the theory--论文代写范文精选

在每种情况下，有一个附加条件的歧视，用来提供控制行为矫正。额外的歧视不提供控制行为修改,它定义一个给定的生产行为。下面的essay代写范文进行阐述。

Abstract
To evaluate the account we need a clear understanding of three kinds of things: (i) the nature of the behavior control taxonomies, (ii) the taxonomic transitions that are possible or likely, and (iii) the factors that drive transitions. This section explores the structure of Sterelny’s theory from an abstract perspective.

Integrative behavior control 
A very simple way to make behavior flexible is to make behavior production sensitive to the recent history of responses, as with habituation and sensitization (figure 5a). In the case of habituation, production of the response declines with repeated presentation of the stimulus, whereas in the case of sensitization it increases. In 5b a second signal is involved in the control of the behavior; B occurs as a consequence of the conjunction of the two signals. In 5c the second signal serves as a contextual modifier. 5c is similar to 5a inasmuch as the effect is modulation of a first order relation, but it differs in that the modulatory signal comes from beyond the immediate signal-behavior mechanism. An initial question to ask is how these forms of behavior control relate to those of figure 1. In particular, can they be thought of as variants of 1b? 5a and 5c seem to fit, because in each case there is the discrimination of an additional condition which is used to provide control for behavior modification. 5b doesn’t quite fit because the additional discrimination doesn’t provide control for a behavioral modification, it refines the production of a given behavior.

These comparisons help to make it clearer that complexification can take some subtle forms. We might simplistically think that increase of behavioral flexibility will occur through the addition of a second complete signal-behavior pathway, but 5a and 5c highlight the possibility that behavioral flexibility can be achieved through the addition of regulation to a given pathway. 5b makes it apparent that controlling behavioral flexibility isn’t the only adaptive reason for making additional discriminations. In the case of 5b the additional discrimination serves to make production of the behavior more focused. The last point has implications for Godfrey-Smith’s analysis of the evolution of behavioral complexification. 

This analysis is intended to specify the conditions in which it is adaptive for the organism to discriminate environmental complexity, or as GodfreySmith puts it, answer the question “[w]hen should environmental complexity bring it about that the organic system will make a distinction, will attend to a difference in the world?” (1996, p. 11). Thus, Godfrey-Smith’s account is based on the model of perceptual complexification shown in figure 6a. However in light of 5b we can identify 6b as a distinct model of perceptual complexification. Godfrey-Smith’s account could be used to predict 5a and 5c, but it doesn’t predict 5b. If 5b is an empirically significant form of behavior control, as it surely is, we should prefer the 6b model of perceptual complexification. This model draws our attention to the problem of behavior targeting as a general issue that can take a number of forms. That a given perceptual signal can adequately target a given behavior is taken for granted by the conceptualization of the problem of perceptual complexification schematized in figure 1. However this isn’t always the case; multiple perceptual signals may be needed for the effective targeting of an action.

Of course, using multiple signals to control behavior is just what Sterelny’s concept of robust tracking is about. However the distinction between 6a and 6b raises questions about the relationship between Sterelny’s account and that of Godfrey-Smith. Although Sterelny describes Godfrey-Smith’s analysis of the evolution of behavioral flexibility as a framework, it isn’t clear exactly how this is so. On the face of it Godfrey-Smith’s stage 2 (1b) is different to Sterelny’s stage 2 (2b). Moreover Sterelny’s stage 2 looks like it’s driven by a different problem — that of targeting a given behavior — than GodfreySmith’s stage 2, which is driven by the problem of flexible control. A plausible interpretation is that Sterelny sees Godfrey-Smith’s account as explaining how detection systems evolve, and his account as covering, not the pressures that add further detection systems, but selection for more complex forms than detection. An explicit analysis of the relations between the two accounts would be helpful, however.

Are the forms of behavior control shown in figure 5 cases of robust tracking? They each involve the use of multiple signals to control a behavior, and at one point Sterelny defines robust tracking as the ability to use several cues to control behavior (pp. 28-29). However Sterelny at another point says that robust tracking systems track “important features of [the] environment” (p. 17), and this is shown in 2b. But there is no requirement that the multiple signals of 5b and 5c have a single environmental source. Indeed, for both 5b and 5c S1 might have an external source and S2 might have an internal source, or vice versa. In the case of 5c contextual control signals will generally come from a different source to the primary signal whose effect is modulated; this is part of their functional value. There are at least two different classes of things that can be tracked: (i) particular environmental (and internal) states and entities, and (ii) the conditions for behavior production. Much of the way that Sterelny talks about robust tracking suggests that he has (i) in mind, but, as discussed below, the account treats (i) and (ii) as effectively equivalent. Yet (i) and (ii) can come apart, and 5c in particular looks like it’s aimed at regulating behavior production rather than tracking some particular thing in the world.

Behavior management 
Figure 7 depicts three forms of behavior management. As evolutionary complexification proceeds, and animals gain an increasing variety of behaviors, the need for coordination between behaviors increases. Behavior management mechanisms reduce this problem by providing coordination and resolving conflict. Figure 7a shows an elementary reciprocal modulation arrangement in which activation of one pathway influences the other; these links can inhibitory or excitatory. Conceptually it parallels 5a, inasmuch as the modulatory signals are internal. In the case depicted in 7b, S3 is a context signal which favors B1 in some circumstances and B2 in other circumstances; this parallels 5c. 7c shows a more complex arrangement involving a specialized arbitration system. One advantage of a specialized behavior management system like 7c, as compared with 7a and 7b, is that the arbitration itself can be context-sensitive. 

Thus, whereas 5a, 5c, 7a, and 7b all show first order control flexibility, 7c shows second order control flexibility. Like 7b, 7c incorporates a context signal, but it also illustrates several more sophisticated behavior management mechanisms. Prospective arbitration receives sensory signals, anticipates response conflict, and inhibits one or more conflicting behaviors. Retrospective arbitration detects the actual occurrence of response conflict and arbitrates the conflict. These two mechanisms can be coupled via learning: retrospective arbitration can manage novel response conflicts, and learning can transfer retrospective to prospective control.

It isn’t clear how behavior management relates to Sterelny’s taxonomy of figure 2. As was noted above, he defines a detection agent as an organism equipped only with detection systems. However he also says that detection agents can learn (pp. 14, 17) and have motivation systems (p. 20). This doesn’t seem to be entirely coherent. If an agent has a motivation system then at least some of its behaviors are being governed by multiple signals. Depending on the interpretation of robust tracking this might count as robust tracking, but it doesn’t fit the definition of a detection system: a modulated sensory-behavior relation (5c) is different to an unmodulated sensory-behavior relation (2a). It is also hard to see how detection agents could learn since all their perceptual discriminations are coupled to behaviors. Learning requires flexibility in the coupling between perception and behavior, and Pavlovian learning will yield robust tracking, not detection, as Sterelny himself notes elsewhere (2001, p. 271).（essay代写)

51Due网站原创范文除特殊说明外一切图文著作权归51Due所有；未经51Due官方授权谢绝任何用途转载或刊发于媒体。如发生侵犯著作权现象，51Due保留一切法律追诉权。
更多essay代写范文欢迎访问我们主页 www.51due.com 当然有essay代写需求可以和我们24小时在线客服 QQ:800020041 联系交流。-X（essay代写)