There are several reasons why the proposed model would be more advantageous than a one-sided dichotomous relationship model. Firstly, a feedback model posits that there are natural laws in which individual and collective categorizations both share (Nowak, Szamrej, & Latane, 1990). As an individual agent is governed by scientific and physiological laws so must the collective be subsumed under a law that would best represent how collective categorizations and representations operate.
Secondly, an integrationist feedback model is continuous model. A feedback model presupposes that both the individual and collective approaches are dynamic systems in which goals and needs are shaped and altered according to present circumstances (Triandis, Bontempo, Villareal, Asai & Lucca, 1988). The proposed feedback model implicates that the group/collective undergoes "learning" processes similar as an individual.
Thirdly, an integrationist feedback approach should be more sensitive and attuned to key changes and alterations in the environment. Gleitman (1995) likens the feedback approach as complex systems that exist in all levels of the nervous system in which the whole is greater than the sum of its parts (Gleitman, 1995).
Two Way Street: Language, Communication and Trust as a Bridge
The possibility of an integrationist perspective can be approached by observing key bridges and associations that encourage a two way interaction. This observational approach was one of the key highlights of a multiagent connectionist model done by Van Overwalle and Heylighen. Accordingly, both authors have emphasized that this model of multiagent systems originate in distributed artificial intelligence (Weiss, 1999) as well as complex, adaptive systems such as animal swarms and most importantly, human societies (Epstein & Axtell, 1996; Nowak, Szamrej, & Latane, 1990).
The multiagent connectionist approach, as proposed by Van Overwalle and Heylighen, place a crucial emphasis on a trust-relationship between individual agents. This connectionist model offers us as to how an integrationist feedback model would operate. The multiagent connectionist model consists of individual recurrent networks that communicate with each other. Although much has been said of individual and collective representations, the bridge that intimately links the two representations would fall on the shoulders of language and communication. Both of whom share well-founded evidences that support an integrationist feedback model.
Firstly, people validate their thoughts and ideas through comparison with others. People carry this on an individual level in which the information being categorized is compared with previous knowledge (Van Overwalle & Heylighen, 2006). From an integrationist perspective, Maass et al. have proposed that existing intergroup biases will ultimately produce a "biased" language which then will significantly contribute to the maintenance of existing biases (Maass, Salvi, Arcuri & Semin, 2000). That is to say that there exists differing levels of language abstraction in attributing characteristics to in-group or out-group behaviours. Consequently, once a specific characteristic has been attributed to the in-group or out-group, the subsequent information greatly influences information processing of both source and receiver of the communication (Maass, Salvi, Arcuri & Semin, 2000). We clearly observe processes that exist in both individual and collective representations, a flowing and dynamic process that propounds assertions of unification between the two representations.
Fundamentally and critically, we must observe that this phenomena occurs on the collective level influencing the individual and vice versa. Key findings would then need to be recognized. First, Van Overwalle and Heylighen have considered that social validation and trust may not become crystallized until the individual realizes that he or she is a part of the greater system of networks (Van Overwalle & Heylighen, 2006). That is, social validation and trust are needed by individuals for a group to exist. Secondly, these interactions, as suggested by multiagent connectionist model, will occur among agents in a framework of long-term, stable and trust factors (Van Overwalle & Heylighen, 2006). Once again, we see an interplay of modal factors (social validation and trust) that occurs both in the collective and individual approach.
Other findings equally generate parallel support to an integrative and dynamic model. Moscovici & Zavalloni (1969) have observed that judgments expressed by the group will often be adopted by individuals as their personal opinions. The extent to which personal opinions sway group dynamics acquires their support in Van Overwalle and Heylighen's (2006) connectionist model which examines trust as a major factor. Travis (1939) explains the biological imperative of the importance of interaction within the neural levels in man. Such implications suggest that physiological processes contribute greatly to individual and collective categorization. Through this physiological process; man finds meaning, relevance and rationality in the environment (Travis, 1939). Arising from an integrative perspective, Stangor and Schaller (1996) have argued that both individual and collective approaches converge towards the same underlying set of cognitive mechanisms. They have made further distinctions, in which, the role of fundamental human needs (explanation and prediction of the social world, and esteem maintenance) greatly contribute to the importance of recognizing individual and culturally shared stereotypes simultaneously (Stangor & Schaller, 1996).
