IBSS-L: Information Networks and the Evolution of Social Organizations

这个跨学科的研究项目将解释为什么不同的社会群体表现出特定的组织形式，通过提供一种严格的方法来确定社会群体的结构在多大程度上是由外部压力（如环境提供的需求和机会）而不是内部压力（如群体成员之间的竞争和不一致的激励）塑造的。 该项目将为社会和行为科学提供一个新的研究方向，探讨组织结构如何限制其处理环境信息的能力，并将提供新的见解，说明组织如何在内部根据行为者能力的限制传递信息。 该项目可以帮助团结社会和行为科学家，从关注灵长类动物群体和觅食社会的科学家到研究跨国公司结构的科学家。 它还将通过建议如何最好地组织现代企业和政府官僚机构来提供切实的回报。 该项目将开发和提供具有广泛学术价值的各种过去和现在社会组织的数据库，从史前社会到商业公司到军事组织，社会团体的组织方式截然不同，从平等的“水平”社会到深刻的“垂直”等级，从分散的模块化团队到集中的指挥和控制大会。 社会群体之所以展现出他们的组织，是因为这些结构优化了信息交换，还是仅仅是受当地历史先例的影响？ 本计画将利用最新发展的数学技巧，将“集体问题解决”与“认知限制”的概念正式化，以解决这个问题。“这些形式化将被用来调查哪些组织在不同的限制和群体规模下表现最好，以及哪些组织在成员或成员间联系被删除或群体压力发生变化时继续运作时最“稳健”。 然后，研究人员将确定这种最佳组织是否反映了真实的群体的组织方式。 最优内部组织的问题已经受到了从人类学到经济学的各个领域的关注。然而，一直没有一个统一的数学框架来研究不同类型的组织之间的关系和组织的压力和约束。本项目将通过结合图论和信息论的技术来开发这样一个框架。 图论研究社交网络的组织，并提供了丰富的词汇来量化组织属性，如层次结构，集中化和模块化。 信息论的方法可以量化群体成员之间处理或通信的信息比特，将社会群体概念化为电信网络，其中网络节点对应于群体成员，网络信道和节点上的“带宽限制”对应于认知约束。 研究人员将确定在不同带宽限制下最好处理信息的网络结构，并将结果与过去和现在人类组织的经验数据进行比较。 该项目通过NSF跨学科行为和社会科学研究（IBSS）竞赛获得支持。

This interdisciplinary research project will explain why different social group exhibits particular forms of organization by providing a rigorous way to determine the extent to which a social group's structure is shaped by external pressures, such as demands and opportunities afforded by the environment, rather than internal pressures, such as competition and misaligned incentives among group members. The project will constitute a new research thrust for the social and behavioral sciences regarding the ways that the structure of organizations constrains their ability to process environmental information, and it will provide new insights as to how organizations internally route information subject to restrictions on actors' capabilities. This project can help unite social and behavioral scientists ranging from those focusing on primate groups and forager societies to those studying the structure of multinational corporations. It will also provide tangible returns by suggesting how to best organize modern firms and governmental bureaucracies. The project will develop and make available databases on a variety of past and present social organizations which will be of broad scholarly value.Social groups ranging from prehistoric societies to business firms to military organizations are organized in dramatically different ways, from egalitarian "horizontal" societies to deep "vertical" hierarchies and from sets of decentralized, modular teams to centralized command-and-control assemblies. Do social groups exhibit the organizations they do because these structures optimize information exchange, or are groups simply channeled by local historical precedent? This project will address this question by using recently developed mathematical techniques to formalize the notions of "collective problem solving" and "cognitive constraints." These formalizations will be employed to investigate which organizations perform best under different constraints and group sizes as well as which are most "robust" in continuing to function when members or inter-member connections are removed or the pressures on the group change. The investigators will then determine whether such optimal organizations reflect the way real groups are (or have been) organized. The problem of optimal internal organization has received attention in various fields, from anthropology to economics. However, there has not been a unifying mathematical framework for studying the relationship between different kinds of organizations and the pressures and constraints on organizations. This project will develop such a framework by combining techniques from graph theory and information theory. Graph theory studies the organization of social networks and provides a rich vocabulary for quantifying organizational properties like hierarchy, centralization, and modularity. Methods from information theory can quantify the bits of information processed or communicated between group members, conceptualizing a social group as a telecommunication network where network nodes correspond to group members and "bandwidth limits" on the network channels and nodes correspond to cognitive constraints. The researchers will identify network structures that best process information under different sets of bandwidth limits and compare the results to empirical data on past and present human organizations. This project is supported through the NSF Interdisciplinary Behavioral and Social Sciences Research (IBSS) competition.