A firm's innovativeness is driven by its ability to recombine existing technologies. Elaborating on this argument, we contend that there exist two distinct types of recombinant capabilities. First, firms may innovate through recombinant creation, i.e., by creating technological combinations new to the firm. Second, they may innovate through recombinant reuse; i.e., by reconfiguring combinations already known to the firm. We study what drives each type of capability by examining two factors: the degree of integration of a firm's intraorganizational network and the diversity of its knowledge base. We test our theoretical predictions using data on 126 semiconductor firms between 1984 and 2003. Our analyses indicate that factors that favor recombinant creation generally hinder recombinant reuse and vice versa; however, combining an integrated collaboration network and a diverse knowledge base may concurrently enhance both recombinant capabilities. Copyright © 2013 John Wiley & Sons, Ltd.

Integrating insights from cognitive psychology into current network theory on the social capital of brokering and closed networks, we argue that cognitive style is a critical contingency explaining the relation between social network position and innovative performance. Based on a "complementary fit" argument, we posit that a social network rich in structural holes enhances the innovative performance of employees with an Adaptive cognitive style; however, individuals with an Innovative cognitive style are most innovative when embedded in a closed network of densely interconnected contacts. Using data on the individual cognitive style and complete workplace social network of all employees within a small design and manufacturing firm (N=68), we show that our theorized contingency mechanism accounts for a large share of empirical variation in employee innovative performance over and above existing social network explanations. AcknowledgmentWe would like to thank Stefano Brusoni, Martin Goossen, Gudela Grote, Eric Quintane, and participants in the 1st Workshop on the Micro-foundations of Social Networks (CBS, Denmark), for their comments and suggestions. We are grateful to Valentina Iannuzziello, who provided research assistance during the data collection and initial phase of this project as part of her final MA thesis. Finally, we thank the editor and three anonymous reviewers for their thorough and constructive feedback. Abstract:Integrating insights from cognitive psychology into current network theory on the social capital of brokering and closed networks, we argue that cognitive style is a critical contingency explaining the relation between social network position and innovative performance. Based on a "complementary fit" argument, we posit that a social network rich in structural holes enhances the innovative performance of employees with an Adaptive cognitive style; however, individuals with an Innovative cognitive style are most innovative when embedded in a closed network of densely interconnected contacts. Using data on the individual cognitive style and complete workplace social network of all employees within a small design and manufacturing firm (N=68), we show that our theorized contingency mechanism accounts for a large share of empirical variation in employee innovative performance over and above existing social network explanations.

SummaryWe examined the role of social embeddedness in creating positive perceptions of organizational support among managerial and staff employees (n = 72) of a large manufacturing firm. We operationalized social embeddedness as the size, density, and quality of employees' networks of multiplex, reciprocated exchange relationships with colleagues. After controlling for support from supervisors and upper management, we found all three aspects of social embeddedness to be associated with perceived organizational support (POS). This research suggests that in addition to the top-down influence of the organizational hierarchy, POS results from the organizational community within which employees are embedded through their social network in the workplace.

Why do certain domains of knowledge grow fast while others grow slowly or stagnate? Two distinct theoretical arguments hold that knowledge growth is enhanced by knowledge specialization and knowledge brokerage. Based on the notion of recombinant knowledge growth, we show that specialization and brokerage are opposing modes of knowledge generation, the difference between them lying in the extent to which homogeneous vs. heterogeneous input ideas get creatively recombined. Accordingly, we investigate how both modes of knowledge generation can enhance the growth of technology domains. To address this question, we develop an argument that reconciles both specialization and brokerage into a dynamic explanation. Our contention is that specializing in an increasingly homogeneous set of input ideas is both more efficient and less risky than brokering knowledge. Nevertheless, specializing implies progressively exhausting available recombinant possibilities, while brokerage creates new ones. Hence, technology domains tend to grow faster when they specialize, but the more specialized they become, the more they need knowledge brokerage to grow. We cast out our argument into five hypotheses that predict how growth rates vary across technology domains.

Organizational network research has demonstrated that multiple benefits accrue to people occupying brokerage positions. However, the extant literature offers scant evidence of the process postulated to drive such benefits (information brokerage) and therefore leaves unaddressed the question of how brokers broker. We address this gap by examining the information-brokerage interactions in which actors engage. We argue that the information-brokerage strategies of brokers differ in three critical ways from those of actors embedded in denser network positions. First, brokers more often broker information via short-term interactions with colleagues outside their network of long-term relationships, a process we label “unembedded brokerage.” Second, when they engage in unembedded brokerage, brokers are more likely than are actors in dense network positions to intermediate the flow of information between the brokered parties, consistent with a tertius gaudens strategy. Conversely, and third, when they broker information via their network of long-term ties (embedded brokerage), brokers are more likely than are densely connected actors to facilitate a direct information exchange between the brokered parties, consistent with a tertius iungens strategy. Using a relational event model, we find support for our arguments in an empirical analysis of email communications among employees in a medium-sized, knowledge-intensive organization, as well as in a replication study. The theory and evidence we present advance a novel, temporal perspective on how brokers broker, which reconciles structural and process views of network brokerage. Our findings substantiate the notion of brokers as a dynamic force driving change in organizational networks, and they help to integrate within a unitary explanatory framework tertius iungens and tertius gaudens views of brokerage.