This form of intermediary exists in all areas of social life. For example, professors write letters of recommendation to prospective employers about students, and persons seeking a job or a loan list other persons who will recommend them. The acceptance of a recommendation by a prospective employer or creditor is a placement of trust in the judgment of the intermediary, which allows a placement of trust in the ability of the prospective trustee to perform as expected. If the latter defaults, then the trustor's trust in the intermediary's judgment is reduced.

A social actor can advertise its position in the social cohesiveness structure of social capital as a valuable resource. Consider, for example, the language used by Mayfield, a venture capitalist firm, in its promotional brochure; as quoted by Freeman (this volume): 'Because of our long association with a large number of successful companies and entrepreneurs, a relationship with Mayfield is highly regarded. It can enhance the credibility of a young company with potential customers, vendors and employees, and with other financial institutions.' Similarly, professional service firms, such as advertising agencies and investment banks, 'sell' access to the social capital inherent in their networks. Indeed, advertising agencies that occupy a central position in the market are likely to be kept by their corporate clients, indicating the value clients place on this structure of social capital (Baker, Faulkner, and Fisher 1998).

The level of trust in the social cohesiveness structure is probably higher, on average, than the level of trust in the tertius gaudens arrangement. This is one reason why some critics argue that structural-holes theory is not a theory of social capital (see above). Because the tertius gaudens exploits the structural hole between two alters, the level of trust is presumed to be low. This is not necessarily so. Each alter may trust ego, even though the alters are unaware of each other's existence; in other cases, the alters may prefer to remain out of contact with each other, relying instead on their trust in ego. Moreover, the issue of trust in the social cohesiveness structure is not as unambiguous as it might seem. Sometimes an 'intermediary in trust' runs the risk that he or she will be circumvented or 'cut out' by opportunistic alters. For example, the risk that the principals in a real estate transaction might consummate the deal in secret, saving the commission owed to the broker, is so high that the standard legal contract between a real estate broker and seller contains protections against such actions.⁵

The distinguishing feature between the two structures of social capital is not trust, but the answer to this question: What does the social entrepreneur do with the gap between alters? While one social entrepreneur exploits the structural hole, keeping alters apart, another social entrepreneur may choose to close the gap, bringing together the two alters. These actions represent the two basic entrepreneurial strategies for accessing and mobilizing the social capital inherent in social networks. We next describe and illustrate these strategies, followed by our analysis of the relationship between entrepreneurial strategies and institutional context.

Two Entrepreneurial Strategies
Strategy is used here to refer to a social entrepreneur's plan of action for using network structure to access and mobilize social capital. Burt's (1992b) conceptualization of social capital as the structured absence of ties favors what we call 'disunion strategies.' In this strategy, the social entrepreneur generates 'profit' by taking advantage of the disconnection of the two parties. The disunion strategy is

illustrated in figure 1. As shown, ego (the tertius gaudens) is linked to two disconnected alters. The hole is represented by the blank space between alter I and alter 2, and enables ego to play the two alters against each other or to secure a valued resource from one and provide it to the other, extracting a profit in the exchange. In such a case, ego benefits from the absence of a connection between the two alters, may act to keep the alters apart, and at the very least, chooses not to introduce the two alters. For disunion strategies, value is created by the exploitation of these structural holes; value, therefore, is a 'private good' — the benefits of social entrepreneurship accrue to the third party (ego).⁶

The view of social capital as social cohesiveness leads to the alternative entrepreneurial strategy we call 'union strategy'—illustrated in figure 1. In this case, ego (Simmel's non-partisan or arbiter) is linked to two alters who are disconnected or in conflict. Ego 'closes' the gap between alters by bringing them together or by resolving their differences. This suggests a sharing or exchange of resources. The union strategy produces what Coleman (1988: 107) calls 'closure' — a social structure that creates the conditions for the enforcement of norms through 'sanctions that can monitor and guide behavior.' The combination of norms and trust that emerge under these structural conditions facilitates additional use of the union strategy.

Disunion Strategies in Action
Disunion strategies are common in situations where the formal differentiation of the organization presents the social entrepreneur with many temptations to play one person (or department) against another person (or department). Burt's (1992b) original study of managers and structural holes was conducted in one such organization. Consider, for example, an entrepreneur with ties to a person in the sales department who collects current information about customer needs, and to another person in the marketing department who is desperate for customer input for a new product. The entrepreneur can position him or herself to sales as someone with influence over the company's product development process and to marketing as a source of new information (Obstfeld 1997). With this disunion strategy, the entrepreneur seeks to benefit without ever introducing the two alters, a move that would eliminate the advantageous position. A similar case is the classic role of the boundary spanner who links two disjoint groups (Friedman and Podolny 1992).

Competitive markets are principal locations of disunion strategies at the inter-organizational level. Simmel (1950) notes that the market is a prime example of the tertius gaudens strategy writ large. Here, disunion strategies are characterized as rivalry,' where two or more sellers vie 'for opportunities of exchange' with a buyer (Weber 1978: 63; Swedberg 1994: 271). For example, Coca-Cola Company maintains relationships with six different advertising agencies (Baker, Faulkner, and Fisher 1998: 149), playing one advertising agency against the other in an elaborate disunion strategy. The practice of competitive bidding is based on the disunion strategy, where multiple sellers are pitted against each other. For example, disunion strategies are evident in the garment industry studied by Uzzi (1996a) in cases where dress manufacturers (buyers) select contractors (sellers) on the basis of price alone to effect discrete, nonrecurring exchanges. Of course, sellers in an industry suffering from intense competition caused by the buyers' relentless use of the disunion strategy may become motivated to collude, employing an illegal union strategy to counterbalance the power of buyers (Baker and Faulkner 1993).

Disunion logic drives the avoidance of ties in a competitive market. Competing companies will not use the same supplier because doing so would put the supplier in the structural position of the tertius gaudens. For example, General Motors avoids using the investment bank Goldman Sachs because its chief American rival, Ford Motor Company, uses Goldman as its main bank (Baker 1990). Similarly, corporate competitors avoid using the same advertising agencies, citing 'conflict of interest' as their rationale (Baker, Faulkner, and Fisher 1998).

The social structure of disunion strategies can be fluid, particularly in dynamic markets. When AT&T and China began negotiations to install an undersea cable system to provide a telecommunication link between China and the U.S., the entrepreneurial leverage associated with the disunion strategy shifted from one actor to another. AT&T initially approached the negotiation as the owner and operator of international marine cables, enjoying an advantageous negotiating position (Glain 1997). Deregulation and other changes in the telecommunication industry, however, unleashed a variety of competitors (such as Baby Bell SBC Communications, Nynex, Britain's Cable & Wireless, and Japan's Kokusai Denshin Denwa Co.) that the Chinese invited into the negotiations. One of China's key negotiators indicated, 'Our general policy is to not engage in projects that exclude other parties. We want to engage as many companies as possible on an equal basis' (quoted in Glain 1997). This approach nullified AT&T's disunion strategy by leveraging confidential information provided by the competing companies and the disconnections between them. Ultimately, the Chinese insisted on a 14-member consortium that included all of AT&T's major rivals.

Finally, direct exploitation of disconnected parties may represent only a fraction of the activities associated with disunion strategies. Some disunion strategies require constant effort and vigilance (perhaps even subterfuge) to keep alters apart and to maintain the structural conditions of disconnection, secrecy, and concealment. Price-fixing conspiracies, for example, require conspirators to deliberately maintain the ignorance of their corporate customers (Baker and Faulkner 1993). Some corporate actors attempt to prevent competitors' actions or the development of new regulations that would reduce or close the structural holes in markets, fighting to maintain the structural conditions that favor disunion strategies. For example, the strategic alliance of IBM and Apple was opposed by many competitors (Baker 1994a) who viewed the potential union as a threat that would reduce their ability to compete for customers — that is, to operate using the disunion strategy.


Union Strategies in Action
The formal differentiation of an organization does not always lead to disunion strategies. For example, Kanter (1983a: 141) describes the union strategy used by a corporate entrepreneur who created an internal alliance between sales, service, and product development:

He first wrote a memo to all of the sales people in his area, copying the district managers for service and products.... He then held a series of sales meetings, inviting commercial and service staff too.... [He] explained and re-explained the benefits of cooperation across the sales/service/products boundaries to people from each function. (Ashkenas et al. 1995: 17)

Similarly, Burgelman (1991) depicts how a corporate entrepreneur at Intel collaborated with two other product champions to develop RISC processor technology and to line up a customer base in advance of Intel's entry into the new market. Ashkenas et al. (1995) describe union strategies used to transform General Electric's Retailer Financial Services into a 'boundaryless' organization. For example, a systems manager employed union logic to streamline a business process: 'Nastasi then brought together a group of systems, marketing, finance, and customer service people and challenged them to complete new customer conversions in a matter of days, not weeks.' In many cases, union strategies are used by senior managers targeting major organizational change. Union strategies are also initiated by relatively low-ranking employees with more modest objectives.

Union strategies exhibit variation in official support. General Electric CEO Jack Welch, for example, officially sponsored GE's Corporate Executive Council to stimulate the exchange of information and collaboration (Baker 1994a,b). In contrast, Hutt, Reingen, and Ronchetto (1988) describe the emergence of a new product through the collaboration of actors from multiple divisions along with key customers — all acting outside prescribed corporate guidelines. Similarly, the successful Intel intrapreneurs described above went against explicit corporate policy and had to disguise the RISC project as just another 'co-processor' until sufficient progress had been made (Burgelman 1991).

Union strategies also occur in markets. Some union strategies involve the introduction of initially unconnected suppliers or customers via common third parties. Starr and MacMillan (1990:86) describe a case of an entrepreneur who introduces two suppliers:

A corporate entrepreneur saw an opportunity to connect two of his major suppliers while developing a new medical products business. One manufacturer, in the south, had cheap labor costs and good employee morale, but was nonetheless losing money due to lack of business. The other, in the north, had high labor costs and an employee shortage. By introducing these two manufacturers the venture manager reduced the costs of one and increased the sales of the other.

Uzzi (1996a: 679) describes a similar union strategy used in the New York garment industry:

One CEO explained how [a] tie formed between him and a manufacturer named 'Diana.' He said that his contact with Diana began when Norman, a close business friend of his and Diana's, asked him 'to help Diana out' in a time of need (cut her fabric at a special price and time), even though he had no prior contact with her.... [The CEO said] 'So why did I help her out? Because Norman asked, 'Help her out.'

The formation of strategic alliances — joint ventures, technology sharing, marketing arrangements, product development, and others — can result from union strategies. For example, Corning introduces alliance partners to each other, fostering creation of ties between them. AT&T Global Information Solutions convenes an annual conference in which its strategic allies meet in 'alliance fests' used to generate new alliances and associations (Baker 1994b). Digital Equipment Corporation (DEC) invites its alliance partners to a four-day conference to share information and foster alliance formation (Gulati 1995a). Similarly, the monthly meetings of the 128 Venture Group were convened by an entrepreneur and venture capitalist to establish a place where venture capitalists, entrepreneurs, consultants, and management team candidates could meet and explore collaborations (Nohria 1992b).

The introduction of disconnected parties is only a fraction of the activities associated with union strategies. Union activities include investing in already established ties, cultivating ongoing collaborations, and maintaining the general structural conditions that facilitate union strategies. Some union strategies are self-sustaining, such as those associated with the trading groups in the industrial districts of north central Italy and southwestern Germany (Powell 1990). In other cases, however, such as real estate brokerage, the repeated cultivation of new alters is necessary to stay in business. Of course, the real estate broker with a good reputation and lots of contacts enjoys the benefits of union strategies in reverse, as satisfied customers refer new alters to the broker.

These examples illustrate the general principle that union logic drives the selection of ties, whereas disunion logic drives the avoidance of ties (as described above). Disunion strategies proscribe the use of common third parties, such as suppliers, while union strategies prescribe the use of common third parties, such as alliance partners. These strategies may exist side by side in the same institutional context, which is a topic we take up in the following section.


INSTITUTIONAL CONTEXT
So far, we have not been very specific about the context in which social entrepreneurs operate. This was intentional, so that we could be clear about the distinctions between the two structures of social capital and their corresponding strategies. However, the relationship of structures and strategies to institutional context is critical. First, we argue that the nature, level, and forms of social capital — and therefore the strategies social entrepreneurs employ — depend on the structure and culture of the institutional context. This relationship holds in the institutional context of an organization as well as that of organizational fields, business sectors, industries, and markets. Second, we argue that the frequency, legitimacy, and success of an entrepreneurial strategy depends on its 'fit' or compatibility with the institutional context in which it is used.

Table 1. Institutional Context and Entrepreneurial Strategies

Institutional Context Disunion Strategy Union Strategy Structural Conditions Size Density Connectivity Formal differentiation Cultural Conditions Rules of exchange Norms Orientation Large size Sparse networks Disconnected networks Many formal boundaries Competition Opportunism, distrust Individualist orientation Small size Dense networks Integrated networks Few formal boundaries Cooperation Reciprocity, trust Collectivist orientation

Nahapiet and Ghoshal (1998) argue that the 'firm' is a better institutional setting than the market for the development of high levels of social capital; because the firm is a 'social community' (Kogut and Zander 1996: 503), it enjoys an 'organizational advantage' over markets. However, some firms are organized and operated as markets,' and some markets are organized and operated as 'firms' (Eccles 1981; Eccles and White 1988; Stinchcombe 1985). Therefore, we make the bold assumption that the characteristics of 'institutional context' can be specified in such a way that they apply to both firms and markets. An advantage of our approach is the generalization of the concepts of corporate social capital across institutional levels — that is, within firms and between firms. The relationship between institutional context and strategy is summarized in Table 1.

Structural Conditions
By structural conditions, we refer to the 'network configuration' of an institutional context, consistent with Nahapiet and Ghoshal's (1998) definition of the 'structural' dimension of social capital. Standard network measures can be used to represent a configuration. Nahapiet and Ghoshal (1998) propose density, connectivity, and hierarchy. We substitute 'formal differentiation' for hierarchy, because an organization can be differentiated along three dimensions — spatial, horizontal, and vertical. Markets, too, are organized along these three dimensions, as economic geographers have documented. Formal divisions such as these, whether in firms or markets, create structural holes. We add 'size,' since the number of social actors influences the fragmentation of an institutional setting, and along with it, the number and extent of structural holes.

The structural conditions associated with disunion strategies are large size, sparse and disconnected networks, and many formal boundaries. The design of most large-scale, traditional organizations favors disunion strategies because it creates so many structural holes between departments, across levels, and between spatially separated operations. Formal differentiation, for example, hampers collaboration across boundaries (Lawrence and Lorsch 1986). Large size disfavors the integrated 'network' organizational design (Baker 1992a). Similarly, markets with many players are more fragmented than markets with few players, producing many structural holes. The alert tertius gaudens in a fragmented market generates profit by arbitraging across these gaps in the social structure of trading (Baker 1984), just as the tertius gaudens in a fragmented firm generates profit by exploiting the disconnection inside the organization (Burt 1992b).

The structural conditions favoring union strategies are small size, dense and integrated networks, and low differentiation (see Table 1). Union strategies are common in network organizations, for example, because these organizations are characterized by flexibility, lateral ties, and a high degree of integration across low formal boundaries (Baker 1992a). A dense network of customers, producers, and suppliers provides a structural basis for cooperation (Perrow 1992), such as the dense social networks of firms, local universities, community colleges, research institutes, financial institutions, trade associations, and regional governments in Silicon Valley (Saxenian 1991, 1994).

The biotechnology industry features dense and well-connected networks of interfirm cooperation as well (Walker, Kogut, and Shan 1997; in this volume see also Stuart, Smith-Doerr et al., and Omta and Van Rossum). These structural conditions are conducive to the formation of new alliances based on the union strategy. For example, the likelihood that two previously unallied firms will form an alliance increases with the number of third-party ties they have in common (Gulati 1995a). Similarly, the greater the number of research and development alliances and other types of collaborations a biotechnology firm has at a given time, the more diverse its future portfolio of ties will become (Powell, Koput, and Smith-Doerr 1996).


Cultural Conditions
Cultural conditions can be defined in many ways. Indeed, the definition of culture itself is a subject of considerable debate and rival interpretations (DiMaggio 1994; Scott 1995). Rather than trying to resolve this debate, we propose a simple definition of cultural conditions — the institutionalized rules of exchange, norms, and social orientation in an institutional context — and focus primarily on the relationship between strategy and cultural context.⁸

Rules of exchange are shared social understandings about 'who can transact with whom and the conditions under which transactions are carried out' (Fligstein 1996: 658). For example, exclusivity (sole-source) is a rule of exchange governing buyer-seller relationships in the advertising industry (Baker, Faulkner, and Fisher 1998: 15 1). Competitive rules of exchange prohibit union triads. The avoidance of common suppliers is a good example (Baker 1990). Disunion strategies are legitimate and successful in firms operated as 'markets' (Eccles and White 1988). In settings such as markets operated as 'firms' (Stinchcombe 1985), however, cooperative rules of exchange discourage disunion strategies (which would be interpreted as self-serving, opportunistic behaviors). Similarly, organizations in small-firm networks share information, establish long-term relationships, and support each other's efforts, in opposition to the competitive rules of exchange in classic markets (Perrow 1992). The search for greater efficiencies through closer ties between customers and suppliers can displace traditional rules of exchange that pit suppliers against one another. New developments in just-in-time (JIT) inventory control, for example, call for close coordination between customers and their suppliers. Honeywell orchestrates close cooperation between its internal buyers and five suppliers who would ordinarily compete for business (Bleakley 1995).

Norms are shared expectations that regulate behavior (DiMaggio 1994), including choice of entrepreneurial strategy. Norms of openness, teamwork, trust, and reciprocity favor union over disunion strategies. For example, Putnam (1993a) argues that norms of reciprocity in the industrial districts of northern Italy foster interfirm cooperation and limit opportunistic behavior — that is, the use of disunion strategies. Similarly, norms of reciprocity support the union strategies evident in such business groups as the Japanese kieretsu and Korean chaebol.

Social orientation refers to the distinction between individualism and collectivism (Triandis 1995: 2). Individualism, for example, is 'a social pattern that consists of loosely linked individuals who view themselves as independent of collectives; are primarily motivated by their own preferences, needs, rights, and the contracts they have established with others; and emphasize rational analyses of the advantages and disadvantages of associating with others' (Triandis 1995: 2). Individualism favors disunion strategies; collectivism favors union strategies. For example, collectivist practices such as cross-functional teams, multi-level management networks, group-level reward systems, and team-building programs, facilitate union strategies within organizations (Baker 1994a). Recent research has shown that a collectivist orientation often increases the odds of alliance formation (Dickson and Weaver 1997). Private economic associations and political organizations foster cooperation in the industrial districts of northern Italy (1993a). The robust collaboration found in these industrial districts is supported by technical colleges, vocational training, supportive banks, and extended kinship ties (Powell 1990).

Distribution of Strategies
We argue that the frequency, legitimacy, and success of an entrepreneurial strategy depends on its 'fit' with the 'design' of the institutional context in which social entrepreneurs operate. This suggests the following proposition: The ratio of disunion to union strategies varies according to the structure and culture of the institutional context. At one extreme, disunion strategies dominate in settings characterized by sparse, disconnected, and differentiated networks, along with competitive rules of exchange, opportunism, and an individualist orientation; at the other extreme, union strategies dominate in settings characterized by dense, connected, and undifferentiated networks, coupled with cooperative rules of exchange, reciprocity, and a collectivist orientation. A mix of strategies occurs in an institutional context located between these two extremes.

The distribution of strategies in a given context can be determined by a triads census (Wasserman and Faust 1994: 556-602). The pattern of strategic alliances in the global automobile industry⁹ illustrates one empirical distribution (Baker 1992b), though the triads approach can be applied in any institutional context. For strategic alliances, a triads census includes only four possible triad isomorphism classes: 1) the null triad, composed of three unallied firms; 2) the dyad, composed of three firms of which only two are connected by an alliance; 3) the disunion triad, composed of two alliance dyads; and, 4) the union triad, composed of three alliance dyads.¹⁰ The first class, the null triad, represents a classic competitive situation — the complete absence of strategic alliances among three firms. The second class, the dyad, represents an isolated alliance between two firms. The third class, the disunion triad, represents the terrius gaudens arrangement (as illustrated in figure 1). The fourth class, the union triad, represents the social cohesiveness structure of social capital (see figure 1).


The empirical distribution obtained in the triads census of the automobile industry is shown in Table 2. For our purposes, we focus here on the two triad types of particular interest, disunion and union triads. Both triad types occur much more often than expected by chance alone, considering all alliance types combined."¹¹ However, if we examine the triads census for each type of alliance, we find an interesting pattern: 1) Disunion triads (but not union triads) occur more often than by chance in joint ventures. 2) Disunion and union triads occur more often than by chance in two types of alliances — technology sharing and manufacturing/assembly. 3) Union triads (but not disunion triads) occur more often than by chance in three types of alliance — supplier ties, marketing/distribution, and equity investments.

This pattern suggests that the structure of social capital varies by alliance type. The structure of joint ventures implies that social capital is accessed and mobilized by exploiting structural holes, using the tertius gaudens strategy. The structure of supplier ties, marketing/distribution, and equity investments implies that social capital is accessed and mobilized by closing structural holes. The mixed structure of technology sharing and manufacturing/assembly suggests the presence of both types of social capital. However, the statistically significant use of both disunion and union strategies may indicate a competition of strategies in which neither dominates.

The general relationship between entrepreneurial strategies and institutional context is illustrated in figure 2. This stylized representation implies the possibility of change or movement, of the transformation of an institutional context and its corresponding strategies. The intersection of the curves in figure 2 represents a balance of disunion and union strategies. This point is an unstable state in which neither strategy dominates, such as the mixed structure of triadic strategies for technology sharing and manufacturing/assembly in the automobile industry (Table 2). If so, then the automobile industry may fall back on earlier, simpler, and over-learned strategies (arms'-length competition), as people and organizations are prone when faced with uncertainty and ambiguity (Weick 1995:102).

For an organization, the point of intersection in figure 2 indicates a particularly risky stage in an organization's transition from one institutional design to another. For example, a firm attempting to improve collaboration and cooperation must foster both the structural and cultural conditions that favor the union strategy (moving from left to right in figure 2). Some consultants claim that a hierarchy can be converted into a network organization simply by adding links (Lipnack and Stamps 1994:72; see, also, Mueller 1986), but structural change is not enough. Failure to change both structural and cultural conditions endangers a change effort. For example, the effort to transform Industrial Computer and Control Group failed because change agents altered only organizational structure (Nohria and Berkley 1995). By replacing hierarchy with a network design, they may have induced more

union strategies, but by itself this structural change was not enough. Without a corresponding cultural change, the effort to foster cross-divisional collaboration was doomed. Employees did not fundamentally change their strategies for action (Nohria and Berkley 1995). The change effort may have collapsed at or near the point of intersection in figure 2.


CONCLUSION
Our chapter attempts to specify and clarify the dimensions and structures of social capital. We offer a set of concepts that encompasses different views of the structural sources of social capital, the basic strategies used to access social capital, and the relationship between institutional contexts and strategies. We argue that social capital 'inheres in the structure of relations between and among actors' (Coleman 1988: 98) in two fundamental ways — one based on the patterned absence of ties, the other on the patterned presence of ties. Structural holes theory (Burt 1992b) emphasizes the absence of ties, where social capital resides in a social structure characterized by sparse networks and few redundancies. We maintain that social capital also inheres in social structure as the presence of ties, which we call social acknowledge cohesiveness, where social capital is found in dense networks with multiple redundancies.

Each structure of social capital represents opportunities to access and mobilize the resources inherent in a social network. The first structure calls for a 'disunion' strategy that exploits the structural holes between alters by keeping them apart. The second structure calls for a 'union' strategy that creates value by bringing alters together, closing the hole between them. Our qualitative and quantitative examples provide ample evidence of the use of these strategies in a wide range of institutional contexts, both inside and between organizations. The concepts of two types of social capital (structural holes versus social cohesiveness) and two types of entrepreneurial strategies (disunion and union) can be generalized and applied to understand the structure and use of social capital in many different institutional settings.

The actual distribution of strategies in a particular institutional context depends on the 'design' of the institutional context in which social entrepreneurs operate. An entrepreneurial strategy that 'fits' the structure and culture of a given institutional context occurs more frequently, enjoys greater legitimacy, and will be more successful in the long run, compared to a strategy that does not fit. Disunion strategies dominate in organizations and markets characterized by sparse, disconnected, and differentiated networks, coupled with competitive rules of exchange, opportunism, and an individualist orientation; union strategies dominate in organizations and markets characterized by dense, connected, and undifferentiated networks, coupled with cooperative rules of exchange, norms of reciprocity, and a collectivist orientation.

Of course, a mix of triadic strategies falls between the pure disunion and union extremes. Thus, one avenue of additional research is to explore further the precise connection between institutional context and entrepreneurial strategy. This line of work calls for demography of social entrepreneurial relationships (see, also, Baker, Faulkner, and Fisher 1998: 173). The statistical methods available for taking a triads census (Walker and Wasserman 1987), which we used in our analysis of the distribution of strategies in the global automobile industry (Baker 1992b), can be used to determine the distribution of strategies in any organization, market, or organizational field. In addition, organizational change can be tracked by measuring shifts in the distribution of strategies over time, using methods designed to model longitudinal change in networks (e.g., Leenders 1995a, 1996; Wasserman and Faust 1994). By coupling new and better measurement methods (e.g., Borgatti 1997; Han and Breiger, this volume; Doreian, this volume) with more precise concepts of social capital and their corresponding entrepreneurial strategies, it is possible to take another step forward in the exploration of corporate social capital.

We are grateful to the editors of this volume for their helpful comments and suggestions. Direct correspondence concerning this chapter to Wayne Baker, University of Michigan Business School, 701 Tappan Street, Ann Arbor, MI 48109 (wayneb@umich.edu).


ENDNOTES
¹ The debate takes place in published works (see citations herein) as well as in informal discussions on the Internet, such as the recent interchanges on social capital in SOCNET that occurred between January and June, 1997. We cite the published literature to support our points, but we also acknowledge the contributions to the debate in SOCNET made by Xavier De Souza Briggs. Robert Putnam, Barry Wellman, and others.

² This point was also made in SOCNET comments by Barry Wellman and Robert Putnam (January 1997).

³ Simmel (1950) describes another triad type, 'divide and conquer,' in addition to the tertius gaudens and 'non-partisan' types. In 'divide and conquer,' the third party deliberately introduces conflict between the other two. This type is not our concem in this chapter.

⁴ Real estate brokerage and literary agency qualify as this type because the ego is not a principal in a transaction, but instead brings together the two principals (alters) who consummate the transaction. In the first case, the real estate broker (ego) matches a buyer and seller (alters) who exchange property for money; in the second, a literary agent (ego) matches a publisher and author (alters) who exchange 'the property' (manuscript) for royalties. Some types of brokerage do not qualify as this type. For example, in a wholesaler arrangement, the broker is a principal in one transaction (buying goods from a manufacturer) and in a second transaction as well (selling the goods to the consumer).

⁵ Consider, for example, the protections contained in the standard 'listing agreement' (officially, the 'cooperative selling contract') used in the state of Illinois to define the legal obligations of the seller to the real estate broker: 'SELLER SHALL: Cooperate fully with Broker; refer all inquiries to Broker. conduct all negotiations through Broker; ... and pay a real estate brokerage commission of X % of the sale price; if 1) Broker provides a purchaser ready, willing, and able to purchase in accordance with this Contract or 2) if the property is sold, exchange, gifted, or optioned by Broker or by or through any other person including the Seller during the period of this contract; or 3) if it is sold directly or indirectly within six (6) months after termination of this contract to a purchaser to whom it was offered during the term thereof.'

⁶ It is possible that the individual pursuit of entrepreneurial profit as a private good produces value at the collective level. This line of reasoning is consistent with traditional management and economic theories. For example, the tournament model of mobility assumes that competition among managers yields better ideas, more satisfied customers, and greater shareholder value. The theory of the market similarly assumes that individual striving maximizes social welfare. Burt (1992) does not emphasize the public-goods aspect of social entrepreneurship, though he mentions the possibility: An entrepreneur is 'a person who generates profit from being between others. A nonprofit player, pursuing entrepreneurial opportunities just for the pleasure of being the one who brings others together to build value, could choose to reinvest it all' and strengthen existing relationships (Burt 1992:34-35). Of course, social capital can be considered a public good in and of itself. Putnam argues, for example, that Coleman's (1988) original concept of social capital incorporates such a view. He adds, however, that the views of social capital as a private and public good are complementary (see discussions in SOCNET; see, also, Putnam 1993a, 1995; Knoke this volume).

⁷ Nahapiet and Ghoshal (1998: 261) briefly mention that some inter-organizational networks may develop an institutional context that is conducive to high levels of social capital, and point to this as an avenue of future research.

⁸ We acknowledge, but do not elaborate here, the importance of the 'constitutive' as opposed to the ,regulatory' view of culture (DiMaggio 1994). In addition, we acknowledge but do not discuss the state as part of the institutional context (Fligstein 1996).

⁹ We obtained data for the triads analysis from a special issue of Ward's Automotive International (1986). This issue reported the alliances known to exist among 37 major automobile companies, classified by the six alliance types reported in Table 2. We created a square, binary, symmetric matrix for each alliance type, where each company is assigned a row and corresponding column, and an entry indicates the presence or absence (0, 1) of an alliance between two companies. Thus, we created six 37 X 37 matrices. We analyzed the triads structure of each matrix using TRIADS (Wasserman and Walker, 1987). These data are treated as symmetric because alliances are naturally mutual ties, and we adjusted the weighting vector in TRIADS accordingly.

¹⁰ A triads census includes sixteen isomorphism classes when asymmetric ties are considered (Wassemian and Faust 1994: 566). Since alliances are mutual ties, however, there are only four classes of interest: triad types 003 (null), 102 (dyad), 201 (disunion), and 300 (union These three-digit numbers [MAN] refer to triad isomorphism classes using standard labeling, where M number of mutual dyads, A = number of asymmetric dyads, and N = number of null dyads.

¹¹ Structural hypotheses are tested using the TRIADS program (Walker and Wasserman 1987) in which the empirical distribution of triads is compared against a theoretical distribution of triads assuming random assignment of ties. The weighting vector used in these tests reflects the number of disunion