Building Bigness: Reputation, Prominence, and Social Capital in Rural South India

Anthropologists have long been concerned with how reputations help people gain the support of others. Here, we study the support ties among adult residents of two Tamil villages, asking how reputational standing in each village mediates access to social support. We find that a reputation for influence has the weakest effect on support ties with others, while a reputation for generosity has the strongest. Further, a reputation for influence is not associated with greater connections to people of “high position” outside the village. Given the weak effects of a reputation for influence, we turn to a network measure of social capital, weighted PageRank centrality. While persons who are recognized as influential often also have an influential network position, there are many others who have similarly high centrality—including, notably, many women. Consequently, we suggest that much of the anthropological evidence for the benefits of prominence may actually reflect the returns to greater social capital and that both may be shaped in large part by acts of generosity and mutual support. By studying social capital, we can achieve a more complete accounting of the many different social strategies employed by all persons, not simply the few who achieve prominence. [reputation, social capital, network analysis, prominence, India] RESUMEN Los antropólogos han estado preocupados por largo tiempo acerca de cómo las reputaciones ayudan a las personas a ganar el apoyo de otros. Aquı́, estudiamos los lazos de apoyo entre los residentes adultos de dos comunidades Tamil, preguntando como la posición en términos de reputación en cada comunidad media el acceso al apoyo social. Hallamos que una reputación por influencia tiene el efecto más débil en lazos de apoyo con otros, mientras una reputación por generosidad tiene el más fuerte. Además, una reputación por influencia no está asociada con mayores conexiones a personas de “alta posición” por fuera de la comunidad. Dados los efectos débiles de una reputación por influencia, recurrimos a una medida de red de capital social, la centralidad ponderada del Page Rank. Mientras personas quienes son reconocidas como influyentes también tienen una posición de red influyente, hay muchas otras quienes tienen similarmente centralidad alta —incluyendo, notablemente, muchas mujeres—. Consecuentemente, sugerimos que mucha de la evidencia antropológica sobre los beneficios de la prominencia puede reflejar actualmente los beneficios de un mayor capital social, y que ambos pueden ser estructurados en gran parte por actos de generosidad y apoyo mutuo. Al estudiar el capital social, podemos lograr una enumeración más completa de las diferentes estrategias sociales empleadas por todas las personas, no simplemente las pocas que logran prominencia. [reputación, capital social, análisis de red, prominencia, India] I n the summer of 2013, Arulprakaash and Selvi celebrated their daughter’s first communion. While some of AMERICAN ANTHROPOLOGIST, Vol. 120, No. 3, pp. 444–459, ISSN 0002-7294, online ISSN 1548-1433. C © 2018 The Authors American Anthropologist published by Wiley Periodicals, Inc. on behalf of American Anthropological Association. All rights reserved. DOI: 10.1111/aman.13100 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. their friends had very simple celebrations for their children’s first communions, Arulprakaash and Selvi went for a more Power and Ready • Building Bigness 445 elaborate affair. They arranged for a special service at the village church with the local priest. Posters and a speaker set loudly announced the event, and more than seven hundred people were invited. Well liked and well regarded, Arulprakaash and Selvi received so many guests that they had to erect a special canopy to provide shade for those who could not fit into their house. A relative stationed at the entrance received the small gifts of money (a few hundred rupees, or five to ten dollars) called moy brought by the guests. Literally, moy means a swarm or throng, and this happened doubly, as people crowded around the relative, and as the money piled up, filling an entire suitcase. The moy from their daughter’s first communion would not only cover the expenses of the function itself but also let them repaint their house and put down a new tiled floor. Arulprakaash was well aware that it was only through the help of others that they were able to do this: For me, the house finishing has been accomplished. This is happiness. Okay, at first, there was a house but not one in good condition. Now, you’ve seen the house. Now the house is good. From now on, with the salary I get, I won’t be able to do any more. I was able to do this because of the function. I got the loan and I repaid the loan. All paid completely. Separate, I cannot do this. If Arulprakaash had been “separate,” without the support of others, his house would still be a weatherworn soot gray. In fact, without support, his house would not exist; it had been constructed years earlier with the moy collected at his wedding. In this article, we ask how Arulprakaash and Selvi manage to get such support. Using social network analysis, we model the flows of support between people in two South Indian villages, focusing on the role that a person’s reputation plays in his/her relations with others. We compare the importance of a person’s reputation for generosity and good character, on the one hand, with their reputations for giving good advice and for wielding influence and authority, on the other. We see the former as capturing local conceptions of personal morality and virtue and the latter as capturing an assessment of a person’s prominence. We find that having a reputation for influence has a limited effect on getting help from others, both within and beyond the village. Instead, it is a reputation for generosity that has the strongest association with accessing support. This correlation is partially due to the strong effects of reciprocity and shared connections: those with reputations for generosity are more likely to be asked for help and so are more likely to get it. Further, a reputation for generosity is more closely associated with having an influential network position, as measured by weighted PageRank centrality, than a reputation for influence. In this setting, influence and authority are almost exclusively attributed to men, and often men of particular castes and wealth. Yet we find that there are many people beyond this narrow category of persons who, while not recognized as influential, actually are influential in terms of their network position. This observation, coupled with the stronger effect of generosity on access to social support, leads us to argue for a wider analytical and theoretical frame, moving beyond a focus on the influence and prestige of a few patrons, leaders, and “big men” to a broader focus on social capital. Such a perspective better allows us to interrogate how people for whom prominence is less accessible—especially women and, in this setting, people of particular castes—may nevertheless be able to build the “bigness” (perumai) that allows them to gain the support of others. With these findings, we speak to two distinct fields: South Asian studies and evolutionary anthropology. These seemingly disparate fields have both suggested that a good reputation can be consequential for people’s livelihoods, and both have been concerned with how people’s relationships with others are implicated in those reputations. Both subfields build from the fundamental observation that people’s relationships and exchanges with one another help shape their standing in society (Bourdieu 1986; Mauss 1954) but perhaps inevitably diverge from there, both in terms of the particular questions asked and in the types of evidence brought to bear on those questions. In South Asian studies, and cultural anthropology more broadly, the recent turn to study the ethical has highlighted how interpersonal relations become a site of moral tension and self-cultivation (Keane 2016; Laidlaw 2014; Pandian and Ali 2010; Robbins 2013). In evolutionary anthropology, there is an increasing interest in exploring the nuances of how people assess others and build relationships with them (Barclay 2013; Bliege Bird and Power 2015; Bliege Bird, Ready, and Power 2018; Lyle and Smith 2014; Macfarlan, Quinlan, and Remiker 2013). We find both to be compelling, if partial, accounts of the ways in which people maneuver within their social worlds. Here, we attempt to offer a rich account of that maneuvering by combining ethnographic observation with social network analysis, a methodological framework that allows for quantitative analysis while fundamentally recognizing the embedded nature of the social person. Our approach not only leads us to reaffirm some core anthropological insights (for example, the importance of reciprocity) but also helps us note new complexities, especially those that operate at the interstices between personal interactions and structural outcomes. This approach also allows us to capture the broader landscape of social strategies employed by persons of very different social and economic positions within the villages, including both men and women, as they attempt to achieve their goals. Having surveyed that landscape, we see generosity as a common driver of both prominence and social capital across all segments of the population. This leads us to argue that both evolutionary anthropology and South Asian studies would benefit from extending their focus from prominence to the more inclusive concept of social capital. REPUTATION, PROMINENCE, SOCIAL SUPPORT, AND THE SOUTH ASIAN PERSON In South Asian studies, it has long been recognized that the flows of support between persons are not only materially 446 American Anthropologist • Vol. 120, No. 3 • September 2018 important but also socially and morally meaningful. Indeed, foundational works have stressed how the South Asian person is fundamentally understood through interpersonal relationships and group identities, so much so that some contend that there is no individual in South Asia but only the “dividual” (Marriott 1976). In this conception, the person is not firmly distinct and bounded but rather is composed of biomoral substances that can be absorbed, dispersed, and combined as persons interact, fundamentally changing their being in the process (Daniel 1984). Ethnographies of South Asia, and South India in particular, have evocatively illustrated the ways in which people work to shape their porous selves (Daniel 1984; Pandian 2009; Prasad 2007) and have further shown how such work on the self involves the careful negotiation of one’s intimate relationships (Busby 2000; Trawick 1992; Venkatesan 2014). This body of ethnographic work, though, does not suggest that there is no individual but instead shows how South Asians navigate the ambiguities and tensions between individual autonomy and the collective. As M. Mines (1994) argues, for example, Tamil persons can be understood as “contextualized individuals,” read through the groups and identities to which they hold. He illustrates this most compellingly through his portraits of “big men” (periyavarkal ., influential local leaders) in Chennai who are able to wield influence through their positions of prominence within caste organizations, temple committees, charitable organizations, and other institutions. Such “big men” are marked by their relationships with others, particularly their patronage and the assistance they render to their constituents. What M. Mines’s (1994; see also M. Mines and Gourishankar 1990) work on “big men” (who are, almost always, men) shows is that a person’s relationships shape not only how he is understood (and how he understands himself) but also how he can work to distinguish himself. The patronage of institutional “big men” expresses a selfless commitment to the group at the same time as it serves to elevate the patron above those followers, establishing the relative difference between them through the flow of giving (Piliavsky 2014). Accounts of such leaders suggest that this is often done strategically, with aspirant leaders working to position themselves as particular representatives of and arbiters between groups (Alm 2010). Earlier works on intercaste relations and ritual prestations by Marriott (1976) and Raheja (1988) also emphasize how relationships can be used strategically, showing how castes employ different strategies of transaction with other castes in an effort to assert and secure relative standing and power. Here, too, people’s relations are sites of negotiation and can establish undifferentiated intimacy as well as difference. Prominence and interpersonal relations, then, are fundamentally entwined, each facilitating and shaping the other. This holds not only for “big men” but for all persons, as “a Tamil’s success in life vitally depends on maintaining good relationships and a good reputation within one’s community” (M. Mines 1994, 31). A person’s ability to call upon the assistance of others is determined in part by how he or she is viewed. If a person is spoken of well by others and has a good “name” (peyar), he or she may accrue a good reputation (perumai), and so find others open to helping him/her. Still, while everyone may be concerned with maintaining a good name, it is primarily men who are afforded distinctions and given authority. Consequently, the work on prominence and patronage in South Asia has largely been a study of men’s efforts at individuating themselves, despite the fact that the benefits of having a “good name” and of being a “big person” are understood to be consequential for all. A very different body of research helps us understand just how consequential prominence and a “good name” can be. Evolutionary anthropologists have long been interested in understanding people’s pursuit of cultural success (Irons 1979), documenting the benefits that result from being a person of prominence. For example, a number of studies have found that people who are known as good hunters have higher fertility and more surviving offspring (Smith 2004), and a similar pattern holds cross-culturally for high-status persons (Von Rueden and Jaeggi 2016). Multiple mechanisms have been suggested as mediators of this empirical pattern (Von Rueden, Gurven, and Kaplan 2011), a number of which suggest that high-status persons will have greater access to better marriage partners and better collaborative partners. Dominant persons (those able to wield power) may be able to coerce others to act in their interest, whereas prestigious persons (those with valued knowledge and expertise) may be able to attract others interested in learning from and associating with them (Henrich, Chudek, and Boyd 2015; Henrich and Gil-White 2001). Like the work on South Asia, nearly all of this work has focused on men and male pursuit of status, dominance, and prestige. Being known as generous may also have productive social consequences. In particular, reputations for generosity, based on past interactions with others, can help maintain cooperation through indirect reciprocity: such reputations can help cooperative persons identify others like them and avoid those who might behave selfishly (Nowak and Sigmund 1998; Panchanathan and Boyd 2004). There is good empirical evidence of the social returns to generosity and prosociality, such as increased help during illness (Gurven et al. 2000), more generous cooperative partners (Bliege Bird and Power 2015), and a greater ability to mobilize others for collective action projects (Lyle and Smith 2014; Macfarlan, Remiker, and Quinlan 2012). These findings suggest that people seek to improve their reputation and their prominence because of the many benefits that may accrue from being well known and well regarded, including being sought out as a partner, being able to access a broader range of resources and support, and being able to influence others. To these bodies of work, we add the tool of social network analysis. In so doing, we heed an early call by Srinivas and Béteille (1964) for South Asianists to take up a network perspective. We view this paper as a demonstration of the potential of this tool for exploring some of the Power and Ready • Building Bigness 447 fundamental questions that are shared across the diverse subfields of anthropology. Our social network approach takes the contention that the self is constituted through relationships as a point of departure, constructing a network to explicitly represent that entanglement. We do this not by studying idealized roles or transactions between representatives of different castes and communities (e.g., Marriott 1968, 1976) but instead by studying the broader multidimensional relationships between persons as relayed by the persons themselves. With this network approach, then, we maintain the tension between the individual and the dividual, simultaneously recognizing both. This approach also gives a more comprehensive and nuanced rendering of the socioecological conditions from which particular persons pursue different behavioral strategies. It allows us to look not solely at the strategies of the select few persons who find themselves in positions of prominence but instead at the efforts of all persons, whatever their particular set of constraints and possibilities. Ultimately, we use these analyses to argue for a more expansive view of people’s efforts at social maneuvering and suggest that the concept of social capital may be particularly helpful in this effort. Our study suggests that the common thread of people’s strategies—across economic, gender, caste, and class distinctions—is not the constant pursuit of prominence but rather a striving for interpersonal connections. Our findings show that generosity and reciprocity occupy a more important place in people’s interpersonal relationships than prominence and that prominence is valuable to the extent that it facilitates those relationships. These are arguments that we see as being fundamentally facilitated by the network approach taken here. Our ability to view the aggregate outcome of dyadic interactions allows us to show how intimate relationships and the generous acts that help define them are crucially formative of people’s social standing, and so of their social possibilities. SETTING AND DATA “Al ̄ akāpuram” and “Ten ̄ pat .t.i” are located in the dry scrublands in the center of the Indian state of Tamil Nadu. Fed sporadically by the irrigation waters flowing from the Vaigai River, the fields surrounding the villages support crops of rice and cotton for a few months each year. During the rest of the year, most villagers engage in wage labor, taking construction jobs, cutting wood, or participating in the government’s “100 Days Work” scheme. Increasingly, younger residents are shifting from agriculture and day labor to skilled and semiskilled work, finding employment in factories, shops, and offices, whether in nearby towns and cities or abroad. Each village has roughly two hundred households with a mix of different religious (Hindu, Catholic, and Protestant) and caste communities (primarily Yātavar, Akamut .aiyār, Pal .l .ar, and  ̄ aiyar castes). The caste composition of the two villages differs (see SI Table S1 for more details of the demographics), with the Pal .l .ar caste being numerically preponderant in Al ̄ akāpuram, and so in many ways dominant within the village despite historical and contemporary marginalization as a Scheduled Caste. Reliance on other people is an inevitable part of life in Al ̄ akāpuram and Ten ̄ pat .t.i. People learn of work opportunities through their peers. They rely on neighbors and relatives to watch their children so they can take that work. They eat at each other’s houses when they do not have a hot meal at home. They borrow money from friends and relatives to cover their children’s school fees. In all of these ways, the residents of these villages are linked by flows of material and immaterial aid. Much aid flows, too, from nongovernmental organizations and from the Indian state, with subsidies providing residents with many basic essentials and various schemes offering work opportunities, training courses, microfinance loans, and childcare. Residents are also often reliant on friends and family living in nearby towns and cities, as well as further afield. With young villagers engaging in migrant labor to countries in the Persian Gulf and Southeast Asia, these networks, and in many ways the villages themselves, extend far beyond their physical confines (Osella and Osella 2012; Velayutham and Wise 2005). The first author undertook ethnographic fieldwork in these villages, residing there for twenty months between October 2011 and August 2013. In addition to extensive participant observation and open-ended, in-depth interviews, this work included a formal survey conducted toward the end of that fieldwork (in February and April of 2013) with 782 of the 809 (97 percent) adult residents. The survey had two parts: a social support survey and a reputation survey. First, in the social support survey, respondents were asked to name those persons who provided them with twelve different types of support (see SI Figure S1), covering material aid (lending household items, giving a loan), behavioral assistance (running errands, babysitting), emotional support (conversation partners, close friends), guidance (giving advice, discussing important matters), and vouched support (help finding work, aid during a conflict, help navigating bureaucracy by someone of a “high position”). The particular questions and types of support elicited were determined through extensive discussions and pretesting with local interlocutors. Respondents could name as many people as they liked for each of the twelve questions. On average, respondents named a total of seventeen people and named some of those persons as providing them with multiple types of support, leading to an average of thirty nominations by each respondent. In total, the villagers reported 23,089 ties, 64 percent of which were with other residents of the villages. With these ties, we construct networks representing the supportive relationships among villagers, as shown in Figure 1 (for more details, see Power 2017b; see also SI Table S7 for summary statistics of the networks). Second, in the reputation survey, respondents were asked to name those persons whom they felt had eight particular qualities (for more details, see Power 2017a), of which four are studied here (see SI Figure S2): (1) having a generous attitude (tārāl .a man ̄ appān ̄ mai ut .aiya napar), 448 American Anthropologist • Vol. 120, No. 3 • September 2018

elaborate affair. They arranged for a special service at the village church with the local priest. Posters and a speaker set loudly announced the event, and more than seven hundred people were invited. Well liked and well regarded, Arulprakaash and Selvi received so many guests that they had to erect a special canopy to provide shade for those who could not fit into their house. A relative stationed at the entrance received the small gifts of money (a few hundred rupees, or five to ten dollars) called moy brought by the guests. Literally, moy means a swarm or throng, and this happened doubly, as people crowded around the relative, and as the money piled up, filling an entire suitcase. The moy from their daughter's first communion would not only cover the expenses of the function itself but also let them repaint their house and put down a new tiled floor. Arulprakaash was well aware that it was only through the help of others that they were able to do this: For me, the house finishing has been accomplished. This is happiness. Okay, at first, there was a house but not one in good condition. Now, you've seen the house. Now the house is good. From now on, with the salary I get, I won't be able to do any more. I was able to do this because of the function. I got the loan and I repaid the loan. All paid completely. Separate, I cannot do this.
If Arulprakaash had been "separate," without the support of others, his house would still be a weatherworn soot gray. In fact, without support, his house would not exist; it had been constructed years earlier with the moy collected at his wedding.
In this article, we ask how Arulprakaash and Selvi manage to get such support. Using social network analysis, we model the flows of support between people in two South Indian villages, focusing on the role that a person's reputation plays in his/her relations with others. We compare the importance of a person's reputation for generosity and good character, on the one hand, with their reputations for giving good advice and for wielding influence and authority, on the other. We see the former as capturing local conceptions of personal morality and virtue and the latter as capturing an assessment of a person's prominence. We find that having a reputation for influence has a limited effect on getting help from others, both within and beyond the village. Instead, it is a reputation for generosity that has the strongest association with accessing support. This correlation is partially due to the strong effects of reciprocity and shared connections: those with reputations for generosity are more likely to be asked for help and so are more likely to get it. Further, a reputation for generosity is more closely associated with having an influential network position, as measured by weighted PageRank centrality, than a reputation for influence. In this setting, influence and authority are almost exclusively attributed to men, and often men of particular castes and wealth. Yet we find that there are many people beyond this narrow category of persons who, while not recognized as influential, actually are influential in terms of their network position. This observation, coupled with the stronger effect of generosity on access to social support, leads us to argue for a wider analytical and theoretical frame, moving beyond a focus on the influence and prestige of a few patrons, leaders, and "big men" to a broader focus on social capital. Such a perspective better allows us to interrogate how people for whom prominence is less accessible-especially women and, in this setting, people of particular castes-may nevertheless be able to build the "bigness" (perumai) that allows them to gain the support of others.
With these findings, we speak to two distinct fields: South Asian studies and evolutionary anthropology. These seemingly disparate fields have both suggested that a good reputation can be consequential for people's livelihoods, and both have been concerned with how people's relationships with others are implicated in those reputations. Both subfields build from the fundamental observation that people's relationships and exchanges with one another help shape their standing in society (Bourdieu 1986;Mauss 1954) but perhaps inevitably diverge from there, both in terms of the particular questions asked and in the types of evidence brought to bear on those questions. In South Asian studies, and cultural anthropology more broadly, the recent turn to study the ethical has highlighted how interpersonal relations become a site of moral tension and self-cultivation (Keane 2016;Laidlaw 2014;Pandian and Ali 2010;Robbins 2013). In evolutionary anthropology, there is an increasing interest in exploring the nuances of how people assess others and build relationships with them (Barclay 2013;Bliege Bird and Power 2015;Bliege Bird, Ready, and Power 2018;Lyle and Smith 2014;Macfarlan, Quinlan, and Remiker 2013). We find both to be compelling, if partial, accounts of the ways in which people maneuver within their social worlds. Here, we attempt to offer a rich account of that maneuvering by combining ethnographic observation with social network analysis, a methodological framework that allows for quantitative analysis while fundamentally recognizing the embedded nature of the social person. Our approach not only leads us to reaffirm some core anthropological insights (for example, the importance of reciprocity) but also helps us note new complexities, especially those that operate at the interstices between personal interactions and structural outcomes. This approach also allows us to capture the broader landscape of social strategies employed by persons of very different social and economic positions within the villages, including both men and women, as they attempt to achieve their goals. Having surveyed that landscape, we see generosity as a common driver of both prominence and social capital across all segments of the population. This leads us to argue that both evolutionary anthropology and South Asian studies would benefit from extending their focus from prominence to the more inclusive concept of social capital.

REPUTATION, PROMINENCE, SOCIAL SUPPORT, AND THE SOUTH ASIAN PERSON
In South Asian studies, it has long been recognized that the flows of support between persons are not only materially important but also socially and morally meaningful. Indeed, foundational works have stressed how the South Asian person is fundamentally understood through interpersonal relationships and group identities, so much so that some contend that there is no individual in South Asia but only the "dividual" (Marriott 1976). In this conception, the person is not firmly distinct and bounded but rather is composed of biomoral substances that can be absorbed, dispersed, and combined as persons interact, fundamentally changing their being in the process (Daniel 1984). Ethnographies of South Asia, and South India in particular, have evocatively illustrated the ways in which people work to shape their porous selves (Daniel 1984;Pandian 2009;Prasad 2007) and have further shown how such work on the self involves the careful negotiation of one's intimate relationships (Busby 2000;Trawick 1992;Venkatesan 2014). This body of ethnographic work, though, does not suggest that there is no individual but instead shows how South Asians navigate the ambiguities and tensions between individual autonomy and the collective. As M. Mines (1994) argues, for example, Tamil persons can be understood as "contextualized individuals," read through the groups and identities to which they hold. He illustrates this most compellingly through his portraits of "big men" (periyavarkal . , influential local leaders) in Chennai who are able to wield influence through their positions of prominence within caste organizations, temple committees, charitable organizations, and other institutions. Such "big men" are marked by their relationships with others, particularly their patronage and the assistance they render to their constituents.
What M. Mines's (1994; see also M. Mines and Gourishankar 1990) work on "big men" (who are, almost always, men) shows is that a person's relationships shape not only how he is understood (and how he understands himself) but also how he can work to distinguish himself. The patronage of institutional "big men" expresses a selfless commitment to the group at the same time as it serves to elevate the patron above those followers, establishing the relative difference between them through the flow of giving (Piliavsky 2014). Accounts of such leaders suggest that this is often done strategically, with aspirant leaders working to position themselves as particular representatives of and arbiters between groups (Alm 2010). Earlier works on intercaste relations and ritual prestations by Marriott (1976) and Raheja (1988) also emphasize how relationships can be used strategically, showing how castes employ different strategies of transaction with other castes in an effort to assert and secure relative standing and power. Here, too, people's relations are sites of negotiation and can establish undifferentiated intimacy as well as difference.
Prominence and interpersonal relations, then, are fundamentally entwined, each facilitating and shaping the other. This holds not only for "big men" but for all persons, as "a Tamil's success in life vitally depends on maintaining good relationships and a good reputation within one's community" (M. Mines 1994, 31). A person's ability to call upon the assistance of others is determined in part by how he or she is viewed. If a person is spoken of well by others and has a good "name" (peyar), he or she may accrue a good reputation (perumai), and so find others open to helping him/her. Still, while everyone may be concerned with maintaining a good name, it is primarily men who are afforded distinctions and given authority. Consequently, the work on prominence and patronage in South Asia has largely been a study of men's efforts at individuating themselves, despite the fact that the benefits of having a "good name" and of being a "big person" are understood to be consequential for all.
A very different body of research helps us understand just how consequential prominence and a "good name" can be. Evolutionary anthropologists have long been interested in understanding people's pursuit of cultural success (Irons 1979), documenting the benefits that result from being a person of prominence. For example, a number of studies have found that people who are known as good hunters have higher fertility and more surviving offspring (Smith 2004), and a similar pattern holds cross-culturally for high-status persons (Von Rueden and Jaeggi 2016). Multiple mechanisms have been suggested as mediators of this empirical pattern (Von Rueden, Gurven, and Kaplan 2011), a number of which suggest that high-status persons will have greater access to better marriage partners and better collaborative partners. Dominant persons (those able to wield power) may be able to coerce others to act in their interest, whereas prestigious persons (those with valued knowledge and expertise) may be able to attract others interested in learning from and associating with them (Henrich, Chudek, and Boyd 2015;Henrich and Gil-White 2001). Like the work on South Asia, nearly all of this work has focused on men and male pursuit of status, dominance, and prestige.
Being known as generous may also have productive social consequences. In particular, reputations for generosity, based on past interactions with others, can help maintain cooperation through indirect reciprocity: such reputations can help cooperative persons identify others like them and avoid those who might behave selfishly (Nowak and Sigmund 1998;Panchanathan and Boyd 2004). There is good empirical evidence of the social returns to generosity and prosociality, such as increased help during illness (Gurven et al. 2000), more generous cooperative partners (Bliege Bird and Power 2015), and a greater ability to mobilize others for collective action projects (Lyle and Smith 2014;Macfarlan, Remiker, and Quinlan 2012). These findings suggest that people seek to improve their reputation and their prominence because of the many benefits that may accrue from being well known and well regarded, including being sought out as a partner, being able to access a broader range of resources and support, and being able to influence others.
To these bodies of work, we add the tool of social network analysis. In so doing, we heed an early call by Srinivas and Béteille (1964) for South Asianists to take up a network perspective. We view this paper as a demonstration of the potential of this tool for exploring some of the fundamental questions that are shared across the diverse subfields of anthropology. Our social network approach takes the contention that the self is constituted through relationships as a point of departure, constructing a network to explicitly represent that entanglement. We do this not by studying idealized roles or transactions between representatives of different castes and communities (e.g., Marriott 1968Marriott , 1976 but instead by studying the broader multidimensional relationships between persons as relayed by the persons themselves. With this network approach, then, we maintain the tension between the individual and the dividual, simultaneously recognizing both. This approach also gives a more comprehensive and nuanced rendering of the socioecological conditions from which particular persons pursue different behavioral strategies. It allows us to look not solely at the strategies of the select few persons who find themselves in positions of prominence but instead at the efforts of all persons, whatever their particular set of constraints and possibilities.
Ultimately, we use these analyses to argue for a more expansive view of people's efforts at social maneuvering and suggest that the concept of social capital may be particularly helpful in this effort. Our study suggests that the common thread of people's strategies-across economic, gender, caste, and class distinctions-is not the constant pursuit of prominence but rather a striving for interpersonal connections. Our findings show that generosity and reciprocity occupy a more important place in people's interpersonal relationships than prominence and that prominence is valuable to the extent that it facilitates those relationships. These are arguments that we see as being fundamentally facilitated by the network approach taken here. Our ability to view the aggregate outcome of dyadic interactions allows us to show how intimate relationships and the generous acts that help define them are crucially formative of people's social standing, and so of their social possibilities.

SETTING AND DATA
"Al¯akāpuram" and "Tenpat . t . i" are located in the dry scrublands in the center of the Indian state of Tamil Nadu. Fed sporadically by the irrigation waters flowing from the Vaigai River, the fields surrounding the villages support crops of rice and cotton for a few months each year. During the rest of the year, most villagers engage in wage labor, taking construction jobs, cutting wood, or participating in the government's "100 Days Work" scheme. Increasingly, younger residents are shifting from agriculture and day labor to skilled and semiskilled work, finding employment in factories, shops, and offices, whether in nearby towns and cities or abroad. Each village has roughly two hundred households with a mix of different religious (Hindu, Catholic, and Protestant) and caste communities (primarily Yātavar, Akamut . aiyār, Pal . l . ar, and Paraiyar castes). The caste composition of the two villages differs (see SI Table S1 for more details of the demographics), 2 with the Pal . l . ar caste being numerically preponderant in Al¯akāpuram, and so in many ways dominant within the village despite historical and contemporary marginalization as a Scheduled Caste.
Reliance on other people is an inevitable part of life in Al¯akāpuram and Tenpat . t . i. People learn of work opportunities through their peers. They rely on neighbors and relatives to watch their children so they can take that work. They eat at each other's houses when they do not have a hot meal at home. They borrow money from friends and relatives to cover their children's school fees. In all of these ways, the residents of these villages are linked by flows of material and immaterial aid. Much aid flows, too, from nongovernmental organizations and from the Indian state, with subsidies providing residents with many basic essentials and various schemes offering work opportunities, training courses, microfinance loans, and childcare. Residents are also often reliant on friends and family living in nearby towns and cities, as well as further afield. With young villagers engaging in migrant labor to countries in the Persian Gulf and Southeast Asia, these networks, and in many ways the villages themselves, extend far beyond their physical confines (Osella and Osella 2012;Velayutham and Wise 2005).
The first author undertook ethnographic fieldwork in these villages, residing there for twenty months between October 2011 and August 2013. In addition to extensive participant observation and open-ended, in-depth interviews, this work included a formal survey conducted toward the end of that fieldwork (in February and April of 2013) with 782 of the 809 (97 percent) adult residents. The survey had two parts: a social support survey and a reputation survey.
First, in the social support survey, respondents were asked to name those persons who provided them with twelve different types of support (see SI Figure S1), covering material aid (lending household items, giving a loan), behavioral assistance (running errands, babysitting), emotional support (conversation partners, close friends), guidance (giving advice, discussing important matters), and vouched support (help finding work, aid during a conflict, help navigating bureaucracy by someone of a "high position"). The particular questions and types of support elicited were determined through extensive discussions and pretesting with local interlocutors. Respondents could name as many people as they liked for each of the twelve questions. On average, respondents named a total of seventeen people and named some of those persons as providing them with multiple types of support, leading to an average of thirty nominations by each respondent. In total, the villagers reported 23,089 ties, 64 percent of which were with other residents of the villages. With these ties, we construct networks representing the supportive relationships among villagers, as shown in Figure 1 (for more details, see Power 2017b; see also SI Table S7 for summary statistics of the networks).
Second, in the reputation survey, respondents were asked to name those persons whom they felt had eight particular qualities (for more details, see Power 2017a), of which four are studied here (see SI Figure S2): (1) having a generous attitude (tārāl . a man appān mai ut . aiya napar), (2) having good character (nalla kun . am pat . aittavar), (3) giving good advice to anyone who asked (yārit . am nallaālōcan ai kēt . t . u peralām), and (4) having influence (celvākku, literally a "voice that spreads") and authority (atikāram pat . aittavarkal . ). We consider all four as measures of "reputational quality" because they result from peer assessments. While the reputational qualities are related (see SI Table S5 for the partial correlations), both local understandings and principal component analyses (see SI Table S6 and Figure S3) suggest that they fall into two distinct domains: reputations for generosity (i.e., being generous or having good character) and reputations for prominence (i.e., giving good advice and wielding influence and authority). We further distinguish between different aspects of prominence, following distinctions in evolutionary theory (Cheng et al. 2013), as well as in Tamil understandings of the moral conception and valuation of a person (M. Mines 1994;P. Price 2006). In particular, we consider being known for giving good advice to be a measure of prominence comparable to concepts of prestige and being known for wielding influence and authority to be a measure of prominence closest in local meaning to ideas about dominance.
For each of the four questions, respondents typically named two or three villagers. The resulting number of nominations each resident received gives a sense of that person's relative standing for each reputational quality. These measures are quite skewed, with many people being named only once or twice (248; 32 percent) across all qualities or not at all (101; 13 percent), and a few people being nominated by many residents. To ameliorate this skew, we log transform the count of nominations and scale it to range from 0 to 1 for each reputational quality (see SI Table S4 for summary statistics).
Additional background data come from a household census conducted by the first author. The data from this census cover a broad range of demographic, economic, and social variables, including gender and age, educational attainment, religious denomination and caste membership, household wealth, kinship relations among villagers, and the physical distance between households. These variables are expected to influence the probability of a relationship between persons, and so they are included in the analyses.

ANALYSES
Guided by both theoretical and local understandings of reputation, we expect a strong relationship between how a person is perceived by his/her peers and his/her relationships with them, with those persons who are most well regarded being best able to call upon others for support. This should hold both for those who are seen as generous and also for those who are seen as prominent. To evaluate this, we draw on records of residents' reputations and of their supportive relationships, assessing how the former relates to the latter. We do this by looking at the support network within the village and at connections to people beyond the village.

Within the Village
We investigate the extent to which reputations for generosity and reputations for prominence play both direct and indirect roles in supportive relationships within the village using exponential random graph models (ERGMs). ERGMs model the probability of a tie between each pair of nodes (here, persons) in a network, in a manner analogous to logistic regression, while also allowing for the inclusion of structurally dependent terms, such as reciprocity and transitivity, in addition to individual-and dyad-level attributes (Robins et al. 2007;Snijders et al. 2006). We conduct the analyses in R (R Core Team 2016) using the statnet suite of packages (Handcock et al. 2008). 3 Throughout the analysis and results, we refer to two hypothetical women ("Uma" and "Mutthu") for clarity in describing the model terms (see Figure 2).  We construct four ERGMs for each of the villages, differentiated by the reputational quality considered (generosity, good character, giving good advice, or influence). In examining the effect of the reputational qualities on support ties, we consider four model terms: (1) "indirect effect, support given," which is the effect of a person's reputation on the probability that others name him/her as providing them with support; (2) "indirect effect, support received," which is the effect of a person's reputation on his/her probability of naming others as providing him/her with support; (3) "direct effect, support given," which models the effect of a person being named as providing support, given that the receiver nominates the provider for that reputational quality (Uma helps Mutthu, and Mutthu nominates Uma); and (4) "direct effect, support received," which models the probability of a person being named as providing support, given that the provider nominates the receiver for that reputational quality (Mutthu helps Uma, and Mutthu nominates Uma). The first two "indirect" terms represent how a person's overall reputation relates to his/her ability to give or receive support, and the latter two "direct" terms represent how a particular person's evaluation of another relates to the support they give to each other (see SI Table S2 for further description of all model covariates and SI Table S3 for summary statistics of those covariates).
The effect of support given on support received is modeled with a term for reciprocity, which accounts for the increased probability of a support tie in one direction when a support tie already exists in the other (Uma helps Mutthu, and Mutthu helps Uma). We also include a term for transitivity (using geometrically weighted edgewise shared partnerships; Hunter and Handcock 2006), which measures the increased probability of a tie between persons who have shared partners (Uma and Mutthu have a supportive relationship, and they both have a relationship with Latha). Each model includes terms for the effect of a person's age, gender, and wealth on the probability of a supportive tie, as well as for the tendency for people to affiliate with others of the same gender and education level, and with those who are kin and neighbors. Finally, we include two terms for caste: a term for the general effect of a person's caste on the probability of a supportive relationship and a term for caste homophily (the effect of two people being of the same caste).
The full ERG models are presented in SI Tables S9 and S10. Table 1 summarizes the effects of the four different reputational qualities for each village along with the effects of reciprocity and shared partners. Figure 3 demonstrates the effect of each reputational quality on the probability of being named as providing support, for both the indirect and direct measures of reputation as well as other forms of social affiliation, using the probability of a tie between two fortytwo-year-old women of the Pal . l . ar caste as an example. Each panel in Figure 3 shows the changing probability of requests for support as the supporter's reputation increases along with the effect of different combinations of other model terms, such as reciprocity and having a shared partner.
The results show that all reputational qualities have similarly small indirect effects on the overall probability of being asked for support, with the odds of a tie increasing only very slightly with greater reputational standing (Figures 3a and b, panel 1). However, as the increasing slopes show in each subsequent panel, greater reputational standing has a more pronounced effect when combined with other terms.
When we examine the direct effect of reputational standing on incoming requests for support, the effects begin to differentiate across the reputational qualities. Increasing reputation for generosity has the largest effect on the probability of a request for support, while influence has a comparatively small effect on the probability of incoming requests for support (Figures 3a and b, panel 2). That is, if Uma believes Mutthu is generous, she is more likely to ask Mutthu for support than if she believes Mutthu is influential.
Other forms of social connection also have a sizable effect on the probability of a support tie, across all models. Having even just one common partner substantially increases the probability of a supportive relationship (Figures 3a and  b, panel 3). Even stronger is the effect of reciprocity: if Uma provides support to Mutthu, Mutthu is generally (averaging between the models) 5.3 times as likely to provide support to Uma in Al¯akāpuram and 3.5 times as likely in Tenpat . t . i (Figures 3a and b, panel 4) than if she did not.
Turning to the effect of reputation on receiving support, we find that the indirect effect of reputation is negative for all reputational qualities and, in most cases, is more negative than the indirect effect of reputation on giving support is positive. That is, persons with greater reputational standing experience, on average, a slight increase in requests for support but are actually less likely than persons of lower reputational standing to ask others for support (SI Figure  S4).
The direct effect of reputation on receiving support is generally small and positive across models, meaning that when Uma nominates Mutthu as having some reputational quality, Uma is also more likely to provide support to Mutthu (Figures 3a and b, panel 5). There is one key exception: nominating someone as influential has no effect on whether the nominator provides that person with support in Tenpat . t . i, but it actually decreases the probability that the nominator will provide him/her with support in Al¯akāpuram.
Across the different reputational qualities, the results consistently show that a reputation for generosity has the strongest effect on the probability of incoming requests for support, while a reputation for being influential has the weakest effect. For support received, although in both cases the probability of receiving support decreases as reputation increases (SI Figure S4), persons with a reputation for generosity still have a much higher probability of receiving support than those with a reputation for being influential. Further, supporting others substantially increases the probability that a person will receive support, thanks to the strong  positive effect of reciprocity. These results hold when we consider multiple types of reputation within a single model: for both villages, when influence and generosity are included in the same model, the effects of generosity remain similar to the single-reputation model, but the effects of influence generally become smaller (SI Table S11).

Beyond the Village
While having a reputation as influential may have a relatively small or even negative effect on accessing support within the village, influential persons may instead be more oriented to others beyond the confines of the village. Many have suggested that "big men" are considered as such because of their privileged access to powerful people beyond their constituents to whom they can connect (e.g., Alm 2010; Srinivas and Béteille 1964). So, we now examine whether people who have a reputation for influence are more likely to have connections to people of "high position" (said in English, or as uyarntu pataviyil ul . l . avarkal . ) beyond the village, where a person of "high position" is someone who is either reported as being such by respondents or is reported as having a position that is locally considered to be so (e.g., government employee, police officer, doctor, lawyer). On average, people reported having ties to two people of "high position" outside the village (SI Table S8). We model the proportion of support ties that each person has with "high position" alters outside the villages using a binomial regression with individual random effects and measures of education, age, gender, caste, and wealth in addition to the person's reputation for wielding influence and authority. We run the models using the rethinking package (McElreath 2017) in R; numeric variables in the models have been centered and rescaled, except for the reputational variable, which is scaled between 0 and 1.
We find that a reputation for influence has only a minimal effect on the likelihood of connections to people of "high position" outside the villages (see Figure 4 and SI Table S13). Those persons who are seen as having influence and authority within the village do not seem to have privileged access to people of "high position" outside it. The strongest effect is instead seen for education, with those who are more educated than most villagers being much more likely to report connections to "high-position" people outside the village.

DISCUSSION
We find that reputation alone does little to influence a person's supportive relations both within and beyond the village. Regardless of the reputational quality, the indirect effect of reputation on giving support is positive but slight, and its effect on receiving support is generally larger and negative. The direct effect of reputation is, not surprisingly, much stronger and positive. It is really with greater social proximity-where persons have shared support partners, reciprocal relationships, or mutual recognition that each is reputable-that reputation begins to more substantively influence a person's ability to give and receive support (as seen in panels 3 to 5 of Figures 3a and b).
It is important to consider how a person's reputation for generosity and reputation for prominence differentially impact his/her ability to give and receive support. The models indicate that incoming requests for help increase more quickly with reputation than outgoing requests. In other words, more reputable people have a greater number of people asking them for support, and they are less likely to ask for support themselves. Yet the strong effect of reciprocity in all of the models means that reputable persons do ask for support from those whom they help. Highreputation persons are also more likely to ask for support from those who consider them reputable (for character, generosity, and advice, though importantly not for influence). Thus, the benefit of a reputation may stem not from a person's enhanced ability to ask unconnected others for support but rather from the fact that others who view that person as reputable may approach him/her for help. The importance of reciprocity means that, when in need of assistance, those of high reputational standing are able to draw on the help of those who have already sought them out. Consequently, even if they do not necessarily call upon a greater number of people on a day-to-day basis, reputable persons may have a wider pool of potential helpers to draw upon. Critically, though, this does not hold as strongly for a reputation as being influential, suggesting that the coercive power associated with dominant persons may not generate these positive feedbacks of mutual recognition and support.
These findings strongly echo ethnographic work in South India that has shown how close relationships are defined by fluctuating flows of support and indebtedness (e.g., Appadurai 1985;Busby 2000;Daniel 1984;Trawick 1992). The effect of reciprocity is not only empirically visible in our models but is also highly valued in Tamil society. A reputation is valuable because of how it is instantiated through supportive social relationships, and it is a reputation for generosity, rather than the more aggrandizing markers of prominence, that we find is most closely associated with both incoming and outgoing supportive ties. This suggests that it is actually the more subtle aspects of interpersonal relationships and mutual assistance that best help people access support themselves. These findings suggest that attempts to understand how people access resources through social channels should move from a focus on explicit markers of status and prestige to a more comprehensive measure of social capital.

From Prominence to Social Capital
The word that is often translated as "reputation" in Tamil is perumai, though it is more accurately translated as "bigness" or "greatness" (Bate 2009;D. Mines 2005). It implies a sort of social and spatial expansion as a person's renown grows. At functions such as Arulprakaash and Selvi's daughter's first communion, perumai is made particularly tangible; the number of people in attendance, the amount of food served, the number and size of gifts given by relatives, the distance from which people traveled to attend, and the growing total of the moy donations are all measures of perumai (D. Mines 2005).
We suggest that, more so than reputation, perumai might be better understood as social capital, the potential material and immaterial resources embedded within a person's social network (Bourdieu 1986;Lin 2001). This is most clearly illustrated in the large vinyl "flex-board" banners that accompany such functions. The banners have studio photos of the honoree, often edited to add in new suits or extra gold jewelry. Markers of caste and religion are also included, with images of caste leaders or of movie stars from the same caste. Most important for us here, the banners typically have a second crowd of faces and names: the family and friends of the protagonist(s). Alongside the photos of all of the supporting characters are not only their names but also often their degrees (BCom, DEEE), their jobs (Airtel, police), or simply their locations (Dubai, Chennai). Many groups make banners, so the entrance to a function hall is often flanked by numerous banners with different constellations of supporting characters. Certainly, these banners include markers of economic and cultural capital, but the "bigness" that is most conspicuous is the visible manifestation of the social capital of the honoree.
As with the English translation of perumai, the line between reputation, prominence, and social capital is often blurred, making it easy to conflate them. Figure 5 shows the close relationship between a network proxy of social capital-weighted PageRank centrality-and a reputation as influential or generous (see also SI Table S5). Weighted PageRank is a measure of network centrality that reflects the strength and number of incoming ties to each node while also more heavily weighting ties from important nodes (those that themselves have more incoming ties; Xing and Ghorbani 2004). Compared to other measures of centrality (such as simple degree or eigenvector centrality), weighted PageRank captures a much more comprehensive set of indicators of the importance of a person's network position, including the number of incoming and outgoing ties, the strength of those ties, and whether those ties are to other structurally important persons. For these reasons, we suggest that, in this case, weighted PageRank calculated from the social support network is a good proxy for social capital: it simultaneously measures a person's investment in social relations and the quality of those relations. Broadly, weighted PageRank and reputational standing for both generosity and influence trend together, with those who are most widely recognized as reputable also having higher weighted PageRank centrality.
While the broad correspondence among generosity, prominence, and social capital may lead us to see them as proxies or reflections of one another, the discrepancies between them should make us wary of doing so. Figure 5 also shows that there are many people who are not explicitly recognized as being influential but who nevertheless occupy central, and potentially influential, network positions. Most arguments in evolutionary anthropology for why people pursue prominence rest on how it can lead to better access to desirable partners, strengthened relationships, or greater influence over others (Cheng et al. 2013;Von Rueden, Gurven, and Kaplan 2011). However, here we find that the returns to prominence are limited, both within and beyond the village, and what benefits exist may be more accurately understood as returns to helping others and investing in supportive relationships. Indeed, much of the literature on leadership and eminence in South Asia explicitly describes leaders as wielding influence only insofar as they are able to actually work as brokers for their constituents and manage to get things done (Alm 2010;M. Mines and Gourishankar1990;Piliavsky 2014). Instead of focusing on prominence itself as facilitating influence and support, then, it may actually be the prosocial acts of those persons that drive the consistent finding of a relationship between prominence and social and biological returns (Barclay and Willer 2007;M. Price 2003;Raihani and Barclay 2016;Willer 2009; see also McFarland et al. 2017;Silk et al. 2009). In sum, our findings suggest that small, subtle acts of support are what give prominence its potency.

From "Big Men" to "Bigness"
We argue that the study of people's attempts to improve their livelihoods should focus not on efforts to become a periyavar, a "big man," but instead on the broader task of building perumai, "bigness," or social capital. By turning to directly study the full range of actions that can build a person's social capital, rather than solely acts of aggrandizement, we can broaden our understanding of the social foundations of people's livelihoods in a number of important ways. Most crucially, it will help counteract the narrow focus on the small subset of people who aspire to or achieve prominence. While a young girl having her first communion may not be a "big person," she still has some "bigness," as seen by the supporting characters on the banners announcing the event. Social capital is not something limited to only a select few.
Treating the study of "bigness" as the study only of "big men" is problematic for a number of reasons. First, it presumes that only the pursuit of obvious eminence is the way to achieve social success (Irons 1979). This excludes most people and, with it, most strategies (Bowser and Patton 2010;Collier 1974). In Tamil Nadu, at least, prominence is often seen as the purview of the wealthy (Dean 2013). Only a person with "means" (vacati) can have prestige, leaving the poor with "prestige problems" (kaurava piraccinai). Similarly, "big people" are traditionally wealthy property holders of the dominant caste in the village. Saravanan (a poorer Dalit man from Tenpat . t . i) put it succinctly: "Influence [celvākku] is nothing but money." Prominence is also almost exclusively the domain of men in these villages. In Figure 6, we break out the relationship between network social capital and reputational standing by gender, showing that women are rarely, if ever, recognized as being influential.
Second, the study only of "big men" hinders the recognition of changes over time. The traditional image of a few "big men" working as privileged brokers to other regional leaders and sources of power is not reflected in our analyses. This is not to say that people of influence do not have these connections but rather that everyone does. While in the past, village "big men" (coming from dominant, landholding castes) may have wielded such authority, recent political, economic, and social changes mean that they no longer have exclusive access to those advantageous connections (cf. Gorringe 2010;Philip 2017;P. Price 2006). As more people have been able to own land, pursue higher education, get loans, enter politics, secure salaried jobs, find work abroad, and so on, both the number of loci of power and the range of people who have access to them have increased. Had we studied only the few obvious "big men," we might have seen their connections to regional leaders as a source of their influence and authority. Instead, we see that others are not dependent on them as brokers, having instead multiple potential avenues to social capital and influence, whether through education (as our binomial regression suggests), opportunities abroad (Osella and Osella 2012), political connections (Alm 2010), or local social connectedness (as our ERGMs suggest).
Perhaps most importantly, the study only of "big men" results in a misunderstanding of the strategies and pathways to building "bigness." It is likely because much of the work on prominence has actually been a study of male prominence, both in evolutionary anthropology (e.g., Casimir and Rao 1995;Gurven and Von Rueden 2006;Henrich, Chudek, and Boyd 2015;Pérusse 1993;Smith 2004;Kaplan 2008, 2011;Von Rueden and Jaeggi 2016) and in South Asian studies (e.g., M. Mines 1994; Piliavsky 2014; P. Price and Ruud 2010) that the focus has stayed solidly on how people work to build prominence instead of on their efforts to build social capital. Our results show, though, that it is not prominence that is most crucial in accessing the supportive relationships that are so foundational to peoples' livelihoods. This directs us, then, to look at additional strategies and aims, ones that are more grounded in peoples' relationships, and ones that are ultimately more widely employed. 4 For example, while a reputation for being influential is mostly limited to men in these villages, an influential network position and a reputation as generous are not similarly constrained. While women are essentially not recognized explicitly as being influential, they often are, in terms of their network centrality. Determining how women, and all people, build social capital will ensure that we have a fuller understanding of the different strategies persons employ to get by.
Kausalya, a former resident of Tenpat . t . i, is one person who clearly demonstrates the importance of social capital in achieving "actual" influence. Villagers often seek her out for help with their particular causes. When a few women decided that their cluster of houses deserved its own bore well, for example, they quickly recruited Kausalya to talk to a local government official on their behalf. "She speaks well," Selvi explained. Kausalya later gave her own explanation, saying that she is so often consulted because she is unique in knowing how to navigate the systems of government (through petitions, requests, strategically placed phone calls, etc.). She has developed this skill, she said, thanks to her job at a government office tasked with overseeing government-sponsored women's self-help groups and to her close relationship with her boss. A district-level government official, the boss holds some authority, as well as the purse strings for some government funds, making Kausalya, with his assistance, uncommonly successful in enrolling young Tenpat . t . i villagers in free training programs, securing additional funding for village facilities, and acquiring loans for forays into new business endeavors. "Atu en nōt . u power," she said, using the English word: "That's my power." Kausalya uses her knowledge of bureaucracy to advance the lives of her friends and relations in the village. Such abilities make her one of the few women who was ever mentioned as having any influence or voice (celvākku) in the reputation survey. Often, that voice is quite literal: she is the only woman to regularly speak up in village meetings. Through her efforts to help the village and its villagers, Kausalya has developed a reputation as generous and influential, to be sure. Still, her "power" is not her reputation per se but her social capital.
Kausalya, of course, remains an exception, especially among women. While her social capital is indeed accompanied by a reputation for influence, the two often do not trend together. In fact, the person with the highest weighted PageRank centrality in Tenpat . t . i was nominated as being influential exactly zero times (she is, however, nominated as being generous eighteen times). This person is Selvi, Arulprakaash's wife. Unlike Kausalya, Selvi does not have any personal knowledge of the workings of government programs. What Selvi does have is a calm, friendly demeanor and an open space behind her home. Each evening, when Selvi brings the dirty dishes from dinner out to the backyard to wash them, the women from the neighboring homes roll out reed mats and sit down to chat and gossip about the day. Selvi also has a number of relatives, former neighbors, and friends who are government officials, NGO employees, teachers, and police living outside the village whom she can call upon. While Selvi is not named and recognized by other villagers as wielding influence and authority, according to her position in the village support network, she is actually one of the most influential people in the village. Her social capital and the ways in which she fosters it are much more in line with how most residents of Al¯akāpuram and Tenpat . t . i maneuver their social worlds.
As exemplified by Kausalya, Arulprakaash, and Selvi, there are multiple pathways that can help a person build social capital. Certainly among them is the pursuit of prominence, but this appears to be of secondary importance; more crucial may be the fostering of trusting, cooperative relationships. Opening your home to others, assisting with the harvest, volunteering time to help prepare for the temple festival, lending tools-all of these acts can potentially help build one's social capital. In addition to focusing on the set of strategies employed by those persons aspiring to prominence, we argue for a study of this wider suite of strategies, as all are relevant to understanding how people are able to gain support and influence others.

CONCLUSION
Neither Arulprakaash nor Selvi was recognized by their peers as being a person of influence. Yet they were able to draw hundreds of people together to celebrate their daughter and to help them refinish their home. Their ability to call upon others rests not on their prominence but on their reputation for generosity and good character and on the actual support that they provide to others. Indeed, as Arulprakaash said, their reasoning for having had such an elaborate function for their daughter's first communion was, essentially, reciprocity: To get back the moy, the collection money, that we had given out earlier. We had to get it back, you know. That's why. We go and give to all of our relations. To get back what we have given, it can only be collected if we conduct this function. Otherwise, it's a waste. That's why we had the function.
Arulprakaash and Selvi's last function had been their own wedding, more than ten years earlier. In the intervening years, they had given out, in total, many thousands of rupees as moy at other peoples' functions. To their minds, it was time to recoup some of that moy. Such relationships of reciprocal support are, for Arulprakaash, what sustain people:

Supporting Information
Additional supporting information may be found online in the Supporting Information section at the end of the article.  Table S2: Description of the variables used in the exponential random graph models. Table S3: Descriptive statistics of the variables used in the models. Table S4: Descriptive statistics of reputational nominations for the two villages. Table S5: Partial correlations between the rescaled reputation variables.  Table S8: Descriptive statistics of villagers' ties to external alters of "high position." Table S9: Summary of ERG models of Al¯akāpuram, for each reputational quality. Table S10: Summary of ERG models for Tenpat . t . i, for each reputational quality. Table S11: Summary of ERG models including both influence and generosity terms, for each village. Table S12: Description of the change statistics used to generate predictions for Figure 3 and SI Figure S4. Table S13: Summary of posterior distributions of parameters in binomial regression models with individual-level random effects, modeling peoples' number of ties with high position individuals outside of each village. Figure S1. The survey (in English and Tamil) used to elicit the support relationships. Figure S2. The survey (in English and Tamil) used to elicit the reputational nominations under study in this paper. Figure S3. Plots of the first and second principal components for the principal components of the four reputational qualities in the two villages. Figure S4. Summary of the effect of the four different reputational qualities and social proximity on the probability of a supportive relationship in Al¯akāpuram and Tenpat . t . i, as the reputation of the individual requesting support increases. Figure S5. Scatterplots with LOESS curve showing the relationship between a person's weighted PageRank centrality in the social support network and his/her reputation for giving good advice and having good character.