Social capital and polarization: The case of Polish think tanks

ABSTRACT In this article, we study polarization within civil society. While earlier research on civil society has shown that civil society organizations can be divisive, research on polarization has only paid scant attention to the role of civil society. We bring these two aspects of the literature together to develop a framework for analyzing social capital in a polarized context. The framework helps identify practices that organizations may engage in when shaping social capital and working with others: facilitating the flow of information; providing credentials for actors; influencing agents; and reinforcing identity and recognition. Importantly, while originally developed for a fundamentally positive analysis of the mechanics of social capital, this framework includes inverted practices. In our analysis, we observe a bifurcation of actions depending on what role they play in the polarization dynamic – integrating relations within the poles or separating relations between the poles. In this sense, social capital contributes to intensified polarization. Empirically, the article is based on a dataset of 30 interviews with 24 policy-oriented civil society organizations (CSOs), here termed think tanks, in Poland.


Introduction
In 2020, in the Polish city of Poznań, a roundtable discussion about recent changes in the country's judiciary system was announced.A broad range of actors were invited under the slogan of 'Justice doesn't divide', including organizations representing judges and several think tanks working in the field of the rule of law.When it became known that advocates of widely adverse positions, including organizations supporting the controversial changes, were among the invitees, one think tank declined participation: 'A major row blew up; quite a lot of nasty words were directed at us "How can you talk to them?"In the end we didn't go' (Interview 17).Another think tank, after careful consideration, decided to take part in the meeting to later face public condemnation from fellow organizations: 'We were attacked, even publicly called out on the internet, and told that we have blood on our hands, because we sat at the same table as Ordo Iuris [an ultraconservative think tank]' (Interview 19).
The general question this event highlights is how polarization affects relations within civil society and how civil society organizations (CSOs) are part of and potentially contribute to polarization.Indeed, as exemplified in the above Polish event, severe polarization among CSOs influences their relations to other organizations and the social capital that these relations entail.While much of the earlier research on polarization has been centred on party politics or voters (e.g., Abramowitz & Saunders, 2008;McCoy et al., 2018;McCoy & Somer, 2021;Somer et al., 2021;Svolik, 2019;Vachudova, 2019), less has been said about civil society actors in polarization (but see Jezierska et al., forthcoming;Özler & Obach, 2018;Rahman, 2019;Roggeband & Krizsán, 2021;Sarkissian & Özler 2013).
In this article, we study polarization within civil society.We analyze how relations between CSOs change in a highly polarized setting, showing how these relations lead to bifurcating social capital practices and intensified polarization.Relational work among CSOs is undertaken to establish, maintain, negotiate, transform, and possibly terminate contacts (Zelizer, 2012).This work may result in social capital, that is, networks of recognized connections (Bourdieu, 1985, pp. 248-249) that may be drawn upon both by individuals and organizations (Rostila, 2008).In this capacity, social capital is a resource that actors may wish for, striving to reach their specific goals (Lin, 2001).
Social capital is often pictured as fostering trust and reciprocity within civil society (e.g., Aldrich & Meyer, 2015;Boeri et al., 2021;Fukuyama, 2000;Putnam, 2000;Putnam et al., 1992;Sommerfeldt, 2013;Taylor, 2011).However, the more critical branch of civil society scholarship has pointed out that this view is overly optimistic, or at least not complete in its coverage of the dynamics and vulnerabilities of social capital within civil society (e.g., Berman, 1997;Chambers & Kopstein, 2001;Gambetta, 1993;Kopecký & Mudde, 2003;Levi, 1996;Portes, 1998;Portes, 2014;Trigilia, 2001;van Deth & Zmerli, 2010).Foremost, critics have voiced that the importance of social capital in civil society varies depending on context; indeed, while social capital may be the source of public goods, it may also turn into public 'bads' (Portes, 1998, p. 14).We wish to contribute to this critical line of research through an analysis of social capital in the context of civil society framed by pernicious political polarization.How does social capital in civil society function in such a context?Conceptually, we combine Zelizer's (2012) notion of relational work, which helps us highlight the continuous effort that organizations put into accumulating social capital, and Lin's (2001) framework consisting of four elements that make social capital work (inside but also outside civil society): information, influence, credentials, and reinforcement.We see these elements as specific practices pertaining to relational work, which are helpful to conceptualize the dynamics of civil society in a polarized context.Our analysis indicates that what works for settling recognition and vouching credibility in a nonpolarized setting may turn into disciplining and distrust in a polarized setting.
Methodologically, we use an abductive approach (Awuzie & McDermott, 2017;Reichertz, 2010).In practice this entails, that in our analytical framework, we depart from both theoretical notions and our empirical observations, going back and forwards between these.Therefore, the theoretical section includes both key concepts and some empirical illustrations.Our analysis is based on a dataset of 30 interviews with 24 policy-oriented CSOs in Poland.The sample consists of foundations, non-for-profit enterprises, and nongovernmental organizations active in policy advisinga group of organizations that we call 'think tanks'.
We ask the following question: How have respondents in our material described shifts in the relations between think tanks in Poland since 2015?How can these changes be conceptualized in relation to civil society and polarization?
The paper continues as follows.First, we present the notions of think tanks, relational work, polarization and social capital.In the subsequent section, we introduce the case of Poland and our empirical undertakings.The two ensuing sections centre around the analysis of relational work and consequential shifts in social capital.In the final section, we reflect on how civil society relations transform in the context of pernicious polarization and how these changes may be conceptualized.

Think Tanks, Civil Society, and Polarization
This article focuses on Polish think tanks in the interest of exploring relations among CSOs in a context of severe polarization.Thus, we address two bodies of literature: scholarship regarding social capital in civil society and literature on polarization.The goal of cross-referencing these two bodies of knowledge is to develop an understanding of how CSOs' relations and social capital are implicated by and contribute to polarization.
Think tanks have come to be consequential civil society actors in large parts of the world (e.g., Abelson, 2002;McGann & Weaver, 2009;Medvetz, 2012;Stone & Denham, 2004) and in Central and Eastern Europe (Buzogány & Varga, 2019;Jezierska, 2021;Jezierska & Giusti, 2021;Plehwe et al., 2018).They constitute a wide array of organizations that are engaged in policy analysis and advocacy in various issue areas.'Think tank' is a vague term, with no common scholarly definition (Garsten & Sörbom, 2023;McGann & Lazarou, 2019;Pautz, 2011;Stone, 2007).We use it as an emic term (Harris, 1976;Mostowlansky & Rota, 2020), which is referred to in our field of study; while some of our respondents use the term, others prefer to avoid it.In both cases, the decision is strategic.Conceptually, we conceive of think tanks as boundary-spanning organizations in the sense that they are situated at the boundaries of various social fields, such as academia, civil society, market, politics, and the media (Medvetz, 2012).To be consequential, think tanks draw upon and combine various forms of capital such as money, academic legitimacy, and contacts.The power they may accrue is to be understood in terms of social relations (of hierarchy, struggle, partnership, and so on) (Medvetz, 2012, p. 35).This relational aspect of think tanks is decisive for our choice of think tanks as an object of study in this paper.Among these organizations, wide relations to politics, markets, academia and other civil society organizations form a key resource.Especially, think tanks position themselves closer to politics compared to other CSOs, which makes them especially vulnerable to polarization dynamics, affecting their relations.Hence, we expect to find good examples of social capital transformation in a context characterized by severe political polarization.Consequently, what happens to think tanks may be applicable beyond this particular group, albeit in less strong expressions.
A useful definition of pernicious polarization is provided by Jennifer McCoy and colleagues (McCoy et al., 2018, p. 16), who identify it as 'a process whereby the normal multiplicity of differences in a society increasingly align along a single dimension and people increasingly perceive and describe politics and society in terms of "Us" versus "Them"'.
Such a process entails that sociopolitical space is simplified into one dominant cleavage and 'alignment of opinions under a single identity' (McCoy et al., 2018, p. 18; see also Vachudova, 2019, p. 690).In a severely polarized context, there are very few connections between the two polarized poles (McCoy et al., 2018).According to existing accounts, polarization is often driven by political parties, which deliberately employ polarizing politics (Lorch, 2021;Mietzner, 2021;Somer et al., 2021;Svolik, 2019).
The role of civil society broadly, and of think tanks more specifically, in relation to polarization has thus far not gained much scholarly attention.Among the wealth of studies on polarization, most have focused on party political polarization and/or voters' preferences (e.g., Abramowitz & Saunders, 2008;Esmer, 2019;McCoy et al., 2018;McCoy & Somer, 2021;Somer et al., 2021;Svolik, 2019;Vachudova, 2019).To the extent that civil society is mentioned at all in this field of study, it is only done in passing and in arguing that polarization among political parties tends to spread to civil society actors, drawing them into the dynamics of alignment along the same cleavage (Mietzner, 2021;Özler & Obach, 2018, see also short notes on this question in Bonura, 2015;McCoy et al., 2018;Vachudova, 2019).In some cases, it has been observed that the polarization of civil society is deliberately designed by parties, for instance through these parties creating new organizations and actively supporting existing organizations that side with the government (Bernhard, 2020;Greskovits, 2020;Özler & Obach, 2018;Roggeband & Krizsán, 2021) while defunding and in various ways stifling organizations that hold an opposing worldview (Kravchenko et al., 2022;Marzec & Neubacher, 2020;Toepler et al., 2020).
Nevertheless, that civil societies can be a divisive force has long been acknowledged.Jeffery Alexander, for instance, argued that there is no civil discourse 'that does not conceptualize those who deserve inclusion and those who do not' (Alexander, 1992, p. 291), which implies that civil society is always to some extent polarizing, creating an in-and out-group.Likewise, Foley and Edwards (1996) submitted that conflicts among CSOs may be sharp and play an important role in autocratizing processes, civil disruption and violence and may thus be polarizing.In a recent article, Ekiert (2021) argued that a 'pillarized' civil society built on vertical divisions facilitates autocratization by enforcing dominant cleavages and transforming politics into a zero-sum game (Ekiert, 2021, p. 2).There are a few other contributions tracing polarization in civil society in various places around the globe (e.g., Abelson, 2020;Aydın-Düzgit & Balta, 2017;Esmer, 2019;García-Guadilla, 2007;Lorch, 2021;Mietzner, 2021;Özler & Obach, 2018;Rahman, 2019;Simsa, 2019).This research indicates that polarizing dynamics are visible in civil society by fewer contacts between groups over the main division line and the CSOs adhering to the other pole being seen as more or less illegitimate (e.g., Esmer, 2019), which further spurs polarization.

Theoretical Framework: Polarizing Social Capital
The dominant, positive connotations of civil society often lean on ideas of social capital as fostering trust and reciprocity within civil society (e.g., Aldrich & Meyer, 2015;Boeri et al., 2021;Fukuyama, 2000;Putnam, 2000;Putnam et al., 1992;Sommerfeldt, 2013;Taylor, 2011).Social capital is essentially understood as based on relations, which require mutual acquaintance and recognition (Bourdieu, 1985, pp. 248-249).Like other forms of capital, social capital is productive, making possible achievements of certain ends that in its absence would not be possible (Coleman, 1990, p. 302).Social capital can be a consequence of social relations both at the individual and the collective level: 'the same dimensions of social capital may be active on different levels of aggregation' (Rostila, 2011, p. 9).For civil society, the main importance of this type of capital is related to the notion that relations, which are based on trust and reciprocity, have the capacity to reduce transaction costs for collective action (van Deth & Zmerli, 2010).Hence, social capital in civil society is assumed to function as a resource that links people to each other, enabling them to pursue their common objectives more effectively (Stolle, 2003, p. 19).
We conceive of the social capital that think tanks construe as a consequence of relational work (Zelizer, 2012).Here, work refers to all activities that organizations undertake to establish contacts within and outside themselves.Through these efforts, they uphold and/or terminate specific relations and negotiate the meanings and implications of who they are and how they are of consequence.The seminars, policy briefs, backchannel communications, TV appearances, and Twitter messages are the media used for this work and transactions between think tanks.The social capital they construe is a temporary fall-out of these interactions.Civil society, and more specifically, the think tank field, is thus seen as a sphere for relational work, in and by which trust and reciprocity can be expressed, negotiated and implemented to varying degrees.
However, as has been stressed by earlier research, social capital may not unilaterally be of positive consequence for civil society.The role that social capital may play largely depends on context.For instance, it is now clear that membership in associations may also result in nonsocial capital and that social capital may be a double-edged sword (e.g., Berman, 1997;Encarnación, 2003;Fiorina, 1999;Levi, 1996;Rivetti & Cavatorta, 2017;Roßteutscher, 2010), at times resulting in trust and networking within specific groups, which is paired with strong dislike of others.Additionally, just as social capital can function for open discourse formation, it can be the tool used by authoritarian regimes to control populations.In that way, it is evident that political institutions may play a significant role in shaping social capital within civil societies (e.g., Encarnación, 2003;Le Van, 2011;Xue, 2021).Moreover, various types of organizing may also have different consequences with respect to social capital, with membership in horizontally ordered associations being more likely to breed trust and civic values when compared to membership in hierarchically ordered organizations; and membership in socially heterogeneous organizationsby providing bridging social capitalbeing more likely to foster generalized trust and tolerance than membership in socially homogeneous organizations (which provides bonding social capital) (Putnam et al., 1992, p. 175;Putnam, 2000, p. 22).Hence, what earlier research successfully has shown relates to contexts and variation, but less has been said at a conceptual and analytical level regarding the cultivation and de-cultivation of trust and confidence.Moreover, practices that make up the functioning of social capital and how they are contextually adapted need clarification.Our ambition is to provide such a conceptualization.Our analytical point of departure is that, essentially, the same practices are drawn upon, but with shifting content depending on context and on who the other actor is.In our case, the context is the polarized political sphere of Poland, in which only actors adhering to the same political pole are seen as trustworthy.
More specifically, drawing upon Lin (2001, p. 19), we identify practices that organizations may engage in when shaping social capital and working with others: facilitating the flow of information; providing credentials for actors that they are trustworthy; influencing agents; and reinforcing identity and recognition.Importantly, while originally developed by Lin for a fundamentally positive analysis of the mechanics of social capital, with all its alleged beneficial effects on individuals and social structures, we abductively (Awuzie & McDermott, 2017;Reichertz, 2010) adapt these practices for a study of relations in a perniciously polarized context.Juxtaposing Lin's conceptualization with our empirical findings, we observe a bifurcation of actions depending on what role they play in the polarization dynamicspecific actions can integrate relations within the poles and separate relations between the poles (see Table 1).For instance, the practice of vouching for those who can be trusted is complemented with disciplining of connections that are crossing the polarization line; some relations provide credentials of trust, but others are a liability and can lead to stigmatizing.In effect, social capital in a polarized context produces both trust and distrust.
Notably, all types of practices are present in a polarized environment.It may even be that all are generally at play in civil societies, more or less polarized.However, our data only permit us to talk of the situation where polarization has become 'toxic', invading what earlier would have been (more) open relations.
Our key concepts are operationalized as follows: facilitating of information refers to actions that think tanks describe as undertaken to promote a flow of knowledge between actors within the pole, whereas controlling flows of information entails limiting knowledge so that only those on one's own side may be reached by it; organizations can provide credentials for others, elevating or discrediting them, by showing their trust in organizations on the same pole, but also by stigmatizing connections with people/organizations on the other pole; influence involves activities by which a person/organization vouches for another, affirming their reliability based on belonging to the same pole, in comparison to disciplining activities that are meant to chastise and scold relations across the polarization divide; reinforcing of identity entails practices that positively acknowledge another person/organization, i.e., confirm a person's/organization's place within the pole, while rejecting refers to practices that mark distance to those on the other side.All these actions have the role of integrating relations withing the poles while separating relations between the poles.
Policy Advice and Advocacy in Poland -Choice of Case, Method, and Data

Think Tanks in Poland
As a single case study (Tarrow, 2021), this article focuses on Poland, with the ambition of yielding propositions on the alterations of social capital relevant for other severely polarized societies.The few previous studies of civil society polarization reviewed above and located in disparate contexts such as Turkey (Esmer, 2019;Özler & Obach, 2018), Indonesia (Mietzner, 2021), Venezuela (García-Guadilla, 2007), Bangladesh (Rahman, 2019), Thailand and the Philippines (Lorch, 2021) suggest that the Polish case is by no means unique and that the findings from our study might have a broader appeal.
After the peaceful Round Table agreements of 1989, Poland was often perceived as a successful case of democratization (Ekiert & Kubik, 1998;Linz & Stepan, 1996).Part and parcel of this success story was Polish civil society, which significantly contributed to the overthrow of the undemocratic state socialist regime and helped keep the transformation on track (Ost, 2005).This given trajectory was reversed, however, when Law and Justice (Prawo i Sprawiedliwość, PiS) won parliamentary and presidential elections in 2015.From then on, moving in a rapid and systematic manner, PiS launched a counterrevolution, redirecting the country towards illiberalism (Bill & Stanley, 2020;Pirro & Stanley, 2022;Zielonka & Rupnik, 2020).
Since 2015, several liberal institutions (courts, the Constitutional Tribunal, media, and Ombudsman) have been under attack.CSOs, equated by the new incumbents with the former liberal elites, were among the targeted institutions (Bill, 2022;Korolczuk, 2022).Funding was cut short for various civil rights organizations that were also legally harassed, and public smearing was orchestrated by the state-controlled media.The space for liberal CSOs significantly shrank (Pospieszna & Vetulani-Cęgiel, 2021).At the same time, parallel structures of newly founded conservative CSOs sympathetic to the ruling party were instituted, and significant financial and symbolic resources were channelled to CSOs supporting the regime.
These deliberate governmental strategies and the openly illiberal orientation of the policies were met with fierce criticism from various actors in Poland.This criticism resulted in deep polarization of the elites and broader society (Vachudova, 2019).The dominant cleavage in civil society, mirroring the political divide, became liberal-cosmopolitan vs. nationalist-authoritarian, anti-and pro-current incumbents.At the time of interviewing and writing, no signs could be found indicating a turn towards depolarization (or democratization).Hence, although we hypothesize that civil society activities would be of equal consequence and importance for depolarization, the paper only discusses how relations of policy oriented CSOs are affected by and contribute to polarization, not depolarization.

Studying Think Tanks
Think tanks are CSOs that engage in policy analysis and advocacy in various issue areas.The operative definition of a think tank used in this article, which was drawn upon in the selection of interviewees, is an organization having a research component and actions directed at influencing public policies (directly, i.e., preparing law proposals, and/or indirectly, i.e., changing the public's attitudes and knowledge about a certain issue) (Jezierska & Sörbom, 2021, p. 399).This definition is intentionally broad, in the interest of capturing the variation that exists among the organizations that make up the think tank field in Poland.According to available directories, there are approximately 70 think tanks in Poland today, which all live up to the above definition.Our analysis is based on qualitative semistructured interviews with representatives from 24 of these organizations, collected in 2020/2021, and follow-up interviews with six of these think tanks conducted in 2021 (see Appendix).Due to coronavirus restrictions, the first wave of interviews was conducted over video conference (in Polish), and the second was conducted face-to-face (in English).Hence, our data in this article amount to 30 interviews.It is important to note that we did not ask for stories on polarization or on social capital.These stories were brought in by the interviewees as a way of presenting the situation for their organization after 2015.All interviews were conducted with CEOs, directors of research or board members to ensure that the respondents were acquainted with the organization as a whole.The sample was purposefully composed to cover a broad range of think tanks in terms of budget, staff size, issue area, organizational age, and position in the polarized landscape.In this way, even though no outright representativity can be claimed, maximum effort was made to include diverse voices and experiences of Polish think tanks, including those highly sympathetic to the PiS-led government and those clearly opposing the government.
The empirical analysis was performed using the NVivo qualitative data application.All interviews were systematically coded on a number of aspects relating to polarization (e.g., perceptions of middle ground, perceptions of Us, perceptions of Them, relations to the government).The coded material was used to inform our theoretical framework.Broadly, these codes were then used to provide answers to the research question regarding changes in relations over time.The quotes used in the analysis are illustrative of the complete material, entailing that not all think tanks turn up as examples.The selected quotes are meant to show the broader patterns revealed through our coding and analysis, and were chosen because they are emblematic of these patterns and categories in our framework.
In the next section, we first introduce and discuss the altering of relations of think tanks in Poland.Hence, this part does not relate to the conceptualization presented in Table 1.Rather, it sets the stage, showing that according to think tanks there has been a transformation of relations since the illiberal shift in Poland.In the consecutive sections we analyze how think tanks' relations changed in the severely polarized context, based on our adapted conceptualization of Lin's four elements of social capital.

Altering Relations
In the interviews, 2015 is described as a critical juncture.Since then, our interviewees have observed changes regarding what relations are feasible as well as the very conditions for creating and maintaining diversified relations.This is not to say that all was well prior to 2015.Even before the illiberal turn, Polish think tankers complained about the Polish public sphere and their own role in it (Jezierska, 2018(Jezierska, , 2021)).Nevertheless, their narratives clearly point to a significant shift oscillating around 2015.
To exemplify, a think tanker describes how the leftist anti-government Political Critique regularly invites a wide spectrum of people and organizations.Those answering the call, however, do not reflect the wide scope anymore: 'Well, with PiS we simply haven't had any cooperation for years.In the sense that these people do not appear at debates.Neither does [the ultra right-wing party] Konfederacja' (Interview 11).
Civil Development Forum, a neo-liberal think tank, which in the strong division between pro-and anti-PiS positions itself in the same camp as the leftist Political Critique, confers framing this change in terms of polarization: In the past, this polarization was weaker, and one could observe more such various interactions […] it is a challenge today to drag someone from one side to the other, and earlier it used to be a bit easier.(Interview 20) In the same vein, the head of the liberal anti-government think tank, Batory Foundation, points to polarization as an obstacle to building relations across the divide.Just as McCoy and colleagues have it (McCoy et al., 2018), the think tanker observes that polarization in the political sphere has trickled down to other spheres of society.
We can see that this polarization in the political space, this party political space, let's call it that, has indeed crossed a certain line, where we can already see that what was possible 10 years ago is no longer possible.So, the fact that for example Jarosław Kaczyński [leader of PiS] used to come to the debates … today we cannot imagine that happening.(Interview 22) This sense of a deep division in the sphere of civil society and among think tanks, mirroring the political divide, is also expressed by pro-government organizations.Talking about the lack of social capital in Poland, the CEO of the conservative New Confederation argues: It [cooperation] is hindered by the general deficits of social capital in Poland.That is, a low level of trust, recently intensified by social media.More and more often, there are situations where people or environments that could cooperate do not cooperate or cooperate less because someone insulted someone on social media.(Interview 13) Communication channels are being blocked as media for relational work.As described in the interviews, think tankers from both camps observe that polarization has changed the way they organize their relations.For instance, the vice CEO of the ultraconservative Christian think tank Ordo Iuris decries the 'discourse that is today dominated by shouting ideologies.A lot of people shout at us too, we are used to it.Sometimes they try to shout down [silence] what we say' (Interview 4).Due to the radical position of Ordo Iuris on issues such as reproductive rights and LGBT + rights, the think tank is commonly blamed for the extreme polarization of the debate on these issues.Nevertheless, it is telling that even Ordo Iuris sees the perils of stark polarization, which hits them as well.
Aware of the negative effects of pernicious polarization, think tanks feel forced to adapt their own relations and ways of interacting with others, arguing that the old, more open type of connection no longer works.As the CEO of the liberal anti-government Court Watch claims, the 'old way of working of think tanks, that they propose certain policies, organize debates, congresses, discussions, and so on, where they discuss, grumble, is simply no longer effective today' (Interview 19).
Thus, the picture that emerges from our interviews is of a deeply polarized civil society, in which the patterns of think tanks' relational work change, affecting their ordinary way of operating.Policy-oriented CSOs in Poland on both sides of the government divide deplore the polarized situation in which they find themselves.Despite acknowledging the need for broad relations, they have largely come to limit their relations only to those sharing the same view of the PiS government.Relations are still based on mutual trust and recognition but have turned towards the poles.

Information
Social capital formation is dependent on a flow of information between actors (Lin, 2001).When civil society is toxically polarized, relations between camps are dismantled, and information flows between think tanks from opposing camps are controlled, so that only those on one's own side may be reached by it.Instead of bridges (Putnam, 2000), think tankers in Poland speak of ditches and trenches and the need for monitoring communication.As the CEO of INPRIS, a think tank critical of the PiS-led government, describes, 'these ditches got ever deeper and there's much less of conversations' (Interview 17), which also means a lesser flow of ideas between the camps.The sharp division certainly limits think tanks' chances of reaching a broader audience across the polarization rift with their message, as CSOs, policymakers and their constituencies will at large only be receptive to the information coming from what is seen as their actors.Controlling the messengers more than the message, all policy knowledge or arguments coming from the other side will be debunked as suspicious.
Conversely, but following the same logic, when the messenger is seen as acceptable, i.e., identified as being on the same side of the divide, her analyses and arguments are granted automatic credibility.Hence, within their bubble-like networks of like-minded organizations that gather around the same pole, the information channels are indeed improved when new and intensified connections within the poles are developed.As a think tanker from the Klon/Jawor Association explains, Organizations, which are somehow united by their dissatisfaction with the current situation for various reasons.[…] this anti-government activity […] sort of cause the flow of information between these subgroups that maybe didn't necessarily always meet [before].(Interview 23) While communication between the dominant line of divisionfor or against the governmentis limited and patrolled, new communication channels are established between think tanks within the same pole.For instance, overriding previously existing differences, neoliberal and leftist organizations have formed alliances and upheld regular meetings about various policy issues such as education, environment, and the rule of law, thus creating connections that were hard to imagine before 2015.In their declared struggle to defend democracy (Jezierska, 2023), differences in economic issues, previously seen as fundamental for the identity of the think tanks, subside.Hence, in the polarized environment, think tanks engage both in facilitating flows of information among the various organizations on the same side of the divide, boosting their social capital formation, and in controlling flows of information between the poles, by building trenches instead of bridges to the other side.

Social Credentials
Another key aspect of social capital formation entails establishing of connections, which can be used to promote social credentials for an actor, providing proof of this actor as competent and trustworthy (Lin, 2001).In a polarized context; however, the inverted sources of social credentials are a lack of connections with the other camp and exclusive connections within the pole.Here, the main activity becomes stigmatizing, and thus shaming for the organizations that dare to keep connections with the opposing side.Such relational work attests to which organizations are to be trusted based on a shared conviction that the government is destroying or improving democracy.It serves the purpose of signalling what is seen as the right values (pro-or anti-government).Likewise, showcasing their association with like-minded organizations, think tanks display their credibility as righteous, which simultaneously gives them the stigmatized, enemy label among representatives from the other camp.
Sometimes the polarizing logic taking hold of how social credentials are granted is not limited to current relations and actions.Past connections might be used to shame an organization and construct it as an enemy belonging to the other camp.For instance, the CEO of Politika Insight describes how past credentials of the organization, i.e., its wide connections to the postcommunist and the liberal side of politics, currently discredit them in the eyes of more hardcore Law and Justice members.
For example, the whole team of Piotr Naimski [Secretary of State in the PiS-government] and Macierewicz [former Minister of National Defence in the PiS government and currently member of Parliament] still believes, and this is what Naimski told us, said it directly to someone from our organization, that for him whatever we do, even if we were 12 years old [then], he would still think that we are the heirs of Rakowski and Urban [symbols of the communist era] and generally he will never talk to us and that's it, period.(Interview 10) This negative assessmentthe historical shamingdoes not mean that Polityka Insight has been completely blocked by the government, but that it has impacted the organization's functioning as an arena for a broad range of actors, including other think tanks and CSOs.Social credentials are thus important on both sides of the polarization rift, and even credentials that are decades old might be used to elevate or discredit a given organization.
Polarization, and the simplification of the social space that it involves do not allow for any middle ground (Jezierska et al., forthcoming); hence, CSOs are identified as belonging to one or the other camp.Such externally drawn identifications are usually associated with stigmatizing and the bestowing of negative credentials.Therefore, what is perceived as opposing think tanks are presented as pro-or anti-government in an attempt to discredit them: It is based on the principle of ranking people, assigning them to some camp, and then they can generally be thrown in the garbage can because we assign to them various traits that serve the purpose of disgusting others to our 'electorate', in quotation marks, or simply to our supporters.(Interview 17) The CEO of the above-mentioned government-critical INPRIS describes how the 'righteous fight' for democracy, which is organized through developing close relations and collaborations on one side of the civil society dividethe anti-government, pro-democracy campmay, in the long run, harm the desired open relational work of the think tank: I think it's very important to maintain such an approach to keep even a minimum of credibility for various parties.If we sort of … if we get involved in a righteous fight in some matter, then probably when we want to talk about something in another matter, we may already have these doors closed.(Interview 17) Uniting in the 'righteous fight' against the government (Jezierska, 2023) and the shifting of relations apparently pose dilemmas for think tanks.At the same time as credentials are crucial for promoting social capital (Lin, 2001), without which think tanks cannot survive, actors on both sides of the divide need to cut relations, strengthening only those credentials that indicate connections to their side.In practice, this entails that connections to organizations sharing the same position, pro-or against government, are tightened, while those to the other side are severed.As the think tankers are well aware, this untethering may close doors and may impact their future attempts to influence or even interact with actors on the other side.

Influence
Actors within civil society can influence each other both in a positive and a negative sense by creating connections or symbolically vouching for each other (Lin, 2001).In the polarized context, vouching is inverted to disciplining through condemnation.One protruding influence that think tanks exercise is the disciplining of other think tanks around the same pole in an attempt to keep the line of division clear and to stop fellow organizations from interacting with those on the other side.Such boundary making is typical of relational work (Zelizer, 2012).Think tanks that make efforts to keep at least some connections with organizations in the other camp are soon met with accusations of not staying sufficiently true to the values of their own pole.
One telling example of such disciplining was introduced in the vignette.When the liberal anti-government think tank Court Watch decided to participate in the same panel as the Christian pro-government Ordo Iuris, it was accused by its fellow antiincumbent organizations of having 'blood on their hands'.Hence, attempts to stay open for interactions across the main line of division are seen as compromising the unity of the camp and the unequivocally negative portrayal of the others.According to this polarizing logic, 'all cooperation [across the divide] is legitimizing' (Interview 19).Relations with the other camp are disciplined and those within the camp are vouched for, even if they entail unconventional alliances.
Importantly, the disciplining logic adopted by most actors in the field thus entails that even those on the same side monitor and attack each other.The CEO of Court Watch gives another example of this type of disciplining influence: We set ourselves the task to monitor, i.e., to control how this National Council of the Judiciary in its new composition functions.This means that it was practically elected by the Law and Justice party and shaped by the Law and Justice politicians.We wanted to control it, how they work.And we were attacked by the absolute opponents of Law and Justice.For what?That by examining this new Council in its new composition, we are 'legitimizing' it.It is such an absurdity, such an aberration.(Interview 19, our italics) The only way to manifest true adherence to the cause is to comply with the polarization logic by severing all interactions with the other camp, here apparently even in the form of scrutinizing them.The same line of argument is presented by the CEO of INPRIS, who bemoans that 'even when it [conversation] occasionally occurs, you can get a nudge from another flank which says "why are you talking to them?".It has become … a very difficult conversation, also between think tanks' (Interview 17).
Ordinarily, maintaining a broader scope of connections, thereby symbolically vouching for each other, would be a consequential aspect of the generation of social capital (Lin, 2001).For think tanks in a polarized civil society, however, the vouching is limited to relations within the pole, and by remaining connections across the divide think tanks risk exposing themselves to being disciplined by organizations guarding 'their' pole.Ambiguity with respect to political orientation and flexibility, qualities that otherwise would be seen as an asset for a think tank, become problematic in the rigid polarized division between two camps.

Reinforcement
Social capital requires continuous reinforcement (Lin, 2001).Manifestations of adherence and distance to one or the other side cannot be done only once; such signals have to be repeatedly confirmed.In the polarized context, social capital is generated by practices that repeatedly acknowledge some organizations' place within the pole, while rejecting those on the other pole and thus marking distance to them.For instance, the CEO of the liberal Shipyard explained that 'At the moment, it's like we're habitually signing [petitions]' (Interview 8).Signing petitions against the government not only serves the purpose of declaring opposing views on given policy issues, it also sediments the identity of a think tank in opposition to the government and reinforces its placement on a given side of the divide, together with the organizations that are seen as desirable allies.
This relational work of continuously reconnecting with and rejecting the actors that provide think tanks with desired credentials, thus avoiding stigmatizing by friends, generates a feedback loopby confirming their resolve in terms of where they stand in the polarized landscape, their image and identity become stronger.This is contrary to the unilaterally positive perception of social capital creation (e.g., Lin, 2001;Putnam et al., 1992).As the literature on polarization has shown, the main line of division becomes the basis for social identity (McCoy et al., 2018), and the social relations that think tanks engage in are a means of upholding that identity.

Concluding Reflection
Drawing on the analysis of policy-oriented CSOs, here termed think tanks, we have analyzed polarization within civil society, raising the questions of how civil society's relational work transforms social capital in the context of pernicious polarization and how these changes may be conceptualized.Starting from the critical branch of social capital research (e.g., Berman, 1997;Chambers & Kopstein, 2001;Gambetta, 1993;Kopecký & Mudde, 2003;Levi, 1996;Portes, 1998;Portes, 2014;Trigilia, 2001;van Deth & Zmerli, 2010) and indications of the importance of context for the way relational work (Zelizer, 2012) may shape social capital (Berman, 1997;Encarnación, 2003;Fiorina, 1999;Levi, 1996;Rivetti & Cavatorta, 2017;Roßteutscher, 2010), we have interpreted shifts in think tank relations in the context of pernicious polarization, in terms of social capital practices.
The analysis shows that the practices of social capital originally developed to conceptualize the beneficial work of social capital (Lin, 2001) are also identifiable in a polarized context but should be complemented with an inverted logic of how social capital works.In this sense, our study contributes to critical approaches to social capital, focusing on how it both builds and dismantles social cohesion.We have identified practices of social capital that have the role of integrating relations within the poles and separating relations between the poles.Hence, we have observed how think tanks guard the cohesion of their respective polarization pole by controlling, disciplining and stigmatizing organizations on their own pole while rejecting organizations on the other pole, including those who take an ambiguous position.On the other hand, the beneficial work of social capital, i.e., the facilitation of the flow of information, the vouching for each other, the development of trust and mutual recognition by acknowledging identity, is also found but is more distinctly present among organizations around the same pole, serving to integrate the pole.Hence, the relational work we have observed is simultaneously limiting and opening relations among think tanks.They cut off and are disciplined to cut off relations with organizations on the other pole, while they expand and deepen relations with those around the same pole.Through these practices, both trust and distrust are produced.
Importantly, one of the drivers for gathering around the poles, and hence contributing to the polarization of civil society, is the need to dissociate one's organization from those on the other side.The policing of these boundaries is mainly done by organizations in the same camp.This means that think tanks taking a more open stance will generally be denounced twice.First, organizations from the other pole criticise them for what they are saying, disagreeing on the content; in a polarized context, essentially every policy issue is politicized and polarized (Bill & Stanley, 2020;McCoy & Somer, 2019).Then, think tanks from the same side of the divide criticise them, disciplining them for even exchanging views with the other pole.There is therefore little to gain by interactions unless they are kept within the pole.While our research design and data cannot provide an answer to why this is happening, our theoretical framework suggests that these processes are related to the logic of belongingthe fear of being ousted as an organization or persona non grata within one's own group.
Additionally, while we cannot establish the causal direction, we do observe that shifts in civil society relations tend to reinforce polarization.In particular, when gathering around the poles social capital works in a bifurcated way.Hence, we have identified a two-vector work of social capital among the polarized policy-oriented CSOs in Poland.We argue that by specific relational work, and engaging in social capital generation, think tanks contribute to polarization.Severing relations with those on the other pole clearly replicates political polarization, transferring it to yet another sphere.Concentrating relational efforts within the poles, which results in new patterns of relations among organizations that did not collaborate before, also deepens the polarized situation.
It may be that the identified practices are generally at work in civil societies, more or less polarized.Think tanks have stronger connections to politics, compared to other CSOs, and are thus more exposed to the polarization dynamic than for instance service providing organizations and social movements.This interest in being close to politics, we argue, makes think tanks especially vulnerable to polarization dynamics, affecting their relations.We therefore suggest that the forms of social capital that we have identified among think tanks may be applicable beyond the organizational type of think tanks, albeit in less strong expressions.This, however, will need to be explored by future studies, which hopefully will find our framework helpful for analyses of CSOs within more or less polarized contexts and among other types of organizations.We also hope that our study will vindicate civil society as an interesting field for polarization scholars.

Table 1 .
The work of social capital in polarized contexts.