AI,social media and echo chambers: From algorithmic prediction to antagonistic politics and the democratic potential

Introduction

The question of whether social media technologies reinforce or undermine democracy is to this day a highly contested issue and one that is at the core of scholarly studies of technology, in their various disciplinary manifestations (Loader and Mercea, 2012; Persily and Tucker, 2020; Tucker et al., 2017). Initially beholding the promise of free and open communication, early iterations of the internet and social media platforms have since evolved into multi-layered socio-technical structures of control (Rouvroy and Berns, 2013; Zuboff, 2020), adding layers of complexity and nuance to discussions about “epistemic democratization.” It was once fashionable to extol the liberating potentials of social media platforms, very often by citing resistance and mobilization efforts in non-Western countries (Diamond and Plattner, 2012). One can recall the celebratory proclamations during the so-called—and Eurocentrically labelled—“Arab Spring,” when Twitter was widely credited for presumably unleashing a long-suppressed democratic impulse that had been brewing in the “Arab world.”

This overly optimistic outlook has since waned, giving way to a critically informed scholarship, highly sensitive to digital technologies’ inherent dualities: of controlling and providing, of dividing and uniting, of ordering and dislocating (Feenberg, 2017). Empirical considerations were at the forefront of this impetus. The introduction of digital systems of control, particularly in the form of algorithmically-driven machine learning systems, transformed what was once a, by and large, unfettered space of communication, to a highly regulated milieu predicated on capital accumulation and surveillance imperatives (Crawford, 2021; Pasquale, 2016; Zuboff, 2020). At the same time, the emergence of novel technological tools, particularly in the form of generative AI, introduced novel communicative possibilities, which in their turn exercised a variety of dislocating effects in the digital public sphere (Calvo and Saura Garcia, 2025).

Consequently, critical technology studies had to contend with the growing array of disruptive effects brought on by (AI-powered) social media, ranging from fake news, disinformation and deep fakes, to political polarization and social fragmentation (Habermas, 2024; McIntyre, 2018). The solutions provided to the issues at hand span the gamut, ranging from the need for expanded digital literacy, to the imperative of restoring the status of rationality and science in the public sphere, to frameworks of effective citizenship and governance, to calls for platform socialism, and so on (Barton, 2019; Hoggan-Kloubert and Hoggan, 2023; Lazer et al., 2018; McIntyre, 2018; Muldoon, 2022).

Yet, there appears to be a substantive absence in the literature on the political effects of social media platforms, particularly with regard to the operational logics they produce—the manner by which they alter the mode of political practices. It is this line of investigation that the present article undertakes. Rather than starting from the question of how social media undermine democracy, or how social media can be deployed as incubators of democracy, we peer into the types of political practices social media facilitate. We then direct this line of inquiry towards strategic considerations, asking the following: Which democratic strategies are most effective given the modes of political practice that social media facilitate?

To do so, we develop a novel analytic framework that brings post-Marxism and Bernard Stiegler’s post-phenomenological account of technics into conversation, examining how AI-powered social media platforms reconfigure the spatial and communicative coordinates of the digital public sphere. Our investigation reveals what sort of spaces social media technologies produce, where space is understood as stable symbolic systems, and what sort of political possibilities they enable (Laclau, 1990).

We argue that AI-powered social media pre-emptively structure future political possibilities by offsetting the political subject’s anticipatory faculties. By digitizing the subject’s anticipatory faculties in the form of automated content recommendation, spaces of ideational homogenization (i.e., echo chambers) are produced. Yet, AI-powered predictive processing simultaneously exacerbates heterogeneity at the systemic macro-level, by hollowing out shared spaces of communication. This configuration—of internally homogeneous and externally heterogeneous echo chambers—facilitate the antagonistic form of politics, which is predicated on pronounced in-group/out-group logics, very often through the deployment of rhetoric. In this sense, rhetorical antagonisms discursively formalize the frontiers that AI-powered social media platforms generate.

We conclude by delineating a viable democratic strategy that is compatible with the operational logic of social media, arguing that, in a fragmented digital public sphere, democratic forces can only manifest a decisive political offensive through multi-tier rhetorical deployments that: (a) enable the forging of diverse democratic alliances; and (b) materialize a formidable offensive to anti-democratic processes through the production of multiple antagonistic frontiers. As a potentially formidable discursive strategy, political deployments of rhetorical antagonisms need to establish “rhetorical symmetry” between disparate fields of conflict, as well as disparate political and economic demands.

Structure, indeterminacy and the political

In the post-Marxist tradition, “space” is theorized as ideal and material configurations characterized by recursive and predictable parameters (Laclau, 1990: 41–42), coinciding with sociological understandings of “structure” (Giddens, 1984: 25). Conversely, time is theorized as structural indeterminacy (Laclau, 1990: 42–43), which enables the possibility of agentic change and freedom. This is premised on the fact that in the absence of structural indeterminacy, social agents would succumb to absolute repetitive conduct, rendering change, as a vital component of freedom, agency and experience, obsolete. Seen from this perspective, space (structure) and time (indeterminacy) do not consist in independent ontological precepts, but dialectic ontological possibilities inherent in social experience (Laclau, 1990: 44). Temporality, as freedom from the structure, presupposes the existence of a structure. Space, as the reduction of indeterminacy to recursive parameters, presupposes, as the very possibility of its articulation, a terrain of freedom and possibility.

Temporality, in the form of structural indeterminacy, is the “ground zero” of “the political” (Laclau, 1990: 50–51). It is the indeterminate character of the social that gives room for political agency to emerge, where subjects exercise choice through practices that partially determine the future (Laclau, 2005a: 55). Political agency hinges on processes of “articulation,” which involve the (re)ordering of symbolic systems, such that they create, reproduce, or transform social and political relations (Laclau and Mouffe, 2001: 105–114).

This understanding of “the political” is premised on the linguistically-oriented paradigm that sees social life as a network of discursive operations (De Saussure, 1966; Foucault, 2002; Laclau and Mouffe, 2001). From this perspective, meaning and, thus, social relations, are not established self-referentially, but contingently, within networks of differential and associated ideas that acquire a degree of stability (De Saussure, 1966; Lacan, 2006; Laclau and Mouffe, 2001: 108–114). Obversely, indeterminacy—and thus political agency—is seen as the flipside of a “discursive excess” that is not fully-subsumed by its relational (i.e., structural) context (Laclau and Mouffe, 2001: 113). In this sense, meanings are at any given point probable yet incompatible with one another, occasioning, at once, potentials for structuration (space) and disruption (time) through the articulating practice.

Articulations (re)situate and (re)signify meaningful terms, resulting in meaningful polysemy and, crucially, the overdetermination of the social, whereby symbolic systems become increasingly complex (Althusser, 1967; Freud, 1955; Laclau and Mouffe, 2001: 111). A liberal articulation of “democracy,” as an example, is very different from a socialist or fascist articulation of the same term. A “location” of potential contestation thus opens up that is centered on the definition of the terms at play—a political language game, if you may, as an outcome of conceptual “incompatibilities.” Politics amounts to a contest over the control of language in a never-ending quest for finality and determination (Freeden, 2015: 22–25), that is, as an attempt to curtail indeterminacy, by fostering stable symbolic systems and, potentially, forms of control (Foucault, 1978; Laclau and Mouffe, 2001). Politics thus oscillates within the confines of structural coordinates and indeterminate possibilities.

This strife, this tug-of-war, between space and time, determinacy and indeterminacy, has also been thematized in the (post-)phenomenological tradition and, in particular, in the work of Bernard Stiegler who clarified its political significance with regard to the way it is manifested through digital technologies and platform media. As Stiegler argued, algorithmically driven social media systems curtail structural indeterminacy (and thus political agency) through predictive automation. This results in the fragmenting of the public sphere into segmented and internally homogenized public spaces (i.e., echo chambers). Paradoxically, this accelerates indeterminacy at the macro level, by undermining shared spaces of communication. As we will argue, this causes a general shift in the form that politics assume, whereby political conflict shifts from agonistic looks, premised on contests over the control of language within shared spaces of communication, to antagonistic outlooks, premised on in-group/out-group logics and processes of symbolic exclusion.

Digital technologies and indeterminacy in Stiegler

Phenomenology identified a crucial asymmetry between the way we experience the future and the way we experience the past. These two modalities (pastness and futurity) are given to consciousness, such that time, in its authentic manifestation, maintains an irreversible direction and a radical openness and indeterminacy towards the future, thus remaining irreducible to space, which is characterized by symmetry and reversibility. According to Husserl (1964), time-consciousness (i.e., the experience of time as a continuous flow) is a necessary condition for the experience of meaning, which is co-constituted by three distinct dimensions that structurally intertwine: (i) primal impression (hereafter “immediate perception”), which is the vivid awareness of the present moment; (ii) retention, which retains in consciousness what has just passed; and (iii) protention (hereafter “anticipation”), which anticipates future events.

This is premised on the fact that immediate perception is not enough on its own to constitute meaning, but needs to be supplemented by an awareness of the past: “any impression that consciousness lives through passes into retention and consciousness continues to be aware of the former impressions retentionally as it lives through new impressions” (Geniusas, 2024: 124). If this were not the case, that is, if a thing is not given in immediate perception as well as in retention, then, we would not be able to register any meaningful object in time—we would not be able to perceive object permanence and semantic identity.

Bernard Stiegler takes this classical phenomenological insight and further develops it in a different direction. According to Stiegler, retentions are not automatic (as Husserl assumed): consciousness does not retain anything and everything, but only certain immediate perceptions (Geniusas, 2024: 125; Stiegler, 2014: 125) and it is anticipations that determine which immediate perceptions are retained. And while the future-oriented part of consciousness (i.e., anticipation) entails a dimension of indeterminacy, it is at the same time partially determined through previous experience. And this partial determination determines which immediate perceptions need to be retained in order to constitute the dynamic meaningful whole of purposeful activity in its temporal duration. There is thus a structural intertwining between retention and anticipation.

Allow us to give a simple example from the first person perspective that will clarify this complex dynamic. I am currently sitting at my desk, typing this text on my laptop. I have a general sense of the direction I want this text to take and the arguments I intend to develop, which involve me outlining Stiegler’s theory and tying it to the overarching argument of this paper. To do so, I must carefully select my words to build sentences that support the overall argument. The precise phrasing, however, has not yet taken shape in my mind; it emerges and becomes fixed as I type. What appears in my consciousness at any given moment is not mysterious or fragmented—it is not merely isolated letters lacking meaning. Rather, I experience meaningful wholes. For instance, when I press the key for the letter “p” (the first letter of the word “press”) my consciousness does not register just the isolated “p.” Instead, it unthematically retains the preceding context—both letters and sentences—enabling me to grasp “p” as part of the word “press” and, more broadly, as part of a coherent sentence. In this way, immediate perceptions are understood through earlier impressions, which are preserved in retentive consciousness. Retention, therefore, provides the structure through which immediate perception becomes intelligible.

But by virtue of what mechanism does retention itself become determined? According to Stiegler, it is anticipation that motivates and gives retention its focus. Let’s return to the example. I can allow my consciousness to wander, becoming aware of everything presented to it beyond the specific elements involved in the purposeful activity I am engaged in. I can perceive the entire desk, various items beside the laptop, my dog sitting nearby, books on the shelf in the background, etc. However, I cannot retain all of these things in consciousness at once; not everything can be held in focus, especially if I am to engage in a purposeful activity. It is intrinsic to purposeful activity that consciousness selectively retains only what is relevant to that activity. Guided by my intention to construct a particular argument, I anticipate the words I need, and these relatively indeterminate anticipations motivate me to emphasize certain retentions, while pushing others to the margins. In other words, these partially determined anticipations passively shape what is retained in consciousness, aligning it with the demands of the task at hand.

But what gives anticipation its partial determination? What gives it a focused direction? According to Stiegler, it is “secondary retentions,” that is, recollections of past experience. It is because of the way secondary retentions afford content that anticipations are only relatively indeterminate. Thus, anticipation in purposeful activity is a “motivated possibility” (Geniusas, 2024: 125), that is, it is a product of a predictive process that is motivated by determinate possibilities that are part and parcel of the recollected past. Going back to our example, it is because I consciously recall Stiegler’s theory and its provisional connections to the overarching argument I plan to make that I am motivated to give some determinate content to my anticipated future (i.e., the text I am about to type).

This mechanism, this reliance of anticipation on “secondary retentions” is a by-product of individuated experienced recollections, that is, personal experience, interpretation and memory: the way I read, understand, and remember Stiegler’s theory. The fact that these are rooted in personal experience explains why when different individuals read the same sentences, they may understand and interpret them differently. Differences in personal recollections guarantee that anticipations remain partially indeterminate, rendering the future an open process of individuated possibility, agential action and, potentially, political contestation.

As Husserl had pointed out, since recollections (secondary retentions) rely on personal experience, they are finite: there is a limit as to how far back a specific subject can remember (see Soueltzis, 2021: 55–59). But it is a constitutive human characteristic, argues Stiegler, to have overcome this limit by way of mnemotechnics, in effect arguing that human consciousness is technically constituted (De Preester, 2021: 105; Geniusas, 2024: 126–127) because it relies on technics that exteriorize retention. It is against this premise that Stiegler introduced his most original and important philosophical contribution: the concept of “tertiary retentions.” Tertiary retention “refers to the exteriorization of memory in technical reality” (Geniusas, 2024: 125), the most important example being the technics of writing. Thus, Stiegler identifies three levels of retention (primary, secondary and tertiary) that he deems to be co-constitutive, while granting primary significance to tertiary retentions in the shaping of human intellect and, crucially, in the shaping of anticipations, that is, the character of intention, interpretation, and, thus, political possibility.

A key point to remember here is that while retention, be it primary, secondary or tertiary, partially determines the future, there still remains an important asymmetry between retention and anticipation, between pastness and futurity, between the determinacy of the past and the relative and incomplete determinacy of the future, which still allows for individuated agential action. In effect, this asymmetry and indeterminacy at the heart of temporality safeguards and complies with what physicists call “the arrow of time” (Carroll, 2022: 127; Kuzemsky, 2023: 283): time seems to move from the past to the future and never the other way around—unlike space, which is symmetric and reversible.

But what happens when the technics of tertiary retention become digitized? A fundamental change takes place, according to Stiegler. When tertiary retentions become digitized, they inaugurate a new era that inter alia radically upsets the aforementioned asymmetry. In effect, when tertiary retention becomes regulated by automated processes, and, in particular, by algorithmically-powered AI systems, it becomes severed from secondary retention and no longer grounded in individual consciousness and collective memory (Geniusas, 2024: 131). What emerges is a new mode of power, which Stiegler designates as “algorithmic governmentality,” which involves the modification of tertiary retention in such a way that it minimizes indeterminacy in anticipation, with the effect of establishing symmetry between the past and the future. The past, which is formalized as “big data” in digital networks, in which concepts and symbolic signifiers have been reduced to formalized representations via the mediation of numeric values and statistical techniques, is used to generate future content.

The root of the problem, according to Stiegler, is that these information systems and the content and connections generated by algorithms are designed in such a way that they negate information entropy, ¹ that is, the degree of “disorganization” of a system. Cybernetic systems reduce information entropy insofar as systemic uncertainty is overcome by probabilistic prediction, as in the case of deep neural networks. To put it in post-Marxist terms, AI predictive systems reduce time to space by curtailing structural indeterminacy. As Stiegler (2018: 36) argues, “the data economy has established an industrial and automatized production of protentions [anticipations] that amounts to guiding them by remote control, or, in other words, it amounts to their annihilation.” Hence, the possibility of disruption, indeterminacy and freedom is curtailed (Stiegler, 2018: 36–37), precisely because algorithmic governmentality imposes a certain epistemic homogenization that undercuts political agency, by precluding “the production of new knowledge, differentiation and bifurcation” (Stiegler, 2018: 142).

As regards non-learning, deterministic and rule-based algorithms generating digital content, this is straightforward: the given input always produces an output on the basis of a fixed sequence of steps. Here, there is no room for any spontaneous or creative generation, or any asymmetry between the future content generated and past content: anticipation is simply and monolithically a function of tertiary retention. Dynamic algorithms, on the other hand, which are stealthily embedded in social media platforms, such as Facebook, Instagram, X, TikTok, Bluesky, etc., generate content on the basis of predictive processing. These algorithms do not always yield the same output for the same input—they incorporate probability to handle uncertainty. Nonetheless, by partially ordering user experience vis-à-vis recommendations and content hierarchization, social media platform algorithms “generate” the future, under an equally regulatory regime that negates indeterminacy in anticipation.

To put this into perspective, in cybernetically-regulated social media platforms, capital is constituted as a power over anticipations, whereby “user experience”—and thus behavior—is nefariously directed, in the service of capital accumulation imperatives. By constructing user “digital twins” through multifarious data harvesting processes, and by strategically leveraging such data to maximize user engagement, social media platforms are enabled in maximizing profit through ad revenue. The novelty of platform capital thus lies in its unique capacity to structurally connect the “attention economy” (Thiele, 2025) to user surveillance (Zuboff, 2020) through the production of predetermined feedback loops (thus controlling anticipation). In this sense, capital should be understood as a structural precondition of social media design and ecology, at least in its current manifestation, amounting to an extension of capitalist logics to novel domains of social experience, whereby “digital behavior” is steered and leveraged for the extraction of surplus value.

The key takeaway from this discussion with regard to the overarching argument of the present article concerns the manner by which cybernetic systems (social media in our case) alter communicative and political possibilities. Stiegler’s ultimate conclusion with regard to the effects of platform networks and, in particular their tendency to incubate spaces of epistemic homogenization, is that they ultimately produce a contradiction. On the one hand, cybernetically-regulated social media systems generate homogenous public spaces with low levels of entropy, in what is commonly referred to as echo chambers. On the other hand, the production of echo chambers accelerates systemic entropy by undermining shared spaces of communication on the macro level. In other words, epistemic homogenization incubates a configuration whereby discursive communication becomes segmented and compartmentalized.

From algorithmic processing to antagonistic politics

How can we make sense of Stiegler’s insights in terms of politics? We propose the following: When a political subject’s futural awareness—its anticipations—become automatically produced by social media platforms that curtail indeterminacy, its political imagination becomes entangled in a “past” that is tendentially rid of its overdeterminations (i.e., symbolic complexity). This reproduces and resituates the subject’s political imagination within a novel spatiotemporal logic that has discernible political effects. AI-powered social media, in essence, occasion a dual movement. On the one hand, social media platforms foster echo chambers by standardizing anticipations through the production of feedback loops that position political possibilities within structurally predetermined coordinates. Indeterminacy has been overcome, reduced to stable recurring coordinates that are facilitated by the rank and filter logic of predictive AI, where user experience is molded according to the logic of statistical aggregates based on a wide collection of data points (Burrell and Fourcade, 2021; Rodilosso, 2024; Zuboff, 2020). On the other hand, a structural distancing is occasioned between echo chambers, as shared spaces of communication dissipate and, with it, the possibility of resolving conflict, in order to arrive, event tendentially, to a social consensus. In its totality, this configuration yields a precise political effect, namely, the possibility of political antagonisms.

Above, we identified the overdetermined character of the social as the raison d’être of politics. Meaningful polysemy yields a particular systemic configuration whereby meaningful terms are at the same time probable yet incompatible with one another, potentiating, as it were, the possibility of contestation centering on the control of language. Here, indeterminacy, freedom and, thus, political articulations germinate within discursive locations where meaning is, at the same time, shared but also contested (Mouffe, 2005). However, in the absence of shared spaces of communication political conflict may more readily shade from an agonistic outlook, characterized by political contestation within common discursive channels, to an antagonistic outlook, characterized by processes of symbolic exclusion and “othering” (Laclau, 2005b: 67–100). “The Other,” as formalized in the referent of competing echo chambers appears as the “distant” enemy. Lacking common discursive tools that could reduce the confrontation to one of “disagreement,” “deliberation,” “contestation,” etc., the only alternative that is left is a frontal collision. Note, that what is effaced here is not overdetermination, as such. Rather, the principle of overdetermination becomes encased within self-referential echo chambers.

Political antagonisms presuppose a very precise political operation, whereby a social cluster’s internal components (identities, demands, narratives) are symbolically consolidated through the production of rhetorical formalisms, very often in the form of “empty signifiers.” In semiotics, rhetoric is understood as a fundamental component of language, belonging to the order of the symbolic (Laclau, 2014). Its highly connotative character works in the excess of literary representation by invoking a plurality of meanings and affects through figurative mediums, such as synecdoche, analogy, allusion, irony, metaphor, etc. (Peña Cervel and Ruiz de Mendoza Ibáñez, 2022). Fundamentally, rhetoric employs the paradigmatic pole of language (De Saussure, 1966) through the use of discursive substitutions, where one meaningful term signifies another—what Laclau designates as the “logic of equivalence” (Laclau, 2014). Thus, in the case of metaphor, a disparaging use of the term “bureaucrats” might operate as a symbolic replacement for “government corruption,” and in the case of synecdoche, where a particularity comes to represent a totality, the term “nation” might operate as a symbolic replacement for “territory,” “state,” “citizenry,” “tradition,” “economy,” etc. (Anastasiou, 2022: 165–180).

The case of empty signifiers involves a particular rhetorical operation, whereby a signifier becomes saturated with signifieds, such that it is unable to referentially exhaust them in signification (Laclau, 2007: 36–46). In this sense, no single signified can operate as the signifier’s principle of unity, owing to the presence of multi-tier conceptual incompatibilities. Laclau (2005b), as an example, employs the notion of the empty signifier in his analysis of populist politics, indicating how the notion of “the people” acquires a dimension of emptiness when it comes to represent a plurality of diverse political demands that correspond to heterogeneous identities. Political employments of the signifier “the people” cannot, at the conceptual level, referentially exhaust this diversity, which corresponds to heterogeneous experiences: “The people” may encapsulate the (demands of the) middle class, the working class, poor lesbians, transgendered Latinos, Muslim blacks, etc. Rather, what comes to operate as an empty signifier’s principle of unity is a modality of exclusion—a “constitutive outside,” a constructed “Other,” to which the empty signifier is symbolically counterposed (Laclau, 2005b: 98; 2007: 37–38). Thus, in the case of “the people,” it is the symbolic negation of “the establishment,” or “the elite” that can operate as a common symbolic ground facilitating the consolidation of heterogeneous identities and demands and thus, the signification of “the people” as a political body.

Empty signifiers play a key role in politics, because they can be particularly effective in symbolically consolidating a heterogeneous political body through their very polysemy. By operating as a symbolic substitute for a wide array of political demands that are tied to a plurality of identities, which display heterogeneous ideological orientations, empty signifiers can signify an emergent political body in the excess of conceptual specificity (Laclau, 2014: 64–78). This is what Laclau means when he says that empty signifiers do not reference the particularity of their signifieds (e.g., political demands, identities, narratives, experiences), but the mere equivalence between them. In this sense, an empty signifier is predicated on a dual rhetorical operation whereby a plurality of heterogeneous elements (identities, demands, narratives) are symbolically consolidated through the articulation of symbolic equivalences and, at the same time, symbolically counterposed to a differential entity articulated as negation (Laclau, 2005b: 72–100). It is this dual logic that underpins the logic of political antagonisms.

The argument here with regard to the spatiotemporal configuration of social media platforms is that empty signifiers discursively formalize the frontiers that predictive AI generates, by imparting a rhetorical platform through which intra-cluster (e.g., echo chamber) identities are symbolically consolidated. In other words, there is an operational compatibility between the logic of AI-powered social media technologies and antagonistic politics—a “mirror effect,” if you may. The important qualifier here, however, is that the production of echo chambers does not, in and of itself, produce a seamless intra-chamber cultural homogeneity, but, rather, a degree of ideational clustering, upon which a potentially diverse range of actors converge. These actors may display disparate identitarian characteristics and disparate politics demands, but in coalescing around ideational spaces that are symbolically contiguous, particularly in terms of what they oppose or exclude, the conditions are enabled for their potential symbolic consolidation as part of a political body. As relevant examples, we may cite how the notions of “gender ideology” and “critical race theory,” have had the effect of consolidating diverse identities in social media, including moderate conservatives, neo-fascists, incels, femonationalists, neoliberals, etc., by incubating common representations of a purported enemy.

What should be highlighted here is that the political logics thus far identified capitalize on, and indeed exacerbate, structural coordinates that pre-exist algorithmic governmentalities, and which facilitate “Manichean” politics through what has been described as “politics of resentment” (Tomšič, 2023). The notion that neoliberal governmentality exacerbates marginalization, inequalities and exclusion, and thus occasions a diversity of unsatisfied citizen demands that translate, among others, to reactionary politics, is well established in the sociological canon (see Davidson and Saull, 2017). The crux of the argument is that neoliberalism incubates citizen grievances, indignation and a profuse sense of insecurity by undermining democratic sovereignty, economic security and, indeed, social bonds (Brown, 2019; Tomšič, 2023) through diverse processes, including: welfare state entrenchment, assaults on collective bargaining, privatization of state resources, occupational precarity, social individuation, and so on. However, these processes are at once supplemented with neotraditionalist outlooks that facilitate resentment through multilayered processes of “Othering” (Tomšič, 2023: 96–98). This has the effect of channeling citizen angst in the direction of scapegoating logics that hark back to homogenized understandings of collective belonging, be it “the nation,” “religion,” “the family,” and so on. It is by no surprise that Friedrich von Hayek, the very “father” of neoliberalism, asserted that there exists a mutual symmetry between “traditional moral codes and modern market rules” (Slobodian, 2025; Tomšič, 2023: 96).

In this sense, one can identify a structural correlation between citizen grievances, as products of the neoliberal order, and the politics of resentment. Where neoliberalism engenders an “anti-social” milieu by dissolving collectivities and social bonds, the politics of resentment “step in” to consolidate social fragments through an equally anti-social outlook that constructs homogenized representations of identity by counterposing them to purported enemies (e.g., “immigrants”) that are seemingly responsible for “the theft of enjoyment” (Stavrakakis, 2007: 197–198). In this context, resentment should be understood as “the inevitable affective surplus of the homogenizing division, the affect that logically results from the determination of the other as a threat of the people’s consistency or one’s own identity” (Tomšič, 2023: 98). This duality, which animates a fine line between the logic of the free market and identitarian outlooks, is personified in the most remarkable way in current Far-Right configurations, which tiptoe between libertarian and neotraditionalist sensibilities (Slobodian, 2025).

The design, operationalization and, indeed, the deployment of algorithmic logics in social media exacerbate these underlying dynamics by monetizing the politics of resentment through modalities of “affective polarization,” whereby the “affective surpluses” associated with antagonistic politics are not only leveraged, but actively reinforced and perpetuated in the service of corporate profits, by capturing and instrumentalizing user attention.

Epistemic fragmentation in the digital public sphere

Ever since the publication of Lyotard’s (1984) The Postmodern Condition, discussions centering on epistemic fragmentation have proliferated. This development is expressed in recent analyses of the digital public sphere (Habermas, 2024), which we understand as digitally-mediated communicative spaces involving information exchanges, contestation and conflict pertaining to issues of a political nature. The death of metanarratives, on account of epistemic particularism, yields competing epistemic spheres that redirect contestation through an infinite unfolding of language games. It seems to us that Lyotard’s formulations are today more relevant than they were at the time they were written. The current socio-political impetus, which is to a considerable extent morphed by the input of AI systems, extends the logics identified by Lyotard in the realm of formal knowledge to the general (digital) public sphere. We, at the same time, however, identify a substantive shortcoming in Lyotard’s formulations, with regard to his inability to predict the dislocating effects of epistemic fragmentation, namely, the incubated drive to re-constitute metanarratives. Lyotard (1984) asserts that

The performativity of an utterance, be it denotative or prescriptive, increases proportionally to the amount of information about its referent one has at one’s disposal. Thus the growth of power, and its self-legitimation, are now taking the route of data storage and accessibility, and the operativity of information. (p. 47)

In other words, performative power rests on information accumulation, as well as on the ability to operationally deploy information in the service of particular ends and interests. Here, it is the instrumentalization of information density that potentiates power potentials. Empirical considerations, however, cast partial doubt on this theoretical position. Politics in the digital public sphere—and especially in their post-truth variant—do not rely solely on information density, control of data, or the complementarity of verificational axioms and methods. Rather, the manner by which politics are unfolding in the digital arena are better understood as a symptom of information density (Kalpokas, 2019).

It is true that an unprecedented amount of power has been amassed by states and corporations by instrumentalizing big data analytics for capital accumulation, surveillance and military purposes (Burrell and Fourcade, 2021; Crawford, 2021; Elliott, 2022). However, political prominence in the digital public sphere does not necessarily hinge on actors’ capacity to amass and control information, but to creatively instrumentalize the logics of social media platforms vis-à-vis a highly selective information eclecticism that is directed for public consumption in the form of highly formalized representations (e.g., empty signifiers, figures of speech, “memetic culture,” rhetorical modulations, etc.) (Anastasiou, 2025).

In other words, in the digital realm of proliferating information flows, as facilitated by the logic of predictive AI, it is the ability to effectively articulate frontier effects through political antagonisms that facilitates discursive power. It is the whim of rhetoric that occasions power differentials in the digital public sphere, at least at the level of “discursive conflicts,” by “cutting through” the complexity and density of information flows through processes of discursive simplification. Consider how in recent years the incredibly vague catch-all notion of the “woke enemy” facilitated a formidable anti-democratic political advance that encapsulated a diversity of actors, ranging from ethnopopulists, to neoliberals, paleolibertarians, accelerationists, independents, as well as Liberals and even parts the Left (Dhoest, 2025: 2; Johansen, 2025: 50-8). Looking at the political logics at play, it is fascinating to note that while the presumable “woke enemy” acquired discursive centrality across these “clusters” as a nominal reference point (i.e., as a common symbol of exclusion), it simultaneously meant different things to each group of actors: To those with liberal, libertarian and Left sensibilities, it came to be associated with censorship, that is, as a threat to freedom of expression (Johansen, 2025); to those with conservative sensibilities it came to be associated with the onslaught of “gender and race ideology;” to those with nationalist sensibilities it came to be associated with “great replacement” anxieties on account of the “immigrant threat,” presumably facilitated by “no-border globalists;” to those with “rationalist” sensibilities, it came to be associated with the presumable “excesses” of LGBTQ pursuits, for example, gender-affirming hormone therapy for adolescents.

In other words, it is the rhetorical ambivalence of the term “woke”—its discursive “emptiness”—that facilitates the articulation of a broad “chain of equivalence” (Dhoest, 2025: 5; Johansen, 2025: 51). This is supported by empirical studies that document a pronounced linguistic polysemy in political deployments of “woke” that nonetheless harks back to Schmittean friend/enemy distinctions (Cammaerts, 2022). Dhoest’s (2025: 5) study of anti-woke discourse on twitter, as an example, identifies a broad linguistic chain associated with woke hashtags, including, among many others: #replacement, #migration, #BLM, #trans, #metoo, #Islam, #Europe, #climateactivists, #tradition, #pedophilia, and so on. In Dhoest’s (2025: 5) own words, “[t]hroughout the tweets, “woke” appears as an amalgam, a monster stitched together from many disparate parts [. . .] represented as a unified entity, a ‘woke culture’, ‘sect’ or ‘mob’.”

The heuristic power of rhetoric, whereby short-cut slogans (e.g., empty signifiers) are provided as diagnoses and cures to an increasingly complex and dislocating sociopolitical system, lies at the heart of the politics of our time. It at once animates antagonisms and occasions a redemptive promise of re-capturing the “meta” in the narrative, by grounding and fixing the flow of information in the (digital) public sphere (Kalpokas, 2019). It is in light of this observation that post-Marxism’s emphasis on the power of rhetoric allows us to move beyond postmodernist theory, which correctly identifies the onset of social fragmentation as a key feature of our time.

For post-Marxists, hegemony involves a re-constitution of the chain of signification at the backdrop of social fragmentation through multi-tier rhetorical operations that consolidate diverse subject positions (Laclau, 2014; Laclau and Mouffe, 2001). In other words, hegemonic politics involve a movement towards “full-constitution,” at a time when the structurality of the system is in flux—it is, in a sense, a battle of space against time. The figurative nature of rhetoric, which operates in excess of conceptual specificity, is what enables the symbolic consolidation of “social fragments” vis-à-vis symbolic attributions of equivalence and exclusions. Consider, as an example, how the elusive notion of “MAGA” has effectuated unprecedented political alliances in the digital public sphere, by incubating multiple frontiers of exclusion, including “cultural Marxism,” “cancel culture,” “woke enemy,” “globalists,” “bureaucrats,” etc. As a central signifier in the run-up to the 2024 USA elections, “MAGA” was adopted and politically deployed by a diversity of factions, including economic nationalists and neoliberals, conservatives and techno-accelerationists, isolationists and neoconservatives, neo-Nazis and pro-Israel Christians, etc. The key thing to consider here is that while these groups often displayed divergent political and economic interests, they nonetheless contingently coalesced around common “enemies,” thus occasioning a formidable political tide that facilitated Trump’s re-election. Looking at the particularities of the coalition, we can cite how ideologically divergent actors, if not outright enemies, such as Stephen Bannon and Elon Musk, facilitated, in tandem, the MAGA ascent, only to find themselves at odds following Trump’s re-election (Bannon had explicitly disparaged Silicon Valley elites as “technofeudalists,” “atheistic 11-year old boys,” and “Lefties,” who “don’t give a flying fuck about the human being”).

The precarious MAGA alignment would only reveal its internal vicissitudes in the aftermath of Trump’s re-election, in what has been characterized as MAGA’s internal rift, which registered notable ideological divergences, particularly on the issues of Israel and foreign intervention. As such, “MAGA” is currently undergoing an “internal” process of contestation that is directly centered on the question of authenticity, that is, on who the authentic American is, what true American interests consist of, and who are the true representatives of MAGA. This rift was manifested in the most pronounced way within the Heritage Foundation itself, after its president posted a video in defense of Tucker Carlson’s interview with neo-Nazi influencer Nick Fuentes (Daniels, 2025).

The prominence of the politics of rhetoric are, as identified, symptomatic of current socio-political coordinates, but they also spiral down to ontological precepts. As Laclau (2014: 90–99) has thoroughly demonstrated, rhetorical figuration, particularly in the form of synecdoche, inheres in the very structuration of communal totalities by occasioning a semiotic dualism: That of signifying a “constitutive outside” to which the community is symbolically counterposed and that of establishing a symbolic chain of equivalence between a community’s internal elements, for example, identities, narratives, political demands, and so forth. As an inherent property of discourse, the totalization of a communal totality through rhetorical operations, invariably accompanies, as a potential, political practice, in general.

This helps explain the recent decisive thrust of reactionary and neoreactionary movements that seem to be capitalizing from, and thriving within, the digital public sphere. The increasing political ascent of identitarian and fundamentalist movements, in their various instantiations, can be read as an attempted buffer against the onslaught of postmodern relativism, which has effectively dislodged modernity’s principles of unity through processes of social fragmentation. Purporting the failures of postmodernity, these movements seek grounding in foundational and totalizing narratives, which they effectuate through highly formalized and affect-ridden discursive mediums that assume the form of antagonistic politics. As a backlash against epistemic particularism, these offensives seem to be oriented towards the restoration of a social order through a rhetorical “fixing” of communal limits, by means of exclusions. It is, therefore, by no surprise that neo-Far-Right gurus, in the likes of Jordan Peterson, Ben Shapiro, and Aleksandr Dugin have advanced a decisive polemic explicitly directed against postmodern relativism.

These empirical manifestations run counter to Lyotard’s (1984: 41) claim that “[m]ost people have lost the nostalgia for the lost narrative” and that “[w]hat saves them from it is their knowledge that legitimation can only spring from their own linguistic practice and communicational interaction.” Rather, what they indicate is the emergence of a novel pole of political antagonism, which involves, on the one end, a sliding towards relativism, and, on the other, a neotraditionalism that seeks to restore foundationality. The political battles of our time, whether in the form of “culture wars,” “identity politics,” etc. seem to express tendential positions within the relativism-foundationalism spectrum—positions that very often materialize in the form of highly formalized rhetorical antagonisms.

Towards a renewed democratic impetus

The above-noted deductions, which in our estimation peer into a hitherto unexplored dimension of contemporary social and political developments, represent a decisive step towards new analytic and political possibilities. Existing literature has correctly diagnosed how the operational characteristics of AI systems facilitate capital accumulation, as well as political and military imperatives, but it has fallen short of comprehensively examining AI’s dislocating effects on political and cultural communication, which, as we have demonstrated, involve identifiable politico-discursive logics.

Our estimation is that this shortcoming stems from underlying theoretical proclivities, which tend to reduce the parameters of analysis to dominant epistemic norms, namely, the power-resistance dichotomy. Consequently, empirical investigations tend to be channeled through the broader question of whether social media technologies facilitate power centralization, or augment modalities of resistance, while aprioristically excluding alternative analytic paths that may afford dialectical readings, by considering how modalities of both power and resistance can manifest through contrapuntal logics that nonetheless empirically cohere with existing hegemonic coordinates.

In our day and age, politics seem to feed off the second order effects of AI-powered social media systems, which are increasingly facilitating antagonistic politics. This configuration, at the same time, occasions a cultural backlash that seeks to restore the order of foundationalism, very often in extremist terms, against the backdrop of an increasingly fragmented social. This intertextuality opens up new analytic and normative questions, enabling a higher order problematization of the socio-political impacts of technology and platform capitalism, in particular. Power, in the age of AI, is not actualized solely through the operational capacities of automated digital systems (i.e., information control), but in the dislocating effects they exercise.

The question of whether social media occasion or undermine democratic potentials should be evaluated against the backdrop of these considerations, for what is essentially unfolding are two parallel but interconnected processes. On the one hand, power, in the form of information control, is being centralized within particular social locations, namely, corporate and state entities. Access to these spaces is limited and relegated to the hand of “the few” who wield command over information systems vis-à-vis economic, political and network power (Burrell and Fourcade, 2021; Castells, 2009; Joyce et al., 2021). Democratic potential within this realm is delimited, as it is neither accessible to the citizen-public, nor subject to robust democratic controls. Its rectification as a potential “arm” of democracy requires radical political interventions that would essentially dislodge its underlying logics (e.g., capital accumulation, surveillance, cloud rentierism), gearing it towards the direction of a “democratic control,” highly responsive to social needs/demands (Muldoon, 2022; Varoufakis, 2023).

On the other hand, power, in the form of discursive efficacy, is increasingly becoming decentralized, vis-à-vis the second order effects of social media platforms. Comparatively speaking, the digital public sphere provides greater access to the citizen-public, as well as the capacity to amass discursive power through the creative use of language and representations. ² However, mere access to the digital public sphere does not necessarily yield democratic effects. Rather—and as we have demonstrated—it leads to the pluralization of the field of conflict, which is increasingly ensnared in antagonistic logics.

Putting two and two together, what this configuration amounts to in its totality is a tightly controlled digital infrastructure that facilitates social dissensus and multiple processes of “othering,” in the absence of any sort of democratic safeguards. In this context, the democratic potential becomes encased within the logic of an increasingly pluralistic field of conflict in the form of antagonistic politics. It is, therefore, not surprising that a plethora of recent works that situate the democratic potential within the logic of antagonistic politics coincide, temporally, with the increasing prominence of social media technologies (e.g., Mouffe, 2018; Stavrakakis, 2014). The former is merely a symptom of the latter. By reconfiguring the spatiotemporal logic of symbolic systems, social media have unleashed and, indeed, made visible, the potent political thrust of antagonisms. The “veneer,” which has for so long facilitated the illusion of a democratic consensus, has been removed.

Given the above, it becomes clear that a formidable democratic offensive can only materialize vis-à-vis an operational logic that is immanent to the coordinates of the systems at play. The challenge, then, consists in fostering a democratic chain of alliance between digital collectivities, through the effective deployment of rhetoric that, through its very polysemy, through its very formalism, can tendentially unite disparate identities with overlapping political orientations. Floating and empty signifiers, rhetoric prose, memetic formulations, etc. can indeed potentiate such alignments precisely because they can provide, through their very polysemy, a common principle of exclusion, be it Far-Right politics, economic injustices, the dangers of technological accelerationism, environmental degradation, and so on.

By implication, signifiers endowed with high levels of polysemy will have a privileged role to play in today’s politics. This is why recent works have highlighted the political efficacy of evocative signifiers such as “the people,” which are endowed with a dimension of “universality”—”the people” can reference both the communal totality and plural (excluded) identity groups (Stavrakakis, 2014; Stavrakakis and Katsambekis, 2014). Calls for a democratic populism, as an example, tie the democratic potential to the broader political project of how “the people” is constructed (Mouffe, 2022).

However, this line of scholarship has fallen short of considering a more comprehensive political strategy that could potentiate a democratic offensive through and beyond populist politics. It is our estimation, given the fragmented character of the digital public sphere, that the actualization of a broad democratic alliance presupposes the production of multi-tier political vocabularies that, by fostering chains of symbolic equivalence between disparate digital collectivities, could potentiate a multifarious offensive against non-democratic forces in the digital public sphere. This would involve a discursive strategy that relies on more than just the deployment of empty signifiers (e.g., “the people”), through the concomitant deployment of diverse rhetorical devices, including analogy, irony, metaphor, hyperbole, etc. In other words, there needs to be a rhetorical “fixing” of the flow of differences in the (digital) public sphere, in the service of democracy, where “seizing the memes of production” (Bown and Bristow, 2019) would assume a key strategic task.

A cursory glance at Far-Right digital spaces—and particularly their accelerationist variants—readily reveals that these communicative logics are already being deployed by opponents of democracy, in a remarkably successful fashion. Today, Far-Right logics are being mainstreamed through diverse rhetorical mediums that make them palatable to broad audiences vis-à-vis a variety of pipelines, including cartoons (e.g., Pepe the frog), anime/manga (e.g., Attack on Titan), films (e.g., profuse references to “redpilling”), cuteness (e.g., cat memes), health and wellness (“granola Nazism”), mythology (e.g., pagan and ancient Greek aesthetics), literature (e.g., HP Lovercraft), and so on (Darwish, 2025; Pinto, 2019; Tebaldi, 2023). On the flipside of the aisle, one can similarly observe effectively deployments of rhetoric, though far more direct in character, that have been instrumental in scaling and consolidating democratic struggles, including “me too,” “Black lives matter,” “I can’t breathe,” “say her name,” “the 1%,” “no kings,” etc. However, there is a qualitative difference in how rhetoric is employed by these opposing “camps.” Whereas in the case of democratic actors one observes rhetorical deployments that are poignant but issue-specific, in the case of the Far-Right, the message is insidiously enveloped in rhetorical mediums that are decisively more diverse in character and very often masked in politically-neutral aesthetics (see Pinto, 2019).

And while we do not claim that one style of communication is more effective than the other, not least more “correct,” our observations do allow us to consider pertinent questions related to the political efficacy of rhetoric. The crux of the matter is that the efficacy of rhetorical devices, whether in the form of empty signifiers, metaphor, analogy, irony, etc. will be commensurate with their capacity to consolidate heterogeneous identities, (digital) collectivities, political struggles, etc. Consider, as a most relevant example, how the notion of “stereotyping” can be equally deployed against racist, sexist, homophobic, ableist, etc., assaults. By invoking a fundamental modality of discrimination that inheres in a variety of “isms,” the notion of “stereotyping” potentiates a “rhetorical symmetry” between otherwise disparate fields of conflict, by means of analogy. In so doing, it enables a consolidating effect that identifies discrimination, as such, as a common “agent” of negation. The notion of “stereotyping” is to this day a most-potent political tool that is, on an on-going basis, enabling democratic advances, by symbolically consolidating diverse democratic struggles, including those of a burgeoning character, for example, trans rights, neurodiversity movement, anti-weight discrimination, etc.

These deductions are part and parcel of a broader theoretico-empirical question concerning the identity of the “agents of social change”—who would the democratic chain of alliance include? In our estimation, the answer can in no way be prescriptive, as this would delimit the breadth of possible alliances through an affirmationist logic (Noys, 2010) that harks back to essentialist conceptions of identity (Laclau and Mouffe, 2001). A democratic chain of alliance will inevitably include diverse subjectivities, which display a diversity of demands, thus precluding the possibility of a pure and unadulterated political alignment. The key political challenge, therefore, consists in rhetorically framing democratic demands in a manner that can “speak to” diverse political subjectivities.

Rather than proceeding from an affirmationist logic that “names” the agents of change, we believe it is more effective, from a political standpoint, to proceed from a politics of negation that names the “enemy,” in the diversity of its instantiations, while at the same time redirecting this line of exclusion towards a common vocabulary and programmatic positions that activate radical conceptions of equality and justice—a horizon of possibility animated by a desire to curtail accumulations of power. While the politics of negation can incubate affective charges by “naming” the sources of citizen grievance, as a necessary component of political change, the “logic of redirection” can imbue these affective charges with content.

Lessons from an adversary are worthwhile considering. Stephen Bannon, who is very often cited as one of the masterminds behind Trump’s political ascent, follows a very precise and recurring argumentative logic. He proceeds from a polemic that names the sources of citizen grievance (“neoliberalism,” “corporate cronies,” “Liberal elites,” “technofeudalists”)—typically by appropriating the discourses of the Left. He then redirects this “calling out” towards discourses that are markedly conservative and nationalist, for example, wage suppression is facilitated by a corporate class of Left elites through an open-border policy that has, as its terminal point of arrival, the undermining of the national working class.

These deductions alarm us to the fact that a broad democratic alliance needs to animate a fine line between a rhetorical formalism that can potentiate affective antagonistic frontiers by capturing a diversity of politico-economic demands, and a propositional specificity that remains grounded in the material conditions that produce such demands in the first place. In this sense, the articulation of a democratic chain of alliance will necessitate a combination of both rhetorical and rationalist discourse. The latter will explicate the mechanisms that produce grievances, while the former will incorporate them as elements of an emergent political body by constructing an agent of negation, that is, “an enemy,” as personified by the very agents and structures of power.

As a potentially formidable political strategy, our proposal for a democratic chain of alliance displays a marked sensibility to the sociopolitical conditions of our time, being, in essence, a deductive strategy that proceeds from our interrogation of AI-powered social media systems. Given current structural limits, the democratic chain of alliance can potentially subvert existing hegemonic coordinates because it operates immanently to their logic—it can confront and circumvent the structural interests of corporate actors by instrumentalizing the communicative logics (i.e., affective polarization/antagonism) upon which their business models depend.

Our deductions align closely with Kriess and McGregor’s (2024: 567), which see polarization, not as a problem in itself, but as a form of politics that can potentiate egalitarian claims, to the extend, of course, that the dynamics of polarization remain sensible to the ideational and material underpinnings of citizen grievances: “From a normative democratic theoretical perspective [. . .] it is a mistake to elevate polarization as the primary concern as opposed to underlying social and political inequality.” Social media technologies have proven time and time again to be effective vehicles for advancing democratic claims, not least through highly polarizing antagonisms. Invoking the example of the Black Lives Matter movement, Kriess and McGregor (2024: 570) note that “Social media—especially the use of the hashtag #BlackLivesMatter—allowed the movement to scale rapidly, as well as to gain significant mass media attention and disrupt counter-narratives.” Relatedly, empirical measures that track shifts in attitudes after the Black Lives Matter movement demonstrate that polarization on issues of racial-economic justice is fueled by shifts in, specifically, White Americans’ attitudes, which are increasingly aligning with those of Black Americans (Kriess and McGregor, 2024: 570). In this context, polarization can be interpreted, at least in part, as “the arc of morality bending towards justice.”

Conclusion

Our article has offered some preliminary considerations on how, in the age of AI-powered social media, a digital democratic strategy can be advanced. While we do not maintain that our analysis is exhaustive, we do believe it brings to the fore pertinent theoretical and empirical considerations that can inform, not only future research, but political practice, more broadly. Democratic theory has, over the past 30 years, offered a plethora of possible solutions to the problem of (non)democracy—solutions that have been framed in terms of participation (Fraser, 2005), deliberative democracy (Benhabib, 1996; Habermas, 1996), communication (Castells, 2009), agonism (Mouffe, 2005), recognition (Butler, 2004; Fraser, 2005; Melucci, 1996), redistribution (Fraser, 2005), subjectivity (Touraine, 2009), social inclusion (Giddens, 1998), cosmopolitanism (Beck and Levy, 2013; Habermas, 2003), and so on. However, what is often undertheorized in these otherwise laudable contributions is the dimension of discursive power as a potential incubator of democracy (Anastasiou, 2022: 200–212). In our estimation, this stems from entrenched theoretical proclivities, which tend to understand democracy as the negation of power—and for understandable reasons.

However, if we are to understand power as something that precipitates from the configuration of symbolic systems, then democracy can and should be understood as a hegemonic project to be—as a potential germinating in the indeterminate cleavages of the social, that is, as an immanent possibility yet to be born. Given this deduction, it becomes clear that a formidable democratic offensive presupposes the steering of indeterminate possibilities towards a democratic horizon—a process that will inevitably hinge on the politics of antagonism, given current spatiotemporal coordinates.

A rhetorical “fixing” in the (digital) public sphere that is at once sensible to the realities of social pluralism and material dispossession could operate as a potential weapon against reactionary foundationalism by fixing the chain of signification in the service of social justice and egalitarianism, under the broader vision of negating power differentials, without recourse to absolute and prescriptive narratives that would delimit the field of sociopolitical participation. In this sense, a formidable democratic offensive requires careful tip-toeing between the tenets of foundationalism and relativism. The challenge here consists in articulating frameworks of community and belonging that effectively subvert power differentials, without recourse to absolute symbolic closures that would enable modalities of exclusion.