Can social systems theory be used for immanent critique?

Untangling claims to reflexivity: Critical theory and systems theory

Nearly every discussion of critical theory begins with an appeal to one of its most basic commitments: a critical account of society needs to be able to account for its own normative viewpoint or foundations. The commitment is meant to distinguish the situated and engaged perspective of critical theory from the ‘bird’s eye view’ of ideal theories, which adopt a specific principle, procedure, or value, in order to formally model social phenomena. For example, economists might rely on decision theory to model social behaviour: they appeal to an abstract notion of (expected) utility or preference to model and evaluate the behaviour of rational agents within the target domain. There are two things to note about this approach. First, the resources required to model the behaviour of actors within the target domain are significantly richer than the ones attributed to the modelled actors. Second, neither the model nor its founding principle need to take into account how things appear to social agents in the target domain: rational agents need not acknowledge or embrace (explicitly or not) the notion of utility attributed to them, nor need they account for their own decisions or behaviours in decision-theoretical terms.

The robustness of the decision-theoretical model, it would seem, is quite independent from its target domain. Indeed, in assessing such a model, one usually appeals to aspects like internal coherence and some set of meta-theoretical virtues it is held to exemplify (e.g. explanatory or predictive power, simplicity, generality, etc.). Assuming an ideal theory is internally consistent, and embodies some meta-theoretical virtues, theorists may then move from merely describing the intended target domain to normatively assessing or criticizing the social phenomena falling within its scope by determining how well these empirical instances of social interaction fit or satisfy the model. Such an assessment, however, proceeds without demonstrating that its starting point is the salient one to adopt in this instance.

From its inception, critical theory has rejected this ‘traditional’ approach to social theorizing, ³ while making sense of its own operations by introducing a reflexive relationship between theory and target domain (historically, in terms of praxis). ⁴ As initially envisioned by Horkheimer, the appeal to a reflexive relationship between theory and its target domain involves reconceiving what a successful theory looks like and does. In the first instance, Horkheimer endorsed the Marxian rejection of ‘ideal’ principles that purport to transcend the phenomenon they describe. Instead, he insisted that we try to extrapolate the salient rules or laws from the target phenomenon itself – i.e. commitments or values that actors within the target domain could in principle recognize and accept. Such an approach guarantees a non-ideal theory, since it does not presuppose that the extrapolated principles are ‘maximally rational’, nor does it posit idealizing conditions (e.g. methodological individualism) to simplify the modelling procedure. It also helps ensure that the theory is appropriately related to its target domain and that the criteria used to make intelligible and assess this domain are in fact endogenous to it. Unlike a descriptive decision theorist, a non-ideal decision theorist would have to extrapolate the preferences or interests, and the mode of deliberating on them, from the target domain such that rational agents could recognize them as their own.

Extrapolating the commitments, values, and modes of deliberation from a target domain, such that they could be recognized and embraced by agents within it, yields critical theory’s second difference from ideal theory. Specifically, by identifying the normative commitments that social agents actually embrace, critical theorists equip themselves (or so the argument goes) with a motivational spring to practical action: if the evaluative criteria of social theorizing are shared by actors who inhabit the social spaces being described and criticized, then diagnosing a social failure ought to motivate actors to remedy it. Finally, appealing to ‘reflexivity’ allows theorists to adjudicate between competing models and contextualize their work, since the salient criteria may not generalize beyond a very specific historical situation.

In sum, a reflexive relationship between target domain and theorizing promises to enable theoretical description while motivating situated actors who implicitly accept the same criteria to respond to the issues identified by the theory. Reflexivity for critical theory, therefore, turns on satisfying three desiderata: ⁵

Offering an account of theoretically possible social transformation, which involves:

identifying a form of social failure (e.g. a structural deformation in the communicative coordination of action; misrecognition; or false [class] consciousness related to misidentification of interests);

Although it is not clear that all three of these aims have ever been achieved by any one individual critical theorist, they nevertheless give us a good grasp of the differences between critical theory and more familiar approaches to ideal social theorizing. The first condition grounds a utopian orientation such that the basic structure of a given social organization, in a specific historical situation, cannot be taken for granted. This insistence to contextualize proposals for social transformation precludes technological utopias and ‘utopias of desire’ (Agnes Heller) ⁶ by underscoring that any transformation must respond to a concrete social failure with the means actually available. The second condition identifies critical theory’s motivational and normative core: the impetus for social transformation is the rational unacceptability of unnecessary suffering (where ‘rational unacceptability’ is to be spelled out in terms of the beliefs, practices, values, etc. that situated agents embrace). This aim ensures that theorizing does not proceed from idealizing suppositions that could not be anchored in a target domain. Finally, the third modulates a Marxist commitment concerning the historical subject of change such that social transformations are not imposed from on high by a political vanguard who know better than the hoi polloi, but are freely undertaken by the people who find themselves subject to the harms identified by the critical theory.

Critical theory’s desiderata and insistence on reflexivity resemble some of the basic commitments of Luhmann’s social systems theory, while diverging from it in several appreciable ways. Like critical social theory, social systems theory insists on ‘reflexivity’, though not necessarily under this moniker. Indeed, where Luhmann explicitly invokes the notion of reflexivity, for instance, we usually find the beginning of a metatheory. ⁷ As Krause puts it in his Luhmann Lexikon, ‘Reflexive mechanisms […] consist in reducing and increasing complexity by turning a process onto itself: love of love (passion), learning to learn (compliance [Lernfähigkeit]), […] researching research (methodology), deciding on decisions (planning), valuing values (ideology), thinking about thinking (reflection)’ (Krause, 2005: 216). Although this may sound similar to critical theory’s efforts to account for its own normative foundations, systems theoretical reflexivity involves a fundamental transformation of an initial process to yield a new object domain. It would be a mistake, after all, to think that love and passion are the same thing, or that one’s research and one’s method are identical. On a Luhmannian analysis, then, reflexivity is transformative. It involves a second-order observation of a first-order process that constitutes a new system. Although critical social theory does involve second-order observation (e.g. Habermas’s ‘transcendence from within’), critical theorists do not take themselves to be in the business of offering a new ideology, or merely developing a new method for engaging in social scientific research (that would be tantamount to ideal-theory construction).

A better way to understand critical theory’s efforts to account for its own normative foundations would be to introduce the systems theoretical notions of operational closure and structural coupling. We can characterize ‘operational closure’ in broad terms as the manner in which ‘the operations of a system make possible the operations of that system’ (Krause, 2005: 219). Despite its tautologous feel, this characterization captures the salient idea: an operationally closed system distinguishes itself from its environment via a finite set of processes or functions that reproduce the system as a whole while remaining open to environmental resources or ‘inputs’. Those familiar with first-order logics will recognize the basic idea at work, which concerns the completeness of a logical system. That is, a logical system is said to be complete when every true statement expressed in the logic’s finite and specifiable resources (e.g. logical operators like material implication and negation, symbols to represent atomic syntactic units, rules for concatenating symbols via logical operators, and rules of inference) can be proved within it. ⁸ As Luhmann puts the point, an operationally closed system ‘uses the same type of operation […] in the dual function of (1) producing subsequent operations and (2) confirming or changing the structure used to select the next operation’ (Luhmann, 1992: 1440). Operational closure thus expresses the same idea as completeness: a finite set of operations defines a system and produces all possible states of the system, while differentiating it thereby from the resources of its environment. Luhmann’s operationally closed system and a formal logic’s completeness thus also remain ‘open’ to their specific environments in the sense that they continue to draw upon and formally reorganize the basic materials in their environment, despite their very differentiation from this environment. To continue the analogy, one might say that natural language (or indeed whatever metalanguage one uses) is the environment for a logical system.

The sense of reflexivity captured by operational closure comes closer to satisfying critical theory’s demand that a theory account for its normative foundations. For, much like the formal-logical demand for completeness or the systems theoretical account of operational closure, critical theory’s insistence that practitioners account for their own norms and commitments turns out to be a request to demonstrate the completeness of the social theory on the basis of which one aims to make concrete, situated diagnoses concerning social suffering. It ensures that the resources attributed to actors within the target domain are as rich as the ones used to formulate the model. If this is the right way to parse the appeal to reflexivity, then there might be more common ground between social systems theory and critical theory than is usually acknowledged. Admittedly, ‘operational closure’ alone is insufficient to bring critical theory and systems theory (closer) together. For, on the one hand, we still require a better sense of how a critical theory can make sense of its own basic operations and affect change (i.e. to be praxis in some sense). From a systems theoretical perspective, on the other hand, operational closure alone is not fully intelligible without the further notion of a system’s (structural) coupling with other systems.

As Hans Georg Moeller characterizes it, two systems are structurally coupled when each one ‘shape[s] the environment of the other in such a way that both depend on the other for continuing autopoiesis and increasing structural complexity’ (2006, 19). Those familiar with Habermas’s work will recognize in Moeller’s description something like Habermas’s own thesis of the rationalization of the lifeworld (Habermas, 1984, 1: 145ff.). Although supporters of Luhmann’s social systems theory tend to construe Habermas’s appeal to the lifeworld as a vestigial feature of German Idealism that theorists need to unburden themselves of, ⁹ the notion functions in much the same way as Luhmann’s own ‘environment’. That is, it furnishes specific systems – economy, polity, culture, and societal community, on Habermas’s Parsonian model – with a context and the resources for their own differentiation and continued development. Habermas discusses this developmental process of systemic differentiation culminating in operational closure as a process of rationalization, wherein the operations proper to a system are distinguished from environmentally contiguous processes. Presumably, one could also describe this process more neutrally in terms of mere functional differentiation. Regardless of description, Habermas remains committed to the idea that systems achieve operational closure by differentiating themselves from their environments. Moreover, he is also committed to the idea that this process shapes the environment for other systems as well. Therefore, to the extent that Habermas introduces analogues for ‘environment’, ‘operational closure’ and ‘structural coupling’, we may be justified in claiming that the critical theorist’s efforts to make sense of her own normative foundations involves a form of reflexivity shared by social systems theory. ¹⁰

The specific sense of reflexivity just outlined seems to be about the only thing systems theory and critical theory explicitly share, however. Indeed, a quick glance over the desiderata outlined above should be enough to indicate how critical theory and systems theory diverge. In the first instance, critical theory’s insistence on social actors is, by and large, alien to systems theory. A social actor is merely another functional element within a system. Similarly, any appeal to a utopian dimension within social theorizing simply strikes the systems theorist as an unbearable flight of fancy. In his own work, Luhmann parses this difference in theoretical orientation by insisting that systems theory is affirmative, rather than critical. ¹¹

The difference in orientation, Luhmann argues in several places, ¹² hinges on an ontological difference: although critical- and systems theorists may both be realists of some stripe, Luhmann insists that critical theorists are committed to a form of epistemological realism in which a performance or behaviour tokens a determinate type. The type of practice informing a particular engagement, moreover, can be extrapolated from its tokening performance, and known. In effect, then, critical theorists are realists about social practices. ¹³ They treat social practices as determinate and independent of their theoretical descriptions (i.e. one’s description of a practice does not interact with or change the practice itself), although they remain changeable (by other means). Critical theorists tend to embrace a commitment to epistemologically real kinds because they make a specific form of criticism possible. Namely, integrated kinds may be used to assess and criticize specific instances or performances, since they allow us to assess how well the instance fits its kind. Furthermore, since the kinds themselves are contingent, one may argue that, although instances of a specific kind fail to achieve the requisite fit, the instantiated performance nevertheless possesses features that should prompt us to change the kind.

An affirmative theory like social systems theory, however, is constructivist in orientation: descriptive observations and social practices do interact to yield new social arrangements. Notice that such a claim is not epistemological, but ontological. The different kinds of realism embraced by social systems theory and critical social theory turn out to provide us with a sufficient condition for differentiating them. For, although critical theorists and systems theorists are all realists in some sense, their respective realisms imply distinct cognitive orientations to their (shared) target domains. Critical theorists adopt a ‘practical’ orientation, whereas systems theorists adopt a ‘theoretical’ one. ¹⁴ The difference in orientation is perhaps best understood in terms of what a successful theory does. A theoretical orientation prompts one to understand and explain its target. From this theoretical point of view, a ‘successful’ theory instantiates some subset of the traditional metatheoretical virtues one finds in scientific theories; theoretical orientations thus produce ideal theories. On the other hand, a practical orientation does not merely aim to explain its target, but to change it. Indeed, as we noted above, critical theory explicitly rejects a purely theoretical orientation to social life in favour of practical interventions in and contributions to the development of its target domain.

The distinct cognitive orientations help us understand the different realist commitments social systems theorists and critical theorists take on. Since critical theorists aim to articulate a theory of social interaction where actors can recognize themselves in the theory, recognize a kind of social suffering, and use the theory’s technical resources to intervene in and change their situations, theorists need an epistemically stable feature to coordinate their goals. Epistemological realism thus becomes a necessary commitment. From this practical perspective, however, more robust forms of realism are unnecessary – and indeed potentially unhelpful, since they make it increasingly difficult for a critical theory to satisfy its desiderata. Systems theorists, on the other hand, do not prioritize practical intervention, nor do they insist that social actors recognize themselves within the theory. Furthermore, since systems theory aims to understand phenomena at a maximum level of generality, the kinds of commitments theorists embrace will be markedly different. Instead of requiring an epistemically invariant kind to coordinate various practical commitments, systems theorists require ontological commitments that can accommodate a robust notion of emergence. Given these needs, it is natural for systems theorists to reject the kind of epistemological realism at the heart of critical theory, since they place a higher theoretical value on understanding the emergence and functional patterns of interaction that characterize a system.

The requirements of immanent critique

It would seem, then, that critical theory and social systems theory do differ at a rather fundamental level. Whereas a systems theorist comes armed with a toolbox full of ideal notions that she uses to schematize objects of interest and impute operations which render the observed phenomena intelligible, a critical theorist hopes to extrapolate and reconstruct the operations that social actors could (in principle) recognize as their own with the expressed intention of addressing these actors. The systems theorist thus seems to proceed in much the same way that our decision theorist did in the example above. ¹⁵ The latter brought her notions of utility, methodological individualism, and rational choice to bear on a given social behaviour to assess its rationality. Mutatis mutandis, the systems theorist brings her rich conceptual repertoire to bear on a given social phenomenon in order to model it by appeal to the autopoiesis of variously related systems. Simply put, what the systems theorist observes need not be recognized by anything or anyone within the observed system(s).

Despite the fundamental difference between these approaches to social theorizing, we may still be able to bridge them. By parsing the distinction between ideal and non-ideal theorization in terms of different cognitive orientations and metatheoretical commitments, we have opened the possibility that the two approaches might be compatible at a methodological level. In other words, that the methods of one might be implemented as part of the practices involved in the other. Thus, even though an ideal theory cannot satisfy the demands of critical theory, it may provide a helpful toolbox to build it up. Indeed, as I showed earlier, Habermas’s own theory of communicative action appeals to Parsonian systems theory in precisely this way. It might therefore be a very small step to introduce Luhmannian systems theory in a similar way. While conceding Habermas’s early contention that systems theory is a technology one can use to describe ‘society’ rather than a true theory of society, we may still be able to formulate a critical theory that draws on the resources of social systems theory.

To examine the feasibility of such a step, let us turn to one of critical theory’s most distinctive procedures, namely immanent critique. For, as I will show shortly, the desiderata of critical theory commit theorists to the concept of immanent critique, but ‘immanent critique’ does not entail a critical social theory. To begin, let us consider a recent definition of this procedure:

Immanent critique is a form of social critique that evaluates both the empirical behaviour constituting social practices and the explicit self-understanding of their members according to standards that are, in some sense, internal to those practices themselves. By doing so, immanent critique aims at a transformation of such practices that encompasses both actions and self-understandings. (Stahl, 2013a: 7)

Stahl’s characterization highlights a methodological relationship between the desiderata of critical theory and the unfolding of immanent criticism. As a strategy, immanent criticism allows us to identify a theoretically possible transformation that is practically necessary, and address socially embedded actors who in principle accept the evaluative standards being used. Immanent critique also promises to motivate individuals to undertake the proposed social transformation, because the base practice and situated actors’ normative understandings come into some kind of conflict. Notice, however, that immanent critique is not synonymous with these demands. As a method or strategy, immanent critique identifies or makes available the resources for a successful critical theory without being identical to it. Most simply put, as a means, immanent critique remains distinct from the ends of a critical theory. We can therefore ask what characterizes immanent critique as a means or method of analysis in its own right. On my view, ¹⁶ ‘immanent critique’ needs to meet three requirements:

An Inherence Requirement: normative content needs to be internal to or implicit in a practice such that the base practice and its normative content are connected in a manner that is necessary yet revisable. ¹⁷

The three requirements allow us to differentiate between the demands of critical theory and the methodological considerations involved in trying to satisfy them. If an effort at social criticism meets these requirements, it has a good chance at satisfying the demands of critical social theory, while nevertheless remaining distinct from it (in much the same way that the means remain distinct from the ends they achieve). Characterizing ‘immanent critique’ in terms of these requirements has an interesting consequence: the manner in which one meets these requirements remains a matter of relative freedom. Hence ‘immanent critique’ does not denote a strict method and is open to various disciplinary or technical approaches. Systems theory may therefore provide us with the salient tools to engage in immanent critique and hence contribute to a critical social theory, despite the difference in cognitive orientation identified earlier.

To see if the tools provided by systems theory might meet the requirements, let me consider each requirement individually, beginning with the inherence requirement. In broad terms, this condition involves securing the right kind of relationship between the evaluative or normative criteria for our critical enterprise and the base practices that are the object of our interest. We can get a sense of what’s involved here by contrasting criteria that ‘inhere’ in a situation or performance with those that are extrinsic to it, or remain otherwise external to the case at hand. For example, every Olympic level diver will perform so as to instantiate the judgeable properties of diving (i.e. starting position, approach, take-off, flight, and entry) without necessarily fulfilling the standards of excellence (i.e. receiving a 10). The criteria of assessment – and the degree of fit between an individual instance and the ideal specified by the criteria – are said to structure the performance and hence are internal to it. Without them, the performance itself would be unintelligible.

Prima facie, social systems theory can satisfy this requirement. Indeed, a virtue of the systems theoretical perspective is that it offers an elegant account of ‘inherence’ in terms of information communicated (indirectly) within a system. On this view, the communicative system of Olympic diving differentiates divers, audience members, and judges according to their respective functions, such that the ‘normative content’ just is the information transmitted among subsystems, which makes the differentiation of these roles possible (or intelligible) in the first place: the diver’s physical performance announces his ability according to the functional specification of the sport; however, the information the diver hopes to convey, and what the judges gather from his performance need not be identical, since the formal features of the performance could have been, say, more efficiently conveyed. In this instance, the ‘signal-to-noise’ ratio was sub-optimal. The difference between the diver’s communicative performance, the information conveyed, and the information understood provides us with a way of making sense of how evaluative criteria can be internal to the system in which the performance occurs. Systems theory thus appears well-equipped to meet the inherence requirement.

What about the contradiction requirement? As critical theorists will no doubt recognize immediately, this requirement is supposed to spell out the normative force and authority of social criticism: if one can show that the basic normative commitments which socially situated actors hold contradict (in some sense) the practices they are engaged or enmeshed in, then actors should be motivated to undertake some form of social transformation in order to resolve the tension. Such is the hope, for instance, of Honneth’s early work on recognition, and also Habermas’s account of collective will formation. ¹⁸ In both cases, socially situated actors experience the incompatibility between the accepted norms structuring their engagements and their own first-person experiences.

Now, let it be noted that the kind of incompatibility critical theorists like Honneth or Habermas point to is not strictly formal or logical. Immanent critics don’t need strict contradictions: some kind of cognitive dissonance will do (say, between the rational unacceptability of unnecessary suffering and one’s commitment to a normative principle that can be shown to engender suffering). ¹⁹ Note that the cognitive dissonance characteristic of social contradictions preserves the necessary relationship between base practice and normative criteria and hence meets the inherence requirement, while leaving room for social transformation. We may, in other words, choose to change the practice so that it becomes compatible with the norms we embrace, or we may change the norms we embrace to fit the practice, or we may reject both the norms and practices in favour of something completely new. Furthermore, the cognitive dissonance identified by the immanent critic needs to be pragmatically anchored in specific performances and experienced by concretely situated agents. That is, for immanent criticism to be successful, the cognitive dissonance identified cannot be an artefact of the method of analysis, since that would imply that an extrinsic standard is being applied (thus violating the inherence requirement). The contradiction needs to be something that actors within the target domain can encounter so that their lived experience corroborates the theoretical analysis. Should one fail to anchor a social contradiction within the target domain in this way, one loses much (if not all) of the motivational force and moral authority that the contradiction is supposed to generate. In short, if agents do not recognize the contradiction, they may not accept the analysis or its implications.

On this characterization, systems theory too seems to have the means to account for the contradiction requirement. It just frames this requirement differently, in terms of the paired notions of ‘irritation’ and ‘resonance.’ Specifically, irritation – understood as ‘an as yet undefined surprise [Überraschung] in the context of inter-system relations’ (Krause, 2005: 169) – is said to prompt functional transformations – resonances – in a system not unlike those that critical theorists hope to engender. Like the notion of ‘contradiction’ (cognitive dissonance) that characterizes critical theory, irritation involves a ‘systemic and perceived noise in the environment of an autopoeitic system that could qua structural coupling become informationally relevant (or not), but that is not yet information according to the standards of the autopoeitic system’s own operational code’ (2005: 169). Transforming noise into information requires a shift in the code – and, one could argue, this is precisely the kind of shift that social criticism aims to accomplish. Hence, the effort to turn the perceptible noise within a system into an irritant that prompts systemic changes identifies a point of convergence between critical theory and systems theory. Although the systems theorist may insist that two distinct systems do not directly interact in any causally determinate way, they are nevertheless committed to the idea that the changes one system makes in the environment of another may ‘interfere’ with or ‘perturb’ the operations of a structurally coupled one. This perturbation may then prompt further transformations internal to the perturbed system, so long as a distinctive observational perspective can be introduced to change the way information is handled within the system.

We can illustrate the abstract point by returning to the Olympic diving example. Although not a perfect analogue, it nevertheless clarifies how the irritation in a system introduces specific transformations without causal interaction. Analogically, then, we may consider the practices involved in judging Olympic diving to constitute a specific system, which is coupled to the system constituted by the operations an Olympic diver may select from in her performances. The processes involved in evaluating Olympic diving differentiate judges from the Olympic divers, coaches, audiences, etc. while also shaping the environment in which the divers, judges, audience, etc. operate. Now, a diver’s performance may instantiate some set of operations more or less smoothly, i.e. with a lesser or greater degree of noise. While the noise is perceptible (e.g. the diver’s tuck is not complete, knees are not straight, and entry causes a huge splash), it is only information from the observational perspective of the judges, who treat it as a sign of imperfection in the performance. They use this noise-qua-information to evaluate the performance, which is communicated as a score for the diver’s performance. In turn, the score – which expresses the observed noise-to-signal ratio of the performance – may irritate the diver to perfect her dive. However, how the diver will go about perfecting her performance is not causally determined by the information communicated by her judges (or even coaches).

Noting this convergence between immanent critique and systems theory is helpful, I think, because it goes some way in situating the ambitions of criticism. If immanent criticism is simply the transformation of noise within a system into information that prompts further systemic transformations consistent with the operations that differentiate it from its environment, many of social systems theory’s prominent criticisms of critical theory turn out to be straw men. ²⁰ For this transformation requires the introduction of novel forms of observation based on the operations and resources of the system itself.

The sole criterion systems theory appears to be unable to satisfy is the access requirement. This requirement helps spell out what it means to pragmatically anchor social criticism within its target domain rather than simply rely on extrinsic principles or formal considerations. As already intimated, the failure to anchor social criticism pragmatically would undercut its motivational force, and engender the kinds of sceptical worries that immanent critique is specifically meant to avoid. The problem is a familiar one. Ian Hacking, for instance, gives us a nice illustration of it in The Social Construction of What? when he writes:

We are not surprised to hear that the results of primatology bear strong traces of their discoverers […]: accounts of the behaviour of primates reflect the societies of the scientists who study them. We all know the bad jokes about British apes with stiff upper lips, ruthlessly enterprising American apes, hierarchical and communitarian Japanese apes, promiscuous French apes. (p. 64)

Unsurprisingly, the problem of response dependence has been a recurring issue for critical social theory. It explains why critical theorists find themselves committed to some form of epistemological realism (e.g. Weber’s ideal-type, Adorno’s ‘concept,’ or Habermas’s ideal-speech situation), where an objective, theoretically integrated kind allows theorists to coordinate the various perspectives that social actors may take up within the target domain with the observational perspective that researchers adopt in analysing social phenomena. If a theorist is unable to coordinate her theoretical perspective with the various perspectives actors take within her target domain, she is developing an ideal theory. And, as we saw, the problem with ideal theorizing is that it does not care about addressing social actors such that they can recognize themselves in the model. It also means that any cognitive dissonance discovered between base practice and structuring norms will have less motivational import because it is not pragmatically anchored in the social activities and self-understanding of the agents involved. When it remains bound up in the model, a contradiction does not ‘irritate’ social actors. Worse, it may simply be an artefact of the modelling techniques. In sum, if one cannot satisfy the access requirement, one is constructing an ideal theory. And ideal theorizing is a sufficient condition for identifying a form of social engagement as ‘external critique’.

Interestingly, social systems theory doubles down on the problem of response dependence. Its notion of observation entails that descriptions of social phenomena do not – and cannot – correspond to the phenomena themselves because observation introduces distinctions that are salient to the observational standpoint but not necessarily operative in the observed phenomenon. In effect, observation constitutes the observed phenomenon. Hence, in response to Hacking’s quip concerning the various kinds of apes researchers have encountered, a systems-theorist might justifiably remark that all of those apes do in fact exist because they have been so observed, and the distinctions introduced have had effects within their relevant systems. Furthermore, the systems theorist might continue, the thought that there is one true description of ‘ape’ presupposes precisely the kind of static ontology that systems theory is supposed to free us from.

Conclusion

The discussion of immanent critique has helped identify a genuine wedge issue between critical social theory and social systems theory, which concerns the kinds of realist commitments one is willing to take on and how these commitments shape one’s methodology. Situating this wedge at the level of ontological commitments is nevertheless productive. On the one hand, it implies that the standard ways of making sense of the difference between systems theory and critical theory do not adequately identify what is really different about them. The very practices of social description and analysis employed by systems theory and critical theory commit theorists to different forms of realism (epistemological or ontological). Construing the wedge issue in terms of the metaphysical commitments underwriting the very practices of theorizing shows us why Luhmann’s distinction between ‘affirmative’ and ‘critical’ theory is insufficient. For the latter picks out a difference in cognitive orientation to a target domain but says nothing about the methods one is going to use to model it. We would be better off then saying that systems theory is an ideal-theory and critical social theory a non-ideal one. Reformulating the claim in this way is less misleading since it is ideologically neutral in a way that the language of ‘affirmative’ and ‘critical’ is not. But it also shows why casting the difference between these approaches in this way does not explain much: non-ideal techniques can be put to use in ideal theories just as easily as ideal theories can be incorporated into non-ideal ones. The difference in cognitive orientation does not entail incompatibility.

The consideration of ‘immanent critique’ allowed us to suss out a genuine incompatibility by showing why systems-theoretical analyses cannot be used as part of an immanent social critique and, hence, why systems theory cannot be used – at least not straightforwardly – to achieve the goals of critical theory. In effect, the ontological commitment of systems theory to emergent phenomena (constructivist realism) makes it impossible to satisfy the access requirement. For methodological reasons, Frankfurt School critical theorists cannot embrace such a robust ontological commitment, even though they too find themselves committed to a form of epistemological realism, which in turn social systems theorists have to reject for their own methodologically motivated considerations.

Interestingly, it appears that systems theory has more robust ontological commitments than Frankfurt School critical social theory. For epistemological realism doesn’t entail anything about the way reality is, only about what must be the case for us to know it. To make matters worse, the kind of ontological commitments systems theorists embrace are not ones that can be falsified or verified. They emerge as part of a transcendental argument: the practices of observation and description depend upon the existence of emergent phenomena, which require a robust ontology that can never be corroborated. This kind of indispensability argument, however, is remarkably fragile, since it hinges on the fact that the initial practice is somehow necessary. Yet, the salient kind of necessity has not been established – at least not to my knowledge. If that is so, then systems theory is significantly more fragile a theoretical apparatus than it may have initially appeared. To adapt a remark once made about Spinoza’s Ethics, the whole edifice of systems theory may just be an elephant balancing on one toe.

Let me leave the metaphysical and transcendental subtleties of systems theory aside. My initial aim was to determine whether systems theory could provide theoretical resources for critical theory. The broad answer seems to be this: if the three requirements of immanent critique spell out the means for satisfying critical theory’s desiderata, then these criteria place a decisive limit on the tools we can use. More specifically, social systems theory does not make the cut, because it cannot meet the access requirement – indeed, it explicitly rejects the very idea of access. Once we admit even the smallest possible gap between theoretical model and social reality, however, we generate a cascading set of problems. I have gestured towards a few already, namely, that the accuracy of the model can be called into question and that an accurate modelling doesn’t yield prescriptions (taken here in a broad sense to include both normative claims about what one ought to do, and in the predictive sense of what will happen) for action, or motivational import. These worries yield a further concern that agents may not recognize themselves in the model, and hence will not be motivated by the diagnoses the model produces. Finally, applying unanchored models for social criticism is by definition ‘external critique’. To put the point slightly differently, systems theory might provide the means to explain some phenomena, but does not seem to have – in principle – the resources to allow theorists to intervene in their target domain. And this appears to be a decisive theoretical disadvantage in comparison to other modes of social theorizing.