Rawls—Too Big to Fail?

In what has come to be diagnosed as the ‘crisis of liberalism’ (Kreide 2016), the writings of Judith N. Shklar have been discovered as an alternative approach to liberal theory, and praised by some (Gatta 2018) as a liberal theory of the political able to deal with the political problems of the twenty-first century.

While liberalism is not a monolithic theory and unites different versions under its tenet, all liberal theories share a core premise: namely, the normative primacy of individual liberty, which serves as a basis for concepts of legitimate political authority (Gaus et al. 2018). Despite its theoretical diversity, liberal theorising has come to almost exclusively focus on John Rawls’s A Theory of Justice (Rawls 2003). This monograph has become the quasi-archimedian point of reference of liberal debate since 1971. Here, Rawls develops the concept of a hypothetical original position, in which rational and autonomous subjects choose two fundamental principles of justice for liberal societies (Rawls 2003, § 40). Despite significant changes Rawls made from the Theory of Justice to Political Liberalism, the Kantian concept of autonomy remains crucial for Rawls’s approach (Freeman 2016) and, as a result, for liberal debate as such. The most recent criticisms of Rawlsian theory, realism and non-ideal theory, both maintain the liberal core premise, but put forward a thorough criticism of Rawlsian theory, especially its conceptual and methodological aspects.

Political Conflict

The realist critique of Rawlsian liberalism, most notably put forward by Bernard Williams (2005a, b), argues that contemporary liberalism is incapable of conceptualising genuinely political notions such as power and conflict due to its reliance on a specific conceptualisation of rationality and the resulting subordination of political philosophy to moral philosophy. Williams is particularly concerned with the fact that while Rawls’s Theory of Justice and its modification in Political Liberalism is in defence of a theory capable of dealing with the political sphere, both conceptions are in fact entirely apolitical. Rawls himself highlights the fact that the core of his theory, the original position, is a ‘procedural interpretation of Kant’s conception of autonomy and the Categorical Imperative within an empirical theory’ (2003, p. 226). It is precisely this collapse of political philosophy into moral philosophy that Williams criticises as ‘political moralism’ (2005a, p. 2). In Williams’s terms, political moralism assumes ‘the priority of the moral over the political’ (Williams 2005a, p. 2). In spite of the modifications in Political Liberalism, realist critics argue that political liberalism continues to be based on a Kantian idea: namely the idea that self-legislation as the only legitimation of political authority that is compatible with individual freedom (Sleat 2013). Political liberalism consequently rests on the idea of public justification based on public reason. This results in liberalism’s ‘consensus vision of the political’ (Sleat 2013, p. 14; see also Sagar 2016, p. 380), because political authority needs to be acceptable to each person subject to it. Hypothetical consensus regarding public justification is liberalism’s ‘moral lodestar’ (Macedo 1991, p. 78). Therefore, the underlying consensus regarding the moral conception of the person as well as public reason must be prior to politics itself, and is ‘in an important sense non-political’ (Sleat 2013, p. 39).

For realists, this result begs the question how this version of political liberalism deals with the simple fact that not all actual persons share the moral premise that all persons are free and equal; a premise on which the liberal project essentially relies. The question arises if their rejection of this premise affects the legitimacy of the liberal state, which nonetheless continues to coerce them (Sleat 2013, p. 35). Since public justification demands only that the liberal principle of justification is hypothetically acceptable to reasonable free and equal persons, the standard liberal answer will be negative. Realist theorists are not content with this answer, due to its incapacity to deal with the dimensions of conflict and power which are seen as ineradicable aspects of the political.

In contrast to Rawls, Williams develops his account of ‘political realism’ (Williams 2005a, p. 3); a genuinely political philosophy that operates independently from moral principles: ‘political philosophy is not just applied moral philosophy, […]. In particular, political philosophy must use distinctive political concepts, such as power, and its normative relative, legitimation’ (Williams 2005b, p. 77). Political realists argue that political legitimacy is not derivable from a rationally justifiable moral principle, but consists both of an answer to the ‘first political question of order, protection, safety, trust and the conditions of cooperation’, i.e. the state of nature, and in addition a justification of this answer. Each such ‘answer’—that is, each instance of political authority that aims to resolve the first political question—must be justified (and justifiable) to those subject to it. This is what Williams calls the ‘basic legitimation demand’, or ‘BLD’ (Williams 2005a, p. 3). Williams introduces the additional requirement that the acceptance of the legitimating account cannot itself be the result of force (what he terms the ‘critical theory principle’). However, what makes sense as a legitimation varies depending on the context (it ‘makes sense’ under specific historical circumstances), and is not definable from a purely hypothetical original position (Williams 2005a, p. 10). The purpose of the BLD is to distinguish the ‘problem’, the first political question, i.e. the state of nature, from the ‘solution’, i.e. legitimate political authority (Williams 2005a, p. 5). Merely successful domination is not a ‘political situation’, and, consequently, does not establish legitimate political authority (Williams 2005a, p. 5).

Since political and moral questions follow different functional imperatives, the mere application of moral categories to political questions is inadequate for understanding the political sphere. This is especially the case for the dimension of conflict that is inherent to the political. Political conflict cannot be fully understood within a primarily moral, consensus-based framework. The difference between political and moral conflict relates back to Williams’s rejection of Rawlsian ‘moralism’: Rawls’s theory relies on a Kantian, i.e. rationalist, framework for solving what, according to Williams, are essentially moral questions. This approach cannot be applied to the political, since political and moral opposition differ in two respects: content and scope. First, political conflicts are not conflicts about moral questions or reasons:

political difference is of the essence of politics, and political difference is a relation of political opposition, rather than, in itself, a relation of intellectual or interpretive disagreement. […] the relations of political opposition cannot simply be understood in terms of intellectual error. (Williams 2005a, p. 78)

Political difference or opposition can therefore not sensibly be resolved through a process of ideal reasoning, since other, non-moral considerations like personal interests will play a crucial role in the decision. Therefore, political conflicts cannot always be solved consensually. Second, political conflicts have a different scope than moral conflicts: ‘Political disagreement is identified by a field of application—eventually, about what should be done under political authority, in particular through the deployment of state power’ (Williams 2005a, p. 78). In contrast to moral conflict, political conflict always takes place against the backdrop of a political authority that is entitled and even required to enforce political decisions through its monopoly on violence. The conceptual framework of political moralism, however, does not offer any space for this kind of opposition or difference, seeing that its moral premises are seen as universally acknowledged and reasonable. Given these premises, the type of conflict that realists point to cannot even arise.

It is important to note that contemporary realism neither rejects liberal values nor the importance of morality for politics per se. Williams acknowledges that liberalism is the only acceptable legitimating story for modern societies (2005a, p. 10), i.e. the only such story that ‘makes sense’ to citizens of modern societies. He admits that the strength of liberalism lies in the fact that it is indeed, first, an ‘acceptable’ answer to the BLD (Williams 2005a: 8), and second, that it raises the standards of acceptable answers to the BLD by ‘rais[ing] the standards of what counts as being disadvantaged’ (Williams 2005a, p. 7)—i.e. the standards that a political authority is supposed to ensure for each individual. Any legitimate political authority must protect its subjects from being ‘radically disadvantaged’ in ‘the basic Hobbesian terms of coercion, pain, torture, humiliation, suffering, death’ (Williams 2005a, p. 4). Liberal states, however, introduce ‘more demanding standards of what counts as a threat to people’s vital interests’ (Williams 2005a, p. 7). In particular, liberalism bans ‘hierarchical structures which generate disadvantage’ (Williams 2005a, p. 7).

So political realism need not reject liberalism or the importance of morality. Nonetheless, realism has often been criticised for its ambiguous normative status for two reasons (see Sleat 2010). First of all, due to this close connection of historical circumstances and political theory, realism has been accused of having a ‘status quo bias’ (Rossi 2016). It blurs the critical distance of theory and mere political circumstances, and hence risks losing the progressive and emancipatory potential that the normative core of liberal theory represents. Second, while Williams rejects the idea of morality as a guide for politics, the BLD and the critical theory principle both seem to have a moral content themselves. Since the mere subjection of persons under a political order is rejected as illegitimate, it can be argued that he implicitly relies on moral principles regarding the question of what free and equal persons owe to each other. In this respect, contemporary realism and liberalism are not that far apart.

Against the second objection, Sagar argues that the normativity Williams suggests is of its own kind, since the BLD is meant to distinguish politics from mere coercion and war. While an answer to the BLD must be given, this answer need not be accepted unanimously or consensually by all who are coerced. However, as soon as there is a BLD, ‘politics is happening’ (Sagar 2016, p. 371). This establishes a genuinely political normativity that does not rely on previous moral considerations, but emerges as a reaction to the ‘first political question’. Consequently, the realist conception opens space for political conflict and negotiation about the BLD without presupposing the necessity of a universal consensus. This makes possible a distinction between political legitimacy, deriving from an answer to the first political question, and moral rightness in Kantian terms. While Rawlsian liberalism and realism both accept that a given political authority can be legitimate despite some people rejecting it, only realism accepts that the people rejecting it can do so for good political (if non-moral) reasons, i.e. without being irrational and therefore ‘wrong’.

In spite of the normative ambiguity of the BLD, Williams’s account puts the relation of the political and the moral back into the centre of liberal political theorising. Williams does not deny that political and moral questions can be closely linked, yet his approach allows us to conceptualise the liberal state as a coercive state: even a liberal state can—and must—coerce its citizens into conforming with certain rules in specific domains of their lives for the purpose of social coordination and in order to guarantee order and stability (see also Sleat 2013). What is more, the enforcing nature of the state creates the potential for the abuse of power, an aspect of the political that Rawls’s political moralism seems unable to conceptualise, despite his efforts to address non-ideal situations, such as when civil disobedience is called for in response to unjust legislation. Therefore, the conceptual gap that realism identifies in liberal theory is the failure to integrate the dimension of power and conflict into its conception of the political, while at the same time political realism wants to preserve liberalism’s core commitment, the normative primacy of individual liberty. However, the realist methodology itself is not sufficient to provide this normative standard. I will argue that Shklar’s liberalism is able to integrate the concerns of realist theory, but avoids the imprecision regarding the normative status of her conception by providing her own, explicitly normative principle, which nonetheless proceeds independent of moral theory.

Political Injustice

The realist concern with the conceptualisation of conflict is complemented by the non-ideal critique of contemporary, ideal liberalism. While realists ask if liberalism offers an adequate conceptualisation of the conflictual political sphere, non-ideal theory mainly concerns itself with the methodological resources for understanding injustice within the liberal paradigm. While realist and non-ideal theory differ in their conceptual approach to liberalism, the conceptualisation of political conflict and injustice are related issues, since it is implausible to assume that something can be unjust without there also being a related or underlying conflict. Despite this proximity of realist and non-ideal thinking, they are marked by a crucial difference in their respective approach to liberalism: realism is critical of political moralism as such, whereas non-ideal theory essentially relies on the Rawlsian distinction of ideal and non-ideal theory and hence defends a version of ‘political moralism’.

Charles W. Mills, a prominent proponent of non-ideal theory, argues that Rawlsian ideal theory is abstracting away from precisely those domains of the social and political life in which injustices are most likely to occur, and about which a liberal theory of justice should not be silent given its normative commitment. In particular, ideal theorising results in an insufficient understanding of sexist, classist, and racist discrimination, and consequently cannot offer any solutions for these problems (Mills 2005. For a similar criticism of Rawls, see Okin 1989). Like Williams, Mills supports the core liberal principles of freedom and equality, but argues that they need to be arrived at differently, namely through non-ideal theory: ‘I would like to suggest that a nonideal approach is also superior to an ideal approach in being better able to realize the ideals, by virtue of realistically recognizing the obstacles to their acceptance and implementation’ (Mills 2005, p. 187). Non-ideal theory still works with abstraction and idealisation, but in a more apt way than ideal theory. This is because Rawlsian ideal theory (or ‘ideal-as-idealized’ theory; Mills 2005, p. 167), develops a ‘model of how P should actually work’, be it ‘a perfect vacuum, a frictionless plane’ (Mills 2005, p. 167), or, as in the case of political philosophy, an ideally just society. Non-ideal theory, on the other hand, develops a ‘model of how P actually works’ (ibid.), by maintaining ‘P’s crucial aspects (its essential nature) and how it actually works (its basic dynamic)’ (Mills 2005, p. 166). Since ideal theory develops a model of how a just society should ideally look like, it abstracts away from actual—and often not perfectly just—societies’ basic dynamic, especially regarding the crucial six aspects: it presupposes an idealised social ontology, idealised capacities, silence on oppression, ideal social institutions, an idealised cognitive sphere, and strict compliance (Mills 2005, pp. 168–169); all of them conditions never found in the ‘real’ world, in which injustice needs to be dealt with. These idealisations lead to an inadequate theoretical focus for the conceptualisation of (in)justice. Non-ideal theory, too, works with abstraction: class, gender, and race as paradigm cases of hierarchical social relationships are abstractions of the non-ideal type which are said to mirror the social realities of societies that are not perfectly or ideally just. It is important to note that none of these cleavages of discrimination are fully understandable without their historical background. For example, urgent political problems such as racism in the United States are not even understandable with the tools of ideal theory, seeing that the original position is conceptualised as hypothetical and ahistorical. Therefore, ideal theory is not sufficient for dealing with concrete instances of injustice. Since it is concerned with the justification of abstract principles of justice, its pre-established grid is too wide for grasping the complexities of deviations from these principles. Here, Mills’s critique mirrors the realist interest trying to recover the relevant information that is made invisible through ideal theory.

Although Mills’s critique of Rawls is less radical than Williams’s, given its continuation of the politically moralist outlook, Mills’s account, too, runs into problems regarding its own normative status. While Mills is not criticising political moralism as such, the moral judgements on which his criticism relies in fact need to fulfil the criteria of universality and impartiality (Hampshire 1978). If universalisation is to be possible at all, it is necessary to abstract away from the specific characteristics of the situation. Therefore, ideal-as-descriptive theory has to rely on the ideal, ideal-as-idealised theory, in the way Rawls suggested, such that non-ideal theory is a second step in which particular obstacles to realising justice are taken into consideration (e.g. Goodhart 2018; Sleat 2013). Giving up Rawls’s ideal methodology is hence not sensible given Mills’s own goals. What is missing from liberal theory thus is an approach that is able to deal with concrete instances of injustice, while preserving liberalism’s normative core commitment. Here, non-ideal theory runs into the same problems as realism does: neither of these approaches provides an independent normative standard based on which political arrangements can be criticised. The liberalism of fear, on the other hand, goes further than these two approaches by developing its own normative principle: the avoidance of the summum malum.

It has become clear that the contemporary critiques of liberalism primarily address conceptual and methodological questions, not the normative core premise of the primacy of individual liberty. The idealising approach to political philosophy has been criticised for not being able to conceptualise the political as independent from morality, resulting in an incapacity to conceptualise conflict and a blind spot regarding political injustice. However, regarding their respective normative standards, both the non-ideal and realist critique face the same difficulty: both depend on a standard that is prior to their theories. Non-ideal theory relies on the normative standard provided by ideal theory, and realism relies on the implicit moral assumption that coercion needs to be justified and justifiable to those subject to it, which it is unable to derive from its own historically embedded assumptions. It is the strength and originality of the liberalism of fear to combine the conceptual desiderata made visible through the two criticisms discussed above with a distinctively normative outlook.

The Liberalism of Fear

I will defend Judith Shklar’s liberalism of fear as a useful way of rethinking liberalism, arguing that it offers a radically different conceptual and methodological approach as well as an independent and explicitly normative viewpoint, which sets it apart from other contemporary critiques of liberalism (see Kaufmann 2019). Relating back to the deficits of liberal theory discussed above, I will argue that the liberalism of fear is able to close the conceptual gaps that Rawlsian ideal liberalism has left liberal theory with. I will highlight that the liberalism of fear represents a genuinely realist version of liberalism, since it proceeds from the ‘first political question’ independent of any moral theory (for a reading of Shklar as a realist see also Sabl (2011) and Sleat (2013)). I will argue that it transcends realism by developing its own normative standard, and hence remains a realist version of liberal theory. Second, due to its non-utopian methodology, it is better able to avoid the blind spots regarding injustice.

Non-Utopian Liberalism

Like all liberal theories, the liberalism of fear embraces the normative core of the primacy of individual liberty. However, compared to the standard theories of liberalism, Shklar employs a radically different methodology. For this purpose, she first spells out and criticises the implicit premises of the paradigm liberal theories—namely, the liberalism of natural rights, associated with John Locke, and the liberalism of self-development, associated with John Stuart Mill. She argues that both implicitly rely on an idealising dynamic that results in an implausible picture of the political. In particular, the focus on natural rights in the Lockean tradition entails ‘that a perfect or optimal society would be composed only of rights claiming citizens’ (Shklar 1989, p. 26). The liberalism of personal development assumes that individuals realise their potential primarily through education, which is why, on this account, ‘politics and government’ would eventually become unnecessary (Shklar 1989, p. 27). Avant la lettre, Shklar starts from the non-ideal claim that the chances of establishing a perfect, rights-based society, as well as a society composed of self-perfecting citizens are rather slim, and thus such idealising approaches to political problems inadequate. She explicitly emphasises that her approach is ‘non-utopian’, as she rejects ideal theory as implausible. Instead of making idealising assumptions about the hypothetical capacities of ideal citizens, Shklar wants to recover the initial thrust of liberal theory by looking at its history. She emphasises that liberalism emerged as a theory for resolving profound political conflict, and protecting the individual from certain kinds of mistreatment (Shklar 1989, p. 23). The reason why persons need to be protected from conflict and mistreatment, however, is not that they are ideally rational, rights-claiming citizens or individuals capable of perfecting themselves, but rather because as human beings, they are fundamentally vulnerable. Hence, the liberalism of fear is ‘entirely nonutopian’ (Shklar 1989, p. 26):

it does not, to be sure, offer a summum bonum towards which all political agents should strive, but it certainly does begin with a summum malum, which all of us know and would avoid if only we could. That evil is cruelty and the fear it inspires, and the very fear of fear itself. To that extent the liberalism of fear makes a universal and especially a cosmopolitan claim, as it historically always has done. (Shklar 1989, p. 29)

The non-utopian thrust entails a negative methodology, which is crucial for understanding the fundamental thrust of the liberalism of fear. Rather than striving for a summum bonum that ought to be realised, the liberalism of fear proceeds negatively—by identifying the summum malum, the worst evil that ought to be avoided. According to Shklar, all human beings want to avoid cruelty and fear:

Of fear it can be said that it is universal as it is physiological. It is a mental as well as a physical reaction, and it is common to animals as well as to human beings. To be alive is to be afraid, and much to our advantage in many cases, since alarm often preserves us from danger. The fear we fear is of pain inflicted by others who kill and maim us, not the natural and healthy fear that merely warns us of avoidable pain. And, when we think politically, we are afraid not only for ourselves but for our fellow citizens as well. We fear a society of fearful people. (Shklar 1989, p. 29)

What is meant by cruelty here? It is a deliberate infliction of physical, and secondarily emotional, pain upon a weaker person or group by stronger ones in order to achieve some end, tangible or intangible, of the latter. (Shklar 1989, p. 29)

The liberalism of fear aims at the avoidance of the summum malum for human beings, which Shklar identifies as ‘cruelty and the fear it inspires’, as well as ‘fear of fear’; notions that she does not define in great detail but which will be analysed more closely in the following. The negative methodology of the avoidance of the summum malum is justified based on the assumption that the political setting creates circumstances in which fear for oneself and for others is an adequate reaction. This focus points to the purpose of the liberal state: the protection of the individual. Hence, the negative, non-utopian methodology paves the way for the realist concern, the solution of the first political question. It can even be said that the first political question only arises given the fundamental need for protection, which can only be given precedence through a non-utopian, negative approach.

The Asymmetrical Structure of the Political

In accordance with classical liberal social contract theory, Shklar defends the liberal state as the best method of protecting individuals, and individual rights as the best tool for realising this protection. Consequently, liberalism ‘has only one overriding aim: to secure the political conditions that are necessary for the exercise of personal freedom. […] That belief is the original and only defensible meaning of liberalism’ (Shklar 1989, p. 21). Other than that, ‘liberalism does not have any particular positive doctrines about how people are to conduct their lives or what personal choices they are to make’ (Shklar 1989, p. 21). As a consequence of the primacy of individual liberty, the distinction between the public and the private sphere becomes a necessary characteristic of any liberal theory. The protection of individual liberty and the private sphere is ensured by the ‘original first principle of liberalism, the rule of law [which] […] is the prime instrument to restrain governments’ (Shklar 1989, p. 37).

The liberalism of fear’s focus on the importance of political power clearly marks its realist thrust. One of realism’s core concerns is the role of power in the political sphere (Galston 2010). Shklar’s proximity to and influence in realist theory have been widely acknowledged (Sabl 2011; Sleat 2013), especially by Williams (2005b), who explicitly draws on her liberalism of fear. However, I will defend the claim that Shklar’s liberalism of fear goes further than realist theories in not only emphasising the relevance of power and conflict in the political sphere. Shklar establishes an independent normative standard that sets her theory apart from realist theory. This normative standard allows her to avoid the status-quo bias of realist theory, while at the same time her focus on power enables her to integrate realist concerns into her own approach. Hence, her approach remains within the liberal paradigm, but incorporates the concerns of realism.

In Shklar’s theory, the focus on power is a logical consequence of starting from the position of weakness in the political setting rather than idealised capacities, i.e. of the negative, non-utopian approach. In agreement with political realism, Shklar draws attention to the fact that political power, even in a liberal state, can always become part of the problem it is meant to solve: ‘The liberalism of fear […] regards abuses of public power in all regimes with equal trepidation’ (Shklar 1989, p. 28. Williams 2005b analyses this aspect of realism in close detail.) What the liberalism of fear emphasises is the fact that state power is not simply a neutral agency that impartially protects rights. On the contrary, in order to be able to fulfil the purpose of protecting the individual, the liberal state must rely on a monopoly of violence, distributed across institutions such as the legal system, the police, and the military, which it needs in the first place to be able to assert political power. This simple fact, while necessary for the solution of the first political problem, also exposes citizens to a potential threat. This threat consists in the fact that the establishment of a monopoly of violence is in itself not an alleviation of the first political problem, since there is the inherent potential of an abuse of power:

Given the inevitability of that military, police, and persuasive power which is called government, there is evidently always much to be afraid of. […] And the freedom it [the liberalism of fear] wishes to secure is freedom from the abuse of power and intimidation of the defenseless that this difference invites. (Shklar 1989, p. 27) [my addition]

The liberalism of fear emphasises that the asymmetry of political power, which structures any political setting, does not cease to exist in a liberal state. The liberalism of fear starts from the first political question of ‘order, protection, safety, trust and the conditions of cooperation’ (Williams 2005a, p. 3), and shares the liberal assumption that this question can be answered. However, it does not share the politically moralist assumption that the liberal state can never become part of the problem; something that realists criticise regarding the consensus-view versions of liberalism. The liberalism of fear argues that even the liberal solution of the first political problem does not eradicate the conflictual dynamic of the political. The asymmetry of power that is a necessary characteristic of any political authority can be part of the solution of the political problem just as much as it can be the problem itself. Given the focus on those whose ‘protection’ and ‘safety’ are at stake, it becomes clear that ‘For this liberalism [of fear] the basic units of political life are not discursive and reflecting persons, nor friends and enemies, nor patriotic soldier-citizens, nor energetic litigants, but the weak and the powerful’ (Shklar 1989, p. 27 [my addition]). This observation about the structure of the political is independent of any moral theory: for example, the liberalism of fear ‘does not rest on a theory of moral pluralism’ (Shklar 1989, p. 29). Since it starts from a functionalist observation about the structure of the political, and does not rely on any predetermined moral standard, it is a genuinely political concept in Williams’s terms. It has already been observed for the liberalism of fear, the political sphere per se is constituted through the dichotomy of ‘weak’ and ‘powerful’, and this an important aspect that sets it apart from other liberal theories; it is a ‘liberalism from the margins’ (Gatta 2018, p. 118). For the liberalism of fear, the distinction of the ‘weak’ and the ‘powerful’ is the result of the necessary distinction between those who actually dispose of the means of coercion and those who do not, and are consequently exposed to the inherent threat of an abuse of power, i.e. the solution turning into the first political problem. I claim that based on this distinction, Shklar’s concept of the asymmetrical political sphere is inherently conflictual, since it does not assume that the solution of the first political problem establishes universal consensus and eradicates the possibility of conflict once and for all. What is more, the functionalist perspective on the asymmetry of the political highlights the normative independence of Shklar’s approach from moral theory; rather than having to rely on ‘moral intuitions’, as some critics assume (see e.g. Sleat 2013, p. 106). Political power, even liberal political power, of necessity introduces a dichotomy of the ‘weak’ and the ‘powerful’—consequently, political power is something that is to be feared for good reasons. Hence, Shklar’s approach is driven by a realist concern.

Fear—A Normative Notion?

Shklar’s focus on the asymmetry of power is reflected in the core concepts of ‘fear’ and ‘cruelty’, which form the summum malum that ought to be avoided. Seeing that they carry the normative weight of Shklar’s liberalism, it is particularly problematic that Shklar does not define them in greater detail. Since it seems like she is deriving normative statements from the description of physical pain, the concept of ‘fear’ has been criticised for representing a naturalistic fallacy (see e.g. Robin 2004, p. 149). However, this allegation lacks plausibility, simply because Shklar does not only refer to fear of physical pain but introduces a more complex notion of fear. Recently, Bajohr (2019) has defended a highly illuminating reading of the terms of ‘fear’ and ‘cruelty’ that proves their normative potential while avoiding the naturalistic-fallacy-objection. Bajohr argues that Shklar in fact distinguishes a ‘primary and a secondary notion of fear’ (Bajohr 2019, p. 169), and which represent ‘two distinct sources of normativity’ (Bajohr 2019, p. 168). The primary notion of fear is ‘empirical’ (Bajohr 2019, p. 170) and explicitly refers to the fear of physical pain, which, for Shklar, has universal status as the ‘summum malum, which all of us know’. The secondary notion of fear, on the other hand, is a ‘formal’ principle (Bajohr 2019, p. 170; see also Forrester 2012, p. 252 and Tamir 1997, p. 298). The recurring notion of the ‘fear of fear’ represents an abstraction away from the mere physical dimension: ‘the first “fear” would here denote the quality of the experience, the second its content’ (Bajohr 2019, p. 170). Consequently, ‘what the fear of fear is about does not have to be bound to any naturalistic constant but can change over time and expand its range beyond’ mere physical cruelty (Bajohr 2019, p. 170; see also Forrester 2012, p. 252). ‘Because the fear of fear is a reflexive argument, it can ensure its universality without resorting to a strong naturalism’ (Bajohr 2019, p. 170), and, hence, avoid the naturalistic fallacy. The precise content of the ‘fear of fear’ can vary, which is why Bajohr argues that the second principle entails a ‘contextual universalism’ (ibid.). Similarly, Gatta (2018, p. 115) argues that fear can consist of different things in ‘different settings, situations, and cultural contexts’, which is why there is an ‘agon about what fear is’. This novel reading of fear as the summum malum is crucial for rebutting unjustified criticisms of Shklar, which overstate the primary notion of fear, and assume that her approach is unable to deal with more complex but politically important phenomena. Bajohr’s and Gatta’s readings prove that these phenomena can easily be seen as adequate objects of fear within Shklar’s approach.

In the following, I will draw on this multi-level analysis of the concept of ‘fear’ as a normative universalist notion in order to defend a ‘political’ reading of the terms ‘cruelty’ and ‘fear’. I claim that the difference between the primary and secondary notion of fear becomes crucial in the political context. The empirical, primary notion of fear of physical pain is deemed ‘natural and healthy’, since it protects all animals from danger. As a result, the primary notion of fear is not necessarily relevant in the political sphere. Rather, in the political context that liberalism is addressing, the secondary notion of fear is specified as something that can be aroused by others, who are perceived as a potential threat. Yet it should be noted that Shklar does not defend a Hobbesian ‘homo homini lupus’—view of human nature or society. Although others can constitute a threat, and consequently must under certain conditions fear each other, Shklar grants that individuals can also fear for others. Therefore, I argue that it is not the simple fact of persons living together that inspires political fear, but the political setting of the unequal distribution and asymmetry of power that inspires the ‘fear of fear’. It is important to note that the ‘political’, secondary notion of fear relates back to the realist thrust of the liberalism of fear: the necessary asymmetry of political power generates the idea that the basic units are the ‘weak’ and the ‘powerful’, rather than idealised, rights-claiming and rational subjects.

Further, I claim that the concept of ‘cruelty’, another essential parameter in her theory, reflects the realist focus on the asymmetry of power:

But public cruelty is not an occasional personal inclination. It is made possible by differences in public power, and it is almost always built into the system of coercion upon which all governments have to rely to fulfil their essential functions. A minimal level of fear is implied in any system of law, and the liberalism of fear does not dream of an end of public, coercive government. The fear it does want to prevent is that which is created by arbitrary, unexpected, unnecessary, and unlicensed acts of force and by habitual and pervasive acts of cruelty and torture performed by military, paramilitary, and police agents in any regime. (Shklar 1989, p. 29)

Like the political notion of fear, public cruelty, too, is explicitly qualified as a possible result of the dichotomy of ‘weak’ and ‘powerful’, which enables the latter to subject the former to their will. While this can also happen in the private (or any non-political) sphere, public cruelty is a political problem that liberalism needs to address. The possibility of public cruelty is a given in any regime, and even within a liberal one it results from the difference in status and power that the liberal state’s monopoly of violence entails. This setting results in the ‘fear of fear’ itself. Therefore, I argue that given its awareness of the enforcing and even violent potential of state power, the liberalism of fear closes the conceptual gap regarding power, evident in Rawls’s ‘ideal theory’.

What is more, due to its focus on fear, Shklar’s negative, non-utopian approach not only avoids the apolitical conclusions of mainstream liberal theories, and proves the inherently realist thrust of her liberalism. The focus on fear also provides her with a normative standard for judging political arrangements. Drawing on this conclusion, I defend the claim that rather than resorting to ideal principles of justice decided on by hypothetical, ideally rational subjects, the normative standard of the liberalism of fear is derived from the structure of the political itself. Since this normative standard is arrived at without resorting to theories of the moral or of rationality, it is, as Shklar claims, compatible with various moral theories, without having to depend on any of them (Shklar 1989, pp. 29–30).

A Theory of Injustice?

The vision of the political on Shklar’s terms seems rather pessimistic, if not ‘dystopic’ (Benhabib 1996). Her suggestion for dealing with the conflictual political situation is, again, negative: The Faces of Injustice analyses the (epistemological) resources of conceptualising injustice. The negative approach—injustice rather than justice—is a result of the focus on unavoidable and irreducible conflict in the political sphere. Injustice and fear are conceptually related, as Bajohr argues: the primary and secondary notions of fear ‘are closely intertwined with a third, transcendental principle, which describes conditions of the possibility for articulating a sense of injustice’ (Bajohr 2019, p. 168). Acknowledging the conceptual status and normative significance of the political notion of fear thus makes possible a liberal perspective on injustice: ‘This is where the third source of normativity comes in; after the empirical and the formal, it is a transcendental argument. It looks at the condition of the possibility for giving voice to one’s sense of injustice’ (Bajohr 2019, p. 171). This condition is, for Shklar, the articulation of experiences of injustice in a democratic system.

As a liberal, Shklar relies on the rule of law as a means of dealing with injustice. Yet she shares the non-ideal intuition that the mere establishment of the rule of law or ideal principles of justice is not sufficient. This is because, as Shklar argues, the boundary between misfortune and injustice is historically variable:

When is a disaster a misfortune and when is it an injustice? Intuitively the answer seems to be quite obvious. If the dreadful event is caused by the external forces of nature, it is a misfortune and we must resign ourselves to our suffering. Should, however, some ill-intentioned agent, human or supernatural, have brought it about, then it is an injustice and we may express indignation and outrage. As it happens, in actual experience this distinction, to which we cling so fervently, does not mean very much. The reasons become clear enough when we recall that what is treated as unavoidable and natural, and what is regarded as controllable and social, is often a matter of technology or of ideology or interpretation. (Shklar 1990, p. 1)

Hence, injustice cannot not be assessed independently of its historical context. It is precisely for this reason that ideal principles of justice are too rigid to grasp the complexity of political conflicts. This should not be understood as a relativist approach that makes injustice dependent on the dominant culture or ideology. What Shklar points out is that something that was a misfortune in the past (e.g. dying in an earthquake or of a disease) can become an injustice due to technological, medical, or social progress of protecting persons from existential risks.

Considering its focus on injustice and on the respective circumstances of an unjust situation, Shklar’s approach mirrors the concern of non-ideal theory. Yet there is a crucial difference between these two approaches: non-ideal theory essentially remains dependent on Rawlsian methodology, which is why injustice in fact remains a secondary phenomenon. Shklar’s non-utopian approach, on the other hand, operates without recourse to an ideal theory; starting with the avoidance of the summum malum hence opens up a perspective on injustice without having to take the detour of ideal theory.

For Shklar, the distinction between injustice and misfortune is made possible through the ‘sense of injustice’:

What, however, is the sense of injustice? First and foremost it is the special kind of anger we feel when we are denied promised benefits and when we do not get what we believe to be our due. It is the betrayal that we experience when others disappoint expectations that they have created in us. And it has always been with us. (Shklar 1990, p. 83)

The sense of injustice mirrors the primary concept of fear as its empirical foundation:

For while we internalize the ethos of inequality and accept it as right and just, we do not lose our natural ability to feel deprived, humiliated and offended when our expectations as human beings are not met, when our claims are ignored, […]. And many of our expectations are rooted in nature, not in culture. (Shklar 1990, p. 87)

Just like the primary notion of fear, the sense of injustice is hardwired into human nature and therefore not reducible to the given circumstances. And like fear, the sense of injustice has a secondary and political notion, too: ‘we are not only aroused on our own behalf but emphatically also when the indignities of injustice are experienced by other people. The sense of injustice is eminently political’ (Shklar 1990, p. 83). In its secondary, political dimension, the sense of injustice has a corrective function: ‘As long as we have a sense of injustice, we will want to understand not only the forces that cause us pain but also to hold them responsible for it’ (Shklar 1990, p. 5). In order to detect shifts from misfortune to injustice, Shklar suggests taking seriously experiences of injustice. Her argument for proceeding in this way lies in the specific and unique knowledge that victims of injustice have:

The perceptions of the victims and those who, however remotely, might be victimizers tend to be quite different. Neither the facts nor their meaning will be experienced in the same way by the afflicted as by mere observers or by those who might have averted or mitigated the suffering. These people are too far apart to see things the same way. (Shklar 1990, p. 1)

This specifically epistemological aspect of injustice has become a central aspect of the debate on injustice. In support of Shklar’s claim that victims of injustice have specific knowledge of injustice, e.g. Medina (2013) analyses the respective epistemic capacities of persons belonging to socially marginalised vs. privileged groups. He argues that there are three epistemic virtues marginalised persons are more likely to possess, namely epistemic humility, curiosity, and open-mindedness; whereas persons occupying privileged social positions tend to develop epistemic vices, notably epistemic arrogance, laziness, and close-mindedness, which can result in ‘active ignorance’ (Medina 2013, pp. 27–55). The articulation of marginalised perspectives can lead to ‘epistemic friction’ (Medina 2013, p. 46) in the public sphere. Hence, the marginalised and the victims of injustice have ‘knowledges from below’ (Medina 2011, p. 21), which can, in Shklar’s terms, eventually lead to a shift in the boundary between misfortune and injustice if they are taken seriously. The articulation of individual experiences of injustice can be a valuable resource for political change, which in turn can benefit other people besides the original ‘victim’ too. Paying attention and taking seriously experiences of injustice can thus help remedy one type of epistemic injustice, namely ‘testimonial injustice’ (Fricker 2007, p. 41), which she describes as the denial of proper credibility regarding the testimony of victims of injustice.

Despite these apparent benefits of the victim-centred approach on the epistemology of injustice, it suffers from a crucial shortcoming: victims are neither always capable nor willing to identify themselves correctly. Two cases are conceivable: first, cases in which someone does not identify themselves as a victim of injustice in spite of their situation fulfilling criteria of injustice, and second, cases in which someone does believe that they have suffered or are suffering an injustice, in spite of there being no good reasons for this assumption. With Fricker, the first case is a case of ‘hermeneutical injustice’. Hermeneutical injustice consists in an agent’s inability to make sense of their own experiences due to a ‘gap in collective hermeneutical resources’ (Fricker 2007, p. 6). For instance, ‘sexual harassment’ first had to be identified as a case of inappropriate behaviour, rather than counting as ‘flirting’ (ibid., pp. 152–153). The second case, in which persons see themselves as victims without there being any good reasons for this belief, has become increasingly important in the political sphere. This phenomenon is assumed to be one of the main reasons for the recent success of populism in Western democracies (Mueller 2016). Shklar’s victim-centred approach is unable to distinguish ‘legitimate’ grievances and expressions of injustice from the unjustified resentment and perceived injustice, for example of voters who feel that they are systematically disadvantaged through processes of globalisation, and who blame the ‘liberal elites’ for their situation (Khazan 2017). In light of these two cases, the assumption that the perspective of victims of injustice can contribute substantial insights on injustice becomes doubtful.

What the latter two cases highlight is the fact that the assessment of something as an injustice must depend on more than just the victims’ perspective, and that an independent standard is required. Starting from concrete experiences of injustice hence is a rather circular approach to understanding injustice. Yet, in spite of the challenge posed through the cases of victim-misidentification, a victim-centred perspective on injustice is more promising regarding the crucial problem that rigid theories of justice entail: their insensitivity to the fact that the boundaries of injustice and misfortune are variable. Therefore, the pre-established grid of an ideal theory of justice might simply be too abstract and even lead to new instances of injustice when it is applied to complex real-world injustice. If the boundaries between injustice and misfortune are variable, theories of justice are necessarily deficient tools for recognising injustice. This is why Shklar claims that justice and injustice are asymmetrical phenomena, and injustice consequently more than the mere absence of justice (see also Heinze 2017). Therefore, the liberalism of fear with its awareness of the conflictual character of the political is particularly sensitive to injustice: ‘Actually, the most reliable test for what cruelties are to be endured at any place at any time is to ask the likeliest victims, the least powerful persons, at any given moment and under controlled conditions’ (Shklar 1989, p. 35). Precisely because the asymmetry of the political always represents a reason for fear, the people occupying the weakest positions within the political community must be taken seriously.

In order to properly process the experiences of injustice, the liberalism of fear ‘is monogamously, faithfully, and permanently married to democracy—but it is a marriage of convenience’ (Shklar 1989, p. 37). Only in a liberal democratic political system it is possible to publicly criticise public officials: ‘democracy does not fulfil its immanent promises quickly, but at least it does not silence the voices of protest, which it knows to be the herald of change’ (Shklar 1990, p. 8):

Without the institutions of representative democracy an accessible, fair, and independent judiciary open to appeals, and in the absence of a multiplicity of politically active groups, liberalism is in jeopardy. It is the entire purpose of the liberalism of fear to prevent that outcome. (Shklar 1989, p. 37)

Shklar’s emphasis on the institutional setting in liberal democracies solves the problem of selection that a purely victim-centred approach would entail: the normative, universalist framework, the avoidance of the summum malum and of the ‘fear of fear’ remains in place in the institutionalised form of a liberal democracy under the guidance of the rule of law. Consequently, she does not advocate simply giving every self-proclaimed victim ‘their due’. In this respect, the liberalism of fear is congruent with any other liberal theory. However, the non-utopian methodology enables the perspective on articulations of injustice and the voices ‘from below’, which must be heard and be given an open-minded evaluation, rather than presupposing universal consensus.

Yet there is an epistemic benefit from taking injustice as the starting point. In fact, injustice is taken to have hermeneutical priority over justice. This can be made plausible from the fact that the claim that something is unjust is made much more often than the judgement that something is just (see also Heinze 2017, p. 364; Mieth 2005). Therefore, the critical as well as emancipatory potential of articulations of injustice should not be underestimated when it comes to identifying grievances that ought to be remedied. In this sense, the perspective on, or the ‘sense of’ injustice transcends theories of justice, since it can denounce and criticise theories of justice and their outcomes themselves (Mieth 2005, p. 66). This does not mean, however, that the sense of injustice leads directly into a theory of injustice. Rather, acknowledging two distinct levels of thinking about justice and injustice explains Shklar’s ‘asymmetry thesis’: that, first, justice and injustice are asymmetrical phenomena and second, injustice can only be insufficiently understood through the lens of a theory of justice. First of all, a theory of justice remains crucial for classifying situations or actions as just or unjust. The criteria for this assessment need to come from a normative standard that is not taken from merely subjective, relative experience. This is why Shklar develops the normative principle of the avoidance of ‘fear of fear’, which is not something that results from an individual victim’s perspective but from a universal statement about power asymmetries in the political setting; and why she relies on democratic institutions for the assessment of individual claims. What the victims of injustice do know about, though, is the way in which injustice presents itself: the object(s) of fear and the means of cruelty can change depending on the respective historical circumstances. Without this universal normative standard, the identification and avoidance of cruelty and the fear of fear would become conceptually impossible. Therefore, rigid normative principles of justice and the unsystematic sense of injustice can mutually inform each other, without their relation having to be symmetrical.

Realist Liberalism Reconsidered

The emphasis and reliance on liberal democratic institutions capable of distinguishing legitimate from illegitimate articulations of grievances illustrate the limits of Shklar’s realism and non-utopianism. In spite of her scepticism regarding the establishment of rigid principles of justice, Shklar herself does employ a universal normative standard, namely the avoidance of the summum malum, i.e. the prevention of cruelty and the ‘fear of fear’. Although so far, I have argued for the independence of Shklar’s realist liberal approach from moral theory, not all realists will agree with this assessment. Original and fruitful as the negative approach of thinking about the summum malum is, it is, exactly like any conception of a summum bonum, something ‘on which people disagree’ (Sleat 2013, p. 103). The conception of a summum malum still stipulates consensus—an idea that contemporary realism rejects as an implausible assumption for any theory of the political. In this respect, Shklar’s approach is not that far from Rawls’s after all (see also Forrester 2012). Moreover, Shklar does make use of idealising assumptions—namely, that democratic institutions are reliable and impartial tools for detecting and remedying injustice. This idealised assumption can still be challenged on non-ideal theory’s grounds. The assumption that the democratic public is sufficiently unbiased and epistemically open-minded to listen to the voices of injustice’s victims is, to be sure, not something that reflects the ‘real’ world. Nevertheless, the liberalism of fear puts the voices of the victims on the agenda of liberal theorising. Its negative, non-utopian methodology and the context-sensitive principle of the ‘fear of fear’ render the liberalism of fear more responsive to grievances and conflict in the political sphere.

I have defended the liberalism of fear as a genuinely political liberal theory that is able to address political conflict and injustice, aspects that are criticised as being either missing or at least inadequately conceptualised in mainstream liberal theory. I have argued that it is a theory highly sensitive to the concerns of marginalised persons, and therefore a reinvigoration of liberalism’s emancipatory and progressive thrust. In these respects, the liberalism of fear mirrors the concerns of realist and non-ideal theory, while providing its own normative standard and hence avoiding the normative dependence on ideal liberal theory that the two criticisms face. However, it is neither a fully realist nor a fully non-ideal theory, but because of its normative standard remains essentially a theory of liberalism. A liberal theory that takes seriously injustice and its victims is not interesting for merely methodological and conceptual purposes. Given the prevalence of conflict and ubiquity of injustice in the political world, the liberalism of fear can contribute to our understanding of liberalism as a political theory for the contemporary world.