What not to expect from the pragmatic turn in political theory

Introduction

The pragmatic turn in philosophy usually refers to attempts made by philosophers like John Dewey, Ludwig Wittgenstein, Hilary Putnam, Wilfrid Sellars, Richard Rorty and Robert Brandom to ‘pragmaticise’ analytic philosophy of language by moving away from a traditional representationalist view towards a view that gives primacy to human practices, encouraging a way of philosophising more apt to dealing with problems of everyday life. ¹ The central ideas coming out of this turn have had a major impact on several debates in philosophy over the last decades, ranging from metaphysics to epistemology. Most recently, they have set in motion what may be described as a pragmatic turn in normative political theory. It has become commonplace among political theorists today to draw on pragmatically influenced theories of language and meaning in their theorising of democracy, pluralism, justice, freedom, political legitimacy and so on. That this is not a marginal phenomenon but occurs on a wide front makes it even more interesting. On the one hand, it is not hard to see the appeal of drawing inspiration from philosophers like Wittgenstein, who has rocked established philosophical presumptions at the roots. On the other, it remains unclear from these accounts what work, more exactly, such an approach of language and meaning could do for normative theory.

The aim of this paper is to explore the ‘gap’ between pragmatically influenced theories of language and meaning and normative political theory, and how normative political theorists have attempted to bridge it by making use of the former theories for normative purposes. What motivates this exploration is not only a suspicion that theorists using theories of language and meaning in this way put too much faith in what they can do with them but also that no systematic attempt has been made so far to unfold where such pragmatically influenced normative projects go wrong. By revealing the mistakes underlying the attempts to bridge the gap between theories of meaning and normative political theories, we aim to make plausible the general thesis that pragmatically influenced theories of language and meaning, however full of insight, cannot be put to substantial normative use in political theory. ²

The argument is developed in five steps. First, we illustrate how Wittgenstein's view is used among prominent political theorists in their respective normative theory. This is done to give an idea of how widespread this phenomenon is in political theory (section 1). In sections 2–4, our thesis is defended in a threefold manner. Section 2 focuses on the nature of language, analysing how central ontological aspects of Wittgenstein's view are utilised in these normative political theories. Section 3 focuses on contextualism and perception, studying how political theorists make use of central epistemological aspects of Wittgenstein's account. In section 4, focus is directed at understanding, and we analyse how central semantic aspects of this account are used in normative theory. We demonstrate that faulty conclusions are drawn from Wittgenstein's view in all three cases. Against what is claimed by these political theorists, it is argued that Wittgenstein's account of language and meaning can be used neither to dismiss some specific normative theories, nor to show the superiority of other specific theories. In a final step, we discuss the scope of our thesis, motivating what may seem like a bold conjecture in light of the fact that the argument is based on how one view of language and meaning is used in normative theory. This is done by pointing to the general form of the argument and by moving beyond Wittgenstein to other philosophers of mind and language, illustrating how similar overextensions are made with regard to Robert Brandom's theory of language and meaning (section 5).

1. The pragmatic turn in normative political theory

There are indeed many theorists in contemporary philosophy who adopt ideas emerging from the pragmatic turn and thus may be described as pragmatists, as well as many kinds of projects that could be considered part of a pragmatic turn. Take, for example, the current debate in political theory in which attempts are made to provide an epistemic justification of democracy drawing on the classical American pragmatists. ³ However, the specific pragmatic turn we look at in this paper is concerned with the philosophy of language rather than with epistemic justification. The common denominator of theorists we categorise as representatives of this pragmatic turn, let us call them ‘p-theorists’ for short, is that they are normative political theorists in the contemporary debate who put pragmatically influenced theories of language and meaning into normative use, both in defending certain positions and in criticising others. Thus, when we talk about ‘the pragmatic turn’, we refer to nothing more or less than this. Apart from this unifying feature, p-theorists may represent a wide range of approaches. For example, while some p-theorists would describe themselves as pragmatists also in other respects, others would be better described as poststructuralists or deconstructivists. Moreover, whether the philosophers of language and meaning with a pragmatist outlook that are utilised by p-theorists also call themselves ‘pragmatists’ is not our concern.

Among philosophers adopting such a pragmatist outlook, Wittgenstein is by far the most influential in contemporary political theory. In this section, we briefly present several prominent p-theorists extracting normative ammunition from Wittgenstein's view of language and meaning. ⁴ In doing so, we cannot do justice to the richness of each theory, nor fully appreciate the diversity among them. Still, as will be evident below, what makes an analysis of these p-theorists interesting is not only that they are distinguished scholars in normative political theory but also that they to a large degree make use of the same Wittgensteinian ideas for a range of different normative purposes, spanning from agonistic and aversive theories of democracy to notions of freedom and democratic self-determination.

In Chantal Mouffe's view, liberal democratic theory is ill-equipped to address the challenges that democratic societies face today. In the last decade, she has not only persistently called for an alternative way of theorising about the political but also responded to this call in the form of the agonistic model of democracy. From the very outset, Mouffe has argued that Wittgenstein is particularly helpful in this endeavor. In Mouffe's view, we can ‘use Wittgenstein's insights to undermine Habermas's conception of procedure’ as well as challenge his idea of dialogue. ⁵ In fact, she even claims that Wittgenstein's view is ‘undermining the very basis’ of the ‘universalist-rationalist’ form of reasoning that underpins the liberal democratic approach as a whole, defended not only by Habermas but also by Dworkin and the early Rawls. ⁶ Instead, Wittgenstein points to ‘a new way of theorising about the political, one that that breaks with the universalising and homogenising mode that has informed most of liberal theory’ as well as ‘renounce its [liberal] claims to universality’. ⁷

The normative source of this criticism is to be found in Wittgenstein's specific form of contextualism, according to Mouffe. In order to have an agreement or reach consensus, Mouffe argues, we must first have an agreement on the language used, which in turn implies agreement on what Wittgenstein famously calls a ‘form of life’. ⁸ In contrast to what liberal theorists allegedly claim, a political procedure cannot be seen as consisting of rules created on the basis of principles and then applied to specific cases. Rather, rules for Wittgenstein are always abridgments of practices and as such inseparable from the form of life in which they exist. ⁹ So, instead of wrongly presuming that there is a correct understanding of a rule that any rational person should accept, in line with liberal democratic theory, Wittgenstein's ‘contribution to democratic thinking’ can help us formulate and ‘valorise the diversity of ways in which the “democratic game” can be played’. ¹⁰ Incorporating into our democratic thinking that we are always held captive in what Wittgenstein calls our ‘language-game’, Mouffe argues, means to become more receptive of the multiplicity of voices in a democratic polity and allow them forms of expression, which is exactly what the agonistic model is able to accommodate. By contrast, universalist–rationalist theories are obstacles to this Wittgensteinian democratic vision since they ‘by necessity, tend to erase diversity’ in their attempt to establish ‘a rational consensus on universal principles’. ¹¹

Also Aletta Norval finds Wittgenstein very useful in theorising democracy. Being concerned mainly with the subjectivity of participants in a democratic process, Norval finds liberal democratic theory such as Habermas' deliberative model impotent, since it focuses too much on rational argumentation, reason-giving and procedures that are supposed to bring about desired outcomes and has very little to say about subjects' political participation. Apart from making use of Wittgenstein's ideas of ‘language-game’ and ‘form of life’, in line with Mouffe, Norval also applies his concept of ‘aspect change’ in her criticism of contemporary democratic theory and in developing her own so-called aversive model. Wittgenstein's approach offers a way of understanding the inaugurating moments of democratic subject formation and the practices involved in its maintenance, ¹² since it ‘resists the idealised theorising and disregard for ordinary political activities’ found in much contemporary political theory, which drags a sedimentary conception of subjectivity with it. ¹³ In contrast to democratic theories that accommodate an abstract conception of the subject tied to what Norval calls the ‘cognitive model of the acquisition and accumulation of facts’, aspect change is not about changing one's way of looking at an object, but about a shift in perspective that establishes a different relationship between objects. We do things differently once we have experienced a change of aspect. ¹⁴ And, Norval continues, ‘as with language, so with politics’. ¹⁵ The human voice is not a preconstituted mode of political subjectification, expressed in terms of reason-giving practices through a set of procedures. Rather, a person both discovers her political convictions and changes political identification through understanding herself in relation to a set of wider practices, in which her political voice is exercised in multiple ways, e.g. through passion, persuasion, rhetoric and contestation. ¹⁶ Being based on a model of rational reason-giving, deliberative democracy is not equipped to address these concerns. ¹⁷

In his attempt to develop a political philosophy ‘in a new key’, by which is meant a public philosophy as a critical activity, James Tully makes use of Wittgenstein's ideas of ‘aspect change’, ‘rule-following’ and ‘family resemblance’. For Tully, public philosophy has the capacity to bring to light relations of oppression and injustices and simultaneously work as an emancipatory practice by giving voice to exploited and oppressed. ¹⁸ It is a ‘critical attitude’ rather than a transcendental Kantian or Habermasian doctrine, which can ‘test and reform dubious aspects of the dominant practices and form of problematisation of politics against a better approach to what is going on in practice’. ¹⁹ Instead of taking a transcendental route, which presumes that we must (and can) identify a set of general principles for undertaking this emancipatory and critical activity, we provisionally follow conventional boundaries as we try to reach an agreement on something in our language-games of critical reflection, questioning and altering the rules of the game as we go along. ²⁰ Since all judgments are made within unique language-games in which the ‘testing of true and false, just and unjust’ takes place, ‘there are countless ways of studying politics and no universal criteria for adjudicating among them’. ²¹ However, even if there is no fixed normative framework of higher-order principles or norms of the kind that is presumed by timeless transcendental philosophy, subjects are still free to criticise every norm, according to Tully. ²² Once we follow Wittgenstein and release ourselves from the convention that we are rational and free only to the extent that we can justify the grounds of any uses we follow, Tully argues, ‘we can see that there is a multiplicity of ways of being rationally guided by rules of use’. ²³

2. The nature of language and norms

Against the backdrop of this outline of different uses of Wittgenstein among p-theorists, let us return to the gap mentioned in the introduction – between pragmatically influenced theories of language and meaning and normative political theory – which underpins our thesis. In order to show that p-theorists draw too strong conclusions about the implications for normative theory from Wittgenstein's presumptions about language and meaning, we approach this gap in a threefold manner, examining in separate sections how ontological, epistemological and semantic aspects of Wittgenstein's view are used for normative purposes.

But before doing so, let us first respond to the possible worry of readers with sympathy for p-theory that the argument pursued misses its overall aim by treating p-theorists as if they are engaged in ‘traditional’ forms of normative theorising, often described in terms of first-order (or substantive) normative theory. To assume this would be to misconstrue their ‘therapeutic’ or meta-theoretical purposes – a purpose wholeheartedly in line with Wittgenstein, who believed a substantial part of his contribution to be therapeutic. ²⁴ Indeed, it is true that p-theorists seem to be skeptical of certain forms of first-order theorising, but we must not let labels get in the way of the problem at hand. First of all, also theorists labeled ‘traditional’ in this way may overextend their use of theories of language and meaning for normative purposes. ²⁵ More importantly, what we are interested in here are the ways in which Wittgenstein's presumptions are utilised to draw substantive normative conclusions about the weaknesses and strengths of specific normative political theories. And to claim that a theory is ‘bloodless’, ‘misguided’, ‘flawed’, ‘mistaken’, ‘problematic’ and ‘undermined’, as p-theorists do, is clearly to move away from a therapeutic and meta-theoretical exercise towards a substantive normative one.

Now, let us take a look at what p-theorists claim with regard to how ontological concerns – concerns about the very nature of language – limit the scope and validity of principles and norms. Mouffe, for example, argues that Wittgenstein's insight ‘undermines the very objective’ of aiming at universal principles and renounces liberal ‘claims to universality’. ²⁶ Since rules are inseparable from the form of life in which they exist, liberal democratic principles or institutions ‘do not provide the rational solution’ but are only ‘defining one possible political “language-game” among others’. ²⁷ Similarly, Tully claims that Wittgenstein's Rule-Following Considerations and remarks about family resemblance demonstrate the futility of ‘developing a definite theory’. ²⁸ Since family resemblances among uses of a concept change over time, ‘understanding political concepts and problems cannot be the theoretical activity of discovering a general and comprehensive rule’, according to Tully. ²⁹ Interestingly, also John Gunnell – who has forcefully criticised attempts to use Wittgenstein for political–theoretical purposes ³⁰ – argues that Wittgenstein ‘subverts the search for the universality of both politics and political inquiry'. ³¹ Thus, for p-theorists, a political theory that puts forward generally applicable and universal conditions for justice, fairness or legitimacy seriously misunderstands the limitations that our different forms of life put on our theorising.

The claim that Wittgenstein's thought in general, and his Rule-Following Considerations in particular, preclude universal principles or rules is based on conflating two very different aspects: the prerequisite for understanding a principle (or rule) and the content of the principle itself (including its scope). In his Rule-Following Considerations, ³² Wittgenstein's target is the idea that a principle (or rule) wears its application ‘on its sleeve’, like a Platonic rail leading onwards into infinity, deciding what is in accordance with the principle and what is not. ³³ He contends that such an image is a chimera. A principle only has determinate meaning given a background practice or custom, a way of applying it that is implicit in practice rather than explicit in the form of further background rules (since this would lead to a vicious regress). ‘Outsiders’, people who do not share enough of the background practice – the web of shared facts about behaviour and sensations that Wittgenstein calls a form of life – will thus not be able to grasp the principle in question. ³⁴

This point about how a linguistic community is grounded in a shared practice is about the prerequisite for understanding a principle or rule, and not about the universality of its content. This is evident from the very example Wittgenstein uses to bring home his point, namely, the correct application of mathematical terms. Wittgenstein imagines a person who is taught addition by instructions and a number of exercises, which he solves to the great satisfaction of his teacher. But when asked to add 2 to 1000, a number larger than any he had until then encountered, the pupil responds ‘1004’, and continues the sequence with ‘1008’, ‘1012’ and so on. ³⁵ The teacher complains that he was supposed to add the number two (not four), to which the pupil responds:

‘Yes, isn't it right? I thought that was how I was meant to do it’. – Or suppose he pointed to the series and said: ‘But I went on in the same way’. ³⁶

Now, Wittgenstein emphatically exposes the idea of a rule entailing its own application as a chimera. But he does not in any way question that the rule for addition is universally applicable. He is not suggesting that ‘plus’ means anything other than the universally applicable rule for addition that we are taught in school. Rather, Wittgenstein's point is that a background practice is essential for the very notion of acting in accordance with a principle to get any grip at all. ³⁷ Hence, the Rule-Following Considerations are about understanding a rule – about the prerequisite of knowing ‘how to go on’ – not skepticism of there being a definite answer to what this amounts to. Consequently, the Rule-Following Considerations cannot be used to argue against the possibility of universal principles. ³⁸

Just as Wittgenstein's Rule-Following Considerations do not preclude universal rules in the political sphere, family resemblance does not entail this either, pace Tully's insistence. ³⁹ In his remarks about family resemblance, Wittgenstein points to the plurality of games (board games, Olympic games, etc.) and notes that rather than sharing some essential property in view of which they earned the label ‘game’, they seem to have similarities with each other – family resemblances. ⁴⁰ While this may rightfully be interpreted as skepticism against believing that all analyses of meaning must involve necessary and sufficient conditions for the application of terms, it neither precludes specific terms from having such conditions (e.g. ‘+’) nor implies that principles cannot be universal. If anything, it says that principles containing terms that cannot be codified by necessary and sufficient conditions may be vague in the sense that it might sometimes be unclear whether they apply or not. But the fact that principles, however general, need judgment to be applied does nothing in itself to undermine suggested theories of justice or democracy (Rawls, Forst and Habermas), however universally applicable.

These erroneous conclusions about the possibility of certain types of principles may be traced to a misconstrual of the function of language-games. No one could reasonably reject that people have to share a language-game in order to understand the meaning of a principle, norm or value, whether it concerns Tully's civic freedom, Mouffe's democratic procedure or Rawls' principles of justice. To this end, liberal principles do not ‘define’ one possible political language-game among others, as Mouffe claims. ⁴¹ Rather, they take form and get meaning within a language-game. Wittgenstein is concerned with the preconditions for meaning. But since the point that we would have to share a form of life in order for a rule to be intelligible applies to all rules (contextual, particular, universal, etc.), it is a mistake to believe that liberals cannot offer ‘a correct understanding’ of what constitutes a good regime because it is only one possible language-game among many. There is no contradiction between offering a correct answer and accepting that the language-game needed to understand it is only one among many. So, if it is true that liberals do not offer a correct answer, as Mouffe sustains, it is not due to the multiplicity of language-games.

3. Contextualism, perception and norms

In the previous section, we demonstrated that Wittgenstein's account does not include any ontological ban on rules or principles with universal application, and hence that it is a mistake to attempt to utilise the Rule-Following Considerations against traditional political theory which often comes in form of such principles. In addition to this negative (prohibitive) argument against general and universalistic accounts, p-theorists have launched a more positive set of arguments for how Wittgenstein's thought suggests alternative accounts of political theory. This section looks at two such suggestions, contextualism and aspect change, which highlight how p-theorists claim that Wittgenstein's thought has important epistemological consequences for political theory.

Let us start with Wittgenstein's contextualist focus. There is a multitude of language-games in place – by which Wittgenstein meant more or less distinct, purposeful activities of situated agents, both linguistic and non-linguistic in nature ⁴² – which makes pluralism an essential feature of language. Each language-game gives rise to a logic of its own – what it means to follow the rules of that game, etc. – and we cannot therefore expect different games to function in the same way. It follows from this that it is in the very context in which we, as situated agents, are placed that every formulated rule and principle must be interpreted. The terms used in one language-game can mean something else in another. ⁴³

We find these Wittgensteinian insights convincing. Our objections are raised when p-theorists move from these observations to a criticism of the traditional way of developing political theory in general, and of other theorists in particular. In the view of p-theorists, Wittgenstein's thoughts about the plurality and context-sensitivity of language lie in opposition to the principalist theories of mainstream political theorists; in particular, the idea that political norms have the form of general principles valid across many different contexts. Political philosophers like Rawls, Habermas and Forst act as if general answers are possible to reach in the political arena, something that is not compatible with our incomplete knowledge. For example, Tully accuses Forst of defending ‘unquestionable higher-order norms’, ⁴⁴ and Mouffe accuses liberal universalists of defending ‘the rational solution’ and calling on general rules and principles. ⁴⁵ Moreover, the fact that such principalist accounts are supposed to apply to all subjects and are argued for on rational grounds also flies in the face of p-theorists. Given the diverse phenomenon of politics, we simply cannot expect to find reasons for one single set of principles of justice or legitimacy. Appreciating the pluralistic nature of language means to move away from such context-free accounts of normative principles. According to Tully, for example, Wittgenstein teaches us that ‘actual criteria for the application of a general political term are too various, indeterminate, and hence open to unpredictable extension to be explicated in terms of an implicit or transcendental set of rules or theory’. ⁴⁶

However, this critique relies on a false dichotomy between (good) contextual rules or principles and (bad) general or universal rules or principles. The main idea underlying the reasoning of p-theorists seems to be that since general principles suggested by traditional political theorists are context-free, they fail to apply to a particular language-game constituting a specific political context. Instead, we must construct context-specific rules for particular societies with their particular language-games. But it cannot be that general principles are context-free and require additional judgment whereas context-sensitive principles already come with a full context and without the need for judgments. As Wittgenstein has taught us, all principles require judgment when they are to be applied. All actual decision-situations are descriptively open-ended and for every rule, however contextual, we will always have to use judgment to apply it in a situation. In other words, pointing to context-sensitivity does not settle the matter. ⁴⁷

Moreover, for most normative political theorists, contextual information about, say, actual institutions and the aims and values of citizens are important for which principles or ways of doing things ought to govern a particular society. In fact, as we have pointed out elsewhere, ⁴⁸ that the context matters for what we should do is a basic premise also in moral philosophy, which is typically viewed as paradigmatically universalistic. ⁴⁹ Furthermore, moral theorists have long stressed the need for judgment. Even Kant, the father of universal principles, was aware that no matter how complete and convincing a normative theory might be, it cannot offer an immediate link between principle and application. The application of normative principles requires judgments, assessments and possible trade-offs against the backdrop of a specific social context in which the action is supposed to take place. ⁵⁰ Hence, it is a significant and explicitly stressed property of normative principles that they do not ‘apply themselves’. ⁵¹ It is for this reason that a theorist such as Habermas stresses that an application of highly abstract principles to complex situations requires that these situations are described in all relevant respects as appropriately as possible. ⁵²

Another epistemic aspect of Wittgenstein's view commonly elaborated by p-theorists pertains to perception and connects to his idea of ‘aspect change’. ⁵³ Norval, for example, claims that one reason why liberal and deliberative approaches are not equipped to offer an account of democratic agency is precisely because they do not take aboard Wittgenstein's idea of aspect change, which concerns how subjects are able to see things in new ways through persuasion, contestation and passion. Wittgenstein makes a distinction between ‘seeing’ and ‘seeing-as’ in his elaboration of the notion of aspect change. The famous picture of the duck–rabbit is meant to illustrate that while we may perceive of this picture as a duck or a rabbit, we cannot make sense of this difference in terms of a forthright perceptual story about two distinct objects such that the difference in the two experiences arise from or is caused to occur to us through the object in the material world itself (the duck–rabbit). Rather, it is better understood as a difference in how the subjects respond to what they see, in their particular contexts of perception. ⁵⁴ One of them may respond that he saw the duck–rabbit as a rabbit.

In Norval's view, aspect change is useful for understanding democratic subject formation. ⁵⁵ The experience of aspect change is a moment of subjective assent that involves a process of identification – a picture grabbing hold of us – that cannot be accounted for by the cognitivism stressed by mainstream democratic theories. For it is not an identification pure and simple, but an ‘identification-as’. ⁵⁶ Since an identification-as (e.g. ‘I am a democrat!’) is an embodied act of passionate engagement, we cannot make sense of it through the linguistic reductionism and rationalism of these theories, according to Norval. ⁵⁷ She illustrates this point with Habermas' deliberative model (but claims that ‘Rawlsian accounts suffer from similar problems’), ⁵⁸ which puts forward ‘a bloodless conception of participation’ because it relies on rational argumentation. Such proceduralism purportedly ‘forecloses the question of the identity-forming ethos of democracy’ since it ‘presupposes that we follow the rules “blindly”’. ⁵⁹

Recall the possible worry that the defence of our thesis may be misdirected since it misconstrues the aim of p-theory, which is not substantively normative but therapeutic and meta-theoretical. Indeed, this concern would be valid in Norval's case had she been faithful to her commitment to use Wittgenstein to ‘reorient democratic theory’ towards a ‘different set of questions and engagements’. ⁶⁰ But this is not all that she does. By pointing to the weaknesses of liberal and deliberative theories as well as to the strengths of her own aversive democratic theory, Norval clearly has normative pretensions: whereas the cognitive model ‘rules out understanding aspect change’, this can allegedly be accounted for by Norval's own theory. ⁶¹ The same goes for the model of rational reason-giving, which ‘forecloses’ questions about identity-formation. ⁶²

Using Wittgenstein's view in this way is precisely what we claim is to overextend its reach. To the extent that Wittgenstein is right, Norval is correct in noticing that the context matters not only for how persons perceive of something in the world but also for the possibility of perceiving something anew. Moreover, it seems plausible to presume, with Norval, that without any possibility of such experiences, the subjects involved would have difficulty engaging in a democratic dialogue. However, these are epistemological conditions for all subjects in all situations and say nothing particular about democracy or democratic agency. It is equally true of a case in which a subject participates in a political meeting and discovers (or rather takes to discover) through passionate engagement with the other members that his ethnic group is superior to all others by ‘identifying-as’ a racist (‘I am a racist!’). In other words, these general conditions do not offer the normative source for drawing any conclusions about the strengths of the aversive model of democracy defended by Norval or about the flaws of liberal and deliberative models.

More importantly, these conditions do not lend us any normative ammunition to argue against an overemphasis on reason-giving practices and in favour of passion and contestation. Habermas' subjects would presumably experience moments of aspect change and identification-as – before, during and after their formulation and reformulation of reasons – similar to Norval's subjects, insofar as they are epistemological conditions for all human practices and interactions, including, of course, the process of providing reasons (as also pointed out by Wittgenstein). ⁶³ Therefore, we cannot infer, as does Norval, that it is when giving reasons (i.e. providing grounds) comes to an end that ‘the act of identification, occurring during aspect change, assumes its relevance’. ⁶⁴ Because so far as they are epistemological conditions – as Wittgenstein sustains – identification and aspect change are with us ‘all the way’, as it were. In fact, the only way for Habermas’ deliberative participants to even begin to follow the rules ‘blindly’, to use Norval's terms, is to understand the context in which these rules apply.

4. Understanding and norms

The third part of the defence of our thesis focuses on the relationship between Wittgenstein's thoughts about semantic aspects such as shared meaning, understanding and belief on the one hand, and normative validity on the other. As we have seen, p-theorists argue in favour of specific normative accounts and against competing theorists by pointing to the limitations set by linguistic concerns. One of these pertains to the limitations of understanding. For example, in elaborating democratic procedure, Mouffe argues against Rawls and Habermas that democratic procedures are complex ensembles of practices rather than rules that are created on the basis of principles. To agree on a procedure, she claims, requires agreement on the language used. ⁶⁵ In Tully's view, resonating Mouffe's criticism, ‘Habermas's mistake’ is precisely that his priority of claims to rightness over other kinds of claims is itself grounded in a convention of a particular language-game, namely the right in certain circumstances to ask for reasons. ⁶⁶

But this line of argument conflates the question of whether a principle (rule, norm) is valid (true, correct) with the question of understanding its content. To say that a person cannot understand a normative principle only says that she does not share the necessary preconditions, in terms of appropriate lifeworld or language-game, for the principle to become intelligible. Consider, for example, the rule ‘You should always divide a cake equally’. If one society uses the term ‘equal’ in terms of merit, such that an equal division of a cake is one in which people get pieces in relation to what they deserve, and if ‘cake’ in this society signifies petite madeleines and only petite madeleines, the rule means something different for people in that society than it does for us. The possibility of agreement as well as disagreement about the content of a principle is premised if not exactly on ‘agreement on the language used’ (Mouffe), at least on the possibility of systematically being able to translate others’ utterances. ⁶⁷ But this commonsensical prerequisite on successful communication reflects only the semantic sense of ‘agree on a procedure’, that is to say, to agree on what the principle means. We can agree on this while at the same time strongly disagreeing about the validity of the principle. As long as we can grasp the statements of a speaker, in our own language or via translation, agreement on the content of a principle is possible. If ‘equal’ in the mouth of the speaker signifies ‘merit’ to us, we may understand that the linguistic item ‘You should always divide a cake equally’, uttered by the speaker, means the same as the linguistic item ‘You should always divide a cake according to merit’, uttered by us.

However, this tells us nothing in relation to the validity of the principle thus grasped. The question of validity of a principle is in practice a matter of whether there are good reasons for it. And whatever our position is concerning theories of meaning, we are all in the same boat when it comes to the practice of justification. Being able to engage in such a practice is premised on the possibility of being able to communicate the content of the principle by following appropriate conventions (or by sufficiently understanding how the interlocutors' idioms deviate from such conventions). Indeed, sometimes this practice is undermined, but it would be a very controversial (and non-Wittgensteinian!) claim of p-theorists to assert that communication about the propositional content of principles is impossible between different societies. While these Wittgensteinian linguistic insights set the limits for the discussion, they do not point in favour of any particular normative theory.

This point has important implications in relation to values. When p-theorists argue that the diversity of values following from the diversity of language-games implies the futility of mainstream principled accounts of justice and democracy, they again conflate two distinct concepts: values held and values we ought to hold. Indeed, it seems to be a commonsensical fact that people in different societies or cultures not only use a different language for expressing their values but also that the content, the held values, differ. But this fact entails neither that they have to have these values nor that they should have these values. Normative political theory is foremost about how we should live and arrange our societies, and as with all (practical) normativity, it is an open question how and to what extent (current) facts matter for what we should do. ⁶⁸ To give an example. The Hmong people believe that children with epileptic seizures should not be treated with medicine. ⁶⁹ This is a normative belief, and it corresponds to a perceived value, namely the potentially beneficial effect of being spiritually possessed, which they believe the seizures indicate. Now, we may acknowledge that the language-game of the Hmong people include references to spiritual forces and beliefs about their causal interaction with living people, and agree that within this system of belief, the value of withholding medical treatment for seizures makes sense. We may thus understand the beliefs and values of the Hmong people. But this does not of course establish the truth or correctness of these beliefs and values. While the Hmong take there to be a value with withholding treatment, there is in fact no such value. The perceived value is based on erroneous beliefs. ⁷⁰

One tempting objection to this line of reasoning may be that we are here simply assuming that some – our – values and beliefs are true or correct, whereas other – their – values and beliefs are not, precisely highlighting the Wittgensteinian point that there is no neutral point of view, but only different language-games, different ways of perceiving the world and our place in it. The claim would thus be that the futility of mainstream political philosophy lies in believing that there is a transcendental standpoint from which we may prove that some values are true and others are false. However, this objection misses the above point about the validity of values. In arguing for a principle or value, we give what we take to be the most convincing reasons for it. Naturally, such reasons, be they factual or evaluative, are in turn based on assumptions, most of them implicit, and somewhere we hit rock bottom. Then, with Wittgenstein's famous quote, justification ends and we say: ‘This is simply what I do’. ⁷¹ But this anti-foundationalist idea of justification is in perfect alignment with the view of contemporary political theorists, who do not believe there to be an Archimedean point of reference on which we may rely. This is the entire point of a coherentist approach such as Rawls' reflective equilibrium. ⁷² And Habermas' insistence on steering away from what he calls ‘the paradigm of the philosophy of consciousness’ is meant to avoid such foundationalism, which in his view ‘conflicts with our consciousness of the fallibility of human knowledge’. ⁷³

Naturally, in order to convince a Hmong of the falsity of her non-treatment value, we would have to convince her of the incorrectness of her beliefs about seizures indicating spiritual possession, which may be a very difficult task since we are potentially questioning an entire system of belief. But in principle, the situation is the same as when someone is convinced that drinking a glass of liquid is beneficial since he believes it is nutritious nectar, when in fact it is poisonous. Since truth is not transparent, when we say ‘in fact’, we are always implicitly alluding to something like ‘…according to my best understanding of how it is’. In that sense, we never move beyond ‘held values’ or ‘held beliefs’. But a central point of making claims of value or fact explicit is to make them available for scrutiny, that is, for giving and asking for reasons. This may not be the only way to change people's views, but it is one important way. Certainly, it is practically impossible to make all implicit assumptions explicit. But this is true across the board, and if the progress of science, both natural and social, has shown us something, it is that despite of the lack of transparency of truth, we may change our views in ways that make society better, both technically and socially.

Consequently, while our justification always ends at some point, ‘this is simply what I do’ entails neither that it is what I must do nor that it is what I should do. Political principles may be questioned, but they may also be convincingly argued for. In this sense, Wittgenstein's point about the multitude of language-games, and the corresponding point about forms of life, is descriptive rather than normative: it shows how language functions. ⁷⁴ Successful communication is possible as long as we share enough of a lifeworld. As long as this is the case, we may argue with each other about alternative ways of doing things. Most importantly, there is nothing in Wittgenstein's thought that stops the theorist from arguing that some ways of life (in this narrow sense) are unacceptable and therefore ought to be changed.

5. The scope of the argument

To sum up our argument thus far, we have approached our thesis in a threefold manner to elucidate the different ways in which p-theorists fail in their attempts to make use of Wittgenstein's view of language and meaning as a normative source to dismiss certain theories and support others. As we have seen, the problem of overextension that follows from not ‘minding the gap’ is a widespread phenomenon among different theoretical approaches. Still, our thesis may seem like a bold conjecture considering that our argument is based on how one view of language and meaning is used in normative theory. However, it is motivated by the fact that structurally, our arguments have been directed at general features of theories of language and meaning rather than Wittgenstein in particular.

The general form of our argument is demonstrated if we go beyond Wittgenstein to other pragmatically influenced philosophers of mind and language who have been called upon by p-theorists. Let us illustrate this with Thomas Fossen's use of Robert Brandom's philosophy of language and meaning. ⁷⁵ In two recent papers, Fossen argues that Brandom's account has political implications. ⁷⁶ In his seminal book Making It Explicit, Brandom develops a socio-pragmatic theory of mind and language where meaning is grounded in social practice. On his explicitly normative account – where it is ‘norms all the way down’ ⁷⁷ – each participant is interpreted as keeping a ‘book’ on the commitments and entitlements they take the performances of participants (themselves included) to involve. Such ‘deontic scorekeeping’ is essentially perspectival, enabling each participant to make sense of important distinctions, such as that between what someone believes to be true and what is true (all according to the participant/scorekeeper). In Brandom's view, this scorekeeping contains the resources to ground an account of meaning on practice (rather than, say, on truth or reference). ⁷⁸

Fossen suggests that Brandom's theoretical framework ‘provides fertile soil for a social-pragmatic approach to social and political philosophy’. ⁷⁹ Specifically, he argues for two political consequences: agonal contestation and an alternative approach to political legitimacy. Since both face the same general problem, we will focus on Fossen's approach to legitimacy. While Fossen's aim is not to develop an alternative normative account of political legitimacy as such, he claims that his pragmatist account, ‘a theory that explicates the use of “legitimacy”, rooted in an account of the social practice in which the concept has a distinctly political point and purpose’, ⁸⁰ puts constraints on such normative accounts and thus suggests ways in which they should develop, hence establishing a ‘genuine alternative’ to traditional accounts in political philosophy. ⁸¹ In particular, Fossen focuses on the perspectival aspects of Brandom's theory, how calling an authority legitimate or illegitimate is taking a stance in a linguistic practice, a move that may be disputed or endorsed by others. He argues that this pragmatist perspective puts several anti-theoretical and anti-generalist constraints on accounts of political legitimacy. For example, Fossen claims that ‘recasting the predicament in a pragmatic way enables us to frame political judgment as an on-going task that calls for practical engagement, rather than a philosophical problem calling for a general solution’, ⁸² that ‘the legitimacy of an authority cannot be determined with certainty, definitively, or from a disengaged standpoint’, ⁸³ and that ‘distinguishing in practice between what is legitimate and what merely purports to be so is treated as a lived experience that can be made explicit in critical moments; it is not an abstract question calling for a general solution’. ⁸⁴

The problem for Fossen is that these claims are based on a faulty inference from the practice of treating something as legitimate, which is certainly a practical engagement in which claims are and should be contested for all sorts of reasons, to the conclusion that normative accounts of legitimacy cannot be general and just as definitive as claims may be in other areas of discourse. This is faulty because the model of deontic scorekeeping and stance-taking that Fossen applies to political legitimacy is altogether general. On Brandom's account, all meaning is conferred by the practice in which our performances are to be found. So while Fossen is completely right in stressing that claims about what is legitimate are essentially perspectival – they are correct according to the speaker (when uttered sincerely) – this perspectival feature of assertion is not special for legitimacy-claims, but goes for all concepts on Brandom's account. So when we make claims concerning natural kind concepts (those figuring in our explanation of scientific phenomena such as mass and water) as well as concerning mathematical concepts (such as addition or imaginary numbers), these are just as perspectival and practically engaged, in Brandom's sense, as claims concerning political concepts. Hence, if Brandom's deontic scorekeeping model entailed the anti-theoretical conclusions that Fossen envisions it does in the case of legitimacy, it would do so concerning mathematical and natural science concepts too. Importantly, however, Brandom's account does not contain any such claims. For him, there is nothing troubling with making explicit theoretical commitments that are general or universal in scope. On the contrary, he puts forward a systematic theoretical account of meaning and content that definitely has general pretensions.

Similarly, Fossen's claim that legitimacy ‘cannot be determined with certainty, definitively, or from a disengaged standpoint’ fails to prohibit or to suggest any type of normative theory of legitimacy, since this is true only in the sense in which all knowledge claims – even those of natural science – are fallible on a theory such as Brandom's. It may in principle always turn out that we were wrong even in what we took to be our most justified beliefs. But just as this fallibilism does not put constraints on our theories in natural science, or suggest that they take on any particular form, it fails to do so in political philosophy as well.

Consequently, Fossen's problem is exactly that which we have seen throughout the paper: the aspects of Brandom's philosophy of language that Fossen utilises are perfectly general, and will not lend themselves to any practical-normative conclusions. Fossen shows in detail how Brandom's account may make sense of the distinction between what is legitimate and what is merely taken to be so, in terms of the intricate socio-pragmatic tools he has developed. But this Brandomian story, while being seminal in the sense of offering a new way of understanding what we do (i.e. how meaning is conferred through practice), is not telling us to do anything differently from what we already do. If Brandom is right, we already, all of us, confer meaning through these score-keeping and stance-taking practices. So when Fossen discusses how important various forms of contestations are for disputing entitlement to authority, and points to the many forms they may take in politics – political ridicule, marching in protest, desecration of public symbols and so on ⁸⁵ – he is making explicit how all competent linguistic practitioners learn and develop their skill with political terms. But insight into our practices with political terms does not in itself say anything substantive – neither in terms of meta-theoretical constraints nor in terms of first order normative conditions – about when a state of affairs is, in fact, legitimate.

The upshot is this: given that Brandom's framework supplies a good model of how we recognise or reject a claim (any claim), it establishes a good model of how we recognise or reject claims of political legitimacy. As such, it may spell out in more detail how, to use a favourite expression of Brandom's, ‘the game of giving and asking for reasons’ concerning political legitimacy is played. But this is on a general level of linguistic competence, and as such it is completely neutral vis-à-vis any substantive normative account of political legitimacy, regardless of the form of such an account (general principles, contextually situated claims and so on). Hence, the attempt to turn Brandom's philosophy of mind and language into normative political theory fails for exactly the same reasons as the use of Wittgenstein fails: they are both views on language and meaning and as such neutral as to which normative political theory has the best arguments in its favour.

Conclusion

We wish to conclude by underscoring that it would be premature to infer from our thesis that theories of language and meaning are of no use for normative political theory. ⁸⁶ Quite the opposite. Wittgenstein and his fellow philosophers with a pragmatist outlook stress the importance of acknowledging that language is an extremely diversified phenomenon such that linguistic and pragmatic rules concerning one sphere (language-game) may not apply to another, and they make clear how justification may take many different forms. This demonstrates the large space available to us for problematising distinctions such as between particularism and principalism, different forms of universalism, abstract and concrete principles and so forth. We should just be careful not to turn this possibility of options into the necessity of any one option.