Abstract
It is often argued that Michael Walzer’s theory of social criticism, which underpins his theory of justice, is not much of a theory at all, but rather an impressionistic collection of historical anecdotes. Contrary to this perception, I argue that Walzer’s method can be accurately described as a version of John Rawls’ well-known method of wide reflective equilibrium. Through a systematic comparison it can be shown that the two methods are strikingly similar. This implies that, far from the critics’ claim, Walzer’s method can be described as a philosophically sophisticated method. This also adds credibility to Walzer’s views on politics and justice.
I
There is a widespread perception that Michael Walzer never cared to develop a philosophical method that can lend justificatory credence to his normative views. Instead he relies heavily on a hermeneutical mix of anthropological and historical anecdotes, or so the critics say. In this article, I will challenge this perception, and I will argue that Walzer uses a well-known and respected method: an equivalent of John Rawls’ wide reflective equilibrium (WRE). This implies that Walzer, contrary to common criticism, can claim a philosophical underpinning to his views on politics and justice.
Walzer’s theory of social criticism is formulated most clearly in his Interpretation and Social Criticism (1987) and The Company of Critics (2002 [1988]). 1 The method of social criticism is best described as a hermeneutical, or interpretative, method of gaining moral knowledge. According to Walzer, a normative standard already exists within a community, but it needs careful interpretation. In each community, morality is locally determined by social meanings; that is, the ways phenomena are perceived in a community. For instance, a loaf of bread can be perceived differently in different settings; that is, its social meaning varies between communities. In an affluent community bread is perceived as an appropriate item to be sold on the free market; in a deeply religious community, where the bread is understood to be the Body of Christ, pieces of the bread are to be equally distributed among the members of the community. 2 Interpreting the social meanings in a community means exploring the already existing web of moral concerns and values, deeply embedded in the shared life of the community.
Hermeneutical approaches to morality are usually criticized for being conservative of already embraced values and practices. It is therefore important to note that Walzer’s interpreter does not simply read the moral standard off the voiced opinions and common practices in the community. Instead, the interpreter must dig deeper. The interpreter – the social critic in Walzer’s terminology – reminds us of our community’s shared moral standard and tries to ‘show us as we really are’: in need of a moral make-over to become what we are supposed to be like (Walzer, 1994: 42; original emphases; 2002[1988]: 231). Moreover, the critic’s job is not only to reveal that there is such a standard, but – if necessary – to revise it (Walzer, 1987: 29, 82). He or she should remove inconsistencies and obsolete components. In short, the Walzerian critic calls the people back to the path they actually, deep down inside, want to walk, and in doing so he or she is continuously (re)negotiating that very moral path.
An interesting feature of Walzer’s social criticism is the role given to the social critic. The social critic is a person connected to that community, but who is not found in its centre. She or he is a member, but at the margins of the community, distanced from its reigning power (Walzer, 2002[1988]: xiii, 26, 226). This means that the critic is familiar with the community and its traditions and values, that the critic speaks the ‘local language’, but that she or he nevertheless is in a position to criticize the oppressive powers in the community. True to the hermeneutic approach, the social critic speaks to the community without providing the final answers – the critic is engaged in a never-ending conversation with the fellow members and the critic’s interpretation of the community’s values and norms must by acknowledged by the rest of the community (Walzer, 1987: 30, 49).
The problem with Walzer’s idea of doing normative philosophy is not its hermeneutical character, however. Rather, the worry is that Walzer’s hermeneutical approach is arbitrary, anecdotic and unsystematic; that Walzer’s method is not really a method in any substantial sense, but rather the philosophical equivalent of impressionistic painting. Anthropological and sociological evidence are arbitrarily piled up so as to show whatever picture the interpreter wants to see or think that he or she sees. The critics claim that the numerous examples and historical anecdotes, which are prominent in Walzer’s seminal work Spheres of Justice (1983) , are never explored or used in a systematic fashion. To quote Chris Armstrong, ‘there is little doubt that methodological issues accounted for a substantial portion of the hostile interrogations the work attracted’ (Armstrong, 2000: 87). Ronald Dworkin, in his now classic review of Spheres of Justice, 3 writes that Walzer provides us with only ‘anecdotal and historical examples’, rather than presenting a coherent theoretical framework that allows for normative criticism. Similarly, Joseph Raz has argued that Walzer’s approach to normative philosophy is inherently flawed and that it is ‘unclear what he refers to when he speaks of moral arguments being “an interpretation of the morality that exists”’ (Raz, 1991: 394). Quite symptomatically, Brian Barry has insisted on the similarity between Rawls’ and Walzer’s conclusions but nevertheless claims that Rawls ‘scores over Walzer most decisively … by having a theory of justice within which these [distributive] criteria can be articulated and defended’ (Barry, 1995: 79). 4 It is thereby hinted that Rawls has a philosophical underpinning to his approach while Walzer does not. It is this picture – the reading of Walzer as lacking a decent philosophical method – I will challenge below.
By underlining the similarities between Rawls’ method and Walzer’s, it can be shown that the accusation of philosophical impressionism is false. Rawls’ method is usually seen as an important advancement in moral philosophy, and by making explicit the similarities between Walzer’s social criticism and Rawls’ WRE, Walzer can be rescued from the criticism saying that he lacks any real philosophical method of how to conduct the interpretation of a community’s morality.
Walzer’s social criticism is seldom mentioned as belonging to the same family of methods as Rawls’ WRE, however. Quite often, Walzer and Rawls are frequently depicted as each other’s adversaries: 5 Rawls as the epitome of a liberal and Walzer as the socialist communitarian; the former influenced by economics and psychology, the latter by history and anthropology. Attempts to place the two in the same camp are likely to cause suspicion. To some extent, Rawls and Walzer are themselves to be blamed for such polarization. Rawls claims that his project starts where Walzer’s has come to a dead end (Rawls, 1996[1993]: 44–5), and Walzer – when asked if there is a convergence between his and Rawls’ position – states that while there are few differences in actual moral opinions, there is a ‘fairly deep intellectual difference’ on the theoretical level (Carleheden and Gabriëls, 1997: 115). 6 So the persistence of the perceived divide between Rawls and Walzer should come as no surprise.
Nevertheless, I will argue that claims stating deep differences between the two do not survive any closer scrutiny. Below, I will defend the claim that Rawls and Walzer use the same kind of philosophical method of doing normative philosophy – that Rawls’ WRE and Walzer’s social criticism is alike and can be said to belong to the same family of philosophical methods. Such claim is perhaps even more controversial than any claim stating similarities in general. While Walzer’s interpretative method has received more than its fair share of criticism, Rawls’ reflective equilibrium has been hailed as a milestone in ethics. Still, I will present a plausible case showing that the methods overlap to a considerable extent, bringing the two considerably closer to each other. I will start by taking a brief look at the comparisons made, where Walzer’s social criticism and Rawls’ reflective equilibrium (either narrow or wide) are actually mentioned as being to some extent similar. This is quickly done, since there seems to be very little in the way of systematic comparisons. Next, I will provide a brief description of Rawls’ WRE in order to elicit its core characteristics. Using the characteristics as a standard, I will then examine how well Walzer’s social criticism measures up. In the end, I conclude that there is a substantial overlap between the two methods – enough overlap to categorize Walzer’s interpretative method as a version of Rawls’ WRE, or vice versa.
II
A few commentators have seen parallels between reflective equilibrium and social criticism. An indirect connection between the two models is drawn by Lyle Downing and Robert Thigpen, who describe Walzer’s method of distilling the social meanings as a process of ‘reflective disequilibrium’, by which Walzer confronts us with ‘alien practices’ and thereby raises our ‘awareness’ (Downing and Thigpen, 1986: 454). However, Downing and Thigpen do not discuss the issue any further.
Joseph Carens describes Walzer’s method as one that
… begins with actual moral problems and moral judgements and moves back and forth between these particulars and general principles, paying attention in a critical and reflective way to the actual features of our moral world. (Carens, 1995: 48–9)
In a later article – ‘A Contextual Approach to Political Theory’ (2004) – Carens elaborates what he calls the ‘contextual approach’, which he associates with Walzer, and identifies some of the characteristics that we also find in the method of reflective equilibrium. But Carens insists, just as Downing and Thigpen do, on a notion of ‘reflective disequilibrium’ since he believes that the components are ‘mutually unsettling’ and because he doubts that the ‘dialectic ever reaches an end’ (Carens, 2004: 123).
Glen Stassen (1994: 379) briefly notes the similarities between Walzer’s method and Rawls’, but he never makes a systematic comparison. The same thing can be said about Seyla Benhabib (whom Stassen quotes), who also claims that Walzer’s method is similar to Rawls’ reflective equilibrium at least ‘in its intentions’ (Benhabib, 1992: 79). While she accurately describes Walzer as ‘[p]roceeding from shared views and understandings of certain goods’ and then as ‘refining, systematizing, making coherent, criticizing and replacing by a “better” understanding the common views’ (1992: 79–80), she nevertheless fails to present a systematic comparison.
The perhaps most prominent philosopher elaborating on the idea of reflective equilibrium is Norman Daniels. In his review of Walzer’s Spheres of Justice, he briefly makes a connection between Walzer’s method and a narrow reflective equilibrium; a less complex version than Rawls’ and his own (Daniels, 1985: 148). However, Daniels misses the similarities to the WRE, and never engages in a systematic comparison.
The writer who has most cogently elaborated on the similarities in method between Walzer and Rawls is probably Georgia Warnke (see Warnke, 1990a: 197–203; 1990b: 204–26; 1992). Warnke explicitly claims that ‘Walzer and Rawls share roughly the same interpretative approach to political theory’ (1990a: 198). However, while she beautifully expounds the interpretative components in their respective approaches, she does not specifically make a systematic comparison between reflective equilibrium and Walzer’s social criticism.
Last, Avner de-Shalit has also made, although very briefly, a connection between Rawls’ reflective equilibrium and Walzer’s method. De-Shalit names Walzer’s approach a ‘contextual reflective equilibrium’, as compared with Rawls’ individually centered version (de-Shalit, 2003: 140–1). But his characterization of Walzer’s method is not entirely correct. He writes:
In Walzer’s model, however, the intuitions considered are those of the philosopher’s and the community’s intellectuals, while the theories are those of the philosopher alone. (De-Shalit, 2003: 140; original emphasis)
De-Shalit finds this approach too narrow, and suggests what he calls a ‘public reflective equilibrium’, where the political theory ‘reflects the actual philosophical needs of the public’ (2003: 141). But Walzer’s method is, as we shall see, not that different from Rawls’ version, or from de-Shalit’s own. 7 Contrary to de-Shalit’s claim, Walzer’s method does not start out by, or is confined to, the intuitions and theories of philosophers. In fact, such an approach was strongly rejected in Walzer’s article ‘Philosophy and Democracy’ (Walzer, 1981).
We can conclude that some critics have indeed picked up on the similarities in method between Rawls and Walzer. However, we still lack a careful and systematic comparison. 8
III
Providing a decent characteristic of Rawls’ WRE is difficult, since Rawls never provides a full description of it. His comments, in A Theory of Justice (1999[1971]) and elsewhere, are always brief. When trying to provide a good sketch of WRE I will therefore consult other philosophers who have tried to reconstruct or analyse the idea. Two articles in particular will be used: ‘Wide Reflective Equilibrium and Theory Acceptance in Ethics’ (Daniels, 1979) and ‘What is This Thing called “Reflective Equilibrium”?’ (Petersson, 2000). 9
Daniels is probably the one who has provided the most thorough account of WRE. He describes it as
… an attempt to produce coherence in an ordered triple of sets of beliefs held by a particular person, namely, (a) a set of considered moral judgments, (b) a set of moral principles, and (c) a set of relevant background theories. (Daniels, 1979: 258)
By ‘considered moral judgments’ is meant moral judgments that a person is (1) ‘confident’ about and (2) that he or she would make under conditions where the risk of making mistakes is minimized; for example, being sober, calm and mentally healthy (Daniels, 1979: 258; cf. Rawls, 1999[1971]: 42; 2001: 29). 10 ‘Moral principles’ denotes principles that are perceived as having various degrees of ‘fit’ with our considered judgments – it is among these principles we are trying to find the most suitable one (Daniels, 1979: 258; cf. Rawls, 1974–5: 7). 11 By ‘relevant background theories’ is meant general and comprehensive theories concerning the human person, moral psychology and social interaction. In the account given in A Theory of Justice they appear to be descriptive: ‘general facts’ and ‘general laws and theories’ such as principles of economic theory, of social organization and the laws of psychology. However, Rawls never explicitly excludes the possibility of background theories involving evaluative components, and Daniels, in contrast, seems to assume that background theories contain normative elements (Rawls, 1999[1971]: 119; cf. Daniels, 1979: 259–60). 12 I therefore assume that relevant background theories can have normative components. 13
According to Daniels, in order to ensure that the principle(s) eventually produced in the equilibrium is not mere ‘reformulations’ or ‘accidental generalizations’ of the considered judgments, we must add an ‘independence constraint’. This constraint means that the relevant background theories must be epistemologically independent from our considered judgments; in this way additional support has been added in the outcome (Daniels, 1979: 259–61; cf. Rawls, 1999[1971]: 19; 1996[1993]: 45; 1974–5: 8). 14 The equilibrium is then characterized by coherence and mutual support (Rawls, 1999[1971]: 507). However, none of the components in either set is immune to revision: there is no such thing as a fundamental background theory, principle, or conviction (for instance: Daniels, 1979: 264–7; Rawls, 1999[1971]: 18; 1974–5: 8; 1996[1993]: 8, note 8, 45). An achieved equilibrium between (a) and (b) and (c) (see below) is not claimed by Daniels to guarantee an objective truth, but is merely evidence pointing in that direction – if there is such a truth at all (Daniels, 1979: 276–7; Rawls, 1951: 177, 190; 1974–5: 7, 9; 1996[1993]: 384; 2001: 31). 15 Furthermore, WRE is usually attributed to an individual (Daniels, 1979: 258, 281; cf. Rawls, 1999[1971]: 43–4; 1996[1993]: 28, 384, note 16; 2001: 31). Daniels claims that while it might be tempting to start with socially recognized considered judgments (consensus judgments), we would then base our reflective equilibrium on the assumption of an ‘initial convergence where there is no good reason to expect it’ (Daniels, 1979: 282). 16
Looking at Rawls’ WRE through the interpretations of Daniels, I believe it can be summarized as follows: A WRE involves three elements: a set of considered moral judgments a set of moral principles a set of relevant background theories. (c) is not entirely epistemologically dependent on (a).
17
None of the components within sets (a), (b) and (c) are internally – within the WRE – immune from revision. A WRE consists in an attempt to adjust (a), (b) and/or (c) in order to achieve coherence and mutual support. The method of WRE starts with the considered judgments of one individual.
Even when taken together (1)–(5) do not constitute a strict definition of WRE, but merely a characterization. In fact, we lack any accepted philosophical definition. Nonetheless, I will let the list of characteristics serve as a critical standard. Having said that, I will argue below that some minor adjustments among the characteristics might, after a closer scrutiny, be required to stay as close as possible to Rawls’ WRE.
IV
Having given a brief description of WRE, I will now compare the above account with Walzer’s social criticism. As a brief outline of social criticism has already been provided, I will not rehearse it here.
Criterion (1)
Turning to the first criterion that stated that a WRE has three components, it seems reasonable that Walzer’s social criticism should also be interpreted as using the same three components: considered judgments, moral principles and background theories. Such use is not obvious, however, even though Walzer describes the ‘text’ to be interpreted by the social critic as consisting of ‘values, principles, codes, and conventions that constitute the moral world’ (Walzer, 1987: 30).
The component that plays only an implicit role in Walzer’s method is the set of background theories. Nevertheless, they are there and do some real work. Two examples of background theories are found in Walzer's, Interpretation and Social Criticism: John Locke’s theology and Freudian psychoanalysis. According to Walzer, in Locke’s A Letter Concerning Toleration the ‘meaning of salvation’ is crucial; an idea anchored in a particular theology:
Locke does not speak here in the new language of natural rights; this is very much the old language of ‘salvation by faith alone.’ But his lines suggest how one might move from old to new – not so much by discovering rights as by interpreting faith, ‘inward sincerity,’ and conscience. (Walzer, 1987: 54) 18
Without the theological idea of salvation as both valuable and a matter of the heart, Locke’s argument is not likely to have been successful. By slightly altering – or reinterpreting – the theory of the humans’ relation to God, it became reasonable to adopt a principle of tolerance. Locke’s background theory – his idea of salvation – forced a change in the way the principle of toleration was conceived.
The second example is Freud’s psychoanalysis. Freud’s psychology, Walzer notes, was not only a theory, but a theory that had implications for how we understood our morality. Freud himself used it as a critique of the sexual morality of his time (Walzer, 1987: 65). Clearly this is a theory that – even though we might want to express reservations about it today – had an impact on our view of our moral principles as well as our considered judgments.
Walzer’s use both of considered moral judgments and of moral principles is arguably easier to spot. In an interview he explicitly states his use both of principles and of particular judgments: ‘I am still mixing arguments from principles and arguments from historical cases in the way that I did in those books [Just and Unjust Wars and Spheres of Justice]’ (Carleheden and Gabriëls, 1997: 113). Walzer often introduces principles but then rejects or revises them when they bump into considered judgments, and sometimes the considered judgments need to be revised in the light of moral principles in order to achieve an acceptable structure or consistency. In the preface to the first edition of Just and Unjust Wars, Walzer writes:
But that’s not to suggest that we can do nothing more than describe the judgments and justifications that people commonly put forward. We can analyze these moral claims, seek out their coherence, lay bare the principles that they exemplify. (Walzer, 2000[1977]: xxi)
In Spheres of Justice general principles such as equality, free exchange, desert and merit are discussed in the light of and in contrast to historical and particular cases and judgments. One such example is the discussion of affirmative action, where the considered judgment that the present situation of unequal outcomes is morally unacceptable clashes with the principles concerning the just distribution of offices (Walzer, 1983: ch. 5). 19 A second example is the doctrine of supreme emergency, where the previously accepted principles of jus in bello are contrasted with the conviction that an incommensurable evil must be stopped at all costs – perhaps an instance where Walzer fails to reach an acceptable equilibrium. 20 One last example could plausibly be made out of Walzer’s discussion in ‘Philosophy and Democracy’ of the Supreme Court judges’ task, where he describes the judges as – ideally – lending one ear to the philosophers and their principles, and the other ear to judgments of the community as expressed in the legal traditions (Walzer, 1981: 390).
Criterion (2)
According to this criterion, the set of background theories used in the WRE must include at least some component which is epistemologically independent from the moral principles and the considered judgments. This criterion seems fulfilled as a theoretical necessity, as the main theme in Interpretation and Social Criticism is that our moral convictions are poorly arranged and sometimes hidden beneath shallow judgments: we cannot simply adopt, as justified, principles that appear to have a ‘fit’ to such moral convictions, or vice versa. An interpretation without the independence constraint would be conservative and lack true critical force. 21 Interpreting epistemologically identical, or nearly identical, judgments or principles in the light of each other would then amount to an uncritical restatement of the existing judgments or principles. Moreover, it should be clear from the examples above – Locke and Freud – that the theories were, in the relevant sense, free-standing from the considered convictions and moral principles in the WRE.
Criterion (3)
Social criticism requires that all three of (a), (b) and (c) are open to revision. In Interpretation and Social Criticism it is made clear that interpretation is necessary (Walzer, 1987: ch. 1), implying that there is no fail-safe justification stretching from objective fact to moral proposition; every component needs careful reflection in the light of the other components; (a), (b) and (c) are all parts of a thick particularistic interpretation of morality that is not justified by some foundational component. Considered judgments and moral principles are, as pointed out above, adjusted in the social criticism. A mutual adjusting is explicitly mentioned and endorsed in Walzer’s short article ‘Political Decision-Making and Political Education’: ‘But if principles determine decisions, decisions in turn modify and refine principles: This is the way both law and morality change over time’ (Walzer, 1980: 170). Even the background theories are prone to change when confronted with the principles and considered judgments in the moral community, as the background theories have to be locally received and understood, and they are therefore not immune to revision (Walzer, 1983: 9; 1987: 39, 65). If this is the case, then no component in any of the three sets can be seen as inherently immune to revision; hence: (3).
Criterion (4)
According to Walzer, one of the purposes of social criticism is to provide a more systemized reading of our morality, and I believe this to imply avoidance of logical inconsistencies. 22 But coherence is more than logical consistency. Coherence, from the Latin cohaerentia, implies a ‘fit’ of some kind, a common thread tying the pieces together. 23 While coherence seldom plays an explicit part in Walzer’s arguments, he does mention coherence as a result of social criticism, in this case the Old Testament prophets: ‘But the coherence of Israelite religion is more a consequence than a precondition of the work of the prophets’ (Walzer, 1987: 82; also, see Walzer, 2000[1977]: xxi, 19).
When it comes to the idea of the components providing ‘mutual support’ or as having the ‘best fit’ to each other, which I regard as a stronger version of coherence, it gets a bit trickier. Walzer himself never discusses the issue in philosophical terms. Still, it seems plausible to interpret Walzer as endorsing (4). Georgia Warnke does exactly that, placing Walzer in the hermeneutic tradition of Schleiermacher, Dilthey and Gadamer (Warnke, 1992: 23), who believed that
… the adequacy of a given textual interpretation depends upon the extent to which it can show the text’s coherence as a unified whole. Literary critics must reconcile all the different parts of a text … and show how these work together to compose a well-integrated meaning. (Warnke, 1992: 21; emphases added)
Attributing the notion of mutual support to Walzer’s method of interpretation by placing him within the hermeneutical tradition is clearly possible, although not the foolproof argument one would wish.
It should be noted that the notions of ‘mutual support’ and ‘best fit’ are problematic for any theorist. 24 It must be asked what kind of support we have in mind, what degree of support is required, and what degree of mutuality we might accept – questions not easily answered. However, in order for Walzer to fulfill (4), it is not necessary to require such rigorous analysis. What we need is something that positively connects the components involved instead of leaving them unconnected but logically consistent. This need not be a very strong connection; a mere perceived relation will suffice. Arguably, we can find such connections in Walzer’s work; it is hinted by his analogy that the best interpretation is one that ‘illuminates the poem in a more powerful and persuasive way’ (Walzer, 1987: 30).
However, we still need to say something about such ‘fit’ or ‘illumination’. But first I need to say what this perceived fit is not. In his article, Petersson writes:
When we explicitly discuss a certain person’s set of beliefs, we could of course relate ‘best fit’ to some purpose or some feeling of his. … Of course we can say that such factors are relevant in explaining why P accepts one coherent set of beliefs instead of another. But should we say that P thereby also has good reasons for the one chosen, that he is justified in his choice? I do not think so. (Petersson, 1998: 133)
This seems like a reasonable view. However, the notion of fit attributed to Walzer’s social criticism does not have subjective perceptions at its foundation. This is a very important point, and therefore we need to proceed carefully here. For Walzer, the components (a), (b) and (c) have prima facie credibility, since there are no other points of departure, no other building blocks at hand (Walzer, 1987: 30, 32; cf. Rawls, 1974–5: 8). The social critic engaged in social criticism investigates and analyses the considered judgments, the proposed general convictions and the background theories, and then proposes an interpretation that is supposed to match the deeper morality in the moral community. Hence, the appropriate fit is not left to isolated subjective feelings or preferences, but depends on the core morality of the moral community. If the fit between the components in the interpretation is rightly modeled after this already existing deeper morality, then the fit is arguably also the ‘best fit’. On the Walzerian model, it is not the usual accounts of internal ‘fit’ or ‘mutual support’ that render the equilibrium justified, but perceived conformity to the already existing shared understandings. What determines whether or not an interpretation is correct is that it is recognized – if it is, it also has the ‘best fit’; not to shallow currently held values and norms, but to the values and norms people believe that they, after due reflection and subsequent remembrance, actually embrace on a deeper level (Walzer, 1987: 30; see also Walzer, 1983: 21; 1994: 27, 52). It is an internal conception of ‘fit’, although it is what we might call a deep conception of ‘fit’, as opposed to a shallow one.
This might be seen as a major break with Rawls. Reading Rawls, it is certainly tempting to understand him as placing all justificatory force on the idea of mutual support (Rawls, 1999[1971]: 19). But, arguably, this would be an interpretative mistake as well as an intellectual mistake. 25 A common criticism of WRE concerns the question of what equilibrium to choose. 26 There is nothing that prevents the existence of two or more equilibriums that are all logically consistent as well as exhibiting the same degree of internal mutual support, although being incompatible with each other. Rawls himself seems to be aware of this possibility: ‘[e]ven should everyone attain wide reflective equilibrium, many contrary moral conceptions may still be held. In fact, there are many possibilities’ (Rawls, 1974–5: 9; see also Rawls, 1987: 427). This means that it is highly uncertain if there can be any rational choice between two incompatible but equally coherent WREs. 27 As Hilary Putnam describes it: ‘“coherence” is not something that we have an algorithm for, but something that we ultimately judge by “seat of the pants” feel’ (Putnam, 1981: 132–3).
This ‘“seat of the pants” feel’ cannot itself be evaluated in terms of coherence. The choice between two WREs must be either external or internal. If internal, it implies that no useful comparison between WREs can be made, as the evaluative standard is a part of the WRE evaluated. If external, we are appealing either to a free-standing standard or to a standard’s being a part of a wider TWRE (total wide reflective equilibrium; an all-inclusive equilibrium). If free-standing, we need a non-coherentist justification for such a standard. God’s point of view (or some other objective source) would perhaps suffice here, but the interpretation of God’s view would certainly belong in a TWRE. On the other hand, if we count the WREs as parts of a TWRE, we are only pushing the problem ahead of us: the adjudication of the two WREs would be internal to some TWREs.
However, not only does the suggestion of an independent standard of evaluating interpretations lack support in Walzer’s work, but it also appears unlikely for empirical reasons. Once equilibrium moves beyond mere consistency, it appears to assume a more particularistic shape. It seems to me that an evaluation of the coherence of a WRE (or a TWRE) to a large extent depends on highly particularistic elements – my hunch is that we evaluate different relations differently within every equilibrium. For instance, some cultures would prefer their religious views to be more heavily enmeshed in supportive relations; other cultures would prefer their beliefs about (secular) rights to life, liberty and property to enjoy more support. On a more everyday level, I would certainly prefer my belief regarding the strength of my climbing rope, or my view on parental responsibility, to be well supported, while I do not care nearly as much whether or not my views on the color composition in a Miró painting are well supported. In short, a particular WRE is preferable for me (or us), while not necessarily for others who might prefer another WRE which also shows both consistency and the same degree of mutual support.
Walzer’s theory of social criticism ends, as we have seen, in a particularistic context. But does the same hold for Rawls’ theory? How are we, according to Rawls, to adjudicate between different WREs? This is no doubt a complicated question, and the opinions on the matter are not entirely harmonious. Here I will argue less controversially that Rawls starts out by way of interpretation, and, more controversially, that he also uses shared moral understandings, just like Walzer, as a means to adjudicate between different WREs.
In A Theory of Justice Rawls writes the following regarding the sense of justice that he believes everyone develops ‘under normal social circumstances’ and which he thinks can be captured by a theory of justice:
We do not understand our sense of justice until we know in some systematic way covering a wide range of cases what these principles are. A useful comparison here is with the problem of describing the sense of grammaticalness that we have for the sentences of our native language. … A similar situation presumably holds in moral theory. (Rawls, 1999[1971]: 41) 28
Notice the small details: the sense of grammaticalness that we have – we already have it; our sense of justice – we already have it. 29 The outcome of a WRE is here meant to be a ‘correct account’ of some deeper communal conception of justice. 30 In Political Liberalism, Rawls abandons the term ‘correct’ and instead uses ‘reasonable’, but a shared starting point is nonetheless still presumed. 31 It is therefore not unreasonable to conclude that Rawls is also walking the path of interpretation. 32
While Rawls begins by assuming shared understandings, it could be argued that such shared understandings then lose their importance and play no further role. This interpretation finds some support in Rawls’ own work, most prominently in Political Liberalism where Rawls writes, with a reference to Joshua Cohen’s review of Spheres of Justice, that to start the project of political philosophy by interpreting shared understandings has been a common thread running through ‘Plato, Kant, and Sidgwick’ (Rawls, 1996[1993]: 44, note 47; Cohen, 1986: 467). However, ‘[t]he difference is where Walzer argues it must end up, namely, with our shared understandings’ (Rawls, 1996[1993]: 44, note 47). This quote from Rawls can, no doubt, be interpreted as an explicit rejection of the idea that political philosophy not only starts but also ends with shared understandings.
However, if we take a closer look at the passage and Cohen’s review referred to by Rawls, it is obvious that the whole passage rests on misunderstandings. Beginning with Cohen’s review, Cohen reads Walzer’s interpretations as relying on the morality found on a community’s surface, rather than its shared moral understandings. 33 This causes Cohen to conclude that in Walzer’s project ‘the notions of community and shared values mark the limits of practical reason, not its point of departure’ (Cohen, 1986: 467; original emphasis). According to Cohen’s interpretation then, Walzer’s theory ends up being conservative of the surface morality.
Unfortunately, Rawls builds on Cohen’s mistaken descriptions of Walzer’s theory. And if Rawls shares Cohen’s view of Walzer’s theory, then it is no wonder he reacts against the Walzerian approach as he understands it. Clearly, if our shared understandings and our practices are as dependent on each other as Cohen claims, Walzer’s account will face irresolvable and deep conflicts. Quite naturally, Rawls sees his project as starting from such conflicts, trying to solve them by invoking ‘abstract conceptions’.
The first thing to note is that Walzer, as explained above, never equates shared understandings and actual practices. Just like Rawls, Walzer wants to continue past the conflicts we encounter in society. Second, Rawls’ appeal to ‘abstract conceptions’ is actually an appeal to values deeply embedded in a ‘tradition of thought and practice’ (Rawls, 1996[1993]: 45). 34 The ‘abstract conceptions’ are then not necessarily non-particularistic, although arguably more abstract and general than the conflicting values found at the surface level of a society. More to the point, Rawls continues:
Political philosophy cannot coerce our considered convictions any more than the principles of logic can. If we feel coerced, it may be because, when we reflect on the matter at hand, values, principles, and standards are so formulated and arranged that they are freely recognized as ones we do, or should, accept. (Rawls, 1996[1993]: 45)
What is said here? Our considered convictions cannot be coerced by political philosophy. We freely recognize values, principles and standards as ones we do or should accept. Now, this is a very Walzerian statement made by Rawls. 35 It means that we can use the original position to further such free recognition, but in the end moral standards have to be accepted by us, the moral community. This is, no doubt, a particularistic – and Walzerian – endpoint.
Perhaps the most telling example of Rawls’ acceptance of a particularistic endpoint is his notion of overlapping consensus. 36 A basic problem with Rawls’ theory as presented in A Theory of Justice was that such a liberal regime is likely to produce a widespread pluralism when it comes to basic and fairly systematic outlooks on life; reasonable comprehensive doctrines. Such pluralism would, Rawls believed, make the society unstable. Rawls, attentive to the problem, proposed the idea of an overlapping consensus as a solution. The core idea (which is well-known, and I will therefore describe it only briefly) is that a free-standing conception of political justice is to be justified from different sources; that is, from different and sometimes incompatible comprehensive doctrines. Since the political conception of justice is justified by the different reasonable comprehensive doctrines in society, the conception will remain stable despite the pluralism.
What is important to note is that the idea of overlapping consensus belongs not to the exposition of justice as fairness, but to the explanation of its stability. The foundation of justice as fairness is basically a WRE together with certain helpful tools such as the original position; the overlapping consensus is the litmus test of whether an interpretation is generally acceptable or not, and whether it will be stable or not. 37 Now, Walzer’s social criticism works the same way: it starts out as particularistic, makes interpretative (and sometimes reconstructive) work, and is presented to the public for a temporary approval. When the interpretation concerns the basic political structure, it will hopefully be accepted as the right interpretation of that basic political structure. It is important to note that a Walzerian interpretation need not be ‘comprehensive’; if so, it would clearly contradict Walzer’s own explicit dedication to a pluralistic civil society (for instance, see Walzer, 1983, 1996, 1997a).
To summarize the above discussion, my argument has been that Rawls – just like Walzer – starts with an account of a shared understanding, albeit very thin, and that he anchors his theory in a shared recognition, just like Walzer. There are, as I argued, intelligent reasons not to leave the equilibrium hanging on the notion of mutual support alone.
Criterion (5)
Finally we have criterion (5), which says that a Rawlsian WRE takes the considered judgments of an individual as the starting point. When turning to this criterion, it appears as if our match between WRE and social criticism once again gets disturbed. Walzer’s social critic is commonly perceived as using collective considered judgments, principles and background theories when conducting the interpretation (Walzer, 1987: 35). In contrast, Rawls’ WRE seems to start with the individual and his or her beliefs: ‘[s]o for the purposes of this book [A Theory of Justice], the views of the reader and the author are the only ones that count. The opinions of others are used only to clear our own heads’ (Rawls, 1999[1971]: 44). 38 But Rawls’ statement must be put into context, for if the preceding sentences are considered, its meaning changes:
If we can characterize one (educated) person’s sense of justice, we might have a good beginning toward a theory of justice. We may suppose that everyone has in himself the whole form of a moral conception. (Rawls, 1999[1971]: 44)
According to Rawls, we might start by one (educated) person’s – the reader’s – considered judgments and principles, because the reader has the same moral conception as everyone else. The individualistic starting point is only a pedagogical tool, irrelevant to the content of the overall approach. 39 The method Rawls (and Walzer) use is best labeled, I believe, as interpersonal. 40
Moreover, a collectivistic interpretation of Walzer’s method glosses over the fact that Walzer sometimes adopts the very same individualistic point of departure. In The Company of Critics Walzer writes about the social critic that ‘[t]hough he starts with himself, he speaks in the first person plural’ (Walzer, 2002[1988]: 230). Hence, both Rawls’ WRE and Walzer’s social criticism takes an individualistic starting point, but the individual is also assumed to speak, or to be able to speak, in the ‘first person plural’. This suggests that criterion (5) should be revised in order not to lead us astray: the method of WRE starts with the considered judgments of the collective, as expressed by one individual.
V
In this article, I have tried to defend Walzer against the claim that he lacks a philosophical method when doing normative philosophy. My own claim was that Walzer uses a method very similar to Rawls’ well-known method of wide reflective equilibrium. I started with a brief survey over the comparisons previously made between Rawls’ reflective equilibrium and Walzer’s social criticism. Lacking a systematic comparison, my ambition was to provide such comparison. After having given an outline of Rawls’ wide reflective equilibrium, I elicited a list of its characteristics. Next, I compared Walzer’s method of social criticism to the characteristics associated with Rawls’ wide reflective equilibrium. I concluded that the similarities, underneath the rhetoric, were substantial. Given such similarities, it seems perfectly plausible to conclude that the two methods are really versions of the same kind; instances of the family of wide reflective equilibriums, or of the family of social criticisms. Given the status of Rawls’ wide reflective equilibrium, and given the similarities between Walzer’s social criticism and Rawls’ wide reflective equilibrium, it seems reasonable to conclude that Walzer does not lack a decent philosophical method of doing normative philosophy. And if the philosophical methods of Rawls and Walzer are much more similar than we usually perceive them to be, we should also be careful when trying to put them in the opposite camps. On closer inspection the two adversaries may actually turn out to be allies. In fact, Rawls may be a Walzerian and Walzer may be a Rawlsian.