Interpretation 1: Phronesis as resistance

Phronesis is usually translated as ‘practical wisdom’ and contrasted with episteme (general, scientific knowledge) and techne (craft and technical skill) (Aristotle, 206–25). Phronesis draws attention to the role played within judgements by character and experience, on the basis that virtue cannot derive merely from the application of rules or obedience to laws. It implies being sensitive to the facts of the case.

For Spicker, rather than reading social phenomena as the surface manifestation of ‘underlying mechanisms’, a phronetic social policy implies a grounded analysis of facts and observations where social generalisations should be built up from particular circumstances. On this interpretation, phronetic generalisations are approximate, fallible, self-reflective and concerned to guide action rather than state what is true or false. They are highly case-specific, therefore.

As noted, Spicker is strongly influenced by Flyvbjerg. Flyvbjerg (Reference Flyvbjerg2001: 98–108) is attracted to Foucault's understanding that the pervasive, ‘non-hierarchical’ and productive nature of power renders the search for universal foundations of knowledge redundant. Instead, we ought to start with the socio-historical context to which we belong and explore its infinite tributaries. Social scientists should constantly immerse themselves in the endless reconfigurations and reinterpretations of power relations, without reference to some absolutist model of truth. According to Flyvbjerg (Reference Flyvbjerg2001: ch. 8), this Foucauldian approach resembles Aristotelian phronesis, as both are concerned with doing and practising rather than rule-following and universal principles. But there are significant problems with Flyvbjerg's approach.

For instance, it relies upon simplified analogies. The intuitive expertise needed to make a bicycle move (e.g. balancing) may not involve following ‘rule-based processes’ (Flyvbjerg, Reference Flyvbjerg2001: 16–23), but interaction with others (cyclists, drivers, pedestrians) and pursuit of a goal (which route should I take?) do require them. The ‘rules of the road’ need to be codified and obeyed through a large degree of conscious deliberation. Cyclists who relied entirely upon habit and instinct would be a menace. Therefore, Flyvbjerg (Reference Flyvbjerg2001: 42–7) gives too much weight to what is pragmatic and situational and not enough to problem-solving within formal systems and codes. His stated intention is to balance rationality and phronesis, yet his examples do not always support this intention (Gereluk, Reference Gereluk2002: 217).

As such, he regards social science as a series of context-dependent narratives that we use to improve our moral practices (Flyvbjerg, Reference Flyvbjerg2001: 57–63). Yet if all we have are context-dependent narratives, then how are we to achieve a rationalist and critical distance from them? We cannot rely upon a narrative to assess itself since it may well be the narrative that risks leading us astray. If any possibility of an independent, impartial standpoint is ruled out then we may lack the basis for normative discriminations. Flyvbjerg (Reference Flyvbjerg2001: 126–7) seems to believe that it is permissible to speak about truth so long as you do not assert what is true, for this would be to take truth-claims too seriously and so invite a ‘thought-police’.

Spicker hints at this problem when he acknowledges the difficulty of distinguishing good generalisations from bad ones. If, as Flyvbjerg argues, generalisations are about experiences, which have to be understood through the circumstances in which they are found, then how can we construct a worthwhile generalisation? Spicker offers three guidelines. Firstly, phronesis should begin with evidence rather than theoretical inference. Secondly, generalised statements should be cross-confirmed by example. Thirdly, we should avoid deduction and, instead of depending on ‘explanations about the underlying factors’, understand how processes operate in their social context. Yet would these guidelines allow us to distinguish between good and bad generalisations?

How can we ‘begin’ with evidence when the status of what is to count as evidence is itself philosophically problematic? If experiences are as context-dependent as Flyvbjerg insists, then how can we confidently interpret x as evidence, y as counter-evidence and z as non-evidence? The context alone would have to supply not only (1) the objects and events but also (2) the standards of measurement, (3) the tools for measuring and (4) the categories for organising the results. The risk of confusing (1), (2), (3) and (4), and so of misreading x, y and z, is higher without impartial standards that are independent from, and so provide a basis for assessing, the immediate context. Sticking close to ‘the evidence’ may risk granting too much prominence to precedent and convention where we count something as evidence because we counted things like that as evidence before. A context cannot assess what counts as good and bad generalisations if we have nothing to rely upon but that context itself.

Certainly, establishing the parameters of that extra-contextual frame of reference is difficult, contentious and offers no guarantee of success. There is always the possibility that what we think of as impartial and independent is just another projection of our context-dependent narratives. But reference to universal frameworks is arguably something of a safety-apparatus against the errors – for example, slavishness to precedent – into which particularism can lead us. Rather than ‘beginning with’ evidence, therefore, we may need to begin at the interface of the particular and the universal so that we can rationally weigh what appears, within a context, to be evidence against analytical frameworks and critical explanations which try to reach beyond it.

As such, Spicker's second and third guidelines basically restate the terms of the first. Yes, ‘generalised statements need to be cross-confirmed’, but bad generalisations may involve such cross-confirmations too. If norms are ‘part of the package’, especially when policy debates are so ideologically charged, how can we be sure that phronetic generalisations are any better at facing ‘inconvenient data’ than non-phronetic ones? Similarly, our understanding of ‘processes in their social context’ may be equally unreliable. If a context cannot comprehensively adjudicate its own evidence, then we may be equally confused about what counts as a relevant process.

At this point Spicker asserts that: ‘Phronesis is not a substitute for episteme and techne, but it is an essential counterbalance.’ Yet Spicker has not defined – and therefore not sufficiently distinguished between – a ‘substitute’ and a ‘counterbalance’. He also states that:

Yet it can also be argued that those interpretations contain radically different views of what counts as ‘deliberation’, ‘translation’, ‘general principles’ and ‘practice’. Besides, if a ‘deliberative process’ is meant to encompass many possible interpretations, then it is unclear why Spicker reacts so strongly against causal explanations. If phronetic social policy really is a deliberative melting-pot, and so long as they are ‘more hesitant’, then why not welcome the contributions positivists and realists can make? Finally, Spicker cautions against ‘interpretative leaps’. But when is a leap not a leap? In proposing that humans are characterised by a ‘function’ (ergon), is Aristotle himself not making a deductive ‘imposition’ on the data? At worst, Spicker's qualification feels like a having-your-cake-and-eating-it defence.

Interpretation 2: Phronesis as solidarity

Other Aristotelians are highly critical of the Nietzschean tradition to which Flyvbjerg appeals. Spicker commends Dunne's recommendation that we get ‘back to the rough ground’, that we should deal on a case-by-case basis with the circumstances before us. Yet Dunne (Reference Dunne1993: 377–8) agrees with MacIntyre (Reference MacIntyre1981: ch. 9) that Nietzsche was an adversary of Aristotle, and thus Dunne's interpretation is far closer to communitarianism. Communitarians offer an ethic of roots and solidity; Nietzscheans wish to tear roots up.

Thus, if Dunne's characterisation is accepted, there are important respects in which his version of Aristotelian phronesis is profoundly different to Flyvbjerg's. Perhaps because he is more interested in what both interpretations challenge, Spicker neglects to acknowledge and debate these differences. This is unfortunate since surely a phronetic social policy inspired by Nietzschean ideas will often contradict one inspired by communitarianism. No cogent defence of phronesis can afford this lack of clarity.

However, even if we set this objection aside we again find a range of philosophical debates that Spicker has bypassed. For instance, like Flyvbjerg (Reference Flyvbjerg2001: 88–107), Dunne (Reference Dunne1993: chs. 6–7) criticises Habermas at great length. For Dunne the essential problem is that Habermas fails to reconcile scientific knowledge (theory) with a practical philosophy capable of guiding us through the ever-multiplying vicissitudes of social life (practice). Habermas is accused of emptying practice of its lived realities and, instead, adopting a formalist approach that is overly concerned with constitutional reforms. Better, Dunne believes, to allow the virtue- and character-led qualities of phronesis to guide us and follow Aristotle in keeping scientific-like knowledge (episteme) detached from moral experience and judgement (phronesis).

What makes this problematic is Dunne's (Reference Dunne1993: 363, 372) admission that he has, firstly, committed ‘violence’ to the Aristotle who was in many respects a determinist and, secondly, ignored the challenge which modern science – ‘more capable of explaining human behaviour and social organisation’ than Ancient science – makes to Aristotelian philosophy. These are admissions that Spicker does not mention. Since humans are natural as well as social beings, does it make sense to confine law-like and deterministic explanations to the non-human realm? Yet if we accept that science cannot tell us everything about the social world, we are not necessarily committed to the view that it can tell us nothing.

By rejecting an ethic which speaks to universals, Aristotelians like Dunne encounter objections similar to those already outlined. Take the most prominent of the communitarian Aristotelians. MacIntyre objects to liberal modernity's claim to be offering a universal vocabulary – capable of speaking across traditions – because, by splitting social life into separate compartments, it fails to offer the self a stable identity anchored around the twin imperatives of moral responsibility and critical reflection. Instead, MacIntyre (Reference MacIntyre2006a: ch. 3; Reference MacIntyre2006b: 39, chs. 10–11) argues that the Aristotelian virtues are more capable of providing the basis for intercultural and inter-communal conversations. Members of different communities and traditions – including policy analysts – can speak to one another as communal beings without the need for an extra-contextual vocabulary.

But as with Nietzschean attempts to dispense with normative universals, this assumes that our immediate contexts can offer a sufficient grounding for our practical judgements. For instance, MacIntyre (Reference MacIntyre2006b: 216–22) would silence – though not necessarily ban – the expression of objectionable views (such as holocaust denial) rather than allow a liberal forum for talking them through. To ensure that such exclusions do not descend into a communal authoritarianism, MacIntyre repeatedly invokes the capacities and judgement of the virtuous character: that into which the community must educate itself by accepting the authority and expertise of those who already demonstrate such capacities. But if silencing the objectionable is an habitual practice, such inter-communal conversations may falter – that is, communities may adopt attitudes and practices that are too quick to object to all differences – and the standards of the community may already incorporate authoritarianism as a norm. As such, orienting practical judgement around character alone is no protection against authoritarianism. MacIntyre (Reference MacIntyre2006b: 6) is potentially trapped within the ‘question-begging circularity’ he warns against elsewhere.

The similarities between these two interpretations therefore give rise to similar difficulties. Those rejecting universalism either have to embrace a philosophy of indeterminacy (phronesis as resistance) or they have to offer substitutes for liberal, Enlightenment attempts to discern extra-contextual frames of reference (phronesis as solidarity). In both cases, relying upon context alone gives us no secure ground from which to critique and normatively assess that context. For the Nietzschean interpretation this might leave us unable to distinguish adequately between good and bad generalisations; for the communitarian interpretation it may leave us appealing to facts and truths with one hand that, with the other, we deny have any objectivity or non-discursive solidity.

I am not claiming that a Nietzschean interpretation of Aristotle and phronesis is irreconcilable with a communitarian one. Yet by allowing Flyvbjerg and Dunne to sit side by side, without acknowledging the considerable differences at work here, Spicker is failing to draw our attention to the very large philosophical issues at stake. Whichever approach is preferred, and if it is to be meaningful, a phronetic social science and policy will have to pay much more attention than Spicker provides to these kinds of philosophical debates.