There is a risk that scholarly communities become inured to a particular intellectual orthodoxy. Challenges to such orthodoxies are therefore to be welcomed. In Social Policy, one such orthodoxy was once identified by Alan Deacon as the ‘quasi-Titmuss paradigm’ (Reference Deacon2002: 22). And it is to this that Andrew Dunn in his recent book mounts a challenge. However, in borrowing a term coined by Deacon, Dunn reduces it to something of a caricature: he applies it as a label for supposedly closed-minded, left-leaning academic mainstream. The focus of the book is on labour market activation (LMA) policy, in relation to which Dunn is an articulate disciple of Lawrence Mead's neo-conservative ‘work first’ approach (e.g. Reference Dean and Melrose1997). The book's argument is constructed around four pieces of original UK-based research, each of which has some intrinsic merit, albeit that their findings might arguably be capable of interpretations other than those offered by the author.

The problem constructed and addressed by the book is that of ‘voluntary unemployment’. The opening chapters consider, on the one hand, the substantial body of research by ‘mainstream’ Social Policy academics, demonstrating that unemployment results primarily from socio-economic and structural factors; and on the other, the critiques of neo-conservative thinkers who suggest that unemployment results primarily from individual behavioural or moral factors. While rendering a generally fair account of both sides in this debate, Dunn's central argument is that the academic mainstream has failed sufficiently to address two elements of the neo-conservative critique: on the one hand, the ‘choosiness’ of unemployed people and their reluctance to take ‘unattractive’ jobs; and, on the other, the fact that legitimate evidence from professionals with first-hand experience of LMA work calls into question the plausibility of evidence obtained directly from unemployed people.

The four projects reported in the book are:

1. 1. A qualitative study based on interviews with a mixed sample of fifty employed and unemployed people. This produced an interesting taxonomy reflecting the different trajectories and attitudes to work disclosed by different members of the sample.

2. 2. A further mainly qualitative study, building on the first, based on interviews with a mixed sample of seventy unemployed benefit claimants and workers in various occupations. Taken together, the two qualitative studies generated a range of findings, including: the finding that more highly educated people were choosier about jobs (as they could afford to be) than those who were educationally disadvantaged; but also that most people would accept unattractive jobs if the remuneration were high enough.

3. 3. A quantitative study based on secondary analysis of longitudinal data from the British Cohort Study and the UK National Child Development Survey. The analysis centred on responses to a particular survey question (also used in the preceding qualitative study) that asked whether participants agreed that ‘having almost any job is better than being unemployed’. The headline finding is that unemployed people were more likely to disagree than agree, which is consistent with findings from Dunn's earlier studies, and, in his view, decisively contradicts past findings founded on differently framed survey questions.

4. 4. A study based on telephone interviews with forty activation workers, employed by agencies sub-contracted to the UK's Department of Work and Pensions to increase labour market participation among unemployed benefit claimants. Respondents reported that some of their clients were too choosy about the jobs they would accept. And a substantial majority attested to the existence among claimants of a ‘dependency culture’.

The book concludes with recommendations for a further toughening of the UK's LMA policy regime.

One element of my own work (Dean and Taylor-Gooby, Reference Dean and Taylor-Gooby1992) is a specific target for Dunn's criticisms and it is, in the circumstances, important to draw a clear distinction between my own socialist (though not exactly ‘quasi-Titmussian’) normative leanings on the one hand, and my commitment to rigorous scholarship on the other. It is as a Social Policy scholar that I would point to four fundamental reasons for questioning the thesis Dunn offers:

• • First, what Dunn fails to acknowledge is that a right to work that is both freely chosen and subject to just and favourable conditions is expressly provided by the Universal Declaration of Human Rights (Article 23). This was a right forged by international agreement and, like the capitalist welfare states that claim to respect it, it was founded ‒ however tenuously ‒ on a middle-ground consensus between social-liberal, social-conservative and social-democratic ideologies. It was not ‘quasi-Titmussian’: Titmuss (Reference Titmuss1971) was not especially comfortable with the idea of rights. Dunn is perfectly entitled to join those who would unpick that historic consensus and to question the extent to with people should be allowed to choose their employment and to which they may be justly required to labour in jobs that are unattractive. But the potential magnitude of the wider implications needs more openly to be recognised.

• • Second, while Dunn makes some perceptive observations with regard to existing evidence pertaining to the work ethic, his favoured metric (the proportion of people who agree that ‘having almost any job is better than being unemployed’) has an ambiguity all of its own. Does it measure attachment to the work ethic, or aversion to state dependency? My reading of the evidence ‒ which Dunn does not address ‒ is that by stigmatising state dependency, aggressive LMA policies can be a part of the problem, not the solution. My own findings suggest that forcing people into unattractive jobs can, in some instances, undermine the self-esteem and sense of identity that people seek from work, and, paradoxically, erode rather than promote the work ethic.

• • Third, while it is important to consider the role of individual agency, the structural context in which people live their lives must not be overlooked. Dunn's discussion of the ideological contestability of LMA policy does not expressly account for the palpable consequences of the evolving economic context: the inexorable polarisation between high-skill/high-paid and low-skill/low-paid labour; and the increasingly precarious nature of employment at the labour market periphery. There is qualitative research that captures the ways in which some unemployed workers will exercise their agency through resistance: by tricking the system, by resorting to the informal economy, by finding satisfaction in unpaid work. Such strategies may be counterproductive (e.g. Dean and Melrose, Reference Dean and Melrose1997), but they are rational responses to adverse labour market conditions and the logical consequences of supply-side economics.

• • Finally, while evidence elicited from activation workers is plainly relevant, it must be remembered that the role and raison d'être of the activation worker has been constructed in accordance with a particular policy orthodoxy (a ‘quasi-Mead paradigm’, perhaps?). Findings ‒ especially when elicited through the sub-optimal method of a telephone interview ‒ must be interpreted accordingly. A conclusion that might be drawn is that such findings speak mainly to the enduring character of the dependency culture discourse and its symbolism. Also, activation workers are only one sort of professional with experience of dealing with unemployed workers. Interviews with professional welfare rights advisers, for example, might have revealed a quite different picture.

Dunn's book is undeniably and intentionally provocative. I am, for my part, a little disappointed by inconsistencies in the rigour of its argument, but the questions it poses are important and deserve serious debate.