Author's response
Dear James,
I hope you don't mind my drafting this reply as a letter, rather than adopting a more conventional academic form, but this seemed the most honest way to proceed given that your review is so much part of an ongoing dialogue between us. Given our many differences, I am obviously pleased, both that you have said so many flattering things about Speaking for the People, and that you believe the book deserves to be widely read. On the other hand, perhaps inevitably (since I am no less sensitive than the next author), I feel that many of your criticisms are unjustified, and that at times they spring from a too casual reading of the text. However, rather than devoting the whole of this reply to a point-by-point correction of these perceived 'misreadings', I want to concentrate on the methodological differences which appear to me to underpin most of the criticisms you raise in the review. But before I do this, I must just correct a few misconceptions that I feel would otherwise create undue confusion.
Firstly, I am not arguing for a new over-arching explanatory model of popular politics based on 'locality', and even less am I arguing that we should 'lament' Labour's failure to exploit the 'politics of locality' during the twentieth century. On the contrary, in raising these issues I am engaging with an existing historiography (especially about early Labour politics), and arguing that we need to think much more carefully and systematically about what we mean by terms such as 'locality', 'community' and 'the politics of place'. Secondly, I do not believe that there is any residual utility in the concept of the 'triumph of party'. You claim that I simply wish to push the moment of party supremacy forward to some date (many dates?) beyond 1918, but I would argue that this merely reflects your own continued adherence to a teleological logic on this point. Thirdly, you suggest that I deny 'the plausibility of measuring the objective interests of a social group'. In fact, as the passage you quote illustrates, my position is to deny the relevance of measuring political aspirations against a check list of supposed 'objective' interests (I am not very interested in the abstract question of whether 'objective interests' can be delineated by social theory). You compound the confusion on this point by conflating the question of 'objective interests' with the very different question of whether it was in the interest of political parties to develop successful languages of 'class' and 'locality'. Here the measure of success is not some abstract 'objective' standard, but the only standard that mattered to the parties themselves: popular acclaim and political success (though I make no apology for suggesting that it is also relevant to examine the plausibility of languages by reference to the broader social context within which they were articulated).
This final 'misreading' touches on the big issue between us: namely your scepticism about my ambition to combine an emphasis on the constructed nature of social and political identities with a (non-determinist) interest in social structures and the material context of politics. Although you claim that my position here is 'wildly uncontentious', almost everything you say about the treatment of social structures demonstrates that this is not so, and that you refuse to accept that historians can say anything useful about past 'social realities'. I suspect that the difference here is an epistemological one. My position rests on a firm belief that there is much to be gained by combining insights drawn from the 'realist' and 'constructivist' intellectual traditions. Far from producing a 'schizophrenic' history, I believe that this approach allows us to address questions of 'representation' without ending up simply cataloguing competing discursive constructions.
Turning now to the substance of your criticism, you suggest that my use of nineteenth-century social survey data condemns me as a naive empiricist who assumes that such materials offer an unmediated reflection of the social world. I certainly don't think this, but nor do I think that such data are so tainted by regulatory fantasies as to be of no use to the historian. My argument is that if we wish to take the material context of politics seriously, then we have no choice but to engage critically with a wide range of nineteenth-century social data. It would appear that you take the opposite view, although since you offer few detailed criticisms of my arguments about social structure I am uncertain how far your scepticism extends. Do you, for instance, contend that census data cannot be used to tell us anything useful about contrasting levels of 'demographic stability' before 1914, or that enumerators' returns offer no insights into patterns of 'residential stability' (both key 'structural' factors discussed in the book)? Clearly, surveys offer an imperfect representation of 'social reality', and their classificatory systems are anything but neutral, but does that mean that they tell us nothing? This appears to be your position, and I can only counter that it is a counsel of despair that will reduce the history of popular politics to a one-dimensional concern with the deconstruction of texts. The aim of Speaking for the People is specifically to argue that there is a way out of this theoretical cul-de-sac that nonetheless remains sensitive to the constructed and contingent nature of political and social identities.
Since my use of social survey data is so central to our theoretical differences I feel it may prove useful if I examine your criticisms in some detail at this point. You make two substantive claims to bolster the argument that I treat survey data as an 'unproblematic' reflection of the social world. Firstly, that I accept census-derived definitions of class in terms of 'occupational status'; and secondly, that I treat administrative units as 'discrete communities'. I refute both suggestions. In the case of the analysis of the geography of occupation in late Victorian Wolverhampton (pp. 133-9), I would have thought it was fairly clear that my aims are essentially pragmatic. Namely to use the data available from census enumerators' returns to explore two very specific historical questions: whether cultural distinctions between 'manual' and 'non-manual' labour were reproduced in the social geography of the town, and whether workers engaged in different industries lived in different parts of the town. Neither question touches on the thorny question of 'what is/was class?' - even less do they assume that the census could tell us. In no sense do the analyses developed here contradict the assertion that Victorian class identities were discursively constructed (or that systems of social classification such as the census may have played their part in the constitution of 'class' as a category of understanding). Unfortunately, you do not appear to be interested in the substantive historical issues addressed in these analyses, and you make no attempt to engage with the conclusions drawn. For your position to hold it would be necessary to demonstrate that census enumerators' returns do not reproduce normative assumptions about the distinctiveness of manual labour in Victorian Britain (quite the opposite is surely true - they seem obsessed to police the distinction). You do not, however, make any attempt to sustain this argument - rather it is apparently enough to appeal to the 'authority' of Foucault and Poovey to condemn as naive any attempt to develop a nondeterminist approach to the analysis of social structure.
Your critique of my use of administrative units is even less well founded. Where the analysis makes use of ward boundaries (pp. 128-33), it does so, not because I imagine these wards to have represented 'discrete communities', but because my subject is municipal politics and their relationship to the geography of labour activism in the town. Again, I note that you offer no engagement with the substantive - and to me still surprising - finding of this analysis: that hardly any pre-war labour activists lived in the town's older, more industrial, and yes, predominantly 'working-class' wards. This was not inevitable, and, significantly, was no longer the case after 1918.
In conclusion, what you see as the weakness of my approach (my 'schizophrenia') is by my reckoning its greatest strength. You suspect that an emphasis on social structures reflects naive empiricism, whereas I claim that it represents an essential dimension in the study of popular politics. You believe that contemporary social surveys tell us nothing about the social world because they were merely the 'representational technologies' of those in power. In contrast, I believe that it remains possible to engage critically with these surveys in order to draw useful conclusions about past social structures - and hence about the context within which popular politics were contested. Despite your strictures, I remain convinced that we must not lose sight of the goal of combining 'realist' and 'constructivist' approaches to the study of the past-as-history. And on this I guess we must simply agree to differ.
In this reply I have tried fairly to summarise the underlying differences between us. Obviously I think that my approach is likely to prove the more fruitful, it is certainly the one that best suits my particular historical temperament. Whether it will be to the taste of others is, of course, an open question. As ever I would be interested to know your views on the matters I have raised, though I guess that we will now have to resume 'normal service', and use email, rather than the web, to debate these things.
