This is an anti-positivist piece. Participant observation can be incorporated as merely one methodological option and this is what happens, according to Willis, with the Chicago School, with Downes and the National Deviancy Conference [the former 'bourgeois' exponents of particiapnt/observation -- the latter a marxist rival] Willis wants to make a methodological claim that participant observation gets at ‘raw data’ rather than the self-circularity of positivism.
As a technique, in conventional uses, though, it is close to positivism still: the subjects of research are still objects not to be ‘contaminated’ by the researcher, and any patterns are believed to come from the real world rather than from theory (p.89).
In fact, there is no pre-theoretical object [so empiricism is denied], but we can still be surprised rather than find data which will falsify (p. 90). [This notion of 'surprise' is what everyone remembers this article for]. Theory takes on a general form rather than specifics [so we need p/o research to find the specifics].
Willis is aware of the dangers -- even surprise can be constructed: ‘there is an art concealing art which precisely obscures the theoretical work’ (p.91). So we must have a self-conscious use of theory. We have to do all the checking and confirmations and so on and then ask why any unexplained [ i.e. ‘surprising’] residue remains. We have to re-think our assumptions, in other words engage in self-reflexivity.
The unexplained residues can show contradictions or uncertainty [which are classically ignored in the write-ups in conventional Sociological pieces] but these arise from the contradictions in ‘structured social relationships’. They show, in other words, patterning by external forces and by ideology. This process is uneven and complex, maybe even reciprocal in terms of effectivity. So what is required is circling between theory and data [it’s only when we find surprising or residual elements that we think of the subject as constrained by external forces?] This kind of methodological circling is a Marxist method (p.93), anti-positivist, and seemingly guaranteed by the contradictions, complexities and patterns [rather than by simple ‘surface’ facts which merely describe phenomenal forms]. Reflexivity guarantees it too, as in the dialectic.
A proper method enables us to achieve penetration to ‘the rich veins of ‘lived’ contraction’ (p.94) and produces some kind of thoughts about one’s own social role and its patterning. [This seems to hold out a hope of the modern day equivalent of Raymond Williams and the thoughtful and critical 'scholarship boy']. So, we have to keep this circling as an open possibility and avoid a shift to mere technique which ‘disguises the creative potential of uncertainty’ – we should be able to use a whole range of methods.
Three quick comments:
1. Willis is also famous for doing something which is the very reverse of seeking surprise and open-ness, in my view – what I have called ‘talking up’ the data. An innocent remark by one of the ‘lads’ about the same-ness of manual labour get turned into a whole discussion of abstract labour in capitalism, it is made to agree with Marx, it is held to guarantee the possibility of working-class revolutionary consciousness, and so on. I think this tendency is found in the work on popular culture too.