The main problem with retrospective interpretations is that subsequent experience can play the trick of laying new meanings on old events. In part this is because the need to make a coherent, seemingly planned story of one's life constantly overwhelms the more honest ambition of describing it as it was. As it was is usually a series of false starts and premature stops, a mix of ill-assorted and conflicting ideas, and of feelings and intellectual insights all jumbled together in an unholy melting-pot. Memory is unfortunately but endearingly opaque, leading us to dissemble instead of revealing.

Bearing this qualification in mind, it seems to me now that the Social Support and Pregnancy Outcome (SSPO) study, as it was ultimately called, had its origins in six sets of observations about social relations. These were that:

• 1. Science, including medical science, may be regarded as a ‘social’ product – its content and practice reflect the social backgrounds and motives of its practitioners, rather than existing in some pure, uncontaminated, ahistorical mode.

• 2. The professional ideologies, status and organization of the medical profession militate against recognition of the universe and impact of the ‘social’ in health care.

• 3. The survival and health of mothers and babies are consistently worse in socially disadvantaged as compared with socially advantaged groups.

• 4. Differences in social position and experience, especially as mediated by stress, are linked with different fates of mothers and babies.

• 5. Social support is good for health.

• 6. Being researched may in this sense be health-promoting.

The rest of this chapter expands these six observations as a prelude to describing the background and beginnings of the study.

‘Knit your own incubator’

For twenty years from 1965 to 1985 I worked as a contract researcher, for the last five years of this twenty-year period with other such researchers in the National Perinatal Epidemiology Unit (NPEU) in Oxford.