To write about older people in a book about the city is to raise the question of whether city living is any different for older people than it is for any other group of adults. Whom do we regard as older, why do we regard them as different and why, above all, do influential groups in society treat older people as problematic?

The first of these questions is indeed seldom asked. A review of studies in the geography of old age (Harper and Laws, 1995) cites a number of papers about older people in rural settings, but none specifically about people in urban situations. Many of the studies that have been done have concentrated on housing design, rather than looking at issues of urban or neighbourhood living. Urban living may then be regarded as normative for older people as it is for anyone else, though we might want to ask whether living in cities differs from living in other urban environments. The city is not just a large town: it also implies something beyond the urban about governance and a certain quality of life.

But the question of who is old comes close to the root of the matter. Being – or rather, being regarded as – old is not a simple matter of chronology, since this bears little relationship to people's abilities and patterns of behaviour. Nor is it a reflection of people's physical or mental capacities or decrepitude, which vary widely through the second half of the lifespan. Rather, being old is a cluster of attitudes, on the part of the individual and of the social contacts within which he or she is enmeshed, and which are formed both by individual stereotypes and by the broader factors that underlie social and economic relationships.

These two patterns of perceptions of ageing are well illustrated in the history of gerontological theory. Most gerontology in the early post-war period was based on the pathology of later life. It was observed that older people's social contacts diminished over time, and this role loss (Havighurst, 1963) was initially treated as a state to be remedied with increased social activity. Further studies, however, led to the development of an alternative theoretical model, that of disengagement (Cumming and Henry, 1961).