Future Urban Habitation

The Singaporean model is based on the principle that spatial proximity breeds social proximity. The experiment with integrated housing has produced a fairly successful social mixing of residents with different ethnic and religious backgrounds. They meet each other in the markets, provision shops, hawker (food) centres, coffee shops, and along the corridors of the public housing slab blocks (Lai 1995). Singapore prides its definition as a multicultural society, and the experiment with integrated housing has paid off in the form of the peaceful coexistence (Housing Development Board 2014).

As I see it, however, this is only a starting point. Scholars in urban design and architecture write about the design forms that best elicit sociability, noting, for example, that semi‐private spaces balance the dichotomous needs for sociability and privacy (Gehl 1986). Because much of their work focuses on the neighbourhood, its more general applicability is limited in the contemporary context. Communities today have extended far beyond the neighbourhood; we need to interrogate the broader social network, considering both long and short ties.

In 1973, the American sociologist Mark Granovetter wrote about the ‘strength of weak ties’, noting that while cliques are bastions of support for in‐group members, a society made up of cliques ultimately fails to be socially cohesive. He envisioned instead a society made up of bridges linking cliques. In this context, weak ties are paradoxically ‘strong’.

Here, the ‘small world’ experiments conducted by Stanley Milgram (1967) are instructive. He began his studies by giving participants an envelope (or parcel) to be passed on to a particular target person (a real person in the USA). The rule required that participants pass the envelope/parcel only to someone they knew – a contact – someone they thought was better positioned to relay the envelope/parcel to the target person. As it turned out, the chain was more likely to be completed when envelopes/parcels were sent to weaker ties such as acquaintances and friends than to stronger ties like family members. Furthermore, assuming the target person was of a different race/ethnicity than the original participant, the envelope/parcel travelled much faster when it entered the hands of a contact/intermediary from the same racial/ethnic group as the target person.

The experiment illustrates the nature of the social world, as organized in terms of clusters ; it also illuminates Granovetter's argument on the strength of weak ties as bridges connecting otherwise inward‐looking cliques. These bridges make inclusivity possible.

Singapore is a city of relations connected by bridges and, as such, is a model global city. But we should be cautious in our optimism. Cliques have been forming. A 2016 study I did on the personal communities of some 3000 Singaporeans, commissioned by the Ministry of Culture, Community and Youth, is illuminating (Chua et al. 2017). The survey was representative of the national resident population and comprised a slight oversampling of ethnic minorities to yield sufficient cases in their categories.

My collaborators and I wanted to better understand the personal communities of Singaporeans and did so by asking our respondents to name their contacts in response to a list of different scenarios, for example, ‘With whom have you discussed important matters?’ and ‘From whom have you borrowed money?’

The study elicited 17 000 names from 3000 respondents, with the average network size about six names. Of the 17 000 ties, the majority were friendship ties (45%) and family ties (34%). Co‐workers represented 15%. Notably, neighbours comprised only 6% of all ties, underscoring the point I made earlier about communities being personal and stretching beyond the neighbourhood. Indeed, while much discussion revolves around life in neighbourhoods (e.g. for work on gentrification, see Lees et al. 2013; Smith and Williams 2013), my study suggests today's relationships go far beyond neighbourhoods.

In our study, we paid special attention to the extent to which ties bridge social divides; this was our measure of inclusivity. We focused on network diversity : the extent to which respondents were able to name contacts with a variety of characteristics and backgrounds. For example, gender network diversity measured the extent to which respondents were able to name male and female contacts. Racial/ethnic network diversity measured the extent to which they were able to name contacts from the four Singaporean ethnic groups. The scores, known as ‘indices of qualitative variation’ (IQV) (Knoke and Yang 2008), ranged from 0 to 1, 0 for the absence of contact diversity (on whichever attribute), and 1 for a completely diverse set of contacts. A network of six, with three male and three female contacts would be a perfectly balanced network on the attribute of gender and have a gender IQV score of 1.

The results were unexpected. As Singaporeans socialized into the ethos of our society, including its discourses, we had expected to see the usual ‘fault lines’ of race and religion as the most pertinent social divisions. But this was not the case.

The median network diversity scores on race and religion were sizeable – 0.37 and 0.47 respectively. The network diversity scores on gender and age were also substantial – 0.75 and 0.67 respectively. To our surprise, the largest divides were class‐based. The median diversity scores for ‘elite’ diversity and ‘housing’ diversity were both 0 ( Table 2.1). The scores suggested that the middle person, the average person, had a network that was not at all diverse with respect to contacts with different schooling and housing backgrounds. The typical network was a closed one on both counts.

To illustrate this more intuitively, consider the dyads that connect egos (respondents) and alters (network members). Table 2.2shows a close correspondence between ego and alter characteristics. Public housing respondents frequently named other public housing residents as contacts; private housing respondents frequently named other private housing residents as contacts. Diverse ties (i.e. those between the different housing types) occurred less often.

We observed the same pattern for schooling. Table 2.3shows respondents from elite school backgrounds naming contacts from elite school backgrounds. By the same token, respondents from non‐elite school backgrounds named contacts from non‐elite school backgrounds. Again, diverse ties (i.e. between elites and non‐elites) occurred much less often.

Table 2.1 Network diversity scores based on IQV measures.

aAge diversity: Below 30/30 to 39/40 to 49/50 to 59/60 to 69/70 and above.