Over bar foods and beer in a Japanese izakaya in the Boeung Keng Kang district in Phnom Penh, I engage in leisurely conversation with Mr Inoue, a Japan International Cooperation Agency (JICA) specialist in urban infrastructure. Our discussion centers on a large project to upgrade the deteriorating sewage system of the inner city and the various problems and controversies surrounding this work. We talk of politicians, moto pathways, flooding, sewage treatment plants, the laying of pipes, and much else. Thinking of roads, digging, and pipes, I recall having read recent news

stories about Chinese cities, in which road intersections suddenly collapsed and massive, deep sinkholes opened up to the underground, right in the middle of busy cities.1 Inoue has not heard these stories, but he is undaunted; he can imagine something of what causes them. He says that roads often follow old drained river ways, since they are natural pathways. However, underground rivers continue to run along the same paths. There may thus have been hollow cavities underneath new infrastructures. If, he says, roads are made of poor-quality materials and transport intensifies as it has done in Chinese cities, then the entire road may cave in, opening massive cracks that can swallow cars and buses. On a smaller scale he has seen this happening himself in the process of laying the new pipes in Phnom Penh. Whereas an increasing and vigorous body of work focuses on the social and

political dimensions of infrastructure (e.g. Anand 2011, Jensen and Winthereik 2013), on their symbols and imaginaries (e.g. Barker 2005, Sneath 2009, Larkin 2013), and on their organizational and technical dimensions (e.g. Bowker 1994, Pollock and Williams 2009), sinkholes indexes another kind of process, which STS scholars and anthropologists have explored to a lesser degree. These are processes that elicit what might be called the multinatural characteristics of infrastructure. In common usage, infrastructures form the material basis for the provision

of social services – as in roads and railroad tracks. Infrastructure has been viewed by conventional engineering and social science as a layer added on top of, or sunk into, nature. As ‘first nature” becomes “covered over” by infrastructure, it is gradually severed from social experience. When people need water or heat, that is, they increasingly interact not with rivers or sun but with

faucets and radiators. Thus, infrastructure turns into “second nature” (Bowker 1995). In the case of the sinkholes, however, it would be as valid to say that nature has become infrastructure as vice versa. In Inoue’s rendition, roads often tend to follow dried riverbeds, and it is because of cavities underneath those beds that some roads are prone to reverting to quasi-natural states – taking the form of holes, uncannily opening up in the midst of urban space. Alberto Corsin-Jimenez has recently developed the argument that infra-

structure is neither a human “entitlement” nor simply a nonhuman object. What he refers to as the “right to infrastructure” defines a certain analytical sensibility, which facilitates an escape from “the human-nonhuman and epistemology-ontology dichotomies … by opening up the agential work of infrastructures as a source (an open source) of possibilities in their own right” (2014: 343). In this chapter I suggest that such an escape demands even more than

heeding the agency of infrastructures. It obliges the researcher to focus on the unstable, emergent interrelations between infrastructures, their human developers, and numerous other entities, such as animals and trees. One way of getting into view infrastructural capacities is by exploring how this motley multinatural array makes, sustains, and disrupts them. More decentered ethnographies and conceptualizations of infrastructure, I argue in the following, might shed light on the complex entwinement between infrastructures, the human and nonhuman worlds they produce, and their possibilities and dangers. The present chapter experiments with such a decentered, multinatural mode

of exposition. It takes the reader on a tour of Phnom Penh’s sewage infrastructure in order to relativize the perspective from which infrastructure is seen or experienced – a relativization that does not refer to different groups of people but to the infrastructural work done by various kinds of nonhuman agent: trees, bacteria, sludge, water plants, and cockroaches. The result is a series of figure-ground switches, in which nature and culture not only begin to hybridize but in which their descriptive and analytical relevance quickly begins to fade. Ultimately, like sinkholes, they collapse.