The social world is not like the natural world. Nature is composed of things, forces, and geometries that have strong determining regularities whose interactions can be formulated with mathematical precision. There are problems of indeterminacy in physics, of course; but fundamentally we can rely on the material properties of steel, the magnetic properties of the sun, or the curvature of space-time to continue to work as expected. Nature constitutes a system of interactions. And this is because, fundamentally, nature consists of atoms and forces -- as some of the pre-Socratic philosophers thought 2,500 years ago.
The social world is different. It is not a system, but rather a patchwork, a mixture, an ensemble, a Rube Goldberg machine, a collage, or a jumble. Its properties arise from the activities, thoughts, motivations, emotions, and interactions of socially situated persons. Outcomes are influenced by a hodgepodge of obstacles and slopes that crop up more or less randomly -- leading to substantial deviations in the way we might have expected things to work out. Agents are not fully predictable or comprehensible; and their actions and interactions are indeterminate as well. We discover that people usually compare costs and benefits when they make choices, and we invent rational choice theory and microeconomics. But these are simply abstract models of one aspect of human behavior and choice, and it is rare indeed to find large social processes that are governed exclusively by this aspect of agency. We see large, somewhat stable social structures that persist over time -- patterns of habitation and social exchange (cities), patterns of racial or ethnic discrimination, rising and falling rates of violent crime -- and we believe there are social causes and influences that help to explain these dynamic configurations. But we should never imagine that social outcomes and patterns are the manifestation of an underlying abstract social order, analogous to laws of nature. Social causes are heterogeneous, probabilistic, exception-laden, and inter-connected -- with the result that we can't hope to have a full model of the workings of a social system.
The heterogeneity and contingency associated with the social world suggested by this set of ideas do not imply that social scientific research and knowledge are unattainable. It implies, rather, that we need to understand the limits on representation, abstraction, and prediction that are implied by the fundamental nature of social things. Our knowledge of any particular snapshot of social reality is inherently partial and incomplete.
A number of sociologists and philosophers have put their fingers on this important problem of social ontology. Here is Bruno Latour in Reassembling the Social: An Introduction to Actor-Network-Theory:
The argument of this book can be stated very simply: when social scientists add the adjective 'social' to some phenomenon, they designate a stabilized state of affairs, a bundle of ties that, later, may be mobilized to account for some other phenomenon. There is nothing wrong with this use of the word as long as it designates what is already assembled together, without making any superfluous assumption about the nature of what is assembled. (1)Here is Norbert Elias in The Society of Individuals:
Society is understood either as a mere accumulation, an additive and unstructured collection of many individual people, or as an object existing beyond individuals and incapable of further explanation. In this latter case the words available to us, the concepts which decisively influence the thought and action of people growing up within their sphere, make it appear as if the single human being, labelled the individual, and the plurality of people conceived as society, were two ontologically different entities. (vii)
What kind of formation is it, this "society" that we form together, which has not been intended or planned by any of us, or even all of us together? It only exists because a large number of people exist, it only continues to function because many individual people want and do certain things, yet its structure, its great historical transformations, clearly do not depend on the intentions of particular people. (3)
What we lack -- let us freely admit it -- are conceptual models and an overall vision by which we can make comprehensible in thought what we experience daily in reality, by which we could understand how a large number of individuals form with each other something that is more and other than a collection of separate individuals -- how they form a "society", and how it comes about that this society can change in specific ways, that it has a history which takes a course which has not been intended or planned by any of the individuals making it up. (7)And now Manuel Delanda in A New Philosophy of Society: Assemblage Theory and Social Complexity:
Is there, for example, such a thing as society as a whole? Is the commitment to assert the existence of such an entity legitimate? And, is denying the reality of such an entity equivalent to a commitment to the existence of only individual persons and their families? The answer to all these questions is a definitive no, but several obstacles must be removed before justifying this negative response. Of all the obstacles standing in the way of an adequate social ontology none is as entrenched as the organismic metaphor. (8)So we should not think of the social world in analogy with examples drawn from what we know about the natural world. We should not think of society as a "thing" or a unified system. The ontological properties of the the natural and social realms are substantially different. This is the primary reason I find some of the basic ideas of assemblage theory appealing: because these theories and theorists deliberately question the naturalistic approach to the social world, and they attempt to advance strikingly different and original concepts for characterizing the social world. They emphasize heterogeneity and composition over uniformity and subsumption.
It is striking to consider the parallel that emerges between this way of thinking about the "social" and some post-Cartesian ways of thinking about the "self". Some philosophers and psychoanalysts have argued that we should question the idea of the unified self that has governed the philosophy of mind since Descartes. Instead, we should consider the notion that the self is not a unified center of consciousness and will, but rather a loose and contingent collage of psychological, physiological, and neurophysiological processes; that the impression of a unified self is a post-facto illusion; and that acting, thinking individuals are coalitions of a heterogeneous and often conflicting group of cognitive, emotional, and practical processes. These are radical challenges to the rationalist theory of the unified self. And they bear a striking similarity to the assemblage challenge to the idea of society as a law-governed structural-functional system.
Here is a word cloud of descriptors that seem accurate in application to the social world.