The following is a extemporised draft of a presentation I need to make in an upcoming methods course, but it also works as a stand-alone post, and that it why it is a stand-alone post.
The proper role of history in social science is contentious. The traditional distinction between the ideographic and the nomothetic remains, in the eyes of many in the disciplines of political science and sociology, appropriate. According to it, the discipline of History is concerned with describing the course of history, while the task of the social scientist is to identify patterns in that history and to describe the operations of cause and effect through law-like generalisations. While history matters in this view, its role is as a source of data. rather than an object or domain of study. But there are a number of research traditions that dissent from this perspective. And they’re right to do so.
Why might social scientists want to get their hands dirty in the sandbox of history? First, there are moral reasons to look at a set of events that took place in the past and try to figure out why they happened as they did. Praise and punishment are closely related to the notion of causal responsibility, and so figuring out who or what caused an outcome of interest will shape our judgements. Second, there are ethical reasons, in that we may want to look at how we arrived at our situation, or how similar situations were arrived at, to clarify the problems we face now. Neither of these is so easily done with simple neo-positivist ‘covering-laws’. Third, and perhaps most saliently, historical enquiries might give us some excellent tools for social science theorising.
A number of themes stand out in looking over the literature on the role of history in political science and sociology: narrative, sequence, path-dependence and contingency, and process. A brief discussion of these themes and their relationship to ontological concepts such as ‘critical junctures’ or social mechanisms is helpful for clarifying just how an historical social science might work.
To study the role of narrative is, in a sense, to treat history as an object rather than a domain of study. Human beings are constantly making and re-making stories of their lives. Identity, community, collective action, and the identification of social problems all proceed through symbolic action and interaction. The way that people construct their common histories is a potent determinant of social outcomes, and those histories are thus highly political, being formed through rhetorical battles and resulting in the allocation of values (to use Easton’s classic definition of politics).
Returning to a more conventional definition of history, the explanation of many social outcomes seems to rely heavily on the linking-together of a number of causal links throughout a period of time—a sequence of causes, in other words—where later causes function only due to the presence (and order) of prior ones. To adequately explain an outcome requires the scientist to show how it is contingent upon a series of steps along a pathway of cause and effect stretching back through history. The presence of remarkable things in the social world, such as revolutions or massive wars, may only make good sense if understood as the product of such a sequence, and explaining them as such typically requires a rich and broad description of context.
But the question still remains, how do we build forward-looking theories that don’t just tell us what already happened, or how we’ve reached the present, but how things might work out in the future? This is where process comes in.
To treat society as set of processes of constant construction is a philosophical ontology; it is a package of first principles as to the nature of social reality and the source of change and stability therein. It is also a sound starting-point for an historical social science. To take a processual view of history is to treat it as something in a continual state of becoming: not only are we constantly affirming, revising, and contending our historical narratives, but we are generating the society as a physical reality through continuous social transactions. On this view, any ‘outcome’ of interest is nothing but a snapshot of a moment in time, and any explanation is, effectively, a description of the salient sequences or paths that led up to it.
What generality can theory have, then, if all outcomes are so unique?
Historical sociologists and political scientists often explain outcomes through reference to mechanisms, and this gives their theories some measure of portability or predictive capacity. While there are many possible definitions for a causal mechanism, the very general idea is that they are semi-generalisable patterns of interactions between factors that link cause to effect. They are the bridges that lead from one historical situation to the next, through in a sense, to identify a causal (/social) mechanism is more to identify a particular form that a bridge can take, with the actual construction materials or location to be determined by unique historical context. Causal mechanisms can be very general, such as ‘brokerage', or more specific, such as the way capitalist economies generate the need for market expansion (according to Marxists). The notion of ‘critical junctures’, where social structures remain relatively stable until certain points of historical instability, at which point insurgent actors gain the opportunity to generate change, fits best within this perspective on theory construction. A critical juncture can more or less be understood as a situation in which the conditions for transformational—sometimes called ‘morphogenetic’—mechanisms are present for a short while. Thus causal mechanisms can explain how things stay the same, and thus how institutions endure over time, or how things change, and if we know the conditions under which they ‘activate’, then we can have some idea what will happen given a particular historical context. Such as the present.
Unsurprisingly, social movement theorists and institutionalists often build their theories around mechanism-based explanations. This is because social movements and institutions are everywhere, and despite being reflections of their social contexts nevertheless show certain patterned regularities in how they operate. By studying when (and by extension, how) those regularities do or do not obtain, social scientists can build a ‘general’ theory that allows them to better appreciate context, rather than to be forced into a compromise between context and covering-law.
Marxists and other ‘materialists’ are also often fond of mechanism-based explanations. By building theories that allow them to explain the causal relationships and interactions of particular situations, they can claim to be doing a ‘realist’ social science. In the same way that biologists can talk of the mechanisms of DNA reproduction or photosynthesis, scientific realists in the social sciences can claim to be theorising how society ‘really works’ in a way that goes beyond correlation or prediction, and to be describing the deeper relationships that generate the reality that most in society take for granted. This entails greater potential for social critique than covering-law explanations.
The foregoing is a shallow and expansive survey of some of the major social science perspectives on why history should be taken seriously and how to go about doing that. But hopefully it teases out some of the implications of treating history as more than just a gigantic dataset, and goes a little way towards selling my own appreciation for historical sociology as a legitimate method for conducting social enquiry.
 If I wanted to be glib, I could simply point out that all description is ‘theory-laden’ and therefore already implies causal linkages. But where’s the fun in that?
 Where a third-party mediates or brings together two other parties who otherwise would not interact in a way that would produce the outcome of interest.
 It should be noted that social movements are pretty new things, whereas bureaucracies or other sorts of institutions have been around a very long time.
 Cf. critical realism