Stratification and generative mechanisms

The intransitive dimension is not flat: reality is stratified. It is the twin ideas of ‘levels’ and ‘emergent properties’ that give critical realism its ‘depth'. Most frequently cited is Bhaskar’s distinction between the actual, the empirical and the real, although this is just a beginning way of characterising thecomplexity of ontological stratification (Collier 1994, p. 44). The ‘actual’ refers to things and events in their concrete historicity, only some of which will ever beknown or experienced by human beings (there must have been events on the other side of the moon which no human being ever experienced). The empirical is a subset of the actual, consisting of those phenomena which arc experienced. ‘Experience’ must here be understood not as given but as socially produced, whether from the concepts of science or those of everyday life: ‘in this sense it is the end, not the beginning of a journey’ writes Bhaskar, in a passage that exemplifies the social constructionist aspect of critical realism (Bhaskar 1998, p. 43).

The actual and the empirical arc, of course, both real, and as such are contained within the third domain. But the domain of the real also includes mechanisms, powers and tendencies. ‘A generative mechanism ... is that aspect of the structure of a thing by virtue of which it has a certain power’ (Collier 1994, p. 62). An army, for instance, by virtue of its structure (its organisation and resources, and the rules which govern their use), has emergent properties and powers irreducible to those of the soldiers and other personnel who make it up. An army can besiege and take over a city, an unorganised crowd could not do so. Changing the organisation will change the properties of the mechanisms -as I write, the British Army is merging some of its smaller regiments because it believes that bigger units have greater power to evoke loyalty in the soldiers who belong to them, which would in turn change what these units could do in war. However, generative mechanisms only operate when set into motion. The army has to be mobilised and set into action for its military powers to be realised, though its powers to evoke loyalty to its regiments could operate in its peacetime barracks. Even when a mechanism is operating, its effects depend on the co-acting of other generative mechanisms - sometimes referred to, by Pawson for instance, as ‘context’ (Pawson 2004). Outcomes at the level of events arc ‘complexly co-dctcrmincd’ (Collier 1994, p. 62).

The term ‘generative mechanism’, then, does not imply that the relationship between cause and effect is ‘mechanical’ in the sense of predictable and automatic (indeed, some realists prefer to talk of ‘causal configurations’ and ‘causal powers’, and those who use the term ‘mechanism’ do not do so in exactly the same way) (Carter and New 2004). A mechanism could be working without any apparent effects, or with effects which arc unexpected or seem paradoxical - because of the existence of other, counter-acting and interacting mechanisms.

[T]hcrc are many mechanisms concurrently active. The outcome of this -that is the events - is therefore a complex compound effect of influences drawn from different mechanisms, where some mechanisms reinforce each other and others frustrate the manifestations of each other ... there are countless combinations of accidental circumstances ... which may influence whether a specific causal power will manifest itself or not ...

(Danermark et al. 2002, p. 56)

For this reason critical realists talk about ‘tendencies’. A tendency is the working of a mechanism (always in conjunction with others) which usually, but not always, gives rise to a specifiable range of outcomes recognisable as the effects of that mechanism. The continual injunction to boys that they should not cry (in order to be distinguishable from girls), and contempt directed at those who do, may or may not have the effect of inhibiting crying. When the mechanism is operating, it tends to reduce frequency of crying in boys. (Nowadays there are counteracting cultural messages to the effect that boys have the right to cry - at least a bit.) If we were to analyse in detail how boys respond to the wealth of messages and practices aimed at preventing them expressing feelings by crying, we would find some cases where the boy went on crying, but felt bad about himself and compromised in his gender identity because of this. Thus a tendency always makes a difference to what actually happens, even when the outcome is not obviously affected by its operation, or when it is ‘counterfactual’ - in other words the outcomes are distinctly not what the mechanism is supposed to bring about.