Author Information: Johannes Persson, Lund University, Johannes.Persson@fil.lu.se
Please cite as:
Persson, Johannes. 2012.Social mechanisms and explaining how: A reply to Kimberly Chuang. Social Epistemology Review and Reply Collective 1 (9): 37-41
Please refer to:
- Persson, Johannes. 2012. Mechanistic explanation in social contexts: Elster and the problem of local scientific growth. Social Epistemology 26 (1): 105-114.
- Little, Daniel. 2012. Social mechanisms and scientific realism: Discussion of “Mechanistic explanation in social contexts” by Johannes Persson. Social Epistemology Review and Reply Collective 1 (3) 1-5.
- Persson, Johannes. 2012. Social laws should be conceived as a special case of mechanisms: A reply to Daniel Little. Social Epistemology Review and Reply Collective 1 (7): 12-14.
- Chuang, Kimberly. 2012. In defense of Elster’s mechanisms. Social Epistemology Review and Reply Collective 1 (9): 1-19.
- Little, Daniel. 2012. More challenges for social mechanisms: Contribution to the Persson-Chuang discussion. Social Epistemology Review and Reply Collective 1 (9): 28-32.
Kimberly Chuang’s detailed and helpful reply to my article (2012a) concerns Jon Elster’s struggle to develop a mechanistic account that sheds light on explanation in social science. I argue that a problem exists with Elster’s current conception of mechanistic explanation in social contexts. Chuang (2012) defends Elster’s conception against my critique. I still believe I have identified a problem with Elster’s conception. In this reply I want to recapitulate briefly Elster’s idea, as I understand it, and then use some of Chuang’s critical points to advance the position I advocate.
1. Social explanations and Elster’s mechanistic surrogate for covering law explanations
Elster roots his perspective in a traditional view of explanation. A traditional view holds that a perfect covering law explanation is the best kind of explanation. Elster also assumes that all explanation is causal (Elster 2007, 7) and that the basic type of thing to be explained (the explanandum) is an event (9). In other words, the ideal explanation is when we can account for an event by citing the relevant causal law that produced it.
The problem, as Elster sees it, is that we know of few such explanations in the social sciences. We do not have enough causal laws to deploy as explanans (that which explains) in our explanations. Thus, a risk follows that we are thrown back on mere description and narrative in social science. To bolster our explanatory resources, Elsterian mechanisms are introduced. Elsterian mechanisms are frequently occurring causal patterns. Elsterian mechanisms are also framed in terms of epistemic uncertainty and indeterminacy. For instance, Elsterian mechanisms “are triggered under generally unknown conditions” (2007, 36). Elsterian mechanisms, then, depend on current epistemic conditions. They depend on our ignorance about triggering factors (or net-effects).
I find that a frequently occurring causal pattern can be local. The pattern can occur within a country, a culture, or a social group. I base my case on these contexts. My interest, in connection with Elsterian mechanisms, is what happens when we learn more about the causal patterns in such contexts. Some day we may come to know the triggering conditions we are interested in; thus, we will no longer have an Elsterian mechanism at our disposal. In Elster’s view this outcome does not matter since we now have something even better — a covering law explanation — to replace mechanistic explanations (Elster 2007, 44).
2. My objection and Chuang’s defense
Here I find that a paradox is generated within Elster’s account. What produces the paradox is this: That we come to know too much about what explains a certain event to be able to deploy an Elsterian mechanism does not imply that a covering law is applicable in its stead. As a consequence, we may be thrown back on mere description and narrative as our causal knowledge grows. A mechanistic view should do better. Little (2012a) and Persson (2012b) suggest ways to improve Elster’s conception.
My suggestion regarding situations where the paradox may manifest itself have been local. For instance, in the explanation of a certain kind of drowning accident we first (at t1) make use of an Elsterian mechanism corresponding to the proverb: “The best swimmers drown”. But later (at t2) we discover in greater detail how, for instance, the risk perception of white middle aged men develop in an unfortunate way — making them take unwarranted risks. By learning more it is possible, and sometimes plausible, that uncertainty about triggering conditions or indeterminate net-effects is resolved. On Elster’s view, in such cases we are first able to explain by Elsterian mechanisms and, later, we are not. I agree with Daniel Little (2012a) that it is strange to have a conception of mechanisms linked in this way with our epistemic condition. At t1 there is truly/actually a mechanism. At t2 there is not. The only thing that has changed is our epistemic state. However, the problem I focus on is the consequences this formulation has for our explanatory abilities. Unless our increased causal knowledge at t2 has produced, or made us discover, a causal law, Elster’s mechanistic conception generates the paradoxical situation that at t2 we do not only lose the mechanism, but also our explanation. We are thrown back into description and narrative.
Chuang defends Elster’s position. I take it that she focuses on cases where causal patterns are more lawlike and widely distributed. She argues that we need to distinguish between the mechanism itself and its application or instances of it (Chuang 2012, 8). The kind of examples I present, she claims, do not show that the Elsterian mechanism vanishes. Indeterminacy can be resolved in the local case, but remain globally. Chuang is clearly right. Some Elsterian mechanisms may correspond to more general causal patterns than we are interested in locally. “The best swimmers drown” may be true for women and other non-white men as well. In such cases, resolving the uncertainty concerning (what was previously) instances of the mechanism does not necessarily affect its global status as an Elsterian mechanism. From now on, I will deploy the terms “mechanistic instance” or “application of the mechanism” to refer to a local mechanism (of the kind discussed in my 2012a). “Mechanistic instance” could have referred only to the causal event to be explained as well, but that is not how I intend it. When I speak of the event I will use the term “event”.
In response to Chuang’s critique I could have been satisfied by restricting my case to locally distributed Elsterian mechanisms. But I am tempted to frame a bolder conjecture. In fact, I think that Chuang’s complaint highlights an equally salient kind of illustration of the problem I find with Elster’s mechanisms. For this reason, I will go along with the kind of example on which Chuang’s objection rests. I admit that some Elsterian mechanisms (the very general causal patterns) do not necessarily disappear when local causal knowledge of one of its previous mechanistic instances is accumulated. My additional hypothesis is that this scenario does not fit Elster’s account any better than the situations I discuss (2012a). The new hypothesis would be that the Elsterian mechanism we deployed at t1 does not play a role in the local explanation at t2. And once again Elster’s conception of mechanism throws us back to mere description and narrative.
To argue for this additional hypothesis, I need to spell out why the general Elsterian mechanism does not play any role at t2. Two reasons come to mind. The first reason is that the Elsterian mechanism is not applicable to the individual case at t2. Uncertainty and indeterminacy are essential features of Elsterian mechanisms and hence of their mechanistic instances. But they are not features of mechanistic explanations where causal knowledge has accumulated in a certain way, as at t2. Since the essence of Elster’s model of explanation has to do with subsumption under a pattern (see below, and see also Little 2012b), the recognition of resolved indeterminacy in the local mechanism disqualifies the remaining, more general Elsterian mechanism as explanans of why the mechanism works locally. The second reason is that, for explanatory purposes, we have something at t2 to replace the Elsterian mechanism with; namely, a local counterpart of the kind of mechanism we relied on at t1 — only without the uncertainty or indeterminacy.
Chuang would probably find problems with my second reason. Again, she rightly claims (Chuang 2012, 12) that something of explanatory importance can be lost in the replacement occurring between t1 and t2. Generality is often considered an explanatory virtue and the local mechanism is “more local” than the original causal pattern, so the explanation we provide of the causal event might be affected. I agree. But I also claim that we must look at the entire explanatory picture Elster is painting. We have in the first (best) case covering law explanation; in the second case we have explanation by Elsterian mechanisms; and in the third (non-explanatory ) case we have mere narrative and description. It is unacceptable, I believe, that the generality loss we face when replacing Elsterian mechanisms with their more local counterparts should be accompanied by a total loss, or loss of most, of the power to explain the event.
I would therefore like to hypothesize that Chuang’s introduction of a distinction between mechanisms and their applications (or instances) helps to highlight how problematic the gap between Elsterian mechanisms and laws are for explanatory purposes. Whether the entire mechanism, or previous instances of the mechanism, disappears as causal knowledge increases, it seems that the Elsterian mechanisms become useless for those events we were interested in explaining.
3. What is it to explain with Elsterian mechanisms?
Chuang’s useful reply spawns several other questions. I would like to end by saying something preliminary about one question in particular. The recognition of Elsterian mechanisms, in addition to laws, as explanans does not seem to be the only way the traditional view of explanation is modified by Elster’s account. What is it to explain with Elsterian mechanisms? It cannot be to produce an argument where the explanandum follows deductively from the explanans. This has been an influential way to frame the traditional view of explanation. The causal pattern in Elster’s case is not universal enough to warrant that kind of inference. But for Elster explanation with mechanisms is still about subsumption:
[…] to subsume an individual instance under a more general causal pattern is also to provide an explanation. To know that the child became an alcoholic as a result of conformism is to remove some of the opaqueness of the outcome, although some will remain as long as we do not also explain why the child was subject to conformism. (Elster 2007, 32)
The explanation why the child became an alcoholic is that an Elsterian mechanism (conformism) was instantiated making some children alcoholics because their parents were heavy drinkers.
It is not necessary, but certainly tempting, to conceptualize this shift — from deduction of the explanandum from universal generalizations (as the traditional view is often framed) to mere subsumption of the event under causal patterns (as Elster suggests) — as a shift from explaining why to explaining how. It seems to me that this is what Chuang suggests:
Once we figure out whether something is causally responsible for our phenomenon of interest, we can then look for how it is responsible. Explaining how something is responsible is accomplished by ‘showing it to be an instance of a general causal pattern, even if we cannot explain why that pattern occurred’ (original emphasis, Chuang 2012, 3).
I am not convinced Elster would be prepared to accept this shift. For the purposes of this reply, however, I am happy to proceed with Chuang’s suggestion that Elsterian mechanisms explain how (rather than why).
4. A quick note on explaining how and Elsterian mechanisms
I think that ‘why’ and ‘how’ explanations are intertwined, but in many cases, they can be usefully separated. For instance, the answer to the question why the Populist Party was created is different from the answer to the question how it was created. In particular, explanations exist how something is possible that do not qualify as ‘why’ explanations, but are very important anyway (Persson 2012c). A certain similarity exists between this relationship between why-explanation and how-explanation, on the one hand, and the relation between covering law explanations and the explanations Elster makes possible by introducing his mechanisms, on the other hand.
However, if we accept — what seems to be Chuang’s view — that explanation by mechanisms are also, or perhaps primarily, explanations how, I would be tempted to argue again that Elster’s current conception of mechanisms should be modified. Whereas it may be right that generality is an important value in why-explanation contexts, it is not at all clear that it is equally important for how-explanation. Faced with the choice between a detailed, but fairly local, mechanistic how-account of the production of an event, and a how-explanation subsuming the event under a general causal pattern (surrounded by substantial uncertainty concerning triggering conditions and net-effects), it seems clear that the local alternative will be rather attractive. But the paradox is generated again. The remedies suggested before (Little 2012a and Persson 2012b) should still be effective.
Chuang, Kimberly. 2012. In defense of Elster’s mechanisms. Social Epistemology Review and Reply Collective 1 (9): 1-19..
Elster, Jon. 2007. Explaining social behavior: more nuts and bolts for the social sciences. Cambridge: Cambridge University Press.
Little, Daniel. 2012a. Social mechanisms and scientific realism: Discussion of “Mechanistic Explanation in Social Contexts” by Johannes Persson. Social Epistemology Review and Reply Collective 1 (3) 1-5.
Little, Daniel. 2012b. More challenges for social mechanisms: Contribution to the Persson-Chuang discussion. Social Epistemology Review and Reply Collective 1 (9): 28-32.
Persson, Johannes. 2012. Mechanistic explanation in social contexts: Elster and the problem of local scientific growth. Social Epistemology 26 (1): 105-114.
Persson, Johannes. 2012b. Social laws should be conceived as a special case of mechanisms: A reply to Daniel Little. Social Epistemology Review and Reply Collective 1 (7): 12-14.
Persson, Johannes. 2012c. Three conceptions of explaining how possibly – and one reductive account. In EPSA Philosophy of Science: Amsterdam 2009. H. de Regt, S. Hartmann, S. Okasha, eds. The European Philosophy of Science Association Proceedings, Volume 1. Springer: 275-286.