This post is inspired by recent discussion on Jacob Stegenga's facebook page, and a draft essay he wrote as a follow up to it; this post is also part of a series of posts I started with yesterday's post (which responded to issues raised by Phillippe Lemoine's essay). Today's post will largely abstract away from the pandemic.--ES
By drawing on ideas about synthetic philosophy (recall), I mean by 'policy-apt synthetic philosophy’ a style of philosophy that brings together insights, knowledge, and arguments from the special sciences and normative theories with the aim to offer a reasonably coherent account of complex systems and, thereby, to connect these to decisions under uncertainty in policy and/or debate about policy. Policy apt synthetic philosophy is made possible by, and a response to problems subsequent the intellectual division of labor within and among rather esoteric academic disciplines and the need to make decisions under uncertainty that ramify across many social domains.
Before I explain what policy-apt synthetic philosophy is, it may be useful to contrast it with two other related, familiar intellectual activities that it may draw from but it is not identical to. (In practice there will be lots of blending, and that is okay.) First, in applied ethics decision contexts are narrowed down or constrained such that it is possible to apply a particular principle or decision rule/practice to it. That is, I am conceiving of applied ethics as an instance of casuistry in a descriptive not pejorative sense. (I have been much influenced by a leading campaigner for abolition of the death penalty, Hugo Bedau's (1997) Making Mortal Choices.) In casuistry cases may be very high stakes (life and death) and may seem intractable because they involve conflicting principles or important trade-offs in values. In the best forms of casuistry, the search for the right principles is a bottom up enterprise, "a multiple triangulation of the region in which the best answer lies, the parameters or boundaries of the region being determined by the relevant ethical principles." Naturally, applied ethics often does more than that confronting constrained and partially standardized choice moments when it, for example, helps develop normative principles for whole processes/fields of activity or decision rules for the allocation of scarce resources. This will involve the ordering and integration of principles and/or a hierarchy of priorities that figure into such constrained decisions that can be made to fall under a principle (etc.). When I use 'applied ethics,' I mean this wider sense.
Of course, professional applied ethicists do a lot more than what I have just called 'applied ethics,' including reconceptualizing what counts as a decision (or, say, a disability), and how we should think about about structural features that may be ignored or invisible if we focus primarily on decisions and not the social institutions/norms/resources (etc.) that constrain and structure our lives. Some important professional applied ethicists even think we should reject the focus on choice altogether. That is fine. That they do a lot more of significance is the intellectual labor that is presupposed in my discussion here.
Second, another intellectual practice that has been developed around constrained choice is decision theory (e.g., practiced in philosophy, economics, game theory, and computer science departments). This intellectual practice is concerned with designing tools to reason in a disciplined fashion in choice settings. In particular, it finds ways of aggregating evidence about the environment and outcomes into a mathematical technology so that a mathematically constrained procedure necessitates a course or range of actions in particular choice circumstances.
Readers who will have noticed my appeal to scarcity in my description of applied ethics, will have already noticed that there can be overlap even identity between decision theory and applied ethics. That is to say, much decision theory is de facto normative. Utilitarians, for example, find it relatively easy to move between applied ethics and decision theory because much of normative decision theory is about maximizing/optimizing expected utility. Focusing on expected utility is not the only way one can do decision theory. But generally important normative principle are already internalized in the mathematical technology or the standard understanding of the decision theoretic tools. Even so, and I return to this, this is just a subset of the normative principles and commitments that are taken as significant in applied ethics. One way to distinguish decision theory from applied ethics, is that the former tends to be about generic choices, while the latter tends to be domain specific, even though the domain can be vast: so, for example, my own work in the ethics of financial debt covers practices in multiple social organizations that impact the whole economy the natural environment, and the rise and fall of states (etc.)
At this point one may wonder why, if we already have applied ethics and decision theory, policy-apt synthetic philosophy is even needed. For in many areas of applied ethics, in particular, ('synthetic') knowledge of the relevant special sciences and complex social activities are presupposed. There are at least three important reasons that I will present as independent, but that also mutually strengthen each other. First, what applied ethics and decision theory have in common is constrained choice. And, what I assume is, that this feature is internalized in the way of thinking (the modeling, the standard examples, the policy advice, etc.) in the fields that focus on decision theory and applied ethics. And for many contexts that is incredibly apt.
But there are a range of policy circumstances, where by treating a choice as constrained in a particular way one may end up doing more harm than good. Paradigmatically these are going to be circumstances in which the very constraints that make it possible to think of a choice as constrained, will make it impossible to track or take into account a range of effects or values that will be harmed or prevented by the decision. (This is why some applied ethicists dislike the focus on choice!) This is a feature not a bug of disciplined decision making in applied ethics and decision theory. But as is well known, in reality there can be huge interactions among different kinds of choices. And so we need a place where the, as it were, 'outputs' of particular areas of applied ethics and decision theory, are themselves integrated and studied and compared with the 'outputs' of different scientific domiains.
Second, one may think that decision theory is well situated to do so this itself. So, for example, recently Liam Kofi Bright and David Bradley position decision theory as the activity ot generating not "just a single probability distribution over the outcomes of interest, but a family of them." And in so doing one can help policy makers "to choose actions that can be expected to yield benefits across a very wide range of scenarios." But notice three features, (i) this privileges expected utility (or close analogues); but in many political contexts one may well wish to involve other important values such as various civil rights, respect, autonomy, dignity, and various freedoms (of association, of movement, etc.), etc. In order to think of choices generically, decision theory flattens the value-landscape and in some policy contexts this is unwelcome.*
(ii) Integrating and comparing probability distributions (the outputs of scientific modeling) does not come automatically from heaven. In particular, it involves comparing many different kinds of models that have many ad hoc but socially salient features. Or where the features and practices are entirely standard but have enormous social ramifications. (Recall yesterday's discussion of the significant differences between modeling with homogeneous or heterogeneous agents.) This comparing and integrating of different kind of models and values, requires, when these are not domain specific, the skilled activity of the sort I have called synthetic philosophy. (So, in one sense the synthetic philosopher is the underlaborer to the decision theorist.)
(iii) Not unlike most decision theorists, Kofi Bright and Bradley assume that one can assign probability distributions. There are two problems here: a) what makes something a crisis, often in the face of very new circumstances, are precisely circumstances where one can reasonably doubt one has a grip on any of the relevant underlying distributions, that is, one is faced with true uncertainty. And b) often when we assign probability distributions, there are genuine possibilities one cannot then model at the same time. (That is, this is an instance mentioned above: the very constraints that make it possible to think of a choice as constrained, will make it impossible to track or take into account a range of effects or values that will be harmed or prevented by the decision.) One reason I mention Kofi Bright and Bradley is that they are transparent about the limitations of decision theory.
Third, often in a crisis, the normal operations of science is attenuated even partially suspended. That is, the ordinary stress testing of the concepts, data, and methods over an extended period of time and give and take of scientific debate is compressed. So, even the best informed scientific aggregators (or what Polanyi called 'influentials') can't rely on underlying community consensus. In a crisis, what Jacob Stegenga calls 'fast science' is really fast. So, policy apt scientific philosophy is needed to do the assimilation, partially in public and in real time, of unfolding sciences so that it, down the road, can be turned into the input for disciplines that work on some constrained choice or a field of study.
One may well think that there is no need for policy apt synthetic philosophy because that's why governments have offices for scientific policy/advice that help coordinate decisions that involve multiple sciences and policy domains. And, indeed, such offices often help prepare the inter-agency and inter scientific crisis meetings that have been in the news. And such offices, in turn, can draw on many kinds of experts including applied ethicists and decision theorists and ad hoc government advisors drawn from a wide variety of scientific aggregators/influentials. Even so, even in extremely well functioning polities, there is a need for a class of individuals that study the principles of such policy coordination and are not government officials so they can contribute alongside the aggregators and applied ethics folk to public discussion in independent fashion.
Okay, let me close here. Today's post was rather abstract and short on normative and scientific detail, and I have assumed that readers can use their imagination in connecting the dots to the unfolding pandemic. For example, yesterday, I privileged a version of the most affected principle combined with a commitment to protecting the most vulnerable. It's not obvious that should always inform policy apt synthetic philosophy. I also have not given any examples of promising policy apt synthetic philosophy (in the past I have praised Rachel Carson). But more important there are also risks to any intellectual activity that becomes oriented toward policy. To be continued.+
*Obviously, some kind of utilitarians will be unmoved here.
+I thank Jacob Stegenga for discussion. The usual caveats apply.
Comments