Against Elite Suburbs Some Reflections on Metaphilosophy
Philosophy is about making distinctions. In the most general sense, it is about the distinction between things that are X and things that only seem to be X. A belief may appear to be knowledge but isn’t. An action may appear to be moral but isn’t. A bodily movement may appear to be intentional but isn’t.
In this post, I want to reflect on two ways of establishing distinctions between Xs and seeming Xs. One might be called the ‘additive’ model, the other the ‘disjunctivist’. Additive models hold that Xs are a subset of seeming Xs or candidates of Xs that are derived from the larger set by adding a further property — hence the label ‘additive’. Disjunctivist models hold that there is a categorical difference between Xs and seeming Xs. They are two separate circles without any overlappings in a Venn-diagram — hence the label ‘disjunctivist’.
The title of the post is borrowed from a phrase of Gilbert Ryle’s that I take to be a disjunctivist dismissal of an additive concept of knowledge. Against the ‘justified true belief’ (JTB) account of knowledge, Ryle held that knowledge is not an ‘elite suburb of belief’. JTB is additive because it distinguishes knowledge from seeming knowledge by starting with a class of objects (beliefs) and then adding first the property of being justified and next the property of being true in order to define the subset of beliefs that are knowledge.
Rule ethics is also additive. Kantian ethics starts with the entirety of maxims that might guide an agent and adds the property of universalizability. Consequentialism adds the property of maximizing some kind of good to the set of potential actions. Contractualism adds the property of flowing from a principle that no reasonable agent can reject.
The problem with additive models is that in order to apply them, we need an idea of what results we should like to have. This is necessary in order to show that the additive model does in fact get us what we want. If we want to show that adding those properties reliably distinguishes Xs from seeming Xs, we need to be able to tell them apart prior to the model. In order to show that in any case in which we add the property of universalizability to a maxim we get a maxim that yields a morally right action, we need a conception of morally right actions independently of this conception. The question than arises why we need a process that helps us ‘filtering’ Xs out of candidates for Xs.
Another point is that additive models amount to seeing seeming Xs as the rule, Xs as the exception because, naturally, subsets are smaller than the sets they are a part of. They are therefore inherently pessimistic about our epistemic and moral situation.
If, on the other hand, we say that knowledge is categorically different from error, and virtue from vice (and even from good-hearted but defective moral reasoning or weakness of will), we do not need to pessimistic about our ability to get things right. There is no reason to assume that our instances of knowledge are more rare than our instances of mere belief. We allow for the possibility that, more often than not, our moral reasoning is correct.
Additive approaches are also prone to a kind of uniformization that tends to obscure more subtle distinctions. If we think that all moral actions share one property that separates them from non-moral ones, we lose hold of the idea that there are a lot of ways in which our reasoning can go wrong; the same holds for our beliefs about facts. Against this, disjunctivist approaches preserve the idea that while all good reasoning is alike—on the most abstract level, in the sense that it manifests the capacity to get things right—, each case of defective reasoning is defective in its own way.
Disjunctivist accounts are ‘transformative’ in the way transformative concepts of reason are. It could be said that virtue is distinct from vice and weakness of will by having a property the former lack. But having this property transforms the habit of virtue into something that is entirely different.