Individual Lectures Abstracts

Lecture 1, May 3

Continuum Argument and Persistent Incommensurability

Parfit (2016) asks us to imagine a sequence of populations that starts with one consisting of a fair number of people with wonderful lives. Each population in the sequence has a slightly lower life quality than its predecessor but is much larger and therefore presumably better overall. In the last, huge population, people’s lives are barely worth living. What Parfit dubs “the Repugnant Conclusion” is that the last population is better than the first. He suggests this “Continuum Argument” for the Repugnant Conclusion isn’t sound: incommensurabilities (“imprecise equalities” in his terminology) intervene at some points in the argument’s sequence of purported improvements. It is an attractive suggestion, I think, well worth considering. (Not only in this case but also in dealing with other similar ‘spectrum arguments’). But for the argument to be blocked beyond repair, the relevant incommensurabilities need to be very thoroughgoing: they need to persist even if the next population in the sequence were improved by making it even larger, indeed, arbitrarily large. While such persistency might well seem highly problematic (as argued in my earlier work with Toby Handfield), I think it can be explained and defended if incommensurability is interpreted on the lines of the fitting-attitudes analysis.

Lecture 2, May 4

Degrees of Commensurability

The definition of incommensurability is purely negative: Two objects of valuation are incommensurable if they are not commensurable – if none of them is better than the other and they are not equally good. This negative characterization is coarse-grained; arguably, it fails to capture the more nuanced structure of incommensurability. In a joint work with Alan Hájek, we suggest that our evaluative resources may be much richer than it has been previously recognized. We model value comparisons with the corresponding class of permissible preference orderings. Then, making use of this model, we introduce a potentially infinite set of degrees of approximation to better, worse, and equally good, which we interpret as degrees of commensurability.

One payoff of this more nuanced approach is an explanation it provides of the intuitions that make the Continuum Argument, and other spectrum arguments, so seductively compelling. Blocking the argument is one thing; explaining why it might seem compelling is another. Developing Parfit’s response, we argue that some of the populations in the sequence are almost better than their immediate predecessors. They seem to be better because they are almost better. ‘Almost better’ is not transitive (unlike ‘better’). A series of almost-improvements might end up in a worsening. There are clear analogies between this error theory for spectrum arguments and the classical voting paradox due to Condorcet.

Lecture 3, May 5

Probability as Value

According to the fitting-attitudes analysis of value (FA-analysis, for short), for an object to be valuable is for it to be a fitting target of a pro-attitude (i.e., an attitude with a positive valence). Different kinds of value – desirability, admirability, etc. – correspond to different kinds of fitting pro-attitudes: desire, admiration, etc. Probability can be interpreted analogously: To be probable is to be credible; it is to be a fitting target of credence. Indeed, many probability theorists, from Poisson to Keynes, did adopt this epistemic interpretation of probability. In earlier publications, I proposed an FA-model of value relations. Here, I present a structurally similar model of probability relations. Indeed, I suggests that probability on this epistemic interpretation is a kind of value. One of the advantages of my model is the account it provides of Keynesian incommensurable probabilities. (Cf. A Treatise on probability, 1921.) It goes beyond Keynes, however, in distinguishing between different types of probabilistic incommensurability, some of which Keynes himself might not have been willing to allow.