Jump to accessibility statement Jump to content

94th Joint Session of the Mind Association and Aristotelian Society

94th Joint Session of the Mind Association and Aristotelian Society site image

Postgraduate Sessions: A Choice Pump Argument for Adaptive Preferences

Annalisa Costella (Erasmus)


Since Elster’s (1983) seminal work on “sour grapes” – the supposed irrationality of adaptive preferences – scholarly contributions on the topic have thrived. The vast majority of scholars agree with Elster that a preference can be rational only if it is autonomous. This consensus, I argue, stems from conflating the rationality of adaptive preferences with their contribution to an individual’s well- being. It is doubtful whether autonomy is a satisfactory criterion of rationality. In the paper, I put forth an alternative criterion for the rationality of adaptive preferences, which I label the “opportunity pump argument”. I argue that adaptive preferences are irrational if they are subject to being opportunity-pumped. An opportunity pump induces the agent to strictly prefer having the opportunities offered by her final restricted set of options to the ones of her initial set. And thus, adaptation is irrational for an individual when, by adapting her preferences over the options, the individual also adapts her preferences over the opportunities provided by those options.


Annalisa Costella is a first-year PhD student in Philosophy at Erasmus University Rotterdam. She works at EIPE, the Erasmus Institute for Philosophy and Economics. Before starting her PhD, she completed a BSc in Economics at Bocconi University Milan, an MSc in Philosophy of the Social Sciences at the LSE, and a Research Master in Philosophy and Economics at Erasmus University Rotterdam. Currently, she is working on the relationship between individuals’ preferences and freedom of choice. She investigates in what ways (if at all) an individual’s preferences affect her freedom (to choose) and, conversely, in what ways (if at all) expansions or restrictions of freedom influence the formation of preferences. The goal of her research is twofold, both descriptive and normative. She aims to offer a descriptive account of how preferences and freedom (of choice) interact as well as investigate how they ought to do so. Her broader research interests include decision theory, axiomatic approaches to freedom of choice, and the welfare implications of behavioural economics


  • Olof Leffler

    11th July 2020 at 1:22 pm

    Thanks for the talk! I don’t have an account of what makes a preference adaptive, but I would, in general, have thought that adaptivity has something to do with how preferences are adapted to subpar or unusual circumstances, and they are problematic at least in part for this reason – the problem with them does to some extent depend on how they are formed. Could you say something more about what you think about that intuition? It’s not clear to me that it’s captured by the case you present here, but maybe I haven’t understood it well enough, or maybe it’s wrong or at least insufficient?

  • Pete Faulconbridge

    12th July 2020 at 7:12 pm

    Thanks for your paper! I’m worried that my question is based on my having misunderstood part of your career-choice example (it’s a bit late in the conference, and I’m not feeling as sharp as I’d like!), so my apologies if that’s so, and please feel free to disregard it if I’m missing the mark. I was not quite clear how the agent in the career-choice example exhibited adaptive preferences, as such. When you initially described the case, he ended up taking the consultancy job only after all the other deadlines had passed, i.e. when it was his only option. As you described it there, this was definitely not his preferred outcome at the beginning of the process. However, it seems to me that we can’t infer from the fact that he ultimately chose C that this means he must have come to prefer C to the other options, because at the point when he chooses C it’s actually the only option still left to him. It looks likely that there would still be some form of irrationality involved in missing the other deadlines, but it’s not clear to me that he must have adapted his preferences.

    I suspect that this might be just a matter of how the case is presented (or, as I mentioned, perhaps it’s just me misinterpreting what you go on to say later in that section of the paper), but I wonder whether taking the temporal dimension out of the case (i.e. not having the deadlines passing) might help to focus the audience’s attention onto the adaptive preference aspects of the paper? Or is the temporal dimension playing an important role for your purposes, which I am overlooking?