How Many Concepts of Preference Are There? A Sequel.

In a recent post I asked how many concepts of preferences are in use and I answered: basically two. Some authors favour a behaviourist concept of preferences and some authors favour a mentalist concept. See the original post for an elaboration of this distinction.

I also asked whether the concept of preference used in philosophy differs from the concept used in the social sciences. Since the behaviourist and the mentalist concept are both employed by philosophers and social scientists, I denied a difference cutting along the disciplinary boundaries.

In hindsight, however, I neglected another distinction: Some authors assume that the concept of preferences implies a selfish motivation and some do not. If we ignore what it means for a motivation to be selfish, we might be tempted to subdivide each of the two concepts of preferences discussed in the earlier post further into a version with and a version without selfishness assumption. The following options seemingly result:






No Selfishness-Assumption



However, as soon as we take into account what it means for a motivation to be selfish, one option becomes dubious: A. If preferences are just re-descriptions of choice behaviour, then I don’t see why we should not allow re-descriptions. Assume Matilda works all her life to give money to needy anonymously. No one knows about it The easiest re-description seems to be that she prefers to help those in need. A standard response might be that she might actually do it for feeling happy about being a good person. This is a possible interpretation, but as a re-description it is actually more complex than just attributing non-selfish preferences. I can describe Matilda as preferring to spend her money on helping the needy, or I can describe Matilda as preferring to spend her money and what makes her happy and helping the needy. The first description is more straightforward.

I don’t want to get into the debates about the possibility of altruism, rather I want to point out that if one accepts a behaviourist account of preferences, the more plausible version of this account does not include a selfishness assumption. I doubt that anyone wants to actually endorse this version, so effectively we end up with three analyses of preferences. (And I find B fairly unhelpful as well, but this becomes more difficult to argue and I therefore leave it for another time.)

How Many Concepts of Preference Are There?

It has been recently suggested to me that there might be a difference between the philosophical concept “preference” and social science concept “preference”. Since I am working in philosophy of the social science, getting such a distinction clear would be important for me. Given my area of work , I am especially in danger of mixing them up.

But I am not entirely convinced that there is such a distinction in the first place. In my notes I mainly distinguish a behaviourist-constructivist concept of preferences from mentalist-realist one. The distinction follows basically the 2016 paper by List and Dietrich, in which they argue for the mentalist-realist concept. You can very the exact formulations of these analyses, but the difference between the two concepts should be clear enough.

According to the behaviourist-constructivist approach preferences are nothing but logical constructions out of choice behavior. Preferences turn out to be mere re-descriptions of choice behaviour. If you always choose the orange over the apple, then this implies that you prefer oranges over apples. Many economists ¬†apparently endorse such a concept of preferences, and so does Simon Blackburn in¬†Ruling Passions. Therefore, the concept is used by social scientists and philosophers. One might argue that Blackburn only intends to reconstruct what social scientists are saying, but reading his text I get very much the impression that he buys into it. (On page 167 he notes that he would prefer to use the word “concern” rather than “preference”. He apparently endorses the concept, and only objects to the hedonist connotations associated with the word.)

On the mentalist-realist account preferences are real mental entities with causal efficacy. If you always choose the orange over the apple, this is evidence that you prefer oranges over apples, but it does not imply it. As soon as one endorses the mentalist position a myriad of questions arise within the ontology of mind. How are preferences realised? How can mental entities be efficacious at all? I am not going to say anything on these issues here and only note that List and Dietrich argued forcefully that the social sciences, and in particular economics, need such a mentalist concept of preferences for their explanatory purposes. Continue reading “How Many Concepts of Preference Are There?”