On Being Bryan

Cursory Thoughts On Hyperbolic Discounting And Self Promises

Recently I have been wondering about hyperbolic discounting (HD) and self-promising. I got introduced to the idea by a friend who is busy with a Masters in Cognitive Science. I'm in no way qualified to even give an overview of HD. However, having recently been engaged in thinking about the efficacy and legitimacy of self-promise-making - it occurred to me that it would be interesting to try and understand the practical consequences of overlapping these two domains.

Wikipedia has an entry on HD, which for this basic overview will serve as good a source as any. I would like to give this a much fuller treatment at a later date. The Wikipedia article (http://en.wikipedia.org/wiki/Hyperbolic_discounting) explains that a person (I'm lead to believe that this was assumed about other species as well) faced with two rewards, will prefer a reward that arrives sooner over one that arrives later. Initially it was assumed that we discount the later reward by a factor that increases exponentially as the amount of time we have to wait increases. But this has been discredited in favor of a hyperbolic model of discounting in which "valuations fall very rapidly for small delay periods, but then fall slowly for longer delay periods".

The way this curve is modeled for individuals is to find out their preference for something in the short-term, and something else in the long term, and then to compare the preference ranking. For example, "would you like 1 free car wash now, or 2 free car washes tomorrow?" and then "would you prefer 1 free car wash in 1 year, or 2 free cars washes in 1 year and 1 day?". A sizable portion of people will apparently prefer the one free car wash today, forfeiting the additional one tomorrow, but will inconsistently be happy to wait for an extra day in one years time. The HD theorists describe this as "individuals making choices today that their future self would prefer not to make, despite using the same reasoning."

HD falls into the domains of behavioral economics or cognitive science, and self-promise-making falls comfortably, I believe, into the realm of moral philosophy. This comparison only occurred to me after quick-reading a paper entitled "Promises to the Self" by Allen Habib of the University of Calgary. So regardless of the accuracy of my taxonomy, for the sake of this post I am going to go ahead and assume that even though these domains are clearly different in a number of ways, the notions effect each other in at least one direction.

Habib's work is useful - in fact I sought it out through a kind of bias - because unlike the majority of work on promise making, Habib is arguing that self-promise-making does introduce some kind of obligation, and is this a legitimate source of obligations we can acquire. The majority of the literature seems to argue that although we use the word "promise" when bargaining with ourselves, we have not actually introduced an obligation per se. There is also the question of if a promise of any kind even introduces any obligation. In addressing these questions Habib moves to understand promissory obligations as being comprised of three key features, that when combined make them quite different from other kinds of obligations. For Habib "promissory obligations are created by the promiser when she makes the promise, informed by the utterance or expression that constitutes the act of promising, and can be waived by the promisee". Therefore a promissory obligation is brought into being through the act of promising, as opposed to other kinds of obligations which are generally acquired through other means (legal, social norms, moral etc). Habib points out that promissory obligations are unique in that the recipient of a promise can cancel the obligation merely by saying so. Pulling this together, Habib has a clear view of promissory obligations as ... "an obligation that the promiser has to do what she promised, because she promised it, and furthermore the promisee can release a promiser from her obligation by simply saying so". He thus contends that an obligation exists when a promiser undertakes the act of promising and thus owes the promisee some promised act, and the promisee is owed the promised act, and the promisee can forgive the debt and waive his right to the promised act.

Having set this framework, Habib goes about laying several scenarios where a person makes a promise either directly to themselves, or to themselves as a member of a group that might acquire the promiser's obligation, where the promiser is part of the group. One example is of a sales manager who promises a cash-prize to the sales-person who sells the most stock in a given time-frame, and where the sales-manager is considered part of the sales force. Should the sales-manager sell the most stock, the question of has the sales-manager got an obligation to pay the money to herself?

There are a number of arguments against Habib's ideas for both self-promissory obligations as well as promissory obligations more generally - and he goes some way to answering potential objections. But what interests me is how, if we grant that Habib is successful in arguing that self-promises are real promises such that they introduce a clear and present obligation on one to one's self, we are essentially trusting ourselves to uphold this obligation, yet we are also the same party that can legitimately forgive the obligation.

It occurs to me that we often undertake acts of self-promise in order to correct for some behavior that we are not happy with, or to ensure that our future self is in a better position than our current self, regardless of our current selfs desires. But from our discussion above we saw that as humans we often do act in a way that is inconsistent with the supposed desires of our future selves. So perhaps self-promise making is an attempt to correct for this problematic situation where one is aware that the more proximate reward will often win over the slightly less proximate one. Perhaps our self-promising can in fact be a way of setting up an obligation to not accept the most proximate reward. The benefit of an obligation is that they are generally not easy to shake. I'm sure you see the problem now though? With a self-promissory-obligation, the party with the right to forgive the obligation, is also the party that stands to benefit the most from the forgiveness of the obligatory debt. Thus should the reward be strong enough, all it would take is a simple internal negotiation and our attempt at overcoming the inconsistent HD behavior would be rendered futile.

So with this in mind, where overcoming proximate reward selection is truly desired, perhaps the best way to achieve the goal would be to take on an obligation that is not as easy to thwart. I'm thinking that this could be in two forms. We could either retain the self-promising approach, but voice the promise so that others might make us aware of our attempted negotiation from under the obligation. Or, perhaps more effectively, it would be best to make the promise to a separate party. This way, assuming promissory-obligations are in fact true obligations, we can at least be sure that undermining our own strategies would not be as simple as an internal rationalization.