Does Antoine de Saint-Exupéry's Little Prince Commit the Sunk-Costs Fallacy?

Saint-Exupéry's The Little Prince contains a simple explanation of what seems like an important aspect of love: history. The Little Prince seeks to explain why his rose, in particular, matters to him. (It seems likely that Saint-Exupéry was trying to understand and justify his tumultuous relationship with his wife.) The Little Prince is deeply troubled by the discovery of many, many roses that seem phenomenally indistinguishable from his rose. He says to them:
You are beautiful, but you are empty. One could not die for you. To be sure, an ordinary passerby would think that my rose looked just like you—the rose that belongs to me. But in herself alone she is more important than all the hundreds of you other roses: because it is she that I have watered; because it is she that I have put under the glass globe; because it is she that I have sheltered behind the screen; because it is for her that I have killed the caterpillars; because it is she that I have listened to, when she grumbled, or bloated, or even sometimes when she said nothing. Because she is my rose. [One might just as well say, “My gardenia” or “My frangipani.”]
As such quite different philosophers as Raymond Gaita and Stephen Clark have noted repeatedly (see, e.g., Stephen R. L. Clark, The Mysteries of Religion: An Introduction to Philosophy through Religion [Oxford: Blackwell 1986] 207-9; Stephen R. L. Clark, A Parliament of Souls, Limits and Renewals 2 [Oxford: Clarendon-OUP 1989] 116-35), history matters (this isn't the way Clark frames the matter). We are not simply interested in enjoying the beloved's phenomenal characteristics: we are interested in our relationship with the beloved her- or himself. To believe that my lover has been surreptitiously replaced with a phenomenally indistinguishable duplicate could easily be the first step on a path toward madness—not because the belief itself is insane (it might or might not be), but because trying to come to grips with the conflicting thoughts and emotions involved could prove devastating. When I love you, I am attached to you, not to anyone who happens to resemble you. (This is a way into one sort of quite serious problem with theories which seek to combine belief in (i) life after death understood as resurrection with (ii) a purely physicalist understanding of the human person.)

So, phenomenologically, we're with the Little Prince. Even apart from the issue of tender, compassionate care for the other that is a crucial part of any attitude we would recognize as loving, we value the beloved not simply because she or he can offer certain goods in the future but also because we are attached to her or him as a self extended over time. (It doesn't matter for present purposes whether the continuity is simply narrative in nature or whether it is rooted in some sort of persistent substance.) Once I have a conception of a relationship with the other as an individual with a history, rather than as a moment-by-moment source of satisfactions, then valuing that relationship will not be compatible with ongoing reevaluation of its investment potential: I value the relationship as such, and not merely for what it can produce.

The sunk-costs fallacy can be committed by someone who makes an inapposite judgment regarding the relevance of past expenditures to present-cum-future investment choices. Suppose I have spent $1,000,000 on a product line that has so far made me $250,000, and which can be disposed of for another $250,000, so that I am (or will be) down $1,000,000 total. I can spend another $100,000 to reinvigorate the line in some way, with some chance of boosting my profits, or I can spend $500,000 on another product line entirely, with a significant chance of substantial new income, while shutting down the first product line. What ought to matter, conventional economic reasoning suggests, is my net income on either of the two options. My past investment shouldn't figure in my decision about the future.

It should be clear that, as an entrepreneur evaluating investment options in a case like this one, I am not in the same position as the Little Prince.

As the example has been framed, I can measure investment, loss, and profit using the same currency, and my objectives can be expressed with exclusive reference to this currency, so that it is quite easy to see (allowing for the reasonable determination of probabilities) what my losses and gains can be expected to be. The Little Prince has invested time and care, but his goal is not to yield more time spent caring for the rose, so the comparison between investment and output can't be made straightforwardly in the way it can in the product-line instance.

Not only is the desired output incommensurable with what has been invested, at least in significant part, but the Little Prince isn't seeking an external good. What he values is precisely the ongoing relationship with his rose. His goal is not to maximize the æsthetic pleasure of encounters with the phenomenal features of his rose or of roses like her. He is not seeking to maximize anything external to his relationship with her. Rather, he seeks to extend and nourish the relationship itself. And he seeks to do this not because the relationship itself is a reliable means of generating certain satisfactions; the values the relationship yields are internal to it. As Robert Nozick (whose discussion of this topic I only read, thanks to a suggestion from David Gordon, after crafting the initial version of this post) observes crisply, “Don't commit the sunk costs fallacy”

may be a correct rule for the maximization of monetary profit, but it is not an appropriate general principle of decision . . . . We do not treat out past commitments to others as of no account except insofar as they affect our future returns, as when breaking a commitment may affect others' trust in us and hence our ability to achieve other future benefits; and we do not treat the past efforts we have devoted to ongoing projects of work or of life as of no account (except insofar as this makes their continuance more likely to bring benefits than other freshly started projects would). Such projects help to define our sense of ourselves and of our lives.
Robert Nozick, The Nature of Rationality (Princeton, NJ: Princeton UP 1993) 22.

That's true now, of course, but, an objector might say, it's not true initially. Initially, the Little Prince simply values and delights in the rose and cares for her accordingly—for her own sake, so that she will persist, and for his sake, so that his relationship with her will persist. At first, this seems likely to be because of its observable features and the characteristics of its relationship with. But, over time, he comes to be attached to the rose as an individual persisting over time, caring about the rose as an other with her own point of view and valuing precisely his connection with her.

Now, it is open to the objector to say that the Little Prince shouldn't value the rose in this way, that he should value her only as a source of future values or satisfactions. But what would be the rationale for this normative claim? The notion that the sunk-costs fallacy is a fallacy depends on the idea that the investor wants value of quantity X but settles for some sub-unity fraction of X simply because of prior investment. The idea of the sunk-costs fallacy as a fallacy presupposes that the investment can be evaluated in virtue of its potential to produce a certain kind of output. The investor acts irrationally, on the standard view of the fallacy, when she fails to maximize the relevant output.

But the Little Prince has become attached to this particular rose. The objector could dismiss his continued commitment to her as irrational only if the Little Prince's goal were to generate some value or satisfaction and if his continued commitment could be shown to generate this value or satisfaction in a lower quantity than some alternative in which he had not previously invested. But to judge the Prince as irrational on this basis would be to misunderstood the logic of his interactions with the rose. He doesn't value the rose in the way he would need to for the objection to find any purchase. And the charge of economic, or instrumental, irrationality thus can't get off the ground: this kind of charge is a judgment about means, not ends.

Perhaps, however, the objector might suggest that there was something irrational about the transformation in the Little Prince's attitudes—in coming to value a relationship with a particular rose, a particular individual, rather than valuing simply what the relationship, or the individual, might be expected to yield on an ongoing basis.

There will, of course, be economic arguments against this kind of position, rooted in the idea that future yields will be greater when sought in ways that don't depend on persistent attention to future yields. Committed engagement yields better outcomes than case-by-case, moment-by-moment, assessment of probable outcomes. As recent analysts have pointed out, it will be useful to have a reputation as someone who persists. (See here, for example.) And there will be informational and resource constraints on pursuing alternate options in ways that make it reasonable to consider sunk costs.

But while this may be correct, it doesn't capture the way we actually experience relationships with valued others, personal and impersonal. Consider the loss of a farm that has been in one's family for eight generations. The point is not that one has, or that one's ancestors have, invested in the farm at a particular level: evaluating the loss of the farm isn't a matter of evaluating investments, profits, or losses at all, but of one's attachment to the property and its role in constituting one's identity. Thus, “[t]hat there must be something which compensates for a finite loss is just a dogma, one which is more familiar in the traditional version to the effect that every man has his price.” (See Bernard Williams, “A Critique of Utilitarianism,” Utilitarianism: For and Against, by J. J. C. Smart and Williams [Cambridge: CUP 1973] 144-5.)

I may develop a relationship with a particular object, place, rose (or gardenia, or frangipani), or person for any number of reasons. But my ongoing interaction with it creates a bond, an attachment, that gives it a special place in my motivational structure and so in my deliberation. (The bond need not be created in this way; my point is simply that it can be.) (i) My story is now a story in which interaction and involvement with this entity plays a significant part. (ii) My self-understanding is as someone defined in relation to this entity. (iii) The entity occupies a significant role vis-a-vis my priorities. (iv) Emotionally, I am attached to this entity, with the emotion reasonably understood as registering its place in my story and in my identity and in my priorities. All this is particularly true if the focus of my relationship is an entity extended through time, not appropriately understood as simply a collection of things strung along a temporal clothesline. (John Searle has argued, without elaborate metaphysical foundations, that selves can be thus understood: see John Searle, Rationality in Action [Cambridge, MA: MIT 2001] 75-96.)

The Little Prince doesn't persist in his relationship with the rose because he mistakenly supposes that doing so will yield more external goods than any alternative. He does so because he values the internal good of the relationship with the rose, a relationship with an individual extended over time, and because the rose, and his relationship with her, have come to constitute (in part) who he is.


Unknown said…
The link to my paper on the sunk cost fallacy (linked to above) no longer works. It can now be found here:
(Or in its published form, here:

Popular posts from this blog

Libertarians for Redistribution

Santa Claus: America’s Most Wanted Fugitive

Socialism revisited