A New and Improved Whole Life Satisfaction Account of Happiness Feldman, F. (2019). An improved whole life satisfaction theory of happiness? International Journal of Wellbeing, 9(2), 1-7. doi:10.5502/ijw.v9i2.762 Fred Feldman University of Massachusetts at Amherst ffeldman@philos.umass.edu Copyright belongs to the author(s) 1 ARTICLE An improved whole life satisfaction theory of happiness? Fred Feldman Abstract: Philosophers and others have been captivated by the idea that happiness may be understood to be “satisfaction with life as a whole.” For a person to be happy, according to this idea, is for that person to be satisfied with his or her life as a whole. The view has been developed in a variety of forms, but has been subjected to serious objection in every form. In a paper published in volume 1, number 1 of the International Journal of Wellbeing, Jussi Suikkanen presented what he took to be a new and improved form of Whole Life Satisfactionism. He tried to show that in his formulation, the theory overcomes objections that I have presented elsewhere. In the present paper, after describing the context in which Suikkanen’s proposal appears, I present what I take to be the central point of Suikkanen’s work. I mention some obscurities. I try to show no matter how these obscurities are resolved, the proposed view is still open to objections similar to ones already in the literature. Keywords: life satisfaction, whole life satisfaction, theory, happiness 1. Background Going back at least to Tatarkiewicz’s magnificent 1962 work1, philosophers and others have been captivated by the idea that happiness may be understood to be “satisfaction with life as a whole.” For a person to be happy, according to this idea, is for that person to be satisfied with his or her life as a whole. The view has been developed in a variety of forms. A simple actualist version of Whole Life Satisfactionism (or “WLS”) would say that a person is happy at a time if and only if he actually judges at that time that his whole life is satisfactory. In this stark form, the theory is seriously implausible. An unreflective person could be very happy at a selected time, but at the same time not be thinking about his whole life and not be making any judgments about it. Indeed, a person could lead a long and by all intuitive accounts happy life without ever making any judgment about his life as a whole. Some advocates of WLS have endorsed versions of the view that do not require actual judgments of satisfactoriness. They maintain that a merely hypothetical, or counterfactual judgment is sufficient. A simple hypotheticalist version WLS would say that a person is happy 1 Analysis of Happiness (1976) Władysław Tatarkiewicz; translated from the Polish by Edward Rothert, Danuta Zielińska, vol. 3 in the Melbourne International Philosophy Series (Melbourne, Australia: University of Melbourne). Originally published as O szczesciui (Warsau: PWN/Polish Scientific Publishers, 1962). Whole life satisfaction Feldman 2 at a time if and only if she is like this at that time; if she were to consider her whole life, she would judge that it has been satisfactory. This may seem to overcome the difficulty of the unreflective happy person. Perhaps we can say that if the unreflective happy person were to reflect on his life as a whole, he would judge it to be satisfactory. That explains how he can be happy, even though he never actually thinks about his life as a whole. But a hypotheticalist version of WLS goes wrong in related cases with slightly different stipulations. Suppose Timmy is an unreflective party animal. He is happy leading a life filled with sex, drugs, and rock ‘n’ roll. His happiness depends essentially on the fact that he never thinks about “deep questions” such as the question whether his life as a whole has been worthwhile. If he were to think about it, he would be dismayed and would judge that his life has been a waste. He’s, in fact, happy, but if he were to make a judgment about his life as a whole, he would not judge it to have been satisfactory2. 2. Suikkanen’s approach In a recent paper in this journal, Suikkanen argues for what he describes as an “improved version of the WLS theory of happiness.”3 Apparently alluding to the so-called “Ideal Observer Theory,”4 Suikkanen distinguishes between the actual subject whose level of happiness the theory is intended to assess, and an idealized version of the subject. The idealized version of the subject is assumed to be rational and well informed. He knows a lot about the preferences and activities of the real-world subject. The theory does not assume that the real-world subject makes any judgment about his own real-world life. Nor does it assume that the idealized version of the subject makes any judgment about his ideal life. Instead, the idealized version of the subject is called upon to make a judgment about the degree of similarity between the actual life of his actual-world counterpart and an ideal life that can be imagined for him. Following Suikkanen, let’s call the real-world subject “S.” Suikkanen’s theory is intended to give an account of S’s level of happiness in the real (or “evaluated”) world5. Suikkanen calls the idealized counterpart “S+.” S+, knowing the important facts about S’s values etc., and also knowing the important facts about what goes on in S’s life, constructs an “ideal life-plan” for S. Presumably the idea here is that the ideal life-plan would include an account of behavior for S; presumably, S+ chooses behaviors such that, if S were to engage in them, doing so would help to make S happiest, given his preferences and values. The core of Suikkanen’s theory is this: S’s level of happiness in the real world is determined by the quality of the fit between the life that S actually lives in the evaluated world, and the life as described by the ideal life-plan that S+ constructs for S. Suikkanen states (p. 157) the theory as follows: 2 I presented this and other examples that Suikkanen discusses in my (2010) What is this Thing Called Happiness? pp. 83-88, as well as in “Whole Life Satisfaction Concepts of Happiness,” Theoria, 74, (2008), 219-238. 3 Suikkanen, J. (2011) “An Improved Whole Life Satisfaction Theory of Happiness,” International Journal of Wellbeing, 1(1), 149-166. Parenthetical page references that follow are all to this work. 4 Discussed by Roderick Firth and many others back in the 1950s. 5 More precisely, the theory is intended to give an account of S’s level of happiness at a selected time, t. To make things simpler here, I will, for the most part, suppress the time relativization. I think this decision on my part will not affect the argument. Whole life satisfaction Feldman 3 A person, S, is happy(g)6 to degree n at time t if and only if: 1) There is a certain life that S has lived up to t in the evaluated circumstances, and 2) if, at t, an idealized version of S, S+, were to form an ideal life-plan for S’s life, and 3) if, at t, S+ were to form a moderately detailed conception of how S’s life in the evaluated circumstances has transpired so far, then, 4) at t, S+ would judge that S’s actual life matches S+’s ideal life-plan for S to degree n. 3. Some details This is just the bare bones of a theory. To fill it in, we must have a clearer conception of the knowledge, values, and motivations of the idealized version of S. We also would need a clear account of the constraints, if any, that restrict S+’s choice of life-plan. Otherwise, it will be impossible to determine what sort of ideal life-plan S+ will construct for S. If we cannot determine the details of the ideal life-plan, it will be similarly impossible to determine whether S’s real life corresponds closely to that life-plan or whether instead it diverges. In other words, without some details about the idealized version of S, the theory does not yield any conclusion about how happy S is. Suikkanen makes use of some ideas derived from Michael Smith’s work when he introduces the concept of the idealized version of the subject. He mentions that Smith had said that the ideal version (as used in his theory about desirability) would be “fully rational” (157). In addition, the ideal version is supposed to have “all relevant information” (156). At the outset, it appears that Suikkanen means to incorporate these features into his conception of the idealized version of the subject. When Suikkanen first describes the differences between the actual version of a subject and the idealized version of that subject, he uses an example involving a person identified as “Ann.” Her idealized version is “Ann+.” Suikkanen says that in addition to being fully rational and well informed, Ann+ shares Ann’s likes and dislikes. She also knows what Ann wants and values during different stages of her life. On the basis of this information, Ann+ can formulate an ideal life-plan for Ann. If Ann’s actual life went according to this plan, she would be engaged in activities that she likes and her desires would be satisfied (p. 157). It appears, then, that the idealized version of a subject can be constructed in this way: first we start with the actual subject, complete with her actual likes and dislikes, preferences and values. Then, while leaving these fixed, we change her epistemic state. We make her “fully rational” and in possession of “all relevant information.” We then imagine that the idealized version, 6 At the beginning of his paper, Suikkanen sketches a distinction between three conceptions of happiness. These are the phenomenological sense (“Happiness(p)”), the global sense (“Happiness(g)”), and the welfare or prudential value sense (“Happiness(w)”). Much could be said about Suikkanen’s claims about these three concepts. But since I think his view is intended to be taken as a view simply about happiness understood in the most familiar sense, I shall suppress this reference to the three-way distinction. The curious reader is encouraged to study the relevant passages. Whole life satisfaction Feldman 4 armed with these conative and intellectual features, forms an ideal life-plan for the real-world version of the subject7. The level of happiness of the real-world subject is determined by the extent to which her real-world life corresponds to the life as described by the ideal life-plan. The closer the correspondence, the greater the level of happiness. If understood in this simple way, the requirements placed on the ideal counterpart would turn out in some cases to be impossible to satisfy. Consider a confused smoker (call him “Smoky”) who desperately wants to light up his next cigarette, but who also wants to avoid doing anything that will endanger his health. Smoky refuses to accept the overwhelming evidence that smoking is a serious health risk. Now try to imagine this smoker’s ideal variant, Smoky+. Smoky+ is stipulated to have all of Smoky’s desires and preferences, but unlike Smoky, Smoky+ is also stipulated to have all relevant information and to have coherent preferences. Obviously, there is a problem here. Either Smoky+ lacks the desire to smoke, or he lacks the desire to avoid risking his health, or he lacks the knowledge that smoking is a risk to health, or he lacks rationality. The stipulated combination of conative and epistemic features is impossible. In later passages in his paper, Suikkanen hints at some versions of the theory that do not require the ideal counterpart to have all of the real-world subject’s conative attitudes (pp. 159- 60). In one variant, all that is required is that the idealized version have the largest collection of these attitudes compatible with being fully knowledgeable and rational. In another version, what’s required is that the idealized counterpart have a maximally large subset of the real- world subject’s conative attitudes weighted somehow for centrality8. Suikkanen seems to be saying something like this in this passage: … one could stipulate A+ to have an informed and coherent version of A’s set of conative attitudes. We could then consider what kind of life-plan A+ would form for A on the basis of that idealised set. The resulting theory would claim that a happy(g) life needs to be a desirable life in Smith’s sense (p. 160). Suikkanen provides a brief account of his reasons for thinking that this theory gives the correct results in the case of Timmy9. As Suikkanen puts it: … we first imagine a more rational version of Timmy who knows Timmy’s likes and dislikes (and his desires, values, and the like). Because of this knowledge, the idealised Timmy (Timmy+) can come to believe that the best life-plan for Timmy in the real world just is to live spontaneously. Any selfreflection from his actual counterpart would merely be counterproductive. For this reason, the life -plan which Timmy+ would have for Timmy would not contain any such reflection. 7 In the final section of this paper, I raise some questions about the details of the ideal life-plan. For now, these details remain a bit vague. 8 Each of these formulations seems to presuppose that there will be just one ideal life-plan for the subject. I cannot see what could justify this assumption. Perhaps in some cases, there will be several different ways in which the subject’s preferences could be made coherent; perhaps none of them is uniquely “best.” In that case, there could be conflicting answers to the question of whether the subject is happy. 9 As discussed above in the 4th paragraph. Whole life satisfaction Feldman 5 According to the ideal life-plan then, Timmy should not think about his life. Rather, to follow this plan, Timmy should simply continue to take part in the activities he enjoys. Timmy+ can therefore come to judge that Timmy’s actual life meets the standards of the ideal life-plan he has for Timmy. As a result, this view can provide the right intuitive conclusion according to which Timmy is happy(g) in real life (p. 158). 4. A puzzle for Suikkanen’s view There are several questions that might be raised concerning Suikkanen’s theory10, but I want to focus here on one that I find especially problematic. As I see it, a crucial feature of the theory remains obscure. This concerns the life-plan that S+ is called upon to construct for S. What puzzles me is the question of whether the life-plan is supposed to be in any sense “possible” for S. Can S+ dream up a life-plan that would be completely impossible for S? Or must the plan describe actions that are – or previously were – live options for S? So far as I can tell, Suikkanen does not make his answer to this question explicit. Since every option I can think of is in one way or another problematic, it may be helpful to consider the possibilities. I will focus on three main options. 1. We might assume that the ideal life-plan for S has to describe a life that is, as of the time when we are evaluating S’s level of happiness, still possible for S. By “still possible” here, I mean to indicate that S would still have it in his power, as of the time at which S’s level of happiness is being evaluated, to make his life conform to the description provided by the life- plan. This conception of possibility will lead to problems in any case in which a person is unhappy due to factors no longer under his control. To see this in a stark example, suppose Smoky smoked like a chimney for many years. Eventually it caught up with him and his health has been ruined. Now, as he lies on his deathbed at t, Smoky is miserable. He does his best to cope with his fatal illness, but he cannot undo the mistakes of his youth. Suppose that when Smoky+ formulates a life-plan for Smoky, he is required to formulate a plan that will still be possible for Smoky. Given the restriction to life-plans that are still possible for Smoky, it would appear that every life-plan would have to include the lifetime of smoking that actually brought Smoky to his unhappy current state. Every still possible plan would include having had this unfortunate habit. A comparison of life according to the best of these life-plans with Smoky’s actual life would be easy: the lives would be pretty much the same. As a result, the implication of the theory would be clear, and clearly not as an advocate of the theory would intend. Under this interpretation, the theory implies that since his actual life 10 For example, why does Suikkanen stipulate that the idealized person who constructs the ideal life-plan for S must be a counterpart of the actual person? Wouldn’t any appropriately motivated and knowledgeable plan-constructor do as well? Furthermore, it appears that nothing in the theory implies that anyone is, or would be, satisfied with any whole life. So far as I can tell, whole life satisfaction plays no role in the theory. Instead, the crucial role is played by degrees of similarity between the subject’s actual life and a life-plan constructed by a rational and well informed observer. So why does Suikkanen describe his view as a form of Whole Life Satisfactionism? Whole life satisfaction Feldman 6 corresponds so closely to the life described in the ideal life-plan, Smoky is very happy at t. But of course this is wrong. He is miserable. He deeply regrets the now unalterable errors of his past. 2. We might consider the idea that the life-plan has to describe a life that contains activities that were, at suitable times before the activities took place, genuine options for the real-world subject. So even if some of these activities are no long available to that subject at the time of evaluation, they were previously available to him. If a robust form of determinism is true, and, in fact, all of a subject’s behavior was settled and unalterable prior to the fact, then aside from the actions he in fact performed, no subject ever has any “genuine options.” In this case, when a subject’s ideal counterpart formulates a life-plan for his real-world counterpart, it will have to describe the subject’s actual life. The implication then is clear: there is no interesting difference between any subject’s actual life and the life described in the life-plan. The theory then implies that everyone is maximally happy. This is deeply implausible. Even if determinism is not, in general, true, the problem described here would still arise in any case in which a subject’s misery is due to factors that were never under his control. Consider, for example, a subject, Luckless, who was born with a congenital condition that has made his life miserable. Suppose Luckless has done his best to cope with it, but to no avail. If we assume that when Luckless+ constructs the ideal life-plan for Luckless, Luckless+ has to leave unalterable components of the actual life in place, the result should be clear: the ideal life-plan will contain all the misfortune of the actual life. In these respects, the actual life will look very much like the life as described by the ideal life-plan. As a result, the theory implies that the miserable Luckless is outstandingly happy. But of course he is not. 3. Perhaps it will appear that we should consider the idea that when a subject’s idealized counterpart formulates a life-plan for that subject, he is permitted to describe a way of life that contains behaviors that never were possible for that subject. Then, when describing the ideal life-plan for Luckless, Luckless+ can include happy activities that Luckless would have enjoyed if he had not been born with a congenital defect. As a result, there will be a wide divergence between Luckless’s actual (and miserable) life, and the happier life described in the ideal life- plan. The theory would then declare (correctly) that Luckless is not happy. But if we allow the ideal life-plan to include actions that were never possible for the subject, the theory will go wrong in any case in which a subject is actually quite happy but could have been happier if he had been able to do the impossible. Suppose Mr. Chipper wants to be healthy, wealthy, and wise. Suppose he knows that he has always been healthy, wealthy, and wise. Suppose, as a result, that he is satisfied with his life as a whole. Mr. Chipper might like to be healthier, wealthier, and wiser than he in fact is. But he knows that in virtue of his extraordinary health, wealth, and wisdom, no improvement in these areas is possible. He is as healthy, wealthy, and wise as a person could be. He is perfectly content with the life he has actually gotten. He finds it satisfactory largely because he takes it to be about as good a life as he could have lived. He smiles and laughs; he views the Whole life satisfaction Feldman 7 idea of being impossibly healthy, wealthy, and wise as nothing more than an irrelevant pipe dream. But if Mr. Chipper+ is allowed to include the impossible in the ideal life-plan for Mr. Chipper, he could include impossible behaviors, such as taking imaginary supplements that would make him amazingly healthy, making imaginary investment decisions that would make him amazingly wealthy, taking brain enhancing pills that would make him amazingly wise. In this case, Mr. Chipper’s actual life is quite unlike the life described in the ideal life- plan. The theory then implies that he is not happy. But as the case has been stipulated, he is very happy. Careful scrutiny of Suikkanen’s paper leaves me unable to determine how he intends to answer this question about the possibility of life-plans. So far as I can tell, Suikkanen never specifies whether life-plans have to be “still possible,” or whether they must contain behaviors that were “possible at the time,” or whether they can contain behaviors that are and always have been impossible for the subject whose level of happiness is being evaluated. It seems to me that however we settle this puzzle, the theory goes wrong. 5. Conclusions In general, I am dubious about Whole Life Satisfaction theories of happiness. Objections have been given all the way back to the time of Tatarkiewicz. It appears to me that the version offered by Suikkanen is not an improvement over versions of the theory that are already in the literature. Acknowledgements I am grateful to Miles Tucker, Chris Heathwood, Owen McLeod, and Brad Skow for helpful suggestions concerning earlier drafts of this comment. Author Fred Feldman University of Massachusetts at Amherst ffeldman@philos.umass.edu Publishing Timeline Received 21 August 2018 Accepted 9 December 2018 Published 9 July 2019