Philosophers have long debated the nature of happiness, with some saying that happiness is just a certain kind of psychological state and others claiming that true happiness is not just a matter of having certain feelings but also requires genuine virtue.

The new field of experimental philosophy may not be able to help us arrive at a definitive resolution of this age-old debate, but at the very least, it does seem to have inspired a very funny interactive video!

(Note: To go through this interactive video, you have to click at the end of each segment to begin the next one.)




28 Replies to “Philosophy of Happiness: The Video

  1. Interesting results, but I disagree with the characterization of these results as revealing some kind of difference between the concepts of happiness and unhappiness. Another perfectly plausible conclusion is that unhappiness is just the absence of happiness, and positive feelings about one’s own life are a necessary but not sufficient condition for happiness.

  2. I agree with Benjamin. The experiment assumed that the descriptions contained in the three bullet points were the only thing that mattered on the happiness evaluation. That way, whatever difference found would be due to personal values. Well, it can be, but the experiment is not enough to make that conclusion.
    Besides, the claim that both had exactly the same mental states is a stretch. Using the same descriptions of the states (in the three bullet points) doesn’t mean they had exactly the same mental states. The descriptions are simply overgeneralizations. When people see both stories told, they take in consideration things that are outside the bullet points. So, in a way, even when the narrator says they share the same mental state, one can doubt it based on the other characteristics seen, and adjust the judgment accordingly.

  3. Rodrigo,
    That definitely does seem like a live possibility. The main question, however, would be why the difference between these two kinds of lives seems to be affecting people’s happiness judgments but not their unhappiness judgments. Do you have any thoughts on the matter?
    (I don’t mean to imply that there is no way to explain these results using the basic strategy you outline here; I just wanted to hear a little bit more about precisely how you propose to do it.)
    That’s a really great suggestion. Phillips, Nyholm and Liao actually conducted a follow-up experiment to try testing it out, and I’m curious to see what you think of their results.
    They presented all participants with the same vignette, but then they asked some participants whether they agreed that the agent was ‘unhappy’ and other participants whether they agreed that the agent was ‘not happy.’ These two questions might at first seem almost identical, but notice that the former uses the concept of unhappiness, while the latter uses the concept of happiness. Sure enough, people’s value judgments impacted intuitions about whether the agent was ‘not happy’ but did not impact intuitions about whether the agent was ‘unhappy.’

  4. Why did these researchers choose to use a female character for the scenarios? Has anyone done a follow-up to see whether intuitions about happiness are tracking with traditional gender norms? The differences in intuitions about happiness might be a case of implicit gender bias. Mommy Maria is self-sacrificially investing in and nurturing others, which tracks well with a traditional view of the role of women (or “natural” feminine behavior). Party Maria, on the other hand, is investing only in herself, and the glamorous, self-centered life is the antithesis of the traditional feminine role (and thereby “unnatural”). I wonder whether participant responses would vary as greatly between the cases if the scenarios were kept identical except the main character is Mike instead of Maria. I would also like to see a comparison between the Mommy Maria and “Mommy” Mike scenarios to see if there is a significant difference in attributions of happiness. If “Mommy” Mike is widely perceived as acting contrary to masculine nature, then fewer people may be willing to make the attribution of happiness.

  5. Jenni,
    Excellent point. The reason we used a case that revolves around traditional gender norms is that we wanted to pick a case in which the value judgment would itself be controversial. That way, we can get a better grip on the relationship between people’s value judgments and their happiness attributions.
    More specifically, what the results show is this: The majority of participants say both (a) that Maria the Wholesome Mother has a good life and (b) that she is happy. However, those participants who think that Maria the Wholesome Mother does not have a good life tend also to say that she is not happy.
    This fact about the relationship between people’s two answers helps to show that the difference we observe between happiness attributions in the two conditions is, in fact, driven by a difference in judgments about the value of her life.

  6. Hi Josh,
    Interesting. Can you say more about the details of the follow up study? What kind of “impact” (and evidence of this) are you referring to when you wrote, “Sure enough, people’s value judgments impacted intuitions about whether the agent was ‘not happy’ but did not impact intuitions about whether the agent was ‘unhappy.'”?

  7. I guess I’m pretty skeptical. Why should we care about what the “folk” think about these cases? I judged all four cases exactly the same way. My value judgments affected neither my attributions of happiness nor happiness — and I think my position is more philosophically defensible than the the judgments of the folk. I can’t help but worry that these types of studies threaten to turn philosophy into something of a popularity contest. There is, after all, the threat of something like the following argument: “You don’t share the intuitions of the majority. So, you’re not using the concept of happiness correctly.” Maybe this is uncharitable, but I’ve faced this sort of argument before in real life as a philosopher.

  8. Brad,
    Thanks for the question. I’ll try to give you a little bit more of a sense of what happened with that follow-up study.
    Participants were told about an agent named Mark. In one case, Mark was described as spending his life hopping from nightclub to nightclub across America (meaningless life condition), and in the other case, he was described as volunteering at an orphanage in Africa (meaningful life condition). In both cases, Mark was described as experiencing a lot of negative emotion and dissatisfaction with his life.
    After reading about Mark living one of these two lives, participants were asked whether they agreed or disagreed with one of these two statements:
    (1) Mark isn’t happy.
    (2) Mark is unhappy.
    What we found was that there was a statistical difference between agreement ratings for the first statement (about happiness) such that participants agreed less when Mark was described as living a meaningful life than when Mark was described as living a meaningless life, but there was no difference in their agreement ratings with the second statement (about unhappiness): they just completely equally agreed that he was very unhappy, no matter what type of life he was leading.
    The difference in the agreement ratings was a bit smaller in this study than in the study in the video, but that’s not too surprising given that Mark’s life is described in such a minimal way.
    Anyway, I would be very interested in hearing your take on this follow-up study – I was actually a bit surprised by the result.

  9. Hi Jonathan,
    I would assume that we will avoid strong agreement with ‘Mark isn’t happy’ as long as we think there is some significant respect in which he is happy, but that we will be happy to strongly agree with ‘Mark is unhappy’ if Mark suffers quite a bit of bad emotional/satisfaction states.
    As you describe the set up, I think this allows for an explanation that appeals to Aristotelian happiness and one that does not.
    To see the non-Aristotelian explanation, just think about how one might naturally “fill out” the story in one’s mind. If I am told that Mark is volunteering at an orphanage in Africa but also experiencing a lot of negative emotion and dissatisfaction with his life, I am inclined to think that he still has *some significant* positive emotional experiences and some degree of self-esteem/self-approval. I am assuming that you do not describe him (1) as deeply depressed and unable to even enjoy conversation or seeing the children laugh or (2) as regretting helping children and thinking it is a worthless activity. These point in the direction of residual positive “subjective” features that will push us, for non-Aristotelian reasons, to resist strong agreement with “Mark isn’t happy”. Now in the other case, it is much easier to fill out the case with few if any significant subjective aspects. Mark was described as spending his life hopping from nightclub to nightclub across America, and it is easily to expand from the “negative mental states” description you give to depression, addiction, joyless repetition, and a judgment that his life is worthless and a lack of self-esteem. So in this case we will be more willing to strongly agree with ‘Mark isn’t happy’.
    In both cases we will agree with ‘Mark is unhappy’ in about the same amount, because (I am guessing) being in a significant negative emotional state is sufficient for strong agreement.
    I predict that if you run the experiment while spelling out the negative subjective condition in a way that blocks most naturally posited residual positive states (e.g. you say do-gooder Mark is deeply depressed, unable to even enjoy conversation or seeing the children laugh, regrets helping children, and thinking it is a worthless activity, etc) then the effect might well disappear. And this would support the conclusion that our thinking about happiness is to some extent “non-Aristotelian”

  10. I also have to second the worry that judgments may be influenced by tacit assumptions “between the lines.”. I, for example, had the distinct introspective experience when watching the one woman do cocaine that she couldn’t *possibly* be happy because no truly content person would go around doing cocaine. Alas, I didn’t let this assumption affect my judgment of her happiness — I actively suppressed the assumption — because I was explicitly instructed to believe that she had exactly the same psychological characteristics as the first woman. I suspect, however, that many of the “folk” being tested are not as rigorous as I am, and this that their judgments plausibly *were* corrupted by tacit assumptions that she couldn’t possibly be as psychologically healthy as the first woman.

  11. Marcus,
    I completely agree that it would be a big mistake to reduce this whole thing to a popularity contest. Our goal isn’t to do anything like that but rather to get at the underlying psychological processes that draw people toward one intuition or the other. Then, once we get a better sense of what is pulling us in these different directions, we can have a better idea of whether it is something we ought to trust or something we should simply dismiss.
    Your recent comment is clearly a paradigm case of this sort of approach. There, you propose a specific psychological mechanism that would generate these results even if people’s concept of happiness had no normative component. The hypothesis thereby gives us a way of ‘explaining away’ the results without adopting a normative theory of people’s concept of happiness.
    But I don’t yet completely understand the hypothesis you are proposing. If people assume that only a discontented person would do cocaine, why aren’t they more inclined to say that the cocaine-using version of Maria is more *unhappy*?
    p.s. I don’t know if people around here are already familiar with Marcus’s own experimental work, but if you aren’t, you should definitely check it out! See:

  12. Thanks for the shout-out, Joshua!
    A quick reply to the question you posed: The hypothesis I was floating is that people might be tacitly *violating* a belief that the experiment asks them to assume: namely, that Maria1 and Maria2 have all of the same psychological characteristics (e.g. both are equally excited about their lives, etc.).
    My introspective feeling when I encountered the two cases was that there was a strong implicit bias in me to fill in Cocaine-Maria’s psychological story in a way that clearly differentiates her psychologically from Housewife-Maria. The implicit thought I felt was something like: “Well, Cocaine-Maria can’t be *as* excited about her life as Housewife Maria — because, if she was as excited about her life, she wouldn’t be driven to do cocaine.”
    As an aside, having been a semi-professional rock guitarist in the not-too-distant past, I encountered a great deal of evidence in favor of this bias. I never met a drug user (yes, I met a few!) who was *truly* excited about life. They tended to “fake” being excited about life (often by their very own admission!).
    Anyway, that was the worry: (1) the experiment asks participants to accept the assumption that Housewife-Maria and Cocaine-Maria have the same psychological characteristics, but (2) participants may in fact, in highly tacit ways, not abide by the assumption, due to beliefs we tend to have about the psychological states of drug users, partiers, etc.

  13. Hi Marcus,
    I think I’ve got a good feeling for the shape of the explanatory strategy you’re proposing here. What I was trying to understand was how this strategy could explain the precise pattern of results observed in the present study. What needs to be explained is that when we compare Cocaine-Maria and Housewife-Maria, we do find a difference in people’s willingness to say that they are ‘happy,’ but we don’t find a difference in people’s willingness to say that they are ‘unhappy.’ How are you proposing to explain this pattern? If there is nothing normative in people’s concept of happiness, why do we get this effect for ‘happy’ that we don’t also get for ‘unhappy’?
    (I’m not at all suggesting that there is no way for your hypothesis to explain this pattern — I’m just asking for further clarification.)

  14. Hi Brad,
    If I understand your suggestion correctly, it really has two separate parts. The first is that the concept of happiness differs from the concept of unhappiness in that it is easier to meet the satisfaction conditions for unhappiness than it is to meet the satisfaction conditions for happiness. (To be unhappy, all it takes is some negative psychological states.) The second part is that the difference between the conditions for happiness can be explained by how participants filled in the cases differently (and this difference won’t affect unhappiness judgments because those conditions have already been satisfied).
    I think both of these points are possible and definitely interesting, and I also think that your suggestion for a further experiment would really help us see whether the second piece of your explanation is on the right track. (Another way to do it, would be to ask participants more directly about how they were filling in the details after they’ve completed the happiness question.)
    I’m still a little bit unsure of the first part of your explanation though. I agree that the satisfaction conditions for happiness are more stringent than the those for unhappiness. However, the happiness claim is negated, so participants agreed with the statement ‘Mark isn’t happy.’ But on your proposal, it seems that if it’s harder to ‘be happy’ than it is to ‘be unhappy,’ then it should be easier to ‘not be happy’ than to ‘be unhappy’ – all it would take is to not have one of the many conditions that happiness requires. If anything then, it seems like it might be more likely for the small ways we fill in the case of Mark the volunteer in Africa to change our agreement with “Mark is unhappy.”
    Perhaps there’re still some modifications that can be made to press forward with this sort of explanation though. I have the sense that you might have been suggesting a somewhat deeper explanation than I’ve given in my short response.

  15. Sorry, I guess I should have addressed that, but I think my hypothesis explains it very well.
    My hypotheses are:
    (1) We *rightly* judge First-Housewife-Maria to be happy in the first case because she feels great and excited about life.
    (2) We *wrongly* just First-Cocaine-Maria to be unhappy due to latent/tacit biases we have about the (hidden) psychological features of cocaine users and partiers.
    (3) We *rightly* just Second-Housewife-Maria and Second-Cocaine-Maria to *both* be unhappy because (A) they both feel terrible, and (B) this only confirms the latent/tacit beliefs we already had about First-Cocaine-Maria.
    That’s why we get the effect.
    In the “happy” case, our biases (only) corrupt our judgments about Cocaine-Maria — so we judge Housewife-Maria happy and Cocaine-Maria unhappy.
    In the “unhappy” cases, our bias (regarding Cocaine-Maria) is supported by further, explicit evidence that apply to *both* Marias (both “feel terrible”) — so we judge both to be unhappy.

  16. Hi Jonathan,
    You wrote: “on your proposal, it seems that if it’s harder to ‘be happy’ than it is to ‘be unhappy,’ then it should be easier to ‘not be happy’ than to ‘be unhappy’ – all it would take is to not have one of the many conditions that happiness requires.”
    I think posing questions about people “being happy” and “not being happy” are bound to mislead. It would be better to talk about being completely or maximally happy, reasonably happy, a least a little happy, and not being happy at all. This is still pretty simplistic but at least it is a step in the right direction.
    My hypothesis here is that since you stick with ‘happy’ and ‘not happy’ the question about strength of agreement ends up tracking “degrees of happiness”, not whatever else it is meant to track (degrees of confidence?).
    The idea is that someone who thinks J is a least a little happy will be reluctant to strongly agree with ‘J is not happy’.
    The truth in your thought that any significant lack of happiness is sufficient for the truth of ‘J is not happy’ is this: any significant lack of happiness is sufficient for the truth of ‘J is not completely happy’. But that does not entail or support ‘J is not a little happy’ or even ‘J is not reasonably happy’.

  17. Marcus,
    Ah, now I understand! That’s a good point about the pattern of data we observe in this study. In this case, participants are actually getting different vignettes in the ‘happy’ conditions from the ones they get in the ‘unhappy’ conditions, so the difference in people’s responses could be driven entirely by the difference between vignettes.
    But Jonathan Phillips has done a bunch of follow up studies (described in his comment above), and those studies are much more well controlled. In those studies, some participants are told about an agent with a good life; others are told about an agent with a bad life. Then participants are asked whether the agent is ‘happy’ or ‘unhappy.’ (So people are getting the happiness question or the unhappiness question about the very same vignette.) The difference between the good life and the bad life then has an impact on people’s judgments about whether the agent is ‘happy’ but not on people’s judgments about whether the agent is ‘unhappy.’
    So in the end, it does seem like there is something about the question as to whether the agent is happy (not about the specific vignette) that allows for this impact of value judgments that we don’t see on questions about whether the agent is unhappy.

  18. Joshua,
    Cool, and thanks for drawing my attention to Jonathan’s case. It’s very interesting — yet I think my hypothesis might still account for it.
    I suspect that one reason people may judge Orphanage-Mark as less happy than Nightclubbing-Mark is that we have a tacit belief that at least Nightclubbing-Mark is having some *fun*, whereas Orphanage-Mark is having none.
    This bias might lead people to judge Nightclubbing-Mark to be more “happy” than Orphanage Mark — but, since they both *explicitly* report being miserable, they are both judged to be positively *unhappy*.
    Let me explain.
    Hypothesis 1 (Biases explain why we judge Nightclubbing-Mark to be happier than Orphanage-Mark): Perhaps we are biased towards associate (the psychological state of) having *fun* with happiness. Then suppose (as seems plausible enough) that we are tacitly biased to believe that Nightclubbing-Mark is having more *fun* than Orphanage-Mark (since, after all, at least the former Mark is clubbing!). These facts would explain why we judge them equally happy.
    Hypothesis 2 (Different biases explain why we judge both Marks to be equally unhappy): We are told that both Marks hate their lives. Okay, but suppose that whereas we identify the psychological state of *hating* life with unhappiness, we tend *not* to associate fun (or lack thereof) with unhappiness. This too is plausible. I’ve met drug-users who I would say have more *fun* than people who work in dead-end jobs, and for this reason I would say they are “happier” then the latter. All the same, if you asked my about who is *unhappier*, I might say that the drug user and dead-end-jobber are both unhappy because, although the one at least has fun, they both hate their lives.
    In short, whereas there might have been *one* bias at work in your cases Joshua (e.g the bias that Cocaine-Maria is secretly psychologically distressed), *two* biases are at work in Jonathan’s case:
    Bias 1: we are biased to believe that nightclubbers secretly have more fun than orphanage workers (so, Nightclubbing-Mark secretly has more psychological fun than Orphanage-Mark).
    Bias 2: we are biased towards associating the psychological state of having fun with *happiness*, but biased towards associating psychological distress with *unhappiness* (thus explaining why we judge Nightclubbing-Mark to be *happier* — he’s having more fun — but both equally *unhappy*, since both are equally *distressed* [note: surely fun is compatible with equal levels of psych. distress]).
    I recognize the overall (“bias”) hypothesis I am floating is more complex (and possibly, strained) in Jonathan’s case — yet I still think it could be what’s generally going on in all of the cases.
    Anyway, this is really interesting stuff — I’m glad I happened upon this thread yesterday!

  19. Hi Marcus,
    in 2008 I ran a study that tried to test whether, in these kinds of studies, there is tacit filling-in of psychological details going on that explains these kinds of effects. Details are here:
    In this study, everyone was given the same vignette and asked whether (a) the person is immoral (b) how much distress they think he experiences and (c) whether they think he is happy. Whereas distress ratings of course affected happiness-attributions, they didn’t fully account for the differences. But the remaining differences correlated with differences in morality-judgments. The study described in the link does most likely not fully answer your concerns, but you might nevertheless find it interesting, given your hypotheses.

  20. Joshua,
    Thanks for this. Great and provocative video.
    The final statement is “Why should the concept of happiness differ so much from the concept of unhappiness?” But I think that is an overly broad assessment of what has happened. The initial judgments of happiness are based on the narratives about the women’s lives, and under these circumstances it is clear that our moral judgments about the unwholesome and wholesome Marias are what affect the judgments of happiness. The Hollywood Maria cannot possibly be happy living such a shallow life. But the judgments of unhappiness at the end of the experiment are not in response to a narrative of activities but to summary statements made by the Marias about how they feel about their lives. I think it is possible that what the experiment proves is that summary statements of happiness or unhappiness are given greater credence, regardless of what the shape of the life looks like. If I say I am unhappy, that trumps anything else that might be going on. Basically, the happiness ratings in the first half are based on different data and are different kinds of judgments than are the unhappiness judgments at the end. A methodological difference, not a conceptual one.
    To test this view, you could create another Maria narrative that suggests a dreadful life. Or better yet, two Marias one virtuous but dreadful and the other despicable and dreadful. Then have people judge whether the person is unhappy or not just as they do in the video. Then have the second part where the Maria puts her chin in her hand an says that when she thinks of her life she thinks that she is remarkably blessed and fortunate. So despite what looks like a dreadful life, she says she is happy. My prediction would be that this person would be judged as being happy because—after all—she says she is and who are we to contradict her? If the results are a mirror image of the results of this study then you have shown that summary statements are taken at face value but narrative accounts of day-to-day life are subject to our moral judgments. But whatever the results, it would be interesting to see what that experiment produced.
    Great stuff. Keep it coming.

Comments are closed.