About Rationally Speaking

Rationally Speaking is a blog maintained by Prof. Massimo Pigliucci, a philosopher at the City University of New York. The blog reflects the Enlightenment figure Marquis de Condorcet's idea of what a public intellectual (yes, we know, that's such a bad word) ought to be: someone who devotes himself to "the tracking down of prejudices in the hiding places where priests, the schools, the government, and all long-established institutions had gathered and protected them." You're welcome. Please notice that the contents of this blog can be reprinted under the standard Creative Commons license.

Thursday, December 13, 2012

The remembering self vs the experiencing self

by Ian Pollock

I’ve just finished Daniel Kahneman’s book Thinking, Fast and Slow, which everybody needs to read. One of the useful concepts that he analyzes, which I believe I first heard from Dan Ariely, is the distinction between the remembering self and the experiencing self.

Following Kahneman’s presentation, think for a moment about where you would go on vacation if you could leave right now and money were no object. Why would you go there?

What comes to my mind is a hike along the East Coast Trail in Newfoundland. It’s been a few years since I’ve done any serious backpacking, so I suppose that I’m itching for a bit of challenge.

Now another question: where would you go if you could leave right now, money were no object, but your memory of the vacation would be erased when you returned?

I’m guessing that your choice under these circumstances is far more simply hedonistic – I’m inclined to choose some tropical beach destination.

According to Kahneman and others, this disparity reflects a difference in our preferences between a here-and-now preferrer — the experiencing self — that wants this pleasure to continue and this pain to cease, and a storyteller — the remembering self — that looks at an experience as a whole and evaluates its worth, with special attention paid to the beginning, climax and ending.

The difference between the two gets even more obvious (and creepy) when we consider suffering.

During a painful medical procedure, a patient is asked at regular intervals (say, every few minutes) to rate their pain level on a scale of 0 to 5. Then, afterward, they are asked to rate the painfulness of the procedure as a whole. [Note: I am making up these numbers, but they conform to the general pattern observed in actual experiments.]

Person A’s ratings:
2 4 5 5 5 5 5 5 5 5 5 5 4 4 3 2 1 1 0
Person B’s ratings:
2 3 3 3 3 3 4 5 5 5 0

Which of these people gives a higher pain-rating for the experience as a whole?

Well, the naïve answer is that Person A experienced more moments of pain, of greater intensity than Person B, so it stands to reason that they will rate the experience as a whole as worse. Readers with calculus could even say that the “suffering under the curve” is higher for Person A.

And yet, in such situations, Person B consistently rates their experience as worse — mainly because it ends on a bad note. Duration is totally neglected. “Total suffering” is thus not simply an integral of moment-to-moment suffering - or at least it’s not reported in that way. As Dan Ariely has noted, this helps explain the commonplace fact that most of us actually prefer to have a band-aid slowly peeled off, rather than “getting it over with.” Our remembering selves see the former as a less bad experience, even as our experiencing selves suffer for twice as long with pain that is not much less.

The creepy comes in when people are asked whether it would be acceptable to them to have an operation without anesthetic, so long as they do not remember it. According to Kahneman, many people (including himself) are all right with that, showing extreme callousness toward their experiencing selves! (I happen to lack whatever intuition is driving this phenomenon — I’m quite protective of my experiencing self.)

Of course, the remembering self is not always cast as the bad guy. I suspect most of us will have had the odd day on which all of our considered plans were dropped because of some hedonistic distraction — say, watching an entire season of Game of Thrones in one’s housecoat while wolfing down snacks (speaking purely hypothetically, of course). Here the experiencing self has triumphed over the remembering self. (Incidentally, there is interesting philosophical work here in figuring out the relationship of the remembering vs. experiencing selves to eudaimonia vs. akrasia.)

So far, this article has been a mere rehash of some cognitive psychology results. I want to start exploring the broader implications of this distinction. One of the areas that it seems worth applying to is ethical philosophy; specifically the contrast between virtue ethical and consequentialist strains of thought.

For virtue ethics, the point of morality is to help you to be a better, happier person. Here, happiness is emphatically not understood in the popular modern way as a mere persistent good mood. On the contrary, happiness (or eudaimonia) involves living an ethically good life, with close ties to friends and family, and strong community involvement. A lifetime of good deeds and fine company could be undone by your child’s turning out to be a villain, even if it were not your fault — hence, Solon says “call no man happy until he is dead.”

Meanwhile, consequentialism (particularly its subspecies, utilitarianism) seeks to maximize welfare or utility across all beings. In utilitarianism this gets defined as the balance of pleasure over pain, or some such concept. The definition of utility is always vexatious, but needn’t concern us overmuch here — the point is that almost all plausible consequentialist theories care quite a lot about moment-to-moment mental states like pleasure and pain.

I suspect you may be able to see where I am going with this. Virtue ethics is speaking directly and pretty much exclusively to the remembering self, while utilitarianism is much more friendly to the experiencing self. Is this a defect in one, or in both of these theories?

My tentative answer is that I am sympathetic to virtue ethics’ regard for the ethically substantial dimensions of human happiness — it stands in flattering contrast to the shallowness of popular culture’s “whatever makes you happy” nostrums.

And yet I am also skeptical about the normative importance of the kind of factors that people focus on when they discuss the “good life.” Recall Solon’s quote above — “call no man happy until he is dead.” To anybody with utilitarian leanings, this sounds pretty absurd. In the drama ‘House, MD,’ a character gives up a virtuoso career that is making her miserable. When a doctor asks her whether she’ll feel regret on her deathbed, she replies “You're going to spend one day of your life on your deathbed; the other 25,000 are the ones we should be worrying about.” Surely there is also something right about this!

Now we change the subject to a new econometrics fad: happiness measurement. This new trend has been dogged with controversy since the beginning. Some months ago, Massimo, myself and a couple of others had a conversation about it, in which I defended the idea as a worthwhile metric, on the grounds that it’s good to know how human misery and satisfaction correlate with other things, while Massimo expressed strong skepticism. Our positions then matched up fairly well with those of Julian Baggini (Massimo) and Richard Layard (your correspondent) in this debate on the subject.

I still think Baggini and Massimo’s fears about totalitarian abuses of the happiness metric are pretty histrionic, but I’m willing to eat a little bit of crow: they have a very good point about how problematic and philosophically indeterminate the measurement itself is. The UK asks the question in terms of “satisfaction with your life overall.” Massimo and Baggini’s point, I think, was that to move from this rating to a judgment of whether people are living good lives or not is an immense exercise in philosophical hand-waving. A person could be very satisfied but living a bad life in an ethical sense, or vice versa.

Now we can add to that worry an additional one — what about the poor, neglected experiencing self?

By tailoring the question exclusively to the remembering self (“satisfaction with your life overall”), the census-takers guarantee the result to be influenced by an inner narrative of respondents’ lives, rather than by any actually experienced mental state. For example, one can imagine someone who is altogether miserable moment-to-moment (say, a mother of 10 with an unsupportive husband), but has so internalized cultural norms of motherhood as the be-all-end-all of happiness, that she reports strong life satisfaction — every item on her “good life” checklist is checked off!

Also of particular interest is the relation of happiness to income, and a very famous result shows that this relation is linear at first, then plateaus; in other words, poor people report being unhappy, middle-income people are happier than poor people, but rich people are only a tiny bit happier than middle-income people. This does not merely reflect diminishing marginal utility of money — the curve is flatter than that consideration alone would lead us to expect.

However, here we have a similar problem. We are asking the remembering self, not the experiencing self, about “overall life satisfaction.” The remembering self reports its level of happiness based on a narrative of life so far (“Well, I’ve got a house, kids, a good job…”).  But the trouble is that a salary of $100,000 as opposed to $200,000 doesn’t fit well into a narrative, so it probably gets neglected. We still need to know whether a year in the life of a rich person contains more pleasurable person-moments than a year in the life of a middle-income or poor person.

Accordingly, I propose this as a useful experiment: use the standard experienced-happiness test of giving people pagers set to go off at random intervals. When they go off, the respondents note their mental state and report it shortly thereafter. Apply this experimental procedure to income levels. I (falsifiably) predict that the correlation between income and happiness will get stronger, though there is probably still a plateau (money really does have diminishing marginal utility).

And perhaps this type of happiness measurement should also be used in addition to, or in replacement of, national happiness measurements such as the one mentioned above. Certainly, it has a claim to be more objective than “life satisfaction,” inasmuch as it asks people about their mental states at the time, rather than the former, much more narrative-dependent question.

I acknowledge that this post is somewhat inchoate; in my defense, I mean it more as a call to conversation than as a well-ordered thesis. Can you think of any other relevant applications of the remembering self / experiencing self distinction? Do you think it’s overblown? What are your feelings on my tentative conclusions?


  1. While I would definitely privilege more objective measures of well-being (like those that factor into the UN's HDI or The Equality Trust's index of health and social problems), that doesn't mean that subjective measures are worthless (particularly given the subjective data that goes into diagnosing mental illness, rates of which factor into the latter index, which has a more epidemiological bent to it).

    A problem is that the criteria that subjects use to judge their own happiness and/or life satisfaction may not be as "portable" across nations and cultures. But that only matters if a society aims to compare numbers with other societies. Yet, internally, the metric can still be highly useful [or at least the nation of Bhutan seems to think so, given its GNH (Gross National Happiness) program, even though it has a relatively low HDI score].

    As for the idea of linking happiness reports to income levels, I have no objections to the experiment. But, if The Equality Trust's thesis is correct (i.e. that income inequality and health and social problems tend to go and hand-in-hand), then we might also find a maldistribution effect in some countries (e.g. the US and UK), where gross income inequality might counteract the hedonic effect of high income, even among the relatively wealthy.

    1. >Yet, internally, the metric can still be highly useful [or at least the nation of Bhutan seems to think so, given its GNH (Gross National Happiness) program, even though it has a relatively low HDI score].

      Yes, I still think it's more useful than not.

      >But, if The Equality Trust's thesis is correct (i.e. that income inequality and health and social problems tend to go and hand-in-hand), then we might also find a maldistribution effect in some countries (e.g. the US and UK), where gross income inequality might counteract the hedonic effect of high income, even among the relatively wealthy.

      That's a good point... I wonder what causes the link between income inequality and low happiness. I think the standard speculation is that it boils down to negative comparison with one's neighbour's income, but is that the only possible mechanism? Maybe it has something to do with the fact that even if most people are relatively affluent, the more affluent may still have power over the lives of those lower down in the pecking order in a way that breeds resentment.

  2. I can't make it past "erasing the memory," either of the vacation or of the surgery without anesthetic. Does that just mean the declarative, event memory of the happening, or the million little things, the physical effects, the mental training, the new synapses, etc? I believe the latter are more important, albeit not as critical to suffering or enjoyment.

    1. The intention is for the declarative event memory to be erased; e.g., if your vacation relaxed you, you would still be relaxed.

  3. Thanks for writing this thought provoking post, Ian!

    I enjoyed Kahneman's book as well. However, it has been a couple of months since I read it, so please don't quiz me on it :) Anyway, I often hear people (especially when talking about behavioral economics and happiness research) conflate happiness with pleasure. I don't think Kahneman made this mistake in the book anywhere, but I thought it was worth mentioning.

    Continuing along that line of thought, I generally think we make two egregious mistakes when we talk about happiness (and the Sisyphean quest of measuring it).

    1) We think we know what happiness is.
    2) Happiness, whatever it really is, can be measured.

    I'd love to hear your thoughts on those two points.

    1. >1) We think we know what happiness is.

      Hm. I'm not sure the problem is ignorance per se, so much as equivocation. You've got to choose some understanding of happiness before you try to measure it: are we talking about persistent good mood, or eudaimonia, or 24 hour chirpiness, or having a car and a mortgage?

      >2) Happiness, whatever it really is, can be measured.

      Well, once you choose a definition, you might be able to get some headway. But sometimes it seems to me that the approach of asking people about their happiness/mental states is the wrong way to go. It might be wiser to look at the correlates of happiness/unhappiness in their lives: failed relationships, missed work, alcoholism, suicide attempts... rather than taking their word for it.

      Clearly you can measure extremes with confidence, especially negative extremes.

  4. This book is fantastic in many respects, but I wonder the degree to which it leans too heavily on dual-process, dual selves, dual facts (inner and outer), dual everything!

    Has anyone done any work on evangelical Christians and self-reports of happiness? Some of the most desperately unhappy people I know are born-again, and seem compelled to 'put on a happy face,' not because Jesus has brought them joy, but to paper over the obligatory claim that Jesus has brought them joy.

    1. Kahneman does make it clear that he doesn't actually think there are "two selves" inside us.

      I am actually more willing to take his talk of "two selves" at face value than he is himself, because it seems to me there might actually be a legitimate ethical conflict between the remembering self and the experiencing self (and also between the (related) planning self and akratic self).

  5. Ian, to extend the bounds of thought of Kahnemann, should we not also talk about a forethoughtful self, an introspective self, or whatever you will?

    On this angle, the vacation issue's a busted analogy, because my forethoughtful self says "what's the diff if I won't remember either one afterward?"

    So, from that, one could then recreate the analogy, and go from there to ethics in general, and say that trying to system-build ethics out of either virtue ethics or utilitarianism is a failure. (As are system-building philosophies in general, in my opinion.)

    Therefore, the truly insightful slow thinker will apply one or another ethical system where he or she thinks system A, or B, or C, or D is best applied.

    Of course, one could say such a principle of when to apply which ethical system is itself system-building. I'd say, "true," and that's why the whole idea of philosophical system-building should be approached in a Zen-like way.

    1. >So, from that, one could then recreate the analogy, and go from there to ethics in general, and say that trying to system-build ethics out of either virtue ethics or utilitarianism is a failure. (As are system-building philosophies in general, in my opinion.)

      I'm not sure that it's possible to discern general features of systematic vs ad hoc theories... I'd be tempted to say something like "systematic theories may reflect a hasty impulse to generalize; ad hoc theories may reflect an insufficient unification of knowledge."

  6. I can't help but wonder if we do little more than bewilder ourselves when we make these artificial distinctions, and whether we multiply (imagined) entities needlessly, to paraphrase the Razor when we do so. It makes for interesting discussions, but may create questions, issues, problems that are contrived, as well.

    1. For me, the usefulness of this distinction is that it unifies a lot of bizarre human behaviours under a single theory, making them appear as different manifestations of the same phenomenon as opposed to just human_behaviour_quirk#35 and human_behaviour_quirk#122.

  7. I see your point about the apparent links between Kahneman's remembering and experiencing selves and the two approaches to ethical theory. An interesting idea, but I don't know that the scientific side is firm enough to build anything on. (You mentioned that you were taking Kahneman's idea further than he did.)

    I'm not so sure about the other (philosophical) side either. You talk about utilitarianism and virtue ethics, but there are, of course, all sorts of other possible approaches, forms of subjectivism, say, or egoism or approaches which reject philosophical ethics entirely. I am not recommending any particular approach, just making the point that there are many ways of seeing and interpreting morality. You may happen to favor utilitarian and virtue-based approaches and wish to discuss them in terms of particular scientific results, but this then becomes primarily a philosophical discussion rather than a scientific one.

    And maybe this is how you see it. It's just that the general tenor of your writing and the reference in a comment to the unification of knowledge suggested to me that you were aspiring to a more science-driven approach than the project you are proposing would in fact be likely to deliver.

  8. I think an important thing to work towards is integration between the experiencing self and the remembering self. As many of us has experienced, sometimes opting for pleasure in the form of present-moment gratification can ultimately lead to negative consequences. If these negative consequences pile up, the remembering self will be less satisfied with its life - it will see a string of regrets. In this moment of remembering, the experiencing self will experience present-moment pain. I believe this pain will be more enduring, since the remembering self carries around multiple experiences. It will take a decent amount of positive experiences for the remembering self to become more satisfied.

    In viewing things this way, I think it is important for the experiencing self to find value in delaying gratification - the mild pain of missing out on pleasure right now is ultimately weighted less than the enduring pain of living a life with many regrets. However, pleasurable experiences are, well, pleasurable! And they are also necessary for the remembering self to look back and say "I have had a happy life (so far)". So I think it is important for experiencing self to balance the needs of itself with the needs of the remembering self, because on our deathbed these selves WILL be fully integrated. Beginning this integration in our present lives will, I believe, ultimately lead to a life of both pleasure and eudamonia - a life well lived.

    1. @JChazU, I think you've really got to the heart of the matter here. Your argument is very compelling and it now seems like a self-evident truth to me. I have been spending too much time with a loose number of ideas in my head which you have elucidated perfectly.


Note: Only a member of this blog may post a comment.