I've read Daniel Gilbert's Stumbling on Happiness back in 2008, just before going to college, and I wanted to read it again now that I'm navigating the transition out of grad school and into the workforce. This book is not a manual for how to be happy -- instead, it presents the reasons why happiness is so elusive, and why we work so hard towards things that ultimately don't make us as happy as we thought they would. The book is engaging, funny, accessible, well organized, and full of citations -- everything that a good popular science book should be. If you like learning about human nature, I highly recommend it. In this post I will summarize the book's main points. The author also has a popular TED talk about synthesizing happiness.

Part I, Prospection, asks why we look into the future in the first place. Well, we like to imagine ourselves being successful and happy in the future. We also like to anticipate potential negative outcomes, so that we can prepare for them. We want to feel in control of our future experiences, we want to matter, to make things happen. (Losing this feeling of control over one's life is a hallmark of depression.) But why is controlling our future so important to us? Here Gilbert provides two answers: First, we want to be in control for its own sake, because exercising control is rewarding and it makes us feel good. Second, we want to be in control because some futures are better than others, and we want to steer our lives towards the futures that will bring us the most happiness. The second answer sounds like common sense, but it is in fact wrong, because the future is fundamentally different from how we imagine it. This is the central argument that the rest of the book lays out.

Part II, Subjectivity, tackles the problems of defining and measuring happiness. It's hard to pin down a subjective experience like emotional happiness, so Gilbert settles for "the you-know-what-I-mean feeling". This is different from moral happiness ("I should be happy because I live according to moral standards") and from judgmental happiness ("I am happy that Larry Lessig is running for president").

Now that we know what we mean by emotional happiness, all that remains is to measure it. For reasons that philosophers like to argue about (language squishing, experience stretching), comparing the happiness of two different people is a difficult proposition. Even comparing your current happiness to your past happiness is tricky, because your memory of the past is colored by your present point of view. Gilbert concludes that the best happiness meter is the honest, real-time report of an attentive person. It's not perfect, but it's the best we have. And by measuring the happiness of many different people, we allow the law of large numbers to cancel out any individual calibration errors.

Armed with a way to quantify and compare happiness, the book goes on to illuminate three types of foresight errors -- ways in which our imagination misleads us about how the future is going to be.

Part III, Realism, exposes imagination's first flaw: the tendency to fill things in and leave things out. This happens effortlessly and without our awareness, so we tend to assume that our imagined scenario accurately reflects reality. For example, when we think of George Eastman's successful career as an inventor, we fill in that he must've been quite satisfied, and so we're baffled to learn that he shot himself. Our imagination leaves out the health problems that he had later in life, and which might explain his decision.

Similarly, when we imagine a future event, we sketch out a few salient features, and leave out many others. Because of this incomplete picture, we can't accurately predict how happy some particular future will make us. We might imagine the fat paycheck but not the stressful work, and thus overestimate our future happiness at a particular job. Or we might imagine the arduous hike but not the glorious sunset, and thus underestimate how happy we'll be after a day outdoors.

Gilbert draws parallels between the filling-in and leaving-out of imagination and similar flaws in memory and perception. When we remember something, it feels like we're replaying a movie, but in fact, we're reconstructing the movie from a compressed representation, and this process of reconstruction can lead to memory errors. Similarly, our visual system "fills in" the blind spot on our retina, making us susceptible to visual illusions.

Part IV, Presentism, examines imagination's second flaw: the fact that when we imagine the future, we borrow liberally from the present. This is why when you're satiated, it's hard to imagine feeling hungry, and when you're depressed, it's hard to imagine ever enjoying life again. (It's also why old sci-fi books seem hopelessly anachronistic.) The present colors not only imaginings of the future, but also recollections of the past, especially recalling how you felt at a particular time.

The same areas of the brain that are used for processing visual stimuli, auditory stimuli, and emotions are active when remembering or imagining visual scenes, sounds, and emotions. The brain has a reality-first policy, which is why it helps to close your eyes when you're trying to imagine a penguin. And we're pretty good at distinguishing between a visual or auditory image that comes from the environment, versus one that comes from our imagination. But not so with emotions -- what's the equivalent of closing your eyes to shut off satiety when you try to imagine hunger? So we mistakenly conclude that we'll feel tomorrow as we feel today.

Other subtle forms of presentism include:

• Failing to account for habituation (the fact that the first kiss feels much better than the one hundredth). Two workarounds: add variety, or increase the time interval between repetitions.
• Anchoring. Since mental images are atemporal, when we imagine how something will feel in the future, we actually imagine how it would feel right now, and then make a correction for the fact that it will happen later. We often don't correct far enough.
• Making a comparison in the present and assuming that we'll be making the same comparison in the future. This includes all the favorite tricks of behavioral economists, such as the fact that a $X loss seems more powerful than a $X gain, even though they are equivalent in expectation.

Part V, Rationalization, explores imagination's third flaw: the failure to recognize that we'll feel differently about something once it's already happened. When we contemplate a negative experience, we tend to think that it will affect us intensely and for a long time. But people who actually go through a traumatic experience adapt pretty quickly, and soon say that they feel just as happy as everyone else, and even enhanced by the experience. We are less fragile and more resilient than we think, thanks in part to our psychological defense system.

This defense system is made possible by the ambiguities pervading the world around us, and it is especially good at making us feel better about intensely negative and inescapable situations. It also has other effects that arguably make us less rational:

• We surround ourselves with people who agree with us.
• We seek evidence confirming our beliefs, and when we find it, we accept it uncritically.
• When we encounter evidence against our beliefs, we examine it more stringently, and often find reasons to dismiss it.
• We tend to evaluate a choice (which job to take, which appliance to buy, etc.) more positively after we've already made it.

Part VI, Corrigibility, asks why we never get better at predicting our emotional futures. Usually, when we're bad at something, we get better in two ways: by learning first-hand from our own experience and mistakes, and by learning second-hand from the wisdom of other people around us. So why does it seem like we never learn, when it comes to predicting our own future happiness?

Experience doesn't help much because it relies on accurately remembering how we felt in the past, which is not something our memory systems do very well. We tend to remember unusual occurrences -- the best of times and the worst of times -- rather than the common, ordinary times. We place more weight on how something ends, which is why a bad ending can ruin an otherwise good movie. And we rationalize how we must've felt in the past ("It was Monday morning, so I must've felt grumpy") , instead of actually remembering.

Cultural wisdom doesn't help much either, because memes about happiness spread not on the basis of goodness or truth, but on the basis of how good the memes are at spreading themselves. For example, there is evidence that once they've reached the middle class, people don't get happier with more money. And yet, the belief that money brings happiness is widespread, because it keeps the economy going, which contributes to a stable society, which in turn helps this belief spread. As another example, there is evidence that married couples are happiest before they have children and after the children leave home. And yet, the belief that children bring happiness is common, because any society where this weren't the case would peacefully end in one generation.

So if we can't learn from our own experience or from the wisdom of others, what can we do? Are we doomed to always be poor at predicting what will make us happy? Gilbert proposes a simple solution, but one that most people find unappealing. The solution is this: When you consider how experience X will make you feel, stop relying on your imagination. Instead, find a surrogate -- someone having that experience right now -- and ask them how they feel. This is deeply unsatisfying because we like to feel special, and we don't think that someone else's experience will transfer to us. But given the limits of our imagination, this might be the most accurate way to predict how happy something will make us feel.