At 35, I can no longer call myself young. College, and even graduate school, are a distant memory. Many of my friends have become tenured professors or partners at firms. For better or worse, I have chosen a different path: starting anew in law school.
But how new is this path, really? As a young child, I was fascinated by criminal law. In high school, I developed a love of philosophy, which led to a concentration in the subject during undergrad in which I learned about the philosophy of law and ethics. As I pivoted to graduate school in psychology, I was motivated by moral topics and questions that bear on the legal system, like whether humans have free will. My more recent interests in rationality and decision-making have obvious connections to the law, as well — particularly jury trials and the way judges interact with jurors.
Perhaps in five years, I’ll be doing precisely the kind of philosophical or psychological applied legal work that fits this grand narrative, and my decision to pursue law school won’t seem like a career U-turn, but the culmination of a long academic journey. Or — and this is not so unlikely given the sheer number of paths people can take out of law school — I will end up doing something much farther removed from my past, such as criminal litigation or antitrust research.
If I choose one of the latter routes, my current life narrative falls apart. At best, the first four decades of my life would have been spent learning some tangentially useful facts about the mind and methods of reasoning — much of which could have been picked up outside of formal education or academia. Although it’s exciting to imagine reinventing myself, it also would be a bit horrifying to discover that the first 40 years of my life could have been spent working more directly towards my long-term goals.
As someone who enjoys thinking about the rationality of decisions, I’ve been considering the following puzzle: How much should I take my past into account when exploring possible career paths? If two similar seeming opportunities present themselves, one of which suits a cognitive psychologist better than the other, should this tip the scales? Should I care about the temporal coherence of my lifetime choices?
The facile answer from pop behavioral economics is “no.” The past is ‘sunk’; our decisions can only influence what happens next. For example, the fact that I’ve bought nonrefundable tickets to vacation on the Cape shouldn’t deter me from staying home or choosing a new destination if the weather forecast looks miserable. The money has been lost either way, and the question to ask is whether I’d prefer a free cold and rainy weekend in Cape Cod to staying home or spending money on an alternative. Gritting my teeth through terrible weather doesn’t retrospectively justify my past purchase of the vacation. Better to have known when I purchased the hotel that the weather is unpredictable and there’s a chance I won’t want to use the room.
It’s tempting to apply the same logic to my career quandary. Even if I’ve paid an enormous opportunity cost to train as an academic psychologist, the past is the past. Nothing I do in the future can change the past; therefore, it shouldn’t figure into my future decisions.
Strictly speaking this is true. But in a more metaphorical and narrative sense, it’s not. Imagine you were given the screenplay to a movie that was half completed and were tasked with finishing it. Although you could choose to ignore most of what was written and completely change course for the second half, the viewers of the full movie would leave the theater perplexed and disappointed. What was the point of the first half of the movie in light of the second half? In contrast, even if the partial screenplay you were given wasn’t perfect, you could instead choose to build a compelling story around it, which leaves your viewers more satisfied. In other words, the interpretation of the beginning of the movie depends on how you complete it. By analogy, my future evaluation of my past depends on how I shape my future. In this sense, my past is not completely sunk.
Then again, who cares? My life isn’t a movie designed for others’ consumption, and perhaps even describing this life as belonging to a unitary self bound across time is — as argued by philosophers like Derek Parfit — a metaphysical illusion. I could instead adopt the polar opposite perspective: treat my life as a series of time slices whose momentary wellbeing I aim to maximize across time, perhaps weighing sooner moments more than later ones. In that case, it doesn’t seem to matter whether I haphazardly pivot careers so long as I maintain short-term satisfaction at each step. Of course, maintaining this short-term satisfaction may require pursuing long-term goals and building on accomplishments from the distant past, but only instrumentally. I may also need to accept that the narrative view of my life could influence others’ appraisals of me, which could indirectly affect my wellbeing. For example, I may worry about developing a harmful reputation as a quitter or a dilettante (though some people may look upon these multifaceted interests favorably). So, even if I take a momentary approach to maximizing my wellbeing, I cannot escape the sunk-cost logic completely. My current choices shape how others view my past, which will in turn shape my experienced future.
Recently, though, this view of selfhood (or lack thereof) appeals less to me than it once did as a high-schooler reading Reasons and Persons. Now that I’m old enough for a possible narrative to have taken shape, I’ve been reflecting more on life as a totality. Granted, I still probably do this less than most people, but I am beginning to understand the temptation. Perhaps the best distillation of this perspective is the “deathbed test,” which recommends that you make choices now that will minimize the regret you’ll feel at the end of life. I’m not sure this test is meant to be taken too literally. You aren’t meant to imagine yourself as an old man with dementia or assume you’ll suffer from peak-end bias or other types of memory loss. The person on his deathbed is meant to be lucid and capable of accurately recalling key decisions and events from the past. Critically, he also remembers times that you avoided taking risks and prioritized short-term comfort over long-term character-building. Your deathbed self will judge you for your cowardice and, more generally, for taking the path of least resistance in your career or relationships.
Intuitively, this decision heuristic encourages the optimization of long-run satisfaction over quick temptation or underwhelming comfort. And, in most people’s theories of what makes a good life, long-run satisfaction is what matters. Or at least it matters if we take a narrative view of life seriously — treating ourselves as more than just a series of disconnected experiences. But this test also incentivizes exploration in a way that my screenplay analogy does not. Suppose my dying self could have given me advice about what to do when I was contemplating whether to stay in a comfortable academic role or take a risk applying to law school. My deathbed self would probably say, “Take the leap! Don’t let your life slip away from you before it’s too late.” Although narrative elements may play a role in this advice, the logic is also inspired by a more experiential perspective: the short-term transitional pain of taking the LSAT and writing application essays has the potential to pay off in future happiness (and income?) that far eclipses what I have now. This imagined distant perspective of the deathbed self helps counter my natural myopic discounting (the overweighting of present and very nearterm pleasure at the expense of later experiences). It also downplays the impact of other people’s short-term judgment. Who cares if my peers think I’m insane to go back to school at this stage? This is my life.
There’s a lot to like about this approach. Indeed, it’s championed by successful tech executives like Jeff Bezos, who credits the heuristic with getting him to quit his job and start Amazon. It’s hard to argue with those results! But perhaps therein lies a wrinkle, which — if you’ve been reading this blog long enough — you shouldn’t be surprised to hear me raise. Bezos is a massive success, and his success is largely the reason we hear about his regret-minimization approach to making decisions. Granted, he attempts to downplay the importance of success as a justification for the heuristic and claims he wouldn’t feel regret even if he had failed. But ultimately, this is still the kind of advice you’ll hear from someone who wins the game of life in the end. A few minor setbacks are understandable as learning experiences on the road to a happy retirement. The most embarrassing outcome is to end up on your deathbed as a loser — dying alone without respect, recognition, or love. The deathbed test therefore discourages choices that increase the probability of this kind of outcome.
But is this really such a bad way to die? It depends on the alternative. Even if you don’t become Jeff Bezos, you could be proud of building a reasonably successful business, starting a family, becoming a decently accomplished professor, or dedicating your life to a cause that helps others. The deathbed thought experiment might help you reach these goals, relative to a life with a boring 9–5 job that ends with Netflix and takeout on your couch. On the other hand, this test may also increase the odds of catastrophe. After repeated failures from risky ventures, you could become broke, depressed, drug-addicted, or universally hated. In fact, the proverbial deathbed might never even arrive should your health take a turn for the worse or you die in a drunken car accident. Most people would prefer a long, somewhat regretful life to alternatives like this.
So selection bias is an issue. But there’s an even bigger issue, which brings us back to sunk costs (sort of). By design, the deathbed thought experiment is retrospective — focused on how you will feel long after you make a decision. Yet actual decisions are prospective. In other words, whatever positive or negative experiences that will come from the decision have not yet materialized. A decision that can be expected to look good in hindsight isn’t necessarily a good decision looking forward. For example, imagine you’re offered the opportunity to end your life in pure bliss so long as you’re willing to endure several decades of grueling labor. This is probably not a worthwhile trade. Your joyous deathbed self might think otherwise, though. He has already put in the work to reach a state of bliss; the terrible costs he’s experienced along the way are behind him and possibly forgotten. In other words, this is a case where you shouldn’t neglect ‘sunk’ costs — because they’re not yet sunk for your present self.
While this is a purposely contrived example to highlight the problem, the issue generalizes to plenty of real-world contexts in which you might be tempted to employ the deathbed test. Is it worth it for you to train for an iron man? Spend years of your life in a library doing research for your magnum opus? Working 100-hour weeks building a startup that’s going to change the world? The Bezos-adjacent, ‘high-agency’ crowd may enthusiastically support these goals, using the deathbed test as a means to rationalize intense ambition. If you’re drawn to this perspective, though, it may be worth remembering how Derek Parfit (who is, ironically, a big hit with Bay Area rationalists) thinks about personal identity. When you grind your life away on a grand project, the person you’ll ultimately satisfy isn’t really yourself. This person you’ll become in old age is more like a stranger, who will enjoy telling tales of all ‘he’ has accomplished due to his younger counterpart’s toils. Of course, he prefers that you work all the time towards the outcomes that will make him most satisfied. Like a boss at your workplace, he doesn’t want you to goof off with your friends or to stream Below Deck Sailing Yacht. But maybe you — or maybe I should say, “I” — do?
In any case, to the extent that I support the underlying logic of this test, I imagine that my deathbed self would be pleased with me. After all, by enrolling in law school, I’ve already taken a big leap out of a comfortable, but unsatisfying career path. All of the options that lie ahead — whether they explicitly build on my cognitive psychology training or not — are genuinely novel and at least somewhat daunting. So, when exploring what comes next, this extremely long-term perspective seems less relevant. My focus now is, rather, to find the right balance of intellectual satisfaction, real-world impact, flexible lifestyle, and so on — a more experiential lens. Should I therefore give up on the goal of narrative coherence when contemplating my next career?
I don’t think so. True, my life isn’t a movie for others’ consumption, and I don’t worry a huge amount about people thinking that my life trajectory is incoherent. Nevertheless, I see the desire for coherence as a useful heuristic to the kind of choices that are most likely to satisfy me. The fact that I’ve spent most of my life excited by topics in cognitive science, philosophy, statistics, and so on suggests that there’s a good chance I would enjoy a legal career in this intellectual space, even if I may opt for a faster-paced lifestyle. This is not dispositive, of course, and the fact that I left psychology academia is some evidence to the contrary. But I continue to love thinking and writing about these disciplines, whereas I’m far less certain that I would enjoy the subject matter of a more traditional legal job. This straightforward logic is, perhaps, what’s really driving my trepidation about an extreme career pivot. By completely throwing away my current life’s story, I would be ‘sinking’ a part of myself that has learned what he most values and enjoys. While the deathbed test has encouraged me to break out of an overly predictable path, the impulse to build a cohesive life narrative for myself can help me home in on the kind of decisions that are likely to continue bringing me joy. Even for someone who is most compelled by the deflationary view of ‘self’ as a mere sum of experiences, these loftier perspectives are therefore helpful — albeit imperfect — guides for living a good life.
