TALKS | TEDX
Autonomous cars are coming — and they’re going to drive better than you. Chris Gerdes reveals how he and his team are developing robotic race cars that can drive at 150 mph while avoiding every possible accident. And yet, in studying the brainwaves of professional racing drivers, Gerdes says he has gained a new appreciation for the instincts of professional drivers. (Filmed at TEDxStanford.)
An autonomous car may seem like a thing of the distant future, but mechanical engineer Chris Gerdes is racing to make it a reality today. Full bio »
Struggling to make your mind up? Interpret your gut instincts to help you make the right choice
DECISION-MAKING was supposed to have been cracked by science long ago. It started in 1654 with an exchange of letters between two eminent French mathematicians, Blaise Pascal and Pierre Fermat. Their insights into games of chance formed the foundation of probability theory. And in the 20th century the ideas were developed into decision theory, an elegant formulation beloved of economists and social scientists today. Decision theory sees humans as “rational optimisers”. Given a choice, we weigh up each option, considering its value and probability, and then choose the one with the “highest expected utility”.
With your experience of making decisions, you have probably noticed some flaws here. There’s the risible idea that humans are rational, and the dubious notion that we would be capable of the on-the-hoof calculations of probability, even if we could access all the necessary information. Decision theory explains how we would make choices if we were logical computers or all-knowing beings. But we’re not. We are just rather clever apes with a brain shaped by natural selection to see us through this messy world.
Decision researchers had largely ignored this inconvenient reality, occasionally patching up their theory when experiments revealed exceptions to their rules. But that make-do-and-mend approach may soon change. Earlier this year, an independent institute called the Ernst Strüngmann Forum assembled a group of big-thinking scientists in Frankfurt, Germany, to consider whether we should abandon the old, idealistic decision theory and start afresh with a new, realistic one based on evolutionary principles. The week-long workshop provided a fascinating exploration of the forces that actually shape our decisions: innate biases, emotions, expectations, misconceptions, conformity and other all-too-human factors. While our decision-making may seem inconsistent or occasionally downright perverse, the truly intriguing thing is just how often these seemingly irrational forces help us make the right choice.
We must start by acknowledging that many of our choices are not consciously calculated. Each day we may face between 2500 and 10,000 decisions, ranging from minor concerns about what brand of coffee to drink to the question of who we should marry, and many of these are made in the uncharted depths of the subconscious mind. Indeed, Ap Dijksterhuis at the Radboud University Nijmegen in the Netherlands and colleagues have found that our subconscious thinking is particularly astute when we are faced with difficult choices such as which house to buy or deciding between two cars with many different features (Science, vol 311, p 1005).
What drives these gut feelings? Being inaccessible to conscious examination, the processes are particularly difficult to fathom. One idea is that they are based on heuristics - mental rules of thumb which, applied in appropriate situations, allow us to make fast decisions with minimal cognitive effort. The “recognition heuristic”, for example, will direct you to choose a familiar option where there is very little information to go on. The “satisficing heuristic”, meanwhile, tells you to pick the first option that meets or exceeds your expectations, when delaying a choice for too long is not in your interests.
Heuristics are shaped by previously successful choices - either hard-wired by evolution or learned through trial and error - so it’s no wonder they tend to work. Peter Todd from Indiana University, Bloomington, has shown, for example, that satisficing is a sound basis for choosing a romantic partner (New Scientist, 4 September 1999, p 32). The recognition heuristic, meanwhile, may underpin some of your better guesses in multiple choice quizzes. However, some critics doubt whether our subconscious choices really are based on heuristics; they argue that this approach to decision-making would be neither fast nor cognitively simple since we would need a complex mental mechanism to select the correct heuristic to use.
Our emotions may instead be the driving force in subconscious decision-making. We now know that far from being the antithesis of rationality, emotions are actually evolution’s satnav, directing us towards choices that have survival benefits. Anger can motivate us to punish a transgressor, for instance, which might help us to maintain social order and group cohesion. So says Peter Hammerstein from Humboldt University of Berlin, Germany, who helped organise the workshop. Disgust, meanwhile, makes us fastidious and moralistic, which should prompt choices that help us avoid disease and shun people who don’t play by the rules. And while fear often seems to lead to overreactions, this makes sense when you consider the dangers facing prehistoric humans, says Daniel Nettle from Newcastle University, UK. On that one occasion where a rustle in the bushes really was made by a predator, the less neurotic peers of our ancestors would have paid the ultimate price, failing to pass their laid-back genes on to the next generation (Personality and Social Psychology Review, vol 10, p 47).
Heuristics and emotions help us subconsciously focus on what matters. This is just as important when we make conscious decisions. Even the most basic everyday situations are too complex for our brains to compute all the necessary information. Instead, we must simplify.
Gordon Brown at the University of Warwick, UK, argues that we rank alternatives based on cognitively easy, binary comparisons. For example, when deciding whether £2.20 is too much to pay for a cup of coffee, you might recall half a dozen occasions when you paid less and only two when you paid more, leading you to place this particular coffee in the “expensive” category, and choose not to buy it. This so-called “decision by sampling” approach simplifies the options, but it can also lead to bad decisions when the limited information used to rank alternatives is incorrect or based on false beliefs (Cognitive Psychology, vol 53, p 1). If, for instance, frequent nights out with boozy friends leads you to conclude that your alcohol consumption is in the top 20 per cent of drinkers, when in fact it falls in the top 1 per cent, you are more likely to decide to ignore the problem. Decision by sampling could even sway your choice when you face more immediate threats: people living in a society with high mortality rates are more likely to decide to put themselves at risk than someone who has had little experience of danger.
That’s not very heartening, but Alex Kacelnik at the University of Oxford takes a more optimistic view of our ability to pick and choose the information upon which we base our decisions. “Natural selection allows us to correct our behaviour to do what works,” he says. Kacelnik believes the main force influencing decision-making is reinforcement learning. In other words, we learn from experience and favour what has worked in the past. Nothing controversial there. But, he notes, we are also swayed by our changing internal states - things like hunger, thirst and libido - so that choices are tailored to our needs. Decision theory has long struggled with the problem that people are inconsistent (see “The logic of inconsistency”), but Kacelnik argues that apparent inconsistencies in choice can arise simply because our preferences change according to our needs. “Utility is a moving target,” he says. We may not show the “economic rationality” of traditional decision theory but our choices have their own logic, which he calls “biological rationality”.
Natural selection can even explain our puzzling propensity to eschew choice altogether and simply follow the herd. Rob Boyd from the University of California, Los Angeles, pointed out at the workshop that we have evolved to learn from others because this is often a wise option. “In most situations it is way beyond an individual’s capacity to know the best thing to do,” he says. But we are good at recognising who to copy, says Laura Schulz of Massachusetts Institute of Technology who has found that even young children assess the expertise of their “teachers”. As a result, our conformist tendencies often lead to surprisingly good choices (New Scientist, 1 May 2010, p 40). They allow us to fit in when we start a new school or job and make wise purchases of the latest products without full information on the alternatives. The flip side is that we can also all fall into line with the immoral or illegal behaviour of those around us or be swayed by manipulative leaders.
Consideration of others is yet another aspect of human behaviour that flies in the face of decision theory. There are many situations in which a rational optimiser should not cooperate, since such actions can use up precious resources that we could use to better our own circumstances. From an evolutionary standpoint, it could be argued that some forms of apparent altruism help us to build alliances and improve our standing on the social ladder, but what about the times we overdo cooperation? An anonymous donation to charity, for example, will not boost your reputation or persuade others to help you in your hour of need. In purely evolutionary terms, it is a bad choice. But we do it anyway because the warm glow of altruism, which is evolution’s reward to team players, makes us feel good. In effect, we are tricked by a mental glitch. And it is not the only such glitch we possess. Researchers in decision theory have uncovered a variety of mental biases underlying some of our more illogical and arbitrary behaviours (see “Mental glitches that make fools of us”).
So what’s going on? Have our brains evolved to direct our behaviour in ways that have become maladaptive in the modern world? That should become clear as more decision researchers consider how we actually make up our minds, rather than how we should. Accepting that we are not rational optimisers will make life difficult for economists and anyone searching for a formula for choice, which is why some members of the Frankfurt group were reluctant to abandon decision theory altogether. But a better understanding of the forces that underpin our decisions should help everyone make better choices.
Conformists, for example, might be persuaded to adopt environmentally sustainable habits simply because others already have. Governments wanting us to save up for retirement need to understand why we are so bad at making long-term decisions. And we could all be more aware of the misconceptions and biases shaping our behaviour. The discovery of “decision fatigue”, for instance, which makes judges four times more likely to grant bail in the morning than in the afternoon, might persuade you to take more time out when facing a string of tough problems (Proceedings of the National Academy of Sciences, DOI: 10.1073/pnas.1018033108). And understanding that the behaviours of your nearest and dearest can bias your view of your own lifestyle might remind you to dig into the facts before you choose to follow or reject a new health regime.
Of all the choices that you face everyday, the decision to try to make better decisions is surely the biggest no-brainer.
Mental glitches that make fools of us
The human brain does not compute options like a rational computer, yet our decisions are often effective. Nevertheless, some of our mental biases are hard to explain.
In novel situations, or ones where information is limited, we have the unfortunate habit of basing decisions on random connections. This so-called anchoring effect was first shown by Daniel Kahneman of Princeton University and the late Amos Tversky, and the consequences can be bemusing. One study found that people asked to write a high number subsequently bid more for items whose value was unknown than people who wrote down a low number.
Kahneman and Tversky also revealed our peculiar attitudes to risk. We tend to be more cautious than is logical when there is the possibility of making large gains or small losses. However, we choose unduly risky options when there is the chance of making a small gain or a large loss. In recent years, our inclination to undervalue rare but catastrophic events has been dubbed the Black Swan effect.
Another factor underpinning some bad decisions is the confirmation bias - our tendency to overemphasise anything that confirms what we already believe. Then there’s loss aversion: it feels worse to lose something than to gain the equivalent amount, making us protect what we have rather than take a chance to make a gain. Also, when choosing whether to continue with a venture we irrationally consider the investment we have already made in it - the sunk-cost fallacy. Meanwhile, our short-term bias - temporal discounting - means we tend to prefer smaller rewards now to bigger ones later.
The logic of inconsistency
If you prefer apples to plums, and plums to pears, then given the choice between apples and pears you will obviously pick apples. Or will you? In reality, people fail to show such logical behaviour. This kind of inconsistency, known as intransitivity, has been a headache for mathematicians trying to understand decision-making. But their mistake may have been to think of the human brain as a computer rather than a biological entity that must solve the problem of how to compare apples, pears and plums.
Admittedly, our understanding of what goes on in a brain when it makes a choice is very hazy, as became apparent at an Ernst Strüngmann Forum on decision-making in Frankfurt earlier this year. It is generally agreed that there must be a mental “common currency” for comparing options. What this is, or how it converts into apples, pears, or whatever, is a mystery. However, Nick Chater from the University of Warwick, UK, argues that because the brain lacks time and computing power, it evaluates only a limited number of attributes for each alternative. This process could explain intransitivity, according to cognitive psychologist Danny Oppenheimerof Princeton University.
He believes the brain uses a kind of voting system: different brain areas weigh up the various attributes of apples, pears and plums, say, and compete with each other to have their preference chosen. If there’s no clear winner, you might decide on any of the fruit, depending on which region happens to gain the upper hand at that moment (see diagram). Intransitivity could be a by-product of the way our brains work, rather than a trait we have evolved for its own advantage.
Kate Douglas is an editor at New Scientist