Alchemy

by Rory Sutherland

  The job of a designer is hence that of a translator. To play with the source material of objective reality in order to create the right perceptual and emotional outcome.

  6.4: Mokusatsu: The A-Bomb, the H-Bomb and the C-Bomb

  Translation errors can be expensive, and at times gruesomely so. The following is from ‘Mokusatsu: one word, Two Lessons’, a declassified article in the National Security Agency’s Technical Journal (Fall 1968):

  In July of 1945 allied leaders meeting in Potsdam submitted a stiffly worded declaration of surrender terms and waited anxiously for the Japanese reply. The terms had included a statement to the effect that any negative answer would invite ‘prompt and utter destruction’. Truman, Churchill, Stalin, and Chiang Kai-Shek stated that they hoped that Japan would agree to surrender unconditionally and prevent devastation of the Japanese homeland and that they patiently awaited Japan’s answer.

  Reporters in Tokyo questioned Japanese Premier Kantaro Suzuki about his government’s reaction to the Potsdam Declaration. Since no formal decision had been reached at the time, Suzuki, falling back on the politician’s old standby answer to reporters, replied that he was withholding comment. He used the Japanese word ‘mokusatsu’, derived from the word for ‘silence’. However, the word has other meanings, quite different from that intended by Suzuki.

  Alas, international news agencies saw fit to tell the world that in the eyes of the Japanese government the ultimatum was ‘not worthy of comment.’ US officials, angered by the tone of Suzuki’s statement and obviously seeing it as another typical example of the fanatical Banzai and Kamikaze spirit, decided on stern measures. Within ten days the decision was made to drop the atomic bomb, the bomb was dropped, and Hiroshima was leveled.

  Depending on the context, mokusatsu can mean many things. It is derived from words meaning ‘silence’ and ‘death’, and can mean anything from ‘I cannot say anything at this time’ to ‘I have nothing to say because I am nonplussed’ or ‘I contemptuously refuse to dignify your proposal with a response.’

  Japanese is a highly context-sensitive language, but then so are all languages. In British English, when said in the right context and tone, ‘You stupid fucking idiot’ can be a term of affection – something that can wrongfoot Americans, who mostly speak the same language but tend to interpret it more literally.*

  In translation, it is an enormous mistake to assume that what the translator conveys is what the speaker intended, and it is equally foolish to assume that what you intended to say is what will be understood. Perhaps the most famous example of mistranslation occurred on a US presidential visit to Poland in 1977. Giving a short speech on the tarmac shortly after landing in Warsaw, President Carter was heard by his hosts to say that he ‘had left America, never to return’ and that ‘his affection for the Polish people was so great that he wanted to have sex with them.’*

  The story is often told as an illustration that the translator was some idiot and not up to the job, but this isn’t so – Steven Seymour was a brilliant translator, who had earlier translated the poems of W.H. Auden into Russian and had a true connoisseur’s appreciation of Polish poetry. However, his study of poetic Polish had unfortunately made him overfamiliar with the more antiquated nineteenth-century (or earlier) Polish vocabulary that modern Poles no longer used, or at least not in the same sense.*

  Russian was Seymour’s first language, while Polish was only his fourth. This would not have mattered, were it not for the fact that Polish is misleadingly similar to Russian in much of its vocabulary and grammar,* but occasionally stubbornly different in its meaning. Translators call these misleadingly similar words ‘false friends’, since it is very easy to assume they mean something they don’t.* Languages that are similar may be prone to greater misunderstandings – in Latin American countries, for instance, Spanish words may take on different meanings: ‘Your wife is a tremendous whore’ would seem an odd way to thank your host after dinner, except that in some countries, the formal word for ‘hostess’ has acquired that meaning.

  Strangely, one of the greatest sources of linguistic confusion arises between British and Dutch speakers of English. The Dutch are, almost universally, fluent in English:* their grasp of idiom is superb, their accents are flawless, and they have a similarly cynical sense of humour to the British. After an evening with a couple of Dutch contemporaries, you would have no consciousness of a language barrier, and would find it hard to believe that any misunderstanding could arise. However, misunderstandings are all too common, because Dutch conversation tends to be astoundingly direct, while British English is oblique and often coded to the point of derangement. In a business context a Dutchman might say, ‘We tried that and it was shit, so we won’t do it again,’ while an Englishman intending to say the same thing might say, ‘I think it might be a little while before we try that again.’

  Eventually the Dutch compiled a sort of phrasebook, which translates British English into Dutch English.

  WHAT THE BRITISH SAY WHAT FOREIGNERS UNDERSTAND WHAT THE BRITISH MEAN

  I hear what you say

  He accepts my point of view

  I disagree and do not want to discuss it further

  With the greatest respect

  He is listening to me

  You are an idiot

  That’s not bad

  That’s poor

  That’s good

  That is a very brave proposal

  He thinks I have courage

  You are insane

  Quite good

  Quite good

  A bit disappointing

  I would suggest

  Think about his idea, but I should do what I like

  Do it or be prepared to justify yourself

  Oh, incidentally/ by the way

  That is not very important

  The primary purpose of our discussion is

  I was a bit disappointed that

  It doesn’t really matter

  I am annoyed that

  Very interesting

  They are impressed

  That is clearly nonsense

  I’ll bear it in mind

  They will probably do it

  I’ve forgotten it already

  I’m sure it’s my fault

  Why do they think it was their fault?

  It’s your fault

  You must come for dinner

  I will get an invitation soon

  It’s not an invitation, I’m just being polite

  I almost agree

  He’s not far from agreement

  I don’t agree at all

  I only have a few minor comments

  He has found a few typos

  Please rewrite completely

  Could we consider some other options?

  They have not yet decided

  I don’t like your idea

  The barrier between the English spoken by the Dutch and the British is a decent metaphor for the relationship between reality and perception – in some respects they are similar, but in other contexts they can diverge wildly. Again, this distinction – the gap between the message we intend to send and the meaning that is attached to it – matters a great deal. Often, we are baffled by people’s behaviour. ‘I told him this, and he did that.’ We think they are being irrational, but the reality is that they didn’t hear what we think we said.

  In the same way, you cannot describe someone’s behaviour based on what you see, or what you think they see, because what determines their behaviour is what they think they are seeing. This distinction applies to almost anything: what determines the behaviour of physical objects is the thing itself, but what determines the behaviour of living creatures is their perception of the thing itself.

  The reason this matters so much is that most models of human behaviour and most economic models are blind to this distinction. It won’t surprise you to know that I am sceptical about the promise of ‘big data’, which is frequently promoted as though it were some kind of panacea. Lik
e many things that emerge from the technology sector, we become so drunk on the early possible benefit of a technology that we forget to calculate the second-order problems.* The evangelists of big data imply that ‘big’ equals ‘good’, yet it by no means follows that more data will lead to decisions that are better or more ethical and fair.*

  To use the analogy of the needle in the haystack, more data does increase the number of needles, but it also increases the volume of hay, as well as the frequency of false needles – things we will believe are significant when really they aren’t. The risk of spurious correlations, ephemeral correlations, confounding variables or confirmation bias can lead to more dumb decisions than insightful ones, with the data giving us a confidence in these decisions that is simply not warranted.

  A large tech company recently developed an AI system to sift applications for jobs, but it rapidly developed extreme gender prejudices – marking people down if their CV mentioned, say, participation in women’s basketball. With AI, of course, you cannot always be sure of its reasoning: it may have noticed that more senior employees were men, and so taken maleness as a predictor of success.

  Another company using a big data approach discovered a variable that was vastly more predictive of a good employee than any other: it wasn’t their level of educational attainment or a variable on a personality test – no, it turned out that the best employees had overwhelmingly made their online application using either Google Chrome or Firefox as their browser, rather than the standard one supplied on their computers. While I can see that replacing a browser on a laptop may be indicative of certain qualities – conscientiousness, technological competence and the willingness to defer gratification, to name just three – is it acceptable to use this information to discriminate between employees? The company decided that it wasn’t, in part because it would have been unfair to less privileged applicants, who may have had to use a library computer to apply.

  Illustration by Greg Stevenson

  The confounding variable here, missing from the data, is the weather, which explains the spurious correlation. To a dumb alogorithm, it might appear from the graph that ice cream consumption drove people to commit crimes. However, the real reason is simple: people consume more ice cream when it is sunnier, and they also commit more crimes on warm evenings.

  Another problem with data models is that they may suffer from a psychophysics problem. They match reality with behaviour as though the one maps perfectly onto the other. But this is wrong. For example, the data might say that people won’t pay £49 for a jar of coffee and that’s true, mostly. However people will pay 29p for a single Nespresso capsule which amounts to a similar cost – without understanding human perception it is unable to distinguish between the two. Will people pay £100 for a pair of shoes? In Walmart, no chance, but in the designer store Neiman Marcus, easily. Will people pay £500 for a mobile phone handset? Nokia’s data said no, but Apple discovered they would. Big data makes the assumption that reality maps neatly on to behaviour, but it doesn’t. Context changes everything.

  Perception may map neatly on to behaviour, but reality does not map neatly onto perception.

  We should also remember that all big data comes from the same place: the past. Yet a single change in context can change human behaviour significantly. For instance, all the behavioural data in 1993 would have predicted a great future for the fax machine.

  6.5: Nothing New under the Sun

  It is arguable that even the ancient Greeks grasped the principles of psychophysics. Study the Parthenon, and you’ll notice that there is barely a straight line in it; the floor curves upwards in the middle, the sides bow out and the columns swell in the middle.* This is because it is not designed to be perfect – it’s designed to look perfect to a human standing a hundred yards or so downhill. And long before the Parthenon, nature learned the same trick.

  Nature spends a great deal of resources on what might be called ‘perception hacking’ or, in business terminology, marketing. Berries and fruits that want to be eaten develop a distinctive colouration and an attractive taste when they ripen. By contrast, caterpillars that don’t want to be eaten have evolved to taste disgusting to their predators. And some butterflies produce what look like eyes on their wings because many animals react more cautiously in their presence. Such are examples of how nature is able to hack perception rather than changing reality.

  6.6: When It Pays to Be Objective – and When It Doesn’t

  If you are a scientist, your job is to reach beyond the quirks of human perception and create universally applicable laws that describe objective reality. Science has developed sensors and units of measurement, which measure distance, time, temperature, colour, gravity and so on. In the physical sciences we quite rightly prefer these to warped perceptual mechanisms: it does not matter whether a bridge looks strong – we need to know that it really is strong.

  A problem arises when human sciences – politics, economics or medicine, say – believe this universalism to be the hallmark of a science and pursue the same approach; in the human sciences, just as in TV design, what people perceive is sometimes more important than what is objectively true. In medicine, the obsession with objectivity leads to neglect of the placebo effect when it is seen as a ‘mere’ perceptual hack. But if a treatment like homeopathy, say, leads people to believe they will get better, and this happy delusion helps them feel less ill, then what’s not to like? Surely we should be researching this rather than decrying it?

  But what about snake-oil salesmen,* the fakers, fraudsters and conmen? Alchemy, precisely because it is not an exact science, has always been rife with charlatanry, and we should be on our guard for this. Many of the remedies proposed by people in advertising and design are wrong, and many of the findings of behavioural scientists have already been or will be proved wrong. Some parts of this book are also undoubtedly wrong – I am conscious that I have written this book from an incredibly optimistic perspective, but my argument is not that alchemy is always reliable, ethical or beneficial. Far from it – it is simply that we should not recoil from testing alchemical solutions because they do not fit with our reductionist ideas about how the world works. The purpose of this book is to persuade the reader that alchemy exists whether we like it or not, and that it is possible to use it for good; besides, if people are more aware of its existence, they will be better at spotting its misuses.*

  In physics and engineering, objective models usually make problems easier to solve, while in economics and politics objectivity might make things harder. Some pressing economic and political issues could be solved easily and cheaply if we abandoned dogmatic universal models; just as TV designers don’t wrestle with the problem of producing the entire spectrum of visible light, policy-makers, designers and businessmen would be wise to spend less time trying to improve objective reality and more time studying human perception and moral instinct.

  Every day, companies or governments wrongly make highly simplistic assumptions about what people care about. Two major US retailers, JCPenney and Macy’s, both fell foul of this misunderstanding when they tried to reduce their reliance on couponing and sales, and instead simply reduced their permanent prices. In both cases, the strategy was a commercial disaster. People didn’t want low prices – they wanted concrete savings. One possible explanation for this is that we are psychologically rivalrous, and like to feel we are getting a better deal than other people. If everyone can pay a low price, the thrill of having won out over other people disappears; a quantifiable saving makes one feel smart, while paying the same low prices as everyone else just makes us feel like cheapskates. Another possible explanation is that a low price, unlike a discount, does not allow people any scope to write a more cheerful narrative about a purchase after the event – ‘I saved £33’, rather than ‘I spent £45’.

  It is worth remembering that costly signalling may also play a role in this: certain things need to be expensive for symbolic reasons. A £200 dress reduced to £75 is fin
e, but women may not feel happy wearing a £75 dress to a wedding. TK Maxx,* a psychologically ingenious retailer, is a brilliant place to buy a present for your wife, provided that under no circumstances you reveal to her that you bought it from TK Maxx.*

  Economic logic is an attempt to create a psychology-free model of human behaviour based on presumptions of rationality, but it can be a very costly mistake. Not only can a rational approach to pricing be very destructive of perceived savings, but it also assumes that everyone reacts to savings the same way. They don’t, and context and framing matter. One of my favourite ever experiments on the perception of price and value came from the father of ‘Nudge Theory’, Richard Thaler. He asked a group of sophisticated wine enthusiasts to imagine that they had bought a bottle of vintage wine (that was now worth $75) some years before for $20. Then he asked them to choose an answer that best reflected the cost to them of drinking the bottle:

  $0. I already paid for it: 30 per cent

  $20. What I paid for it: 18 per cent

  $20 plus interest: 7 per cent

  $75 – what I could get if I sold the bottle: 20 per cent

  -$55. I get to drink a bottle that is worth $75 that I only paid $20 for, so I save money by drinking this bottle: 25 per cent

  These results reveal that some people do think like economists – but they seem to be a 20 per cent minority.

  Notice, too, that they are the people who will enjoy the wine least. (There’s a reason it’s called ‘the dismal science’.)