A Point of View: Which comes first - trust or trustworthiness?
Ten years after her Reith lectures on trust, the philosopher Onora O'Neill revisits the subject, and wonders if the often-asked question, "How can we restore trust?" is missing the point.
During the past year, the headlines have publicised lots of cases of untrustworthy behaviour, some of it lurid, even criminal, some of it merely squalid and routine.
Some bankers and some care home workers, some journalists and some politicians and assorted others have been exposed as untrustworthy. Once again the slightly plaintive question "How can we restore trust?" is on everyone's lips.
Taken in one way, the answer is pretty obvious. First, be trustworthy. Second, provide others with good evidence that you are trustworthy. As Abraham Lincoln pointed out, "You may fool all of the people some of the time; you can even fool some of the people all of the time; but you can't fool all of the people all of the time."
To be sure, in a few cases such as that of the aptly named Mr Madoff, who made off with a lot of other people's money, that deception lasted a very long time. But duplicity and betrayal are often discovered, sometimes soon after they begin.
However, most people know this and when they ask "How can we restore trust?" they implicitly think that lack of trustworthiness is a problem others create. They don't want to know what to do to make it more likely that others will trust them, but how to place and refuse trust in others without being duped or deceived. And this is a harder task, because the untrustworthy have to pretend to be trustworthy, so they will try to withhold or fiddle the evidence, or to smother it in glossy publicity.
Placing and refusing trust intelligently is not a matter of finding guarantees or proofs. We often have to assess complex and incomplete evidence, which the masters of spin and PR may be massaging, to make things look better than they are.
But we are all pretty good at placing trust with discrimination in everyday contexts, where we know or can find out enough to discriminate. We are pretty good at checking and challenging the evidence and at reaching intelligent judgements about others' honesty, reliability and competence.
We place trust in particular people for specific purposes quite successfully. There are people whom I would trust to look after young children, but not to post an important letter - and conversely. There are people I trust to respect confidences and others who, I suspect, think that keeping confidences is a matter of sharing them with one person at a time.
Things are harder when we have to decide whether to trust experts, or complicated institutional processes. Here we often can't find or judge the evidence by ourselves.
When we are in this situation one of the least useful things we can to do is to look at opinion polls. The polls ask generic questions about attitudes: do you trust doctors, or politicians, or teachers, or journalists? Specifically, do you trust them to tell the truth?
The only sensible answer to such questions would be that most people trust some of them, but not others, in some respects, but not others. I may trust a teacher to teach my children maths, but not to drive the school minibus - and another teacher to drive safely, but not to teach maths well. We place trust intelligently only when we can grasp relevant evidence and distinguish cases, and in particular get some sense of others' competence, honesty and reliability in the relevant matters.
But the polls aren't designed to distinguish cases: they require us to bracket and ignore differences. They ask about generic attitudes to types of profession or office holder.
Attitudes can float free of evidence, and what we actually think or do, need not align well - or at all - with these generic attitudes.
I once heard someone complain that she didn't trust surgeons and, what was worse, her operation had been postponed for three weeks. Had she really mistrusted surgeons, she would surely have welcomed the delay and wanted it extended. Presumably she in fact judged that her particular surgeon would operate competently, but still had a critical and suspicious attitude to surgeons.
Sometimes it is useful to know about generic attitudes. A company that sells branded goods, where each bar of chocolate or tin of soup is just like others with the same wrapper , may learn something useful by polling customers to discover their generic attitudes to these products.
But we don't learn much of practical use by finding out about generic attitudes when people's judgements are highly differentiated. Even those with suspicious attitudes to journalists may trust those, whose work they judge reliable and convincing. Even those with trusting attitudes to doctors, may mistrust one who has been too casual or incompetent.
For practical purposes, it is not generic attitudes, but judgements about particular cases that matter. We are not lemmings, and we need not defer to others' reported attitudes of trust and mistrust.
Although the polls are of little practical use to us in deciding whom to trust, they are hugely popular, and regularly make the headlines. As a sporting nation we like to know who comes bottom and who comes top. Polls tell us who is rising and who is falling in the trust stakes. Yet this is not a very dramatic race - its yearly ratings are surprisingly stable across time.
On the whole, the types of profession or office-holder who had low trust ratings in the past still have low ratings. And those who have long come high in the trust stakes still get high scores. Many people said they mistrusted journalists and politicians in the past, and they still do. Many said that they trusted doctors and judges in the past, and they still do.
Of course there are lurches in reported trust levels when flagrant misdeeds hit the headlines, but a 30% fall in reported trust in those who were already widely mistrusted is hardly earth-shattering. It looks large only because the absolute numbers are small. And when the headlines are gone, the previous trust ratings tend to return, which is hardly surprising, given that polls collate generic attitudes.
If the polls don't help us in placing and refusing trust with discrimination, where else can we turn? The received view for the last couple of decades has been that we need more accountability and more transparency. Both can be helpful, but both can also obfuscate rather than make it easier to judge others' likely trustworthiness.
Systems of accountability won't make trust easier unless people have reason to trust these systems. If they are too complex or designed for other purposes - as they often are - most of us will find it difficult to follow them, and it's hard to know whether they are trustworthy guides. They may damage professional performance - as a midwife commented recently, it takes longer to do the paperwork than to deliver the baby and that's surely the wrong way around.
Transparency is another fashionable remedy, and has become technically easy. It can be achieved merely by pushing information into the public domain. But as lots of people will not find the information, or will find it obscure, or will not be able to work out whether to trust it, transparency is no guarantee that others will be more likely to trust.
When we are wondering whether to trust others, the most useful evidence of trustworthiness is usually non-verbal communication. A shop that announces that it will take back unwanted purchases, no questions asked, communicates that it is confident about the quality of its products, and is likely to be trusted. A business that deals well with complaints (as opposed to sending customers to a so-called helpline, to talk to people who inflict scripted conversations) is more likely to be trusted.
Professionals who take the time to listen, who use plain language, who open themselves to check and challenge, who offer others opportunity to judge their honesty, competence and reliability, are more likely to be trusted.
Where others are visibly risking their own money or reputation - where they have "skin in the game" - they are more likely to be trusted. Where statistical evidence is competent and well-communicated, it is much more likely to be trusted. Conversely, communication that is evasive, over-complex, incomplete or incompetent, defeats trust.
Good, assessable communication cannot guarantee others' trust: that trust is theirs to give or to refuse. But good communication can make it easier to judge others' trustworthiness. It is important and often indispensable for placing or refusing trust intelligently.
I think it's perverse to think of trust as more basic than trustworthiness. To place and refuse trust intelligently we need first to judge others' trustworthiness, or their lack of trustworthiness, in specific matters. Only when we can do so, will we be in a position to place and refuse trust intelligently.
What matters is not the plaintive question, "How can we restore trust?" but the practical question, "How can we make it easier to judge trustworthiness?"