Hinton Lecture: Can trust be rebuilt?
Speaker: Baroness O'Neill
Introduction
The thought that trust has declined or even collapsed has become one of the clichés of our times. It usually evokes one of two responses. The most common response is, or at least sounds, optimistic: trust has collapsed, so it must be rebuilt. The less common response is, or at least sounds, pessimistic: trust has collapsed, so we must do without it. I think that both reactions are too quick, and that there is a lot to be thought about before we consider, let alone adopt, either of them. The assumption that trust has declined or collapsed itself deserves close scrutiny. I shall argue that it reflects a mistaken focus on trusting and mistrustful attitudes, rather than a useful focus on the epistemic and practical requirements for placing and refusing trust intelligently.
Trusting attitudes and dispositions may indeed have had their day, as have other attitudes and dispositions that were once thought virtues. We know that virtues and vices are buffeted by change, and that their reputations go up and down. Today the martial virtues are not generally admired (don't think only of courage, still much admired, but far more than a martial virtue: think also of obedience to orders and self sacrifice – both now widely questioned). Nor are there many enthusiasts to be found for Hume's "train of monkish virtues" (celibacy, fasting, penance, mortification, self-denial, humility, silence, solitude …) which he thought should be transferred to "the catalogue of vices"1. Equally, some vices have come up in the world. Greed was once seen as a deadly sin, but in the early modern period came to be seen as the motor of capitalism and so as a public virtue, even if still seen as a private vice2. Even charity, the greatest of the Christian virtues has, as the new Catholic Encyclopaedia puts it, come down in the world, and is now often equated with sporadic philanthropy.
Perhaps trust too has had its day; perhaps it is no more than an outmoded form of deference or credulity. Some people think that its collapse would be no bad thing, because it is not a virtue at all, but rather a foolish and risky attitude to others' claims and commitments. Those who think this often claim that suspicion of others' claims and commitments is a safer and more intelligent attitude3. If we had to choose between suspicion and credulity, it might be reasonable to go for suspicion because it seems less risky than credulity. However, the thought that blanket credulity and wholesale suspicion are the only possible attitudes to others, or to their claims and commitments, as well as the underlying assumption that trust is only a matter of having trusting attitudes, seem to me large mistakes with damaging consequences. Placing trust is not simply a matter of having a trusting attitude, and trust need not be a matter of blind credulity. Trust can be placed or refused intelligently, and doing so matters.
1. A Costly Mistake about Trust
How has this costly mistake come about? I think that it arises in large part because many contemporary discussions focus on trusting and mistrustful attitudes, rather than asking more practically what it takes to place and refuse trust well. As a consequence, a lot of public and media discussion of trust, and of so-called trust-building, marginalises practical questions about trust. It does not ask how we can place and refuse trust intelligently, or what we should do if we want others to trust us intelligently. Rather it focuses on empirical questions about the attitudes people typically report, and specifically on their reported attitudes of trust and mistrust towards certain types of professionals or office holders, or towards types of institution, such as public bodies, banks, companies or charities.
Polls provide a principal source of evidence for answering these rather diffuse empirical questions. Pollsters zealously and regularly record trusting and mistrustful attitudes towards politicians, journalists, doctors, judges, and other office holders and towards institutions, noting in particular how much we trust them to tell the truth. The answers given are then cited as empirical evidence of trusting and mistrustful attitudes, and are often supposed to provide evidence that trust has collapsed.
Such polls have become standard fare in contemporary democracies. Their findings are often well publicized and widely savoured. In the UK, for example, in 2009 92% of us said we generally trusted doctors to tell the truth, but only 22% that we trusted journalists to do so. Politicians usually 'score' close to journalists (generally a little better, but worse in 2009 scoring only 13%), and professors reasonably close to doctors (but generally not as good, scoring 80% in 2009). Incidentally, pollsters scored 45%, which is a little lower than the ordinary man or woman in the street scored (54%)4.
The sort of evidence that pollsters collect is sometimes seen as providing reasons to trust or mistrust. Yet it is wholly obscure how evidence about third parties' attitudes of trust or mistrust could provide reasons for placing or refusing trust in others: it would provide reasons for placing or refusing trust only if we knew that attitudes of trust reliably track trustworthiness, and attitudes of mistrust reliably track untrustworthiness. But that is exactly what we do not know, and why placing trust well is difficult.
The evidence that we actually need in order to place trust well is evidence of the trustworthiness - or lack of trustworthiness - of those whom we might trust, or refuse to trust. We cannot infer others' levels of trustworthiness by looking at third parties' attitudes of trust and mistrust towards them, since these need not vary together. Trusting isn't like following or ignoring fashion, where looking at what suitable third parties are wearing gives you the right sort of evidence if you want to dress fashionably - or, for that matter, unfashionably. When we place or refuse trust we don't need to know how third parties place their trust -or their mistrust. Rather we need evidence of the trustworthiness or untrustworthiness of those in whom we might place or to whom we might refuse trust. Reputational evidence of third parties' attitudes towards them is the wrong sort of evidence.
As I see it, much of what is cited as evidence for a crisis of trust mainly reflects this poorly articulated focus on empirical questions about reported attitudes of trust and mistrust, coupled with a persistent disregard of the practical demands of placing and refusing trust intelligently. But our practical aim is to place trust well, not to place it lemming-like as others do. To place it well we need to judge the trustworthiness of those in whom we place it or to whom we refuse it, rather than to discover what attitudes third parties hold towards them. And here is the rub: judging others' trustworthiness is much harder than discovering third parties' attitudes, whether trusting or mistrustful. The practical task of judging where to place and where to refuse trust is more important than discovering the generic attitudes of trust or mistrust of third parties, but is also much more difficult.
2. Has trust collapsed, or even declined?
The basic reason why it is hard to tell whether trust has declined, let alone collapsed, is that the evidence for levels of trust is very heterogeneous, ranging from judgements people make about others' claims and commitments, to the action they take, to the attitudes they evince when polled. Often empirical evidence from these sources points in divergent directions. A person may claim to mistrust banks, but lodge his savings with them; an institution may claim to trust its employees, but monitor their work obsessively.
And, if we did attach weight mainly to the reputational evidence assembled by pollsters we would have to conclude that trust has not collapsed, and even that in many cases it has not declined much. Polls provide empirical evidence about generic attitudes of trust and mistrust at particular times, but they do not show that trusting attitudes have generally declined, let alone that they have collapsed. A remarkable, but generally unremarked, feature of pollsters' findings on trust levels is that where we have time series they suggest that trust levels in types of professional or office holder have been pretty stable across time. In the UK across the last 25 years the percentage of affirmative answers to the favoured diagnostic question ‘Do you trust x's to tell the truth?' has not varied greatly. Journalists have persistently languished between 10% and 20% on trust ratings, politicians between 13% and 21%; government ministers between 11% and 24 %. On the other hand, doctors have flourished between 82% and 92%; scientists between 63% and 72%. Insofar as trust levels have changed in the UK across the last two decades they have mainly increased a bit.
This, of course, is not what is widely believed, or what the media report, yet polling evidence supports neither the claim that we now face an unprecedented crisis of trust, nor the claim that we have an urgent need to 'rebuild' trust5 . Polling evidence offers no reason, to think that we face a spiralling crisis of declining trust that we did not face 10 or 20 years ago.
I do not think this is as surprising that poll results are rather stable. Pollsters mainly record attitudes to typical holders of professional and public roles or typical institutions, and generic attitudes may well respond sluggishly (if at all) to specific evidence, even if it is available and known. Stable generic attitudes are wholly compatible with people holding highly differentiated views of the trustworthiness of individual holders of roles, or of particular institutions, and even more differentiated views of their trustworthiness in specific matters. There can be clear reasons for trusting particular office holders or institutions more (or less) than average, or for trusting them more (or less) for specific tasks and purposes. I may have a mistrustful attitude to journalists or politicians, but trust a journalist or politician whom I judge honest and reliable; you may have a trusting attitude to doctors or teachers, but mistrust one whom you know to be less than competent. I may mistrust a charity that exaggerates its independence or its influence, but you may have reasons to trust their specific claims about a project they undertook. Trust in particular office holders and institutions need not be aligned with the generic attitudes expressed by the same informants to the generality of those holding such positions or the generality of institutions of a given type.
A focus on attitudes to typical cases marginalises or brackets precisely the differentiating evidence we use when we place and refuse trust intelligently. Attitudes need not, indeed often cannot, differentiate between cases: but judgement aims to differentiate between cases intelligently. Placing and refusing trust is a matter of judgement, not a matter of having or lacking trusting attitudes. The fact that intelligent trust in particular cases is often highly differentiated suggests that sweeping claims that trust has declined (or alternatively that it has not declined) cannot be established on the basis of attitudinal evidence. Nor is it easy to see how such broad claims could be settled on the basis of the wider but very incomplete range of more disparate evidence that we can assemble.
3. Trust is not eliminable
So it is no surprise that placing and refusing trust remains, as it always has been, problematic and demanding. It is demanding because trust can be misplaced: the trustworthy may be mistrusted, and the untrustworthy may be trusted. Either type of mismatch can have severe costs. When we misjudge matters and refuse to trust the trustworthy we incur needless worry and cost in trying to check them out and hold them to account, while they may feel wrongly questioned, undermined, even insulted - and ultimately less inclined to be trustworthy6. When we misjudge matters the other way, the costs can be even higher: If we place trust in the untrustworthy we may find our trust betrayed, and lose whatever we staked, be it love, friendship, political aims or money.
Yet we cannot, I believe, eliminate the need to judge where to place and refuse trust, so need to take seriously the epistemic and other demands of doing so intelligently. Trust is not merely pleasant and efficient if it can be well placed; it is indispensable. Any idea that we could replace trust entirely with formalised systems of accountability simply invites the question: why should we place trust in one or another system of accountability? Attempts to rely solely on systems of accountability lead only to a spiralling regress of checking up on others - unless at some point we trust the system of accountability. Fortunately it is often possible to place and refuse trust intelligently.
In judging whether to trust others' claims or commitments we need, and can often find, some evidence for their honesty, reliability and competence, or their lack of these, in the relevant matter. Often we manage to do so quite smoothly. Every day we manage with some success to place trust or refuse trust in others - in motorists and pedestrians, shopkeepers and colleagues, friends and family - in particular contexts. Every day we place or refuse trust reasonably intelligently in the light of evidence, and achieve a reasonable match of trust to trustworthiness and of mistrust to untrustworthiness. We do not manage this practical task to perfection, but we often manage it quite adequately.
Still, there are hard as well as easy cases. It's one thing to judge honesty, reliability and competence in daily life, but harder in public, institutional and professional life. There we may need to judge institutions that are opaque and expertise that is arcane, and strangers with whom we have minimal interaction. These realities suggest that the widespread assumption that trust has declined, or perhaps collapsed, may in fact largely reflect our experience that placing trust intelligently can be hard. We may not have much good attitudinal evidence of the decline of trust, but we all experience the practical difficulty of placing and refusing trust in a world in which we interact with and depend on distant others whom we do not know, experts whom we cannot assess, and institutions that we do not understand. Yet despite these difficulties, we often manage to place and refuse trust intelligently even in these less tractable situations. We buy goods online, accept information supplied by people we do not know and cannot question, and make payments through complex systems that we do not understand. How can we do any of these intelligently?
The standard answer is that where we lack time or expertise, so have no direct ways of judging others' honesty reliability and competence, we construct or use indirect ways of doing so. Systems of accountability can provide indirect indicators of honesty, reliability and competence, and so a surrogate basis for placing or refusing trust intelligently. Intelligent shoppers may decide to trust computer firms because they can judge product specifications, recognise brand names and gauge the damage that a company could suffer if it marketed defective products. Intelligent savers may refuse to trust retail banks if they make too many mistakes that they fail to correct, market incomprehensible products or can't answer sensible questions. (I say nothing about trusting investment banks!). At their best, systems of accountability provide some indirect evidence by which to judge trustworthiness in public, institutional and professional life.
Unfortunately many of the systems of accountability introduced to secure greater trustworthiness in public and professional life, in the UK and elsewhere, across the last 25 years have proved cumbersome, and some of them counterproductive. Some are dysfunctional and create perverse incentives; others incentivise trustworthy performance, but do little to help the inexpert of us to place and refuse trust more intelligently. Company and institutional reports and accounts are dry and complex and ever more voluminous, and the addition of pretty pictures has not made it easier for those with little time or expertise to spot underlying problems. Simplified extracts from the information assembled by formal systems of accountability – press releases, school brochures, ratings of public services, testimonials from satisfied customers - may be more comprehensible to the less expert of us. But unfortunately this clarity is often achieved by introducing 'simplifying' assumptions that limit, distort or misrepresent evidence for judging untrustworthiness. Even when accountability supports trustworthiness, it may not provide usable evidence for placing and refusing trust intelligently.
4. Giving and Building Trust
I have argued that the pessimistic response to lack or decline of trust is a non starter. We cannot eliminate the need to place and refuse trust. Yet this does not show that the optimistic view that we need to 'rebuild trust' is the right response. The difficulty is not only that this formulation of the problem wrongly assumes that we know that there has been a decline or collapse in trust, which therefore needs rebuilding. It is also that the image of 'rebuilding trust' takes a poorly focused view of the placing and refusal of trust, and of the epistemic and practical demands it places both on those who place trust and on those who seek trust. Trust is relational or transactional - it is given or refused by one party to another. If we characterise dealing with a lack of trust as a call for trust to be 'rebuilt' we make several questionable assumptions.
Our aim, after all, is not to have more trust whatever the evidence, but specifically to align trust with trustworthiness and mistrust with untrustworthiness. Those who seek to place their trust intelligently need first to discover and understand what the other party claims is the case or undertakes to do, and then to judge the honesty, competence and reliability that they bring to their claims or commitments. By contrast, those who seek to receive others' trust need first to communicate what they claim and which commitments they make, and then to try to offer others assessable evidence that their claims and commitments are indeed trustworthy. The tasks of those who give and those who seek or receive trust are therefore quite distinct, and characterising both as a matter of building or rebuilding trust elides the difference.
The only case in which it might be apt to speak of 'trust building' is where both parties seek both to give and receive trust from one another. That, I suggest, is why the phrase 'trust building' sits most happily in discussions of conflict resolution, where all parties may seek both to place trust better and to be more trusted. For present purposes I set aside the special case of efforts to achieve greater mutual trust, and focus on the more general case in which the two parties may have asymmetric ambitions, one seeking to place trust and the other to be trusted. (I also set aside the conman's dream of receiving trust without being trustworthy).
5. Trust and Communication
Communication is essential if evidence that of trustworthiness is to be provided or received. This demand cuts deeper than may appear. It is common ground that it is not enough for those who seek to be trusted to provide obscure or unintelligible evidence, or for those who seek to trust to rely on obscure and unintelligible evidence, however copious. But beyond that there is deep disagreement.
There are those who now lay great weight on openness or transparency as an adjunct to accountability that both shows and enables others to judge trustworthiness. But while transparency limits secrecy, it is not enough to support the intelligent placing and refusal of trust. Putting information in the public domain is useful for professionals or experts with the time, resource and experience to find, read and use what is made available, and to assess how it has been selected and edited. But mere transparency is not enough to ensure that wider audiences will find or can use information. Transparently disseminated information often fails to communicate, whether by using technical jargon, small print, or simply by swamping others with irrelevant or distracting detail, or by demanding time, knowledge and attention that they lack. (terms and conditions apply!).
This difficulty cannot be overcome merely by offering and receiving simpler and more accessible digests of the relevant information. Simpler and more accessible extracts or digests of information will help only those who can judge whether the intermediaries who select, edit and purvey the information are trustworthy. Those who cannot judge the intermediaries will not be able to assess the information they provide, so will not be helped to judge the trustworthiness of those to whom the information applies.
Those who seek to be trusted or to place trust need respectively to find ways of grasping and of providing evidence that reveals trustworthiness. This evidence can be given by direct communication (speech content), but is often conveyed through speech acts and by the ways in relations and interactions are structured. Those seek to be trusted can convey their trustworthiness by indirect moves that show that they confidently make themselves vulnerable to the other party, for example by providing opportunities to withdraw from activities or commitments at no cost (returning defective goods; withdrawing from medical research (exit)). Alternatively, they can help others to judge their trustworthiness by providing opportunities to challenge and check their claims and commitments (genuine and usable complaints procedures rather than ones designed to fob complainants off or drown them in process (voice)). They can also provide forms of recognition and esteem that allow those whose trust they seek to tell that their trust is prized and taken seriously (loyalty). Each of these approaches is well explored in commercial life, and in some areas of public life, but less evident in others. It would, I think, be useful to explore more systematically which indirect approaches are effective ways of communicating evidence of trustworthiness that is intelligible, checkable and useful, and which do not.
Those whose trustworthiness is to be judged must offer intelligible and assessable evidence that their claims are true and that their commitments will be honoured. Those who judge others' trustworthiness must grasp and weigh that evidence. We cannot give place trust intelligently without receiving evidence, and we cannot receive trust that is intelligently given without giving evidence.
6. Conclusion
Many of the claims for which I have argued this evening may seem negative, and I have been conscious of swimming against various currents of received opinion. I have suggested that polling evidence does not show that we face a decline, let alone a collapse, of trust, and that the other available evidence does not point in any single direction. I have further suggested that the pessimistic view that we can do without trust, provided we create systems of accountability that secure a reasonable measure of trustworthiness, is untenable. And I have suggested that the supposedly optimistic view that we need to 'rebuild' trust is no more than a slogan, which obscures the complexity of the evidential requirements for placing trust intelligently. I have also suggested, I hope more positively, that once we start thinking about those evidential requirements, demands both on those who hope to receive trust and those who may give it become clearer. The undramatic conclusion is that there is no need to rebuild trust, because it has not collapsed, but that to keep the placing and refusal of trust in good repair needs careful attention to the specific epistemic and practical requirements both for placing trust and for receiving others' trust.
© Baroness Onora O’Neill, 2010
[Draft Only]
Footnotes
1 "Celibacy, fasting, penance, mortification, self-denial, humility, silence, solitude, and the whole train of monkish virtues; for what reason are they everywhere rejected by men of sense, but because they serve to no manner of purpose; neither advance a man's fortune in the world, nor render him a more valuable member of society; neither qualify him for the entertainment of company, nor increase his power of self-enjoyment? We observe, on the contrary, that they cross all these desirable ends; stupefy the understanding and harden the heart, obscure the fancy and sour the temper. We justly, therefore, transfer them to the opposite column, and place them in the catalogue of vices...." (EPM, §9, ¶3)
2A. O Hirschmann 1977. The Passions and the Interests: Political Arguments for Capitalism before Its Triumph. Princeton, NJ: Princeton University Press
'Suspicion is a virtue as long as its object is the public good, and as long as it stays within proper bounds. ... Guard with jealous attention the public liberty. Suspect everyone who approaches that jewel.'
4 Summary of Ipsos-Mori findings at www.ipsos-mori.com/Assets/Docs/Polls/poll-trust-in-professions-topline-2009.pdf
5See the table in Anthony Seldon, Trust: How we lost it and how to get it back, Biteback, London, 2009.
6 This is a classical theme in anthropological literature, often linked to claims that 'gift' relationships are undermined by cash incentives. Mauss, Titmuss. The issue is likely to resurface in current work on using cash incentives to promote health (payments for smoking cessation).
