As I explained in a previous review,
I am interested in trust in the context of agent computing. The blurb made
this book sound ideal:
what, precisely, is trust?
How can it be achieved and sustained?
... how can it be regained once it has been broken?
The authors distinguish three kinds of trust: naive trust, that
unreflective innocent trust of a small child with no concept of betrayal;
blind trust, that self-deceptive wilful denial of any evidence of
betrayals; and authentic trust, reflective and honest, the mature
trusting relationship that the book is concerned with. The first two are
what most people think of as trust, and why they then have problems in how
to sustain, and regain, trust. Additionally, because people use the
metaphor of trust as a fragile thing, it can consequently be too
easily broken, and difficult or impossible to mend. But
authentic trust is an ongoing process, a relationship that
incorporates the concept of distrust; such authentic trust can be
negotiated, built, and rebuilt.
They are careful to point out that this mature relationship requires
honesty, which can be painful, rather than the destructive cordial
hypocrisy that smilingly says everything is fine while knowing it is
not, leading to the impossibility of trust. Authentic
trust is created when you come to be unafraid of the negative assessments
of people you respect. By the same token, trust is destroyed by flattery
or cordial hypocrisy. These are simply lies, and when
people lie to you, you cannot trust them, their assessment of the
situation, or their promises.
The authors emphasise that trust is so valuable because it provides the
freedom to engage in projects one could not, or would
not undertake on one's own. But much of the current
philosophical focus is on trustworthiness, which leads to the
paradox of not being able to rationally trust until trustworthiness has
been established, yet being unable to establish that trustworthiness
without first trusting. So how can we rationally trust? However, the
paradox rests on a misconception of trust: if there are guarantees that
the trusted party is indeed trustworthy, there can be no authentic
trust, simply reliance or predictability. To be real trust, it must
entail the possibility of betrayal. This is because people are not
reliable or predictable in the way that laws of nature are, but they do
have interests and desires, which may conflict with our own.
Trust is restricted to agents, beings (usually people
or human institutions) who have a choice, who make decisions, who have
attitudes, beliefs, and desires, who respond to our acts and gestures with
feelings, acts, and gestures of their own. So the authors move
the focus to the trusting, rather than the trusted, side of the
relationship, and explain how it is possible to perform the act of
trust rationally. To break the paradoxical deadlock, the trust
relationship must begin with the act of trust, not the establishment of
trustworthiness, and must explicitly include recognition of the
possibility of breaches and betrayals of that trust. Trust is not earned,
it is given.
The authors are also careful to point out that not all breaches of trust
are equal. The fragile thing metaphor encourages us to believe any
breach will shatter trust beyond repair. But there are several forms of
breaches of trust. Since not all these breaches are equal, the responses
should not be equal. And, with the focus on the entire relationship, not
the single breach, the response should be context dependent.
What counts as a ... breach of trust should always be
considered within that larger framework [of the entire relationship]. It
is a question of keeping in mind what really counts. In
particular, they point out that circumstances can change, and so promises
may need to be renegotiated in the light of new understanding. They
emphsise this such requests for renegotiation are not a breach of trust,
but actually a confirmation of the trusting relationship.
Breaches occur for many reasons. Sometimes things just don't work out,
and it is nobody's fault. Sometimes someone makes a mistake, someone is at
fault, yet focusing on the relationship rather than
the outcome gives us the navigational tools to overcome, if not overlook,
mere mistakes. Next are the real breaches of trust, maybe due
to arrogance or insincerity, rather than being mere mistakes. These
require apologies to maintain the authentic trust relationship.
An apology is a statement of intention to redeem
oneself, and the beginning of a conversation about how this can be done.
Then we come to indifference, or lack or caring, and finally to genuine
betrayal. Focusing on the relationship, rather than the metaphor of fragile
thing, even full betrayals can be overcome: this requires the act of
forgiveness. Forgiveness is an action, a ritualised undoing of the
act of betrayal. But in the same way that an act of trust changes both
parties and the relationship, so an act of betrayal and forgiveness
changes things; the relationship does not return to what it once was. This
is part of the property of authentic trust: it is a dynamic, changing
process that must be worked at and maintained.
This focus on the trusting party also helps focus on the essential
property of self-trust, trust or confidence in one's own ability
to trust wisely. Trust is a skill that can be learned by doing, but this
skill begins at home: If you do not trust yourself or
your ability to behave correctly in a difficult situation, you may well
find it difficult to have faith that others can do so. You also must trust
yourself to have a good sense of people, to choose to trust people who are
likely to be trustworthy rather than, for instance, because you find them
charming or attractive. ... If you do not trust yourself to choose wisely
the people you trust or with whom you form relationships, you may have a
good deal of trouble trusting people.
All these points are important, and interesting, and act as an essential
a first step in clarifying terminology and metaphors, and setting the
stage for building authentic trust. It has certainly clarified many
concepts for me, and given food for thought. But I am also interested in
the promised descriptions of how this authentic trust can be achieved, and
regained, in practice, and there was very little concrete said on this
topic. I would have liked to see a further section of examples (and
advice?) on building such relationships.
Some longer quotes:
It is often suggested that widespread
distrust is stimulating the current interest in trust. We think that the
truth is more subtle and interesting: there is more trust in the world
than ever before, and the increasingly global dependence on trust spurs
both our interest in and our need for trust. In situations of distrust,
people do not talk about trust but rather develop strategies for coping
with its absence.
Both trust and distrust tend to be
self-confirming, and it is easy to see why. If one person trusts another,
the second person, knowing that he or she is trusted, will be more likely
to be trustworthy, thus confirming the trust on the part of the first
person. The psychological reward of trust is that it is gratifying to be
trusted. It is also gratifying, in a more profound way, to trust. Trust
indicates respect, and trust creates a bond (if only, at first, the bond
of trust). The problem with thinking about trust as an attitude toward
other people is that it ignores the reciprocal nature of trust.
Most people respond to trust by being trustworthy, making further trust
all the more likely.
Simple trust, the sort of trust that is
celebrated in a well-cared-for dog, is not the paradigm to follow, but
neither is the lawyer's fantasy of an "ironclad" contract.
... Blind trust is denial. ...
Too often, trust as such is confused with blind trust ..., and trust
as such is taken to exclude criticism, scrutiny, and "objective"
consideration of the evidence. If "objective" means simply "impartial,"
this is true. To trust is to be committed, and thus to be partial. But
trust need be neither blind nor simple.
Trust can be prudent, measured,
reflective, and conditional and still be authentic, not blind. Part of the
problem with trust is that too many people refuse to consider as trust any
trust that is prudent, measured, and conditional. We suggest, to the
contrary, that blind trust is not really trust at all. In religious
contexts, it might better be called faith, but in secular contexts, it is
best identified as foolishness. To confuse such uncritical acceptance and
willful denial of all possible counterevidence with trust is to
misunderstand trust in the most profound way. The equation of trust with
blind trust leads to the conclusion that it is never wise to trust. Trust
becomes a vice instead of a virtue, a liability instead of a strength.
predictability requires a high degree of
probability, even a kind of (psychological) certainty. Trust requires
something else: a reciprocal relationship in which questions of
probability take a back seat to questions of mutual expectations,
responses, and commitments. Analyses of trust in terms of predictability
miss this essential aspect of trust, the element of reciprocity. In what
we call reliance, predictability (combined, perhaps, with a certain amount
of control) is definitive. One must still make decisions about whether to
rely or not rely on something, given the possibility or probability of
failure or disappointment. But what is not present in reliance that is
crucial in trust (and distrust) is that the person trusted has intentions
and motives and makes his or her own decisions, with or without regard for
the other person's decision to trust.
We do not merely predict and control the
behavior of other people. We reason with them. We appeal to their
emotions, their sympathies, their fears, their hopes and desires. ...
corporations have interests and strategies, and this leads us to consider
them in human terms.
Machines ... do not have interests and
strategies. ... But if corporations have interests, even if they are only
narrow economic interests, and they behave strategically, then they can be
appealed to, negotiated with, depended on (or not) to fulfill their
Failing to trust someone is not merely an
omission. It is unethical not to trust people when they are
plausibly trustworthy, just as it is unethical to treat them unfairly. In
fact, refusing to trust people may be more damaging to them than treating
them unfairly, for the latter fails only to give them what they deserve.
The former limits their capacity to act as full human beings.
Simple trust is unreflective. Blind trust
is self-deceptive. Authentic trust is both reflective and honest with
itself and others. All forms of trust involve counting on other people,
and, as such, they all are vulnerable to betrayal. But whereas simple and
blind trust experience betrayal as earth-shattering, betrayal is neither
surprising nor devastating to authentic trust. All trust involves
vulnerability and risk, and nothing would count as trust if there were no
possibility of betrayal. But whereas simple trust is devoid of distrust,
and blind trust denies the very possibility of distrust, authentic trust
is articulated in such a way that it must recognize the possibilities for
betrayal and disappointment. It has taken into account the arguments for
distrust, but has nevertheless resolved itself on the side of trust.
Authentic trust is thus complex, and it is anything but naive. Authentic
trust is not opposed to distrust so much as it is in a continuing
dialectic with it, trust and distrust defining each other in terms of the
We are all painfully familiar with
bureaucratic moods, which are often conflated with "policy."
Such moods dictate a retreat from personal responsibility and judgment,
typically in the name of following rules.
our moods and emotions do not happen to
us. We choose them. To think of trust as an emotional phenomenon is to
accept that trust begins (and ends) with care, and it is also to embrace
the idea that trust is a personal choice and within our realm of
responsibility. This suggests a philosophy of life. Human life is a series
of emotional engagements and projects, in which we invent a shared future
through our moods and emotions. To believe otherwise is to cut ourselves
off both from other people and from the power we each can bring to our
lives by working together and trusting one another.
Sometimes trusting ourselves is actually
much like trusting another person. We find ourselves waiting to see how we
will behave in some emergency or emotionally charged situation, for
instance. ... Some people think that self-trust means the absence of all
anxiety, total self-confidence, but they are wrong. Fear and anxiety are
indications of uncertainty, and trust necessarily involves uncertainty. It
is the absence of fear and anxiety that may mark the lack of true
self-trust. Their total absence more likely indicates indifference or
ignorance, or perhaps a simple trust that is bewildered and devastated
when we do fail ourselves.
Just as care is an essential ingredient of
trust of every kind, lack of caring and indifference stand as antitheses
to trust. Cynicism, even when it presents itself as serious and sincere,
is often a self-deceived form of indifference. One pretends not to care
when one really does care, or one intends not to care because one does not
want to be responsible for doing anything about the situation. But between
cynicism and indifference, there is only a philosophical difference: the
cynic claims to have a philosophy of life to justify his or her
Lying is in itself a breach of trust;
indeed, by some standards, the ultimate breach of trust. Lying embodies a
wholesale insincerity---stating as truth what one fully knows not to be
true---and it may also manifest a profound lack of caring, even when the
lie (a "white" lie) is intended to protect the feelings of the
person to whom the lie is told. In such cases, one may well care about the
feelings of the person, and that is a form of care. But it is a
shortsighted, limited notion of care, and it may cause violence to the
longer-term relationship. ... But many lies are not so white, and not
intended to protect the feelings of the recipient. They are rather
designed to protect the liar from the consequences of his or her actions.
Thus Kant, in a judgment that captures the viciousness of some lies, says
that lying is a violation of the very humanity of the person lied to, a
denial of his or her human dignity.
Compensation is repayment, where the
betrayal consisted of some quantifiable good. Money is the most obvious
example. When the betrayal is financial, as in many securities fraud
cases, the idea of "measure for measure" would seem to have a
precise meaning. Compensation in this sense could thus be assimilated to
the more general concepts of debt and repayment. That would be misleading,
however, because what must be compensated for is not merely the financial
loss but also the betrayal. Thus the apparent quantitative objectivity
gives way to something much more subjective. And when the loss due to the
betrayal is entirely made up of such intangibles as pride, status,
self-confidence, and trust, the notion of compensation begins to look