Dynamics of Trust - Trust Theory: A Socio-Cognitive and Computational Model - page 144

6.2 Experience As an Interpretation Process:
Causal Attribution for Trust
It is commonly accepted ((Jonker and Treur, 1999), (Barber and Kim, 2000), (Birk, 2000))
and discussed in another work of ours (Falcone and Castelfranchi, 2001)) that one of the main
sources of trust is direct experience. It is generally supposed that, on the basis of the realized
experiences, to each success of the trustee (believed by the trustor) there is a significant
increment or a confirmation of the amount of the trustor's trust towards him, and that to every
trustee's failure (believed by the trustor) there is a corresponding reduction of the trustor's
trust towards the trustee itself.
There are several ways in which this qualitative model could be implemented in a repre-
sentative dynamic function (linearity or not of the function; presence of possible thresholds
(under a minimum threshold of the trustworthiness's value there is no trust, or vice versa, over
a maximum threshold there is full trust), and so on).
This view is very naıve, neither very explicative for humans and organizations, nor useful
for artificial systems, since it is unable to discriminate cases and reasons of failure and success
adaptively. However, this primitive view cannot be avoided until trust is modeled just as a
simple index, a dimension, an all-inclusive number; for example, reduced to mere subjective
probability. We claim that a cognitive attribution process is needed in order to update trust on
the basis of an ' interpretation' of the outcome of X 's reliance on Y and of Y 's performance
(failure or success). In doing this, a cognitive model of trust - as we have presented - is crucial.
In particular we claim that the effect of both Y 's failure or success on X 's trust in Y depends
on X 's 'causal attribution' ((Weiner, 1992)) of the event.
Following 'causal attribution theory', any success or failure can be either ascribed to factors
internal to the subject, or to environmental, external causes, and either to occasional facts, or
stable properties (of the individual or of the environment).
So, there are four possible combinations: internal and occasional ; internal and stable ;
external and occasional ; external and stable .
Is Yody's guilt or merit based on whether he was failing or successful on ? Or was the real

responsibility about the conditions in which he worked? Was his performance the standard
performance he was able to realize? Were the environmental conditions the standard ones in
which that task is realized?
The cognitive, emotional, and practical consequences of a failure (or success) strictly depend
on this causal interpretation . For example - psychologically speaking - a failure will impact
on the self-esteem of a subject only when attributed to internal and stable characteristics of
the subject itself. Analogously, a failure is not enough for producing a crisis of trust (see
Chapter 9); it depends on the causal interpretation of that outcome, on its attribution (the same
for a success producing a confirmation or improvement of trust). In fact, we can say that a
first qualitative result of the causal interpretation can be resumed in the following flow chart
(Figure 6.1).
Since in agent-mediated human interaction (like Computer Supported Cooperative Work or
Electronic Commerce) and in cooperating autonomous Multi-Agent Systems it is fundamental
to have a theory of, and instruments for 'Trust building' we claim that a correct model of this
process will be necessary and much more effective. However, this holds also for marketing and
its model of the consumer's trust and loyalty towards a brand or a shop or a product/service;
and for trust dynamics in interpersonal relations; and so on.

τ