Socio-Cognitive Model of Trust: Basic Ingredients - Trust Theory: A Socio-Cognitive and Computational Model - page 32

worst between them for a brief period, looking for new candidates at the same time; or iii) to
manage for a period without any secretary; and so on.
In situation (3) Xania expresses her trust through an (official) delegation , in fact an act of
communication (or more in general, through observable external behavior). In general, this
kind of relationship does not necessarily have to be official, it is simply known to the two
agents (as we will see in the following, in special cases, it could also be unknown to the
trustee).
In the case of autonomous agents who have no external constraints conditioning their
freedom of making a reliance, 3 we can say that a sufficient value 4 of core trust is a necessary
but not sufficient condition for a positive decision to trust (see Chapter 3); vice versa, a freely
chosen delegation (Castelfranchi and Falcone, 1998) implies the decision to trust, and the
decision to trust implies a sufficient value of core trust. We will specify which beliefs and
which goals characterize the trust of the Trustor ( X ) in another agent (the Trustee, Y ) about Y 's
behavior/action relevant for a given result p corresponding to (or included in) the goal of X ,
g X . Given the overlap between trust and (mental) reliance/delegation, we need also to clarify
their relationships.
Before starting the analysis of the basic kernel, let us analyze the relationship between
p and g X .

2.1.2 Goal State and Side Effects
As we said: g X ⊆ contains not only X 's goal but also a

set of side effects). In other words, sometimes the achievement of the goal g X is conditioned
(given the specific trustee and his own features, the contextual situation, and so on) by the
realization of other results in the world that could compromise other interests or goals of X .
I can trust my dentist to be able to solve my problem with my tooth, but - while going
there - I know (expect) that although he will also be careful and honest I will feel some pain
and will spend a lot of money. Or, I know that if I trust John - which is actually reasonable
and enough for what I need -he will become too familiar, he will take liberties with me, and
maybe I won't like this.
So the analysis and the knowledge of these side effects are really relevant when deciding
upon trusting agents, and often a merely qualitative analysis is not sufficient (see Chapter 3).

p (the set of the results of Y 's action α

2.2 Trust as Mental Attitude: a Belief-Based and Goal-Based Model
Our main claim is that: only a cognitive agent can trust another agent; only an agent endowed
with goals and beliefs.
First, one trusts another only relative to a goal, i.e. for something s/he wants to achieve,
that s/he desires or needs. If I don't potentially have goals, I cannot really decide, nor care

3 Absence of external constraints is an ideal condition: in fact, also in case of autonomous agents some constraints
are always present. In the case of the previous example, a constraint is given from the impossibility to evaluate all the
potential candidates available in the world.
4 In fact, trust is also a quantitative notion: it is constituted by different ingredients to which it is possible/necessary
to attribute a value.