Dynamics of Trust - Trust Theory: A Socio-Cognitive and Computational Model - page 177

If we call KT a class of tasks, we can say that it is characterized from a set of (quite
abstract) properties. The tasks belonging to class KT are all those whose core properties are
specifications of the abstract properties of KT . In more formal terms, in addition to (6.17) we
have:

KT ≡{ ap 1 ,..., ap n } (6.24)

τ is a task belonging to the class KT if and only if:

( ∀ ap i ( with n ≥ i ≥ 1) ∃ p j ( with k ≥ j ≥ 1) | ( p j = ap i ) ∨ ( p j is a specification of ap i ))

(6.25)

where ap i ( with n 1 ) is an abstract property.

For example, we can say that the class of tasks of 'taking care of people' is constituted from
the following abstract properties:

≥ i ≥

a) to avoid dangers to the people;
b) to satisfy peoples' main physical needs;
c) to satisfy peoples' main psychological needs.

In the case of the task 'taking care of children', we can say that it is included in the class
of 'taking care of people' because all of its core properties are specializations of the abstract
properties of the main task class:

a) to avoid dangers to children;
b) to satisfy childrens' main physical needs;
c) to maintain a peaceful and reassuring climate by playing.

At the same time we could consider 'to take care of elderly people' another task included
in this class, because the core properties in this case are also:

a) to avoid dangers to the elderly;
b) to satisfy the elderlies' main physical needs;
c) to maintain a peaceful and reassuring climate;

are specifications of the same abstract properties of the class.
The situation is more complex when a task is just partially included in a class. Consider for
example the case of 'to take care of a house'. The core properties are:

a) to clean and order the things in the house;
b) to avoid damage to the things.

In these cases the analysis has to consider the potential trustee's features and the classes
of attitudes they have. The match among classes of agents and classes of tasks can inform us
about positive or negative attitudes of the agents belonging to that agent class with respect to
the tasks belonging to that task class. Of course, going towards specifications of both agent
classes and task classes permits us to establish better matches and confrontations about which
agent is more adequate for which task.