The Glue of Society - Trust Theory: A Socio-Cognitive and Computational Model - page 267

We consider these 'motivations' and these 'commitments' not all equivalent: some are
stronger or more cogent than others. As we claimed in (Castelfranchi, 1995):

This more cogent and normative nature of S-Commitment explains why abandoning a Joint
Intention or plan, a coalition or a team is not so simple as dropping a private Intention. This
is not because the dropping agent must inform her partners -behaviour that sometimes is even
irrational-, but precisely because Joint Intentions, team work, coalitions (and what we will call
Collective-Commitments) imply S-Commitments among the members and between the member and
her group. In fact, one cannot exit a S-Commitment in the same way one can exit an I-Commitment.
Consequences (and thus utilities taken into account in the decision) are quite different because
in exiting S-Commitments one violates obligations, frustrate expectations and rights she created.
We could not trust in teams and coalitions and cooperate with each others if the stability of
reciprocal and collective Commitments was just like the stability of I-Commitments (Intentions).

Let us analyze this point in more detail by comparing five scenarios of delegation:

Intention Ascription

X is weakly delegating Y atask (let's say to raise his arm and stop the bus) on the basis

of the hypothetical ascription to Y of an intention (he intends to stop the bus in order to take
the bus).
There are two problems in this kind of situation:

τ

The ascription of the intention is just based on abduction and inferences, and to rely on this
is quite risky (we can do this when the situation is very clear and very constrained by a
script, like at the bus stop).
This is just a private intention and a personal commitment to a given action; Y can change
his private mind as he likes; he has no social obligations about this.

Intention Declaration
X is weakly delegating Y atask (to raise his arm and stop the bus) on the basis not only of

Y 's situation and behavior (the current script) but also or just on the basis of a declaration of
intention by Y . In this case both the previous problems are a bit better:
the ascription of the intention is safer and more reliable (excluding deception that on the
other hand would introduce normative aspects that we deserve for more advanced scenarios);
now Y knows that X knows about his intention and about his declaring his intention; there is
no promise and no social commitment to X , but at least by changing his mind Y should care
about X 's evaluation of his coherence or sincerity or fickleness; thus he will be a bit more
bound to his declared intention, and X can rely a bit more safely on it.

τ

In other words, X 's degree of trust can increase because of:

either a larger number of evidences;
or a larger number of motives and reasons for Y doing τ ;
or the stronger value of the involved goals/motives of Y .