Socio-Cognitive Model of Trust: Quantitative Aspects - Trust Theory: A Socio-Cognitive and Computational Model - page 89

3

Socio-Cognitive Model of Trust:
Quantitative Aspects

So far in this topic, we have analyzed trust from the qualitative point of view: we have
carefully discussed the cognitive ingredients of trust, their relationships and formalizations.
Before going on (describing trust dynamics, trust sources, trust generalization, etc.) we have
to evaluate, understand, describe and formalize the quantitative nature of the trust concept.
It is true that when Xania decides to trust Yody , she has considered the different aspects
of trustworthiness, like ability, willingness, context, and so on, and the various reasons and
causes these aspects are based on. But, at the same time, she has also evaluated their amount:
if the weight of each element is enough, if the quantity of their complete composition (also
considering potential overlapping and interferences) can be evaluated as sufficient for trusting
Yody . Every day we participate in discussions where judgments like: 'John is really trustworthy,
because he knows his work and is very competent and serious with the customers', are
expressed, where quantitative evaluations are expressed in an approximate, colloquial way:
what does it mean 'really', 'very'? How much does know about John his work? How much
is serious? How much are we sure about this? In fact, directly or indirectly, there is always
a judgment of quantification over these properties. And we always test these quantifications:
'How much do you trust him?', 'Is your trust in him sufficient?', 'Is he so trustworthy?' 'Are
you sure?', and so on.
So although the qualitative analysis of the trust components is fundamental for getting the
real sense of trust concept, the quantification of its ingredients and an adequate composition
of them will permit the results of its application to be effectively evaluated and simulated.

3.1 Degrees of Trust: a Principled Quantification of Trust
The idea that trust is measurable is usual (in common sense, in social sciences, in AI
(Snijders, 1996), (Marsh, 1994)). In fact, in the majority of the approaches to the trust study
the quantification aspect emerges and prevails over the qualitative and more analytic aspects
(that are considered less relevant and sometimes useless). Because of this, in these approaches