Dynamics of Trust - Trust Theory: A Socio-Cognitive and Computational Model - page 168

Trust Atmosphere

All the previous mechanisms (Section 6.5.1 through to Section 6.5.3) are responsible for that
celebrated 'trust atmosphere' that is claimed to be the basis of a growing market economy
or of a working political regime. They are also very fundamental in computer mediated
organizations, interactions (like electronic commerce), cooperation (CSCW), etc. and even in
multi-agent systems with autonomous agents.
The emergence of a 'trust atmosphere' just requires some generalization (see later Section
6.6): the idea (or feeling) not simply that 'X trusts Y' or/and 'Z trusts X/Y', but that 'everybody
(in this context) trusts Y' or 'Z trusts everybody', or even 'here everybody trusts everybody';
or at least that 'a lot of people trust Z', 'a lot of people trust a lot of people here'.

6.6 Trust Through Transfer and Generalization
Would you buy a 'Volkswagen Tomato sauce'? And what about a FIAT tomato sauce? Why
is it a bit more credible? And what about a Volkswagen water pump? Why is it appealing,
perceived as reliable? How marketing managers and advertising people decide to sell a new
product (good, service) under a well known brand? This is a serious problem in marketing,
but, not based on a real model.
Would you fully trust a surgeon to recommend you a medication? Or a simple medical
doctor to perform a serious surgical intervention? In general: If X trusts Y for a given

τ , will

τ ?
In this section, we analyze how it is possible to predict how/when an agent who trusts
something/someone will therefore trust something/someone else, before and without direct
experience. This is different from the models of trust just based on (or reduced to) a probability
index or a simple measure of experience and frequency, we are interested in analyzing the
trust concept so that we are able to cope with problems like: given X 's evaluation about Y 's
trustworthiness on a specific task

she Trust Y also for

τ , what can we say about X 's evaluation of Y 's trustworthiness

τ ?Given X 's evaluation of Y 's trustworthiness on a specific on a different but analogous task

task , what can we say about X 's evaluation of the trustworthiness of a different agent Z on

the same task

τ

?
In fact, in our view only a cognitive model of trust, with its analytical power (as showed
in Chapter 2), seems able to account for the inferential generalization of trustworthiness from
task to task and from agent to agent not just based on specific experience and/or learning.
In general, trust derives, directly or indirectly, from the experience. 9 There are computa-
tional models of trust in which trust is conceived as an expectation sustained by the repeated
direct interactions with other agents under the assumption that iterated experiences of success
strengthen the trustor's confidence (Witkowski et al. , 2001), (Jonker and Treur, 1999). In
the case of indirect experience, the more diffused case of study is the trust building on the
basis of the others' valuations (reputation) ((Sabater and Sierra, 2001), (Jurca and Faltings,
2003), (Conte and Paolucci, 2002). A different and also interesting case of indirect experience
for trust building (in some cases we can speak of attempts to rebuild, by other tools than
observability, the direct experience), not particularly studied in these years, is based on the

τ

9 We have to say that there is also a part of trust that derives from some personality-based or cultural 'disposition'
not based on previous experience.