The Glue of Society - Trust Theory: A Socio-Cognitive and Computational Model - page 271

Y(BelXp)) . These two ' Bel ' cannot be represented in the same way; I cannot use (Bel X p) to
build a belief in X 's mind.
3) In sum, if it was true that any possible expectation, for its 'prediction' part, is based on some
'inference' and that an 'inference' is based on some 'rule' (about 'what can be derived
from what'), it would be true that any trust is based on some 'rule' (but a cognitive one).
However, even this is too strong; some activated expectations are not based on 'inferences'
and 'rules' but just on associative reinforced links: I see q and this just activates, evokes the
idea of p . This is not seriously a 'rule of inference' (like: ' If (A is greater than B) and (B is
greater than C), then (A is greater than C) '). So we would not agree that any expectation
(and trust) is rule-based. However, one might expand the notion of 'rule' even to this
simple and reactive 'mechanism' (mixing up the observed regularity that they produce,
with a 'regula'/rule that should generate it). With such a broad and weak notion of rule,
we might agree that any trust - being prediction based - is in some sense 'rule-based', it
reflects some regularity and 'norm'. But not in the strict social or moral or cognitive sense;
this holds only for social trust in its 'genuine' sense, based on goal-adoption and (implicit)
commitments or on social norms and prescriptions.
4) Moreover, regularity is also about bad events; we also have 'negative' expectations (based
on the same rules, 'norms' of any kind). Now, it is a real act of violence against the current
notion of 'trust' that we are supposed to model, reducing it just to 'expectations based on
perceived normality' (Garfinkel's claim).
We may have the expectation that the author of a horrible homicide will be condemned to
die (given the laws of our states, and the practice of our government), both if we wish this
to happen and expect it out of revenge, or if we are the killer condemned to die, or activists
against the death sentence. However, if we are in favor of the death sentence, and we desire this,
actually we 'trust' our authorities over this; if we are the condemned man, or the adversaries
of the death sentence, we don't trust the authorities at all over this! This would be a serious
distortion of the concept. This is why in our chapter about third party trust (Chapter 7) we say
that this is a 'paradoxical', not true form of trust. There is a basic common mental ingredient
(the belief about the future event), and this explains why the same belief becomes trust or not
while just changing my role and goal. But it is not the right solution to reduce trust just to such
a belief, and to call 'trust' fear and opposition.
Thus, in sum, normality and regularity are not sufficient for trust, and probably are not
even necessary, if we do not extend conceptually the notion of 'rule' to cover any possible
prediction device.
9.6.1 Does Trust Create Trust and does There Exist a Norm of
Reciprocating Trust?
We have made it clear (Chapter 6) that it is not out of reciprocation that Y does the expected
action after we have trusted him and decided to rely and depend on him; and also that trust is not
always 'reciprocated' (even when Y performs the entrusted action). However, we acknowledge
that there exist a property of trust to elicit trust, and we wonder about the idea that there might
even exist a norm of trust reciprocation . Since trust is not just a behavior, but a mental state