Trust, Control and Autonomy: A Dialectic Relationship - Trust Theory: A Socio-Cognitive and Computational Model - page 190

This captures a very intuitive and common sense use of the term trust (in social interaction).
In fact, it is true - in this limited sense - that if you control me 'you don't trust me!'; and it
is true that if you do not trust me enough (to count on me) you would like to monitor, control
and enforce me in some way.
In this view, control and normative ' remedies ' 'have been described as weak, impersonal
substitutes for trust' (Sitkin and Roth, 1993), or as ' functional equivalent mechanisms'

(Tan and Thoen, 1999): 'to reach a minimum level of confidence in cooperation, partners can
use trust and control to complement each other' (Beamish, 1988). 5
We have some problems with respect to this view:

...

on the one hand, it is correct: it captures something important. However, in such a com-
plementariety, how the control precisely succeeds in augmenting confidence, is not really
modelled and explained.
on the other hand, there is something reductive and misleading in such a position:
- it reduces trust to a strict notion and loses some important uses and relations;
- it ignores different and additional aspects of trust also in the trustee;
- it misses the point of considering control as a way of increasing the strict trust in the
trustee and his trustworthiness.

We will argue that:

firstly, control is antagonistic to strict trust;
secondly, it requires new forms of trust including broad trust to be built;
thirdly, it completes and complements it;
finally, it can even create, increase the strict/internal trust.

As the reader can see it is quite a complex relationship.

7.1.4 Trust Notions: Strict (Antagonist of Control) and Broad
(Including Control)
As said we agree on the idea that (at some level) trust and control are antagonistic (one
eliminates the other) but complementary. We just consider this notion of trust - as defined by
Mayer - too restricted. It represents for us the notion of trust in a strict sense, i.e. applied to
the agent (and in particular to a social agent and to a process or action), and strictly relative
to the 'internal attribution', to the internal factor. In other words, it represents the 'trust in Y '
(as for action

and goal g ) (see Section 2.6.1). But this trust - when enough for delegation -
implies the 'trust that '( g will be achieved or maintained); and anyway it is part of a broader
trust (or non-trust) that g . 6 We consider both forms of trust. Also the trust (or confidence) in
Y , is, in fact, just the trust (expectation) that Y is able and will do the action

α

α appropriately

5 Of course, as (Tan and Thoen, 1999) noticed, control can be put in place by default, not because of a specific
evaluation of a specific partner, but because of a generalized rule of prudence or for lack of information. (See later,
on the level of trust as insufficient either for uncertainty or for low evaluation).
6 Somebody, call this broader trust 'confidence'. But in fact they seem quite synonymous: there is confidence in Y
and confidence that g .