Concluding Remarks and Pointers - Trust Theory: A Socio-Cognitive and Computational Model - page 354

'measures' of some aspect of trust, or some ad hoc technical notion of very limited use; we
need a real theory of trust. This is the (hard) mission of science.

13.2 Neuro-Trust and the Need for a Theoretical Model
In the last few years a set of studies on the neurobiological evidence of trust have been de-
veloped. One of the main claims from Kosfeld et al. 's neurobiological approach (Kosfeld
et al. , 2005) about trust is that trust 'cannot be captured by beliefs about people's trustworthi-
ness and risk preference alone, but that social preferences play a key role in trusting behaviour'
(Fehr, 2008). In other words, on the basis of their studies (Kosfeld et al. , 2005) there is an
important and significant distinction between risk constituted by asocial factors and that based
on interpersonal interactions. The thesis would be: by analyzing the neurobiological bases
of trust it is shown that a trustor does not decrease the general sensitivity to risk, but on the
contrary, this sensitivity seems to be decreased in social interaction.
However, how did they determine the neurobiological bases? They experimented with the
trust game (involving real monetary exchanges between two actors playing the roles of trustor
and trustee) with two groups of students, one that had received neuropeptide oxytocin and
the other that had received an inert placebo. As Fehr writes: 'the rationale for the experiment
originates in evidence indicating that oxytocin plays a key role in certain prosocial approach
behaviours in non-human mammals. (

) Based on the animal literature, Kosfeld et al. (2005),
hypothesized that oxytocin might cause humans to exhibit more behavioural trust as measured
in the trust game.' (Fehr, 2008). In these experiments they also show how oxytocin has a
specific effect on social behaviour because it impacts differently on the trustor and the trustee
(only in the first case is there a positive influence). In addition, it is also shown that the trustor
does not reduce the sensitivity to risk as a general behaviour but as a consequence of the
partner nature (human versus not-human).
Following this paradigm we could say that when some social framework is given and
perceived by the subject, oxytocin is released in a certain quantity, so modifying the activity of
precise regions of the brain and consequently producing a more or less trusting behavior. But,
what are the features of that social framework? Is this framework just based on the presence of
other humans? And what is the role played by past experiences (for modulating the oxytocin
release)? How are the unconscious and spontaneous bias (characterizing the emotional, hot,
not rationale trust) related to the conscious deliberative reasoning (characterizing the rational
and planned trust)?
The results of this discovery are without doubt quite relevant. But they must be comple-
mented and interlinked with a general cognitive theory of trust. Without this link, without
a mediation of more complex structures and functions, we risk that an articulate and multi-
dependent concept is trivially translated only in a chemical activation of a specific brain area
cutting out all the real complexity of the phenomenon (unable to explain, for example, why,
before taking a very difficult decision we think on this for hours, sometimes for days. And
how the contribution of the different beliefs impacts on that choice). The fact of individuating
and establishing some precise and well defined basic neuro-mechanisms can be considered
an important advancement for the brain studies and also for founding the cognitive models
of the behaviour. In any case, in our view, they cannot be considered as the only descrip-
tion of the external behaviour (based on more sophisticated, articulated and complex notions

...