UNIVERSITY OF HERTFORDSHIRE
		COMPUTER SCIENCE RESEARCH COLLOQUIUM

			     presents


       "Bio-cognitive Processes and Social Rules:
The Mutual Shaping of Cognition, Uncertainty and Normativity in Humans,
                Animals and Machines"

                   Dr. Stavros Anagnou
              (University of Hertfordshire)


                    18 November 2026 (Wednesday)
			            F309
		   Hatfield, College Lane Campus
			      13:00 - 14:00


		   Everyone is Welcome to Attend

                   Refreshments will be available



Abstract:

Norms are informal rules that govern interactions within collectives. Among their
many functions, they play a critical role in regulating greed in common-pool re-
source dilemmas, typically through punishment: individuals or institutions disin-
centivise overexploitation by imposing costs on norm violators. Extensive research
has examined the human psychological capacity for norms and its evolutionary ori-
gins. However, it is becoming increasingly clear that normative pressures are far
more widespread across biological phyla than previously recognised. Many ani-
mals face analogous collective-action problems and resolve them through similar
mechanisms, including organisms as simple as insects enforcing reproductive norms
and bacteria punishing free-riders who benefit from collective biofilms without con-
tributing to them. Despite this ubiquity, little research has examined how normative
pressures shaped the evolution of cognition broadly construed, not just in humans,
with most definitions of norms presupposing human-level cognitive capacities. This
thesis addresses that gap by investigating the bidirectional relationship between nor-
mative processes and bio-cognitive mechanisms, asking how each shapes the evolu-
tion of the other. We also stipulate the functional role of noise and non-veridicality,
typically dismissed as defects, as part of cognition in normative contexts. In the
first direction (normativity shaping cognition), we present two contributions. First,
an evolutionary agent-based model of norm emergence shows that behavioural and
perceptual noise is maladaptive at the group level yet does not evolve away under
individual selection, suggesting noise as a novel operationalisation of norm tight-
ness/looseness. Second, a model with evolvable affect in a shared resource dilemma
demonstrates that normative pressures (punishment) and non-normative pressures
(competition) produce qualitatively different mood mechanisms: punishment yields
a "bad mood → consume less" response that promotes resource conservation, while
competition yields the opposite, leading to collapse. This offers a new account of
how social preferences become encoded in phylogenetically ancient affective sys-
tems. In the reverse direction (cognition shaping normativity), we address the insti-
tution bootstrapping problem—the paradox that individuals find institutional par-
ticipation beneficial only once a critical mass of participants already exists. Using
a modified evolutionary game-theoretic framework incorporating cognitive biases
and perceptual noise, we show that heterogeneous, bounded perception among
agents reduces the cooperation threshold needed for institutional emergence, even
when noise is unbiased at the group level. This counterintuitive result arises from
asymmetric boundary conditions in how agents estimate small probabilities. Finally,
we present a fully developed cognition-agnostic behavioural definition of norms
that does not presuppose human-level cognition, enabling the study of normativ-
ity across humans, animals, and artificial agents. This definition is centred on three
elements: behavioural regularity, social maintenance, and the presence of multiple
equilibria under constant environmental conditions. We demonstrate its application
in distinguishing normative from non-normative regularities in AI multi-agent sys-
tems, addressing a growing need as AI collectives become increasingly prevalent.
Together, these contributions demonstrate that studying bio-cognition and norma-
tivity in conjunction yields insights inaccessible to either field alone, opening new
directions for cognitive science, social simulation, and AI governance.



---------------------------------------------------
Hertfordshire Computer Science Research Colloquium
http://cs-colloq.cs.herts.ac.uk