IEEE Technology and Society Magazine - June 2020 - 57

anticipated trajectory leaves the ZoD (see lower right
quadrant of Figure 2).5

Excessive

Quasi-Stability
For institutional dynamics, we propose quasi-stability
for controlling unrestricted self-modification. This
notion of quasi-stability is derived from Rawls' theory of
justice [8], Ashby's design for a brain [4], and Leibensteins's definition of economic equilibrium [23]. Rawls
asserted that "a well-ordered society is quasi-stable with
respect to the justice of its institutions and the sense of
justice needed to maintain this condition." His concepts
of equilibrium and stability of a system were derived
from Ashby, who defined, inter alia: polystable systems,
as systems whose parts have many equilibria and random parts have been joined at random; ultrastable
-s ystems, as a particular type of reactive systems
embedded in an environment; and multistable systems,
a system whose state and behavior increasingly
depends on what happens to it, not the state in which it
was initialized. From these definitions, the concept of
quasi-stability was defined by Leibenstein as a system
in which, after a period of disruption, the system re-stabilizes in an equilibrium except that only some, but not
all, of the control variables return to their prior state
(control variables being those variables controlled by
the system regulator, choosing values for those variables for which the system gives the "best" or "preferred" output).
Quasi-stability for institutional dynamics has been
investigated, in part, in [14] and [24]. Open systems with
self-determined rules, whether social or electronic, have
to deal with the expectation of error (e.g., in a social system, excess appropriation from a common-pool
resource; in an electronic system, interference in a wireless network), dissent, and potential incongruence. In
[14], these issues led to reorientation of either the rules
or replacement of the rulers, depending on the nature of
the error or incongruence (i.e., either the rule was
wrong, or the ruler was applying the rule wrongly). In
[24], incongruence led to a re-assignment of roles to
ensure that all "citizens" share the burden, in one way
or another (e.g., equally), in the making, adjudicating,
and enforcing the rules. In both cases, some form of
quasi-stability could be observed: the rules would be

5

We note in passing that understanding the trajectory of a system is important to overcome the cognitive bias of "just noticeable difference" within
a changing population. A series of "small enough" changes over a "long
enough" period are effectively imperceptible to those that experienced
them; while new entrants to the system think that "it was ever thus." However, the end state is far from the original state and much less preferable.
The climate emergency and the centralization of academic administration
are two examples of when the rate of change in a system synchronizes with
the rate of change of the population, its trajectory can approach and even
leave the ZoD, terminating with systemic crisis and collapse.

JUNE 2020

∕

Devaluation

>δ
Minimal

Exploitation

<δ
Maximal

Zone of Dignity

Fragmentation

Stagnation

Inadequate
FIGURE 3. The Zone of Dignity in two dimensions.

re-configured and the system would be stable for a period, with a different setting for the "control variables",
until another self-examination reset them again.
The key outcome of quasi-stability for ecological
systems is that while a shift in social or environmental
circumstances may cause an institution to become
unjust - for example, a reaction to crisis may temporarily cause an institution to be perceived as unjust, in
due course the institution is reformed as the crisis
abates, and justice is restored. Therefore, it is a critical quality of a system that ensures reversion to equilibrium, i.e., the (current) equilibrium is a product of
what happened to the system, not the equilibrium
from which it started.

Socio-Techno-Ecological Systems
We observe that while reflection requires human agency,
ecosystemic reflexivity also requires the ability to listen to
and interpret signals from the non-human (natural) world.
The challenge for algorithmic reflexive governance is then
two-fold. Firstly, there is a challenge to demonstrate that
the deliberative mechanisms/principles we advocate for
resolution of the four tensions are not only consistent with
each other, but also mutually reinforcing, such that these
systems are sustainable. Secondly, there is the challenge
of using this solution to create socio-techno-ecological
systems, by extending the non-human world with signalgenerating sensors, and enriching the human world with
intelligent software agents capable of both signal processing and participation in the deliberative processes - but
at certain levels of collective decision-making in social
systems, "participation" at such levels means "supportive
of," not "determinant of." In other words, there is an associated risk of, for example, surveillance capitalism [25].
There is a need to ensure our framework does not

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

57
IEEE Technology and Society Magazine - June 2020

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - June 2020
