Systems, Man & Cybernetics - April 2015 - 20

is also called new cybernetics [12], [13]. However, as noted
by Kenny, there is a need to define higher-order cybernetics [14]. The basic concern is that Type_001 cybernetics is
not enough to cover governance in a human society or in
a brain. For example, in a human society, the object is not
a single self but a collection of many individuals (human
beings). It seems that governance of human societies can
be classified as Type_001 because the subject is relatively
independent of the object, like a driver governs a car. However, since the object is not a single self, the governance
is much more difficult. Classifying the governance of a
society as Type_001 is an oversimplification. To make a
distinction, in this article, the governance of a society or
other distributed systems is called Type_011 cybernetics.

a distributed traffic network from outside using a distributed monitoring network belongs to Type_110, the
governance of an unmanned airplane from outside using
a computer network belongs to Type_100, and the governance of an unmanned spaceship from inside using an
onboard computer network belongs to Type_101 cybernetics. Note that so-called governor-in-the-loop cybernetics
generally contains not only Type_001 but also Type_011,
Type_101, and Type_111 cybernetics. However, we think
it is better to study these four types of cybernetics separately to make governance more efficient and effective.

Fundamental Properties of Governable
or Self-Governable Distributed Systems
Distributed systems are often very complex and difficult
to govern. For example, a human brain contains approxiType_111 Cybernetics
mately 1011 neurons, and each neuIn a human or an animal brain,
both the subject and the object are
ron is connected to roughly 10 4
distributed systems, and it is difneurons. If we allow all paramficult to separate the subject from
eters to change freely, the posDifferent societies
the object. We think it is necessible number of states of a brain
sary to make another distinction
can be larger than the number of
have different
by defining governance in a brain
observed stars in the universe.
ontologies because
as Type_111 cybernetics. Since
Governing this kind of complex
of different historical
the subject is at the same time the
system seems to be very difficult.
object, self-governance is a necesSurprisingly, however, human
backgrounds.
sity in Type_111 cybernetics. In a
brains and human societies are
brain, self-governance is an aggrevery simple [15], and it is possible
gated or collective behavior of
to govern them easily. To undermany neurons. Not all systems that
stand the reasons, we investigate
have the same physical structure as a brain are self-govsome fundamental properties of governable or self-governable. A system is self-governable only if certain condiernable systems.
tions are satisfied. Normally, only self-governable systems
are called cybernetic systems [5].
The Networked Structure
For a distributed system to be governable or self-governable, it must not be a simple collection of many individuOther Types of Cybernetics
al components. The components must be connected, and
Similarly, we can define Type_010, Type_100, Type_101,
the whole system must have a networked structure. For
and Type_110 cybernetics. For example, the governance
example, both brains and human societies are networks.
of a distributed power supply network from outside using
Each neuron in a brain is connected with other neua single server belongs to Type_010, the governance of
rons via dendrites, synapses, axons, and so on. Each
person in a huma n sociEnvironment
Subjective
ety is connected to other
Targets
people through different
Vehicle
A
Speed
social relations, via differObjective
Acceleration
Powertrain
System
Targets
ent media. The networked
S
Control
Chassis
System
structure should be a funSystem
B
damental property of govClimate Control
.
ernable or self-governable
HMI
Infotainment System
distributed systems.
It is interesting to note
Subjective Inputs (Behavior, Intentions, Emotional State, Wellness)
that the Internet has the
sa me physica l str uctu re
Cognitive Inputs (Cockpit Data)
as a brain. If we consider
Figure 1. A driver as a governor in the loop in a driver-vehicle system.
each terminal along with its
20

IEEE Systems, Man, & Cybernetics Magazine A pr il 2015

Table of Contents for the Digital Edition of Systems, Man & Cybernetics - April 2015