Professional Documents
Culture Documents
Mario Bunge - Ontology (II) - A World of Systems PDF
Mario Bunge - Ontology (II) - A World of Systems PDF
Volume 4
Ontology II:
A WORLD OF SYSTEMS
ACKNOWLEDGMENTS xv
I. SYSTEM
1. Basic Concepts 3
1.1. Aggregate and System 3
1.2. Concrete System: Definition 5
1.3. More of the Same 8
1.4. Subsystem 11
1.5. Level 13
1.6. Systems Association 14
1.7. Other Kinds of System: Property and Functional 15
1.8. Concluding Remarks 15
2. System Representations 16
2.1. Coupling Graphs and Matrices 16
2.2. The State Space Representation 20
3. Basic Assumptions 24
3.1. Structural Matters 24
3.2. Assembly and Emergence 27
3.3. Selection 30
3.4. Evolution 33
4. Systemicity 35
4.1. Integration, Cohesion, Coordination 35
4.2. Holism, Atomism, Systemism 39
5. Concluding Remarks 43
2. CHEMISM 45
1. Chemical System 46
1.1. Atom 46
viii CONTENTS OF ONTOLOGY II
1.2. Molecule 48
1.3. Chemical System 52
1.4. Chemical Reaction 54
1.5. Chemical Control 59
2. Biochemical System 61
2.1. Biomolecule 61
2.2. DNA Replication, and RNA and Protein Synthesis 67
2.3. Biochemical System 71
2.4. Concluding Remarks 73
3. LIFE 75
1. From Chemism to Life 76
1.1. Self-Organization of Biochemical Systems 76
1.2. Definition of a Biosystem 79
1.3. Cell and Organism, Biospecies and Biopopulation 82
1.4. Biolevel 84
1.5. Concluding Remarks 85
2. Biofunction 86
2.1. Health and Death 86
2.2. Function and Value 88
2.3. Biocontrol 92
2.4. Development 95
2.5. Reproduction and Heredity 98
2.6. Development, Heredity, and the Metaphysical Tradition 101
3. Evolution 103
3.1. Adaptation 103
3.2. Adaptation Mechanism 105
3.3. Evolution and Coevolution: Basic Ideas 107
3.4. A Schema of the Evolutionary Process 108
3.5. The Underlying Mechanism 111
3.6. Bioprogress 115
4. MIND 124
1. Central Nervous System 125
1.1. Philosophical Background 125
1.2. Neural Units 127
1.3. Initial Definitions 129
1.4. Basic Assumptions 132
5. SOCIETY 186
1. Human Society 187
1.1. Definitions 187
1.2. Subsystem, Supersystem. Specific Function 191
1.3. Institution 193
BIBLIOGRAPHY 292
This volume continues and concludes the task begun in Part I, titled The
Furniture of the World - namely the building of an exact and systematic
ontology consistent with contemporary science. However, it can be read
independently by anyone willing to take for granted the basic notions
analyzed and systematized in the companion volume, namely those of
substance, property, thing, possibility, change, space, and time.
The three main themes of this book are wholeness (or systemicity),
variety, and change. These three notions are analyzed and systematized,
and they occur in some of the main assumptions of our ontology. One of
these hypotheses is that the universe is not a heap of things but a thing
composed of interconnected things - i.e. a system. This supersystem is
composed of subsystems of various kinds: physical, biological, social,
etc. Only physical systems may be composed of things that are not them-
selves systems, such as elementary particles and field quanta. However,
even nonsystems are components of some system or other, and every sys-
tem but the universe is a subsystem of some system: there are no strays.
Ours is, in sum, a world of interconnected systems. Moreover it is the only
one.
Another postulate of this system of ontology is that concrete systems
are not all alike except insofar as they are systems and therefore tract-
able with the help of a unifying systems-theoretic framework. There are
different kinds of system and each is characterized by its own peculiar
properties and laws. Surely we sometimes succeed in accounting for the
emergence and the history of a system in terms of its composition, en-
vironment, and structure. Nevertheless, explaining need not be explaining
away: explained systems are not heaps, explained emergence is no mere
resultant, and explained novelty is not old. Systemicity, emergence, and
qualitative novelty and variety are as genuine as they are capable of being
explained. Far from being incompatible with reason, wholeness and
emergence can be understood.
A third major thesis of this work is that no system, except for the world
as a whole, lasts forever. Systems get assembled, change, and break down.
If natural, systems emerge as a result of self-assembly processes - often
from debris of former systems. Even modest accretion processes can ensue
in systems possessing emergent properties. Order can thus emerge from
Xlll
xiv PREFACE TO ONTOLOGY II
MARIO BUNGE
xv
SPECIAL SYMBOLS
XVI
CHAPTER 1
SYSTEM
I. BASIC CONCEPTS
If a thing acts upon another, and the latter does not react back, the former
is called the agent and the latter the patient. If neither action nor reaction
are nil, the things are said to interact. Finally, two things are connected (or
coupled, or linked, or bonded) if at least one of them acts on the other.
The bondage of a set A £ 6J of things is the set B A of bonds (or couplings
or links or connections) among them. So, the total set of relations among
the components of a complex entity may be decomposed into its bondage
BA and the set BA of non bonding relations.
We can now introduce the notion of the A-environment of a thing x
with A-composition ~ A(X). It will be defined as the set of all things, other
than those in ~ A(X), that act on or are acted upon the latter:
tS' A(X) = {y E 6J I ~(y E ~ A(X)) & (3z)(z c x & (y I> z V z I> y)}.
Finally the structure of a thing will be defined as the set of all the
relations among the thing's components as well as among these and the
things in the thing's environment.
We now have all we need to define the notion of a concrete system:
(a) (b)
Fig. 1.1. Two systems with the same composition but different structures and
environments.
(ii) the A-environment of 0' at time t is the set of all things of kind A, not
components of 0', that act or are acted on by components of 0' at t:
(iii) the A-structure (or organization) of 0' at time t is the set of relations,
in particular bonds, among the components of 0', and among them and the
things in the environment of 0', at t:
where lBA(O', t) is the set of bonding relations, and lBA(O', t) that of non-
bonding relations, defined on Ct? A(O', t) U {f A(O', t).
Example The simplest possible system is one composed of two connected
things, a and b, in an environment lumped into a single thing c. That is,
Ct?(0') = {a, b}, {f(0') = {c}. This system can have either of the following
internal structureses: a t> b, b t> a, or a 1><1 b: see Figure 1.2. (These are
the conceivable internal structures. But some of them may not be nomo-
logically possible, let alone technically feasible or even desirable.) As for
8 CHAPTER 1
Fig. 1.2. A two component system with three possible internal structures.
the external structures, they can be any of these or their unions: {a I> c},
{b I> c}, {c I> a}, {c I> b}.
An exhaustive knowledge of a system would comprise the following
items: (a) the composition, the environment and the structure of the sys-
tem; (b) the history of the system (particularly if this is a biosystem or a
sociosystem), and (c) the laws of the system. Such complete knowledge is
seldom attainable, particularly with reference to complex systems. But in
order to be able to speak of systems at all we should know at least their
composition, their enviroI\ment and their structure. Thus we can say that
the model constituted by the ordered triple
THEOREM 1.1 (i) There are concrete systems. (ii) Every thing is a com-
ponent of at least one system.
Surely the identification and modeling of a concrete system may be an
extremely difficult task. Thus it is not always clear what the composition,
SYSTEM 9
DEFINITION 1.3 Let ()' be a system with environment 6"(()" t). Then ()' is closed
at t iff 8'(()', t) = 0 - otherwise ()' is open.
Since every thing but the universe interacts with some other things, we
infer
COROLLARY 1.1 The universe is the only system closed at all times.
10 CHAPTER 1
This holds whether or not the universe turns out to be spatially infinite,
for the universe may be defined as that thing which has a void environ-
ment (i.e. which is self-contained).
So much for the concept of total closure. We need also the notion of
partial closure, or closure relative to a given property, since a system may
be open in some respects and closed in others. (Thus all systems are gravi-
tationally open, but some are electrically closed, others are closed to the
exchange of matter, still others to cultural influences, and so on.) We
make then
have a spatial configuration since they are made up of living beings that
stand in definite spatial relations to one another.) Every system has both
a system structure (or bondage) and a spatial structure (or configuration).
On the other hand aggregates or assemblages have spatial structures but
no system structures.
For ease of reference we collect some of the above elucidations in
1.4. Subsystem
DEFINITION 1.6 Let 0" be a system with composition Yf(O", t), environment
<ff(0", t) and structure Y(O", t) at time t. Then a thing x is a subsystem of 0"
at t, or x -< 0", iff
(i) x is a system at time t, and
(ii) 0'(x, t) f; (C(O", t) & <ff(x, t) ;:2 6'(0", t) & Y(x, t) f; Y(O", t).
By definition, the subsystem relation -< is an order relation, i.e. it is re-
flexive, asymmetric, and transitive. So, in particular, if O"t -< 0"2 and 0"2 -< 0"3,
then O"t -< 0"3' We shall make use of this property when defining the notion
of a system of nested systems (Definition 1.7).
Example 1 Factories, hospitals and schools constitute subsystems of
any modern society. On the other hand the persons composing them are
not themselves social systems: they are biosystems. Example 2 A foetus is
a subsystem of its mother; it becomes a system in its own right after birth:
before that it does not fall under any of the laws, natural or social, that
hold for independent systems.
Systems of different kinds have different compositions or different struc-
tures. (A difference in composition induces a structural difference but not
conversely, as shown by the existence of isomers, i.e. systems with the
same composition but different structures.) However, all systems of the
same genus seem to have the same ove~all structure or "general plan" -
12 CHAPTER I
DEFINITION 1.7. Let 0" be a system and call ~ the totality of systems, and
Fig. 1.3. An imaginary system of Chinese boxes or hierarchy of systems. The primary
structure, i.e. the amino acid sequence, is not shown. The secondary structure is the
helix, the tertiary the Z shape. And the quaternary structure is the way the individual
Z's are assembled together, i.e. the double staircase.
SYSTEM 13
1.5. Level
Talk of levels of organization (or complexity, integration, or evolution)
and of a hierarchy of such has been rampant in science, particularly in
biology, for the last half century or so. Unfortunately there is no consensus
on the significance of the terms 'level' and 'hierarchy', which are used in a
variety of ways and seldom if ever defined (Bunge, 1959b, 1959c). This
fuzziness must be blamed not only on scientists but also on philosophers -
on the inexact philosophers who despise clarity and on the exact ones who
are not aware of the philosophical problems raised by scientific research.
Let us attempt to remedy this situation by clarifying one concept of level,
and the corresponding one of hierarchy, that are widely used in contem-
porary science.
The intuitive idea is simple: the things at any given level are composed
of things belonging to preceding levels. Thus biospheres are composed of
ecosystems, which are composed of populations, which are composed of
organisms, which are composed of organs, which are composed of cells,
which are composed of organelles, which are composed of molecules,
which are composed of atoms, which are composed of so-called elementary
particles. One way of exactifying this notion is by means of
0;
o-i-o
0;
(a) ( b )
o
Fig. 1.4. (a) Before fusion, tt'(a,) = (a,), tt'(a,) = (a,). After fusion, tt'(a, + a,) = 0.
(b) Before merger, .'I'(a,) = (Linear link) and .'I'(a,) = (Triangular link). After merger,
.'I'(a, + a,) = (Pentagonal link) .
SYSTEM 15
DEFINITION 1.9 Let p(x) be the set of properties of a thing x, and n(x)
the set of processes occurring in x. Then
(i) the subset Po(x) c: p(x) is a property system of x iff every property
in poex) is lawfully related to at least one other property in Po(x);
(ii) the subset no(x) c: n(x) is a functional system of x iff every process
in no(x) is lawfully related to at least one other process in no(x).
Because in a system all properties and processes are lawfully interrelated
we conclude that, for every x, x is a concrete system iff p(x) is a property
system or n(x) a functional system.
2. SYSTEM REPRESENTATIONS
z 3 0 0
H 0 0 0 H
0 (Jj 1 0 1
0 1 0
z 0 1 0
011
-1
'0'
0 0 0 1
=
~
(J2
0 0 0
I -1 0 0 I:
4
0 0
11
Ii 0
0
~
(J3 =
2 110
d
(i) a set ofn directed graphs over the composition of (J, one for each kind
of connection, with a total of m nodes (vertices), in such a way that (a)
the nodes represent the components and (b) the edges represent the con-
nections; or
(ii) a set of n m x m matrices PM, where 1 :s; p :s; n, such that (a) the
matrix element PM rs of the pth matrix represents thc strength of the action
of component r, in the pth respect, upon component s, and (b) the matrix
element PM rr stands for the action of kind p of the rth component on
itself (feedback).
The off-diagonal elements PM rs , with r -# s, represent connections other
than self connections. There are m 2 - m = m(m - 1) such elements per
matrix, and a total of nm(m - 1) per system with n different kinds of
connection. This number is called the coupling capacity of the system.
So far we have represented the composition and the internal structure
of a system while neglecting its environment, hence its external structure.
An open system, i.e. one connected with its environment, can be repre-
18 CHAPTER 1
T =I ~:: ~:: II
with unknown Ti/S and perform the indicated operations:
F,
Fig. 1.5. The lawful state space SL(K) of systems of kind K is a subset of the cartesian
product of the ranges of the components of the state function. Only two such com-
ponents, F, and F •• are depicted here. F(a) = <F,(a), F.(a» is a (really) possible state
of a particular system of the given kind. As time 'goes by', the tip of F(a) moves about
within Sxf,K).
22 CHAPTER I
contributions of the latter are all on the same footing, we may take the
total state space to equal the union of the partial state spaces. In particular,
let SL(K) and SL(M) be the lawful states spaces of things of kinds K and
M respectively. Then the state space of the association k + m of two non-
interacting things of kinds K and M respectively, relative to the same re-
ference frame, is SL(K) U SL(M). Not so in the case of a system: here the
state of every component is determined, at least partly, by the states other
system components are in, so that the total state space is no longer the
union of the partial state spaces. Thus in Example 4 above the state space
of the two component system must be constructed ab initio rather than on
the sole basis of the state spaces for the individual biopopulations. In sum,
a thing is an aggregate if and only if its state space equals the union of the
state spaces of its components - otherwise it is a (concrete) system. (Cf.
Bunge, 1977a, 1977b.)
Every event occurs in or to some concrete thing and it consists in a
change of state of the thing - the change being merely quantitative, as in
the case of motion, or qualitative as well, as in the case of the coming into
being or the metamorphosis of a thing. A light flash, the dissociation of a
molecule, a storm, the growth of a bud, the learning of a trick, and the fall
of a government are events - or rather processes, for they are complex
and therefore analyzable into further events. Being changes in the states
of things, events and processes are representable as trajectories in the state
spaces of changing things. (An unchanging thing, if there were one, would
have a state space consisting of a single point.) Different trajectories in
a state space may have the same end points. That is, there are cases in
which one and the same net change can be effected along alternative
routes. See Figure 1.6.
The functions g and gf occurring in Figure 1.6 are not supposed to be
arbitrary: they must be lawful if we are to allow only lawful events and
discard lawless ones, i.e. miracles. In other words, the g occurring in the
event representation e = <s, Sf, g) must be compatible with the laws of
the system(s) concerned. Equivalently: a lawful event or process occurring
in a system of kind K, with end points s and Sf, is representable by a triple
<s, Sf, g), where g : SdK) ---> SL(K) is compatible with the laws of the
K's. If we disregard the intermediate states between the end points of
the processes, we are left with arrows or ordered pairs <s, Sf) E SL(K) x
SL(K). The collection of all such pairs of states, i.e. the set of all net events
(for a given g) constitutes the event space of the K's (for g). Symbol:
SYSTEM 23
F,
Fig. 1.6. Two different processes resulting in the same net change. The net change
from state s to state s' can be represented as the ordered pair (or arrow) <s, s'). Since
the change along curve g may be distinct from the change along curve g' ,p g, we
must represent the full events (or processes) by <s, s', g) and <s, s', g') respectively.
EdK) ~ SllJK) x SdK). In general the inclusion is proper: not all state
transitions are lawful.
Since we shall make ample use of the state space representation in this
work, we may as well compress the preceding into the following semantic
assumption. For each kind K of system possessing n properties there is a
property-representing function F: A ~ VI X V2 X ••• X Vn with n
components, called a state function for systems of the kind. Furthermore
(i) the totality of general properties of systems of kind K is representable
by the set of all the components (or coordinates) of F, i.e. p(K) = {F; I
l:s; i:s; n};
(ii) every particular property of a system of kind K is representable by
a value of a component of F, i.e. by Fi(a) for a E A and some 1 :s; i :s; n;
(iii) the state of systems of kind K at a E A is representable by the value
of F at a, i.e. s = F(a) :=: (FI(a), F2(a), ... , Fn(a»;
(iv) the collection of all such states of things of kind K, i.e. the range of
F, is the lawful state space of the K's, or Sl,(K) for short;
(v) every event occurring in a system of kind K is representable by an
ordered triple (s, s', g), where s, s' E Sl,(K) and g is a lawful map of SL(K)
into itself;
(vi) the collection of all really possible (lawful) events occurring in
systems of kind K is the event space of K, or El,(K) for short;
(vii) for a system in a given environment, and relative to a given re-
24 CHAPTER 1
ference frame, the state function takes often the form of a time-dependent
function F: T -+ Rn, where T £ R is the set of instants relative to the
given frame;
(viii) If F: T -+ Rn, then the totality of processes occurring in a system
x of kind K during the time interval 'r £ T is representable by the set of
states x is in during 'r:
n(x, 'r) = {F(t) I t E 'r};
(ix) the history of a system x of kind K representable by a state function
F : T -+ Rn, during the interval 'r £ T, is representable by the trajectory
hex) = {(t, F(t) I t E 'r};
(x) the total action (or effect) of a thing x on a thing y equals the dif-
ference between the forced trajectory and the free trajectory of the patient
y:
A(x, y) = hey I x) n hey)·
A detailed treatment of these concepts is given elsewhere (Vol. 3, Ch.
5). We shall presently use them to advance a handful of general principles
concerning systems.
3. BASIC ASSUMPTIONS
(i) the totality of inputs of (7 is the set of all the environmental actions on
(ii) the totality of outputs of (7 is the set of all the actions of the system
on its environment:
Our first hypothesis is that all systems receive inputs and are selective,
i.e. accept only a (small) subset of the totality of environmental actions
impinging on them. More precisely, we lay down
POSTULATE 1.2. Let V«(7) be the total output of a system (7. Then
(i) V«(7) =1= 0;
(ii) the (out-selection) function
DEFINITION 1.11 The function f that composes with the out-selection func-
tion ).! of a system to yield its in-selection function fI' i.e. such that fI =
fa).!, is called the transfer (or transducer) functionf: U(O') ---+ V(O') of 0':
Example The retina transforms (or maps or codes) light stimuli into
nervous signals.
We close this subsection with a batch of general principles that will be
stated in an informal way:
POSTULATE 1.3 Let 0' be an arbitrary system other than the universe. Then
(i) every input to 0' is an output of some other system (i.e. there are no
inputs out of the blue);
(ii) 0' receives inputs of several kinds (i.e. at some time or other every
one of the components of the state function of 0' is bound to be affected by
environmental changes);
(iii) for every action upon 0' there is a threshold below which 0' does not
respond;
(iv) the total input of 0' has a nonvanishing random component;
(v) there is a delay, however small, between every input and the cor-
responding output if any.
So much for our general structural assumptions. Let us now look at
systems from an evolutionary perspective.
SYSTEM 27
3.2. Assembly and Emergence
Any process whereby a system is formed out of is components is called
'assembly'; if the process is spontaenous, it is called 'self-assembly'. An
assembly process can occur either in a single step or, more likely, in a
number of steps: see Figure 1.7. We can express the idea formally with
the help of the concept of bondage, or set of bonds among the compo-
nents of a system, introduced in Sec. 1.2. Indeed we make
Unstructured
complex thing
o
6. D 6..
6. 0 0 o 0
o D
o DOD
t o 6. 0
6. o 0 SYSTEM
tJ
6.
0 + o
o
0
SUBSYSTEMS
Fig. 1.7. Assembly of a system out of previously unconnected units, either directly or
step-wise.
28 CHAPTER I
proceeds on its own - i.e. by virtue of its inner dynamics - once suitable
reactants and physical conditions have been supplied. For example, pro-
teins and even ribosome units will self-assemble in vitro in a matter of
minutes if suitable precursors and physical conditions are provided. In
this case man only lends a hand to nature, by reproducing conditions that
have or could have occurred spontaneously.
All chemical and biochemical syntheses are of course assembly processes
and moreover processes accompanied by the emergence of new properties.
However, self-assembly occurs on all levels in a variety of manners. Per-
haps the most conspicuous of all self-assembly process is the aggregation
or clumping of atoms and cosmic dust brought about by gravitation. This
is how planets and other celestial bodies are believed to have formed.
Moreover, far from always leading to disorganized lumps, gravitational
attraction can give rise to complex clustering on all astronomical scales, in
particular star clusters, galaxies, and galaxy clusters. (See, e.g. de Vau-
couleurs, (1970).) Similar examples of self-assembly by condensation of
units of the same kind are the formation of molecular complexes (such
as the haemocyanine polymolecule), of polymers such as the polypeptides
and of crystals out of solutions. (See Calvin (1969); Eigen (1971); Lehning-
er (1975).) Needless to say, self-assembly, and in particular self-organiza-
tion, occurs also on the social level: witness the formation of families,
bands, communities, and social organizations of various kinds. In sum
self-assembly and self-organization are not exclusive to life. What is pec-
uliar to biotic self-organization is that it happens to ensue in living sys-
tems rather than in biochemical systems or in systems of some other kind.
(In other words, it is an emergence mechanism leading to a new level of
organization.) We shall assume that all systems short of the entire
world have formed by assembly:
supersystem was the result of some assembly process. Remark 4 Our axiom
is far from obvious and it could not have been formulated before systems-
theoretic thinking became pervasive, if only because the general concept
of a system had not been elucidated until recently. Remark 5 Our axiom is
of particular relevance to the problem of the origin of biomolecules and
biosystems. Until recently both were assumed either to have been created
by divine fiat or to have existed from all eternity. Darwin's Origin of
Species appeared only in 1859, and scientific research into the evolution
of molecules started only one century later.
The components of a self-assembled system are called its precursors, an
apt name suggesting that the system had not always been around but
emerged from pre-existing things. Interestingly enough, (a) the precursors
of a system are not blended but keep their individuality to some extent,
and yet (b) they give rise to a thing possessing emergent properties. For
example, the two hydrogen atoms that combine into a hydrogen molecule
are distinct components of the latter, yet the spectrum of the molecule
is radically different from that of its components. The first characteristic
is formalized by saying that the composition of the self-assembled system
equals the set of parts of its ancestors - i.e. the set of its precursors. And
the notion of emergence can be elucidated as follows.
Call x a thing and t E T an instant of time, and introduce a function p
that assigns the ordered couple (x, t) the set p(x, t) of all the properties of
x at t. That is, p is a function p: exT ---> 2IP, where e is the set of all things,
T the set of all instants, and 2iP the family of subsets of the set P of all
general properties of things. A change in thing x can be viewed as a certain
change of state of x. Since x is held fixed throughout that change of state,
we can use the simpler function
Px: T ---> 2IP such that pit) = p(x, t).
In short, pit) is the collection of properties of thing x at time t. (For details
onpx, see Vol. 3, Ch. 2.)
Now let t and t' be two distinct instants, such that t precedes t'. The cor-
responding values of Px are of course Px(t) and Px(t' ). If these two sets of
properties of x are the same, then the thing has not changed qualitatively.
If they are different, then the thing has gained or lost some properties. If
the latter is the case, the newly gained properties will be said to be emer-
gent relative to the given thing, even though they may be possessed by
other things as well. In sum, we make
30 CHAPTER 1
DEFINITION 1.13 Let x be a thing with properties pit) at time t, and pro-
perties pit') at a later time t' > t. Then
(i) the total qualitative novelty occurring in x during the interval [t, t'] is
the symmetric difference
nit, t') = pit) 6 U
/«:5,/'
pi,.);
(ii) the emergent properties that appear in x during the interval [t, t']
are those in
3.3. Selection
New systems are being formed all the time, but not all of them are viable
in the environment in which they emerge. In fact many are unfit, hence
ephemeral, and this either because they are internally unstable or because
they cannot cope with environmental aggression. In the latter case we have
SYSTEM 31
I AEI IAEI
peS, E) ""'1ST = 1 - TST'
where 'lXI' designates the numerosity of the set X.
Clearly, the p values range between 0 (maximal fitness) and 1 (maximal
unfitness).
Our promised hypothesis reads thus:
lal (bl
Fig. 1.B. The selective action of E followed by E' (diagram (a» differs from the result-
ing selective action in the reverse order (diagram (b».
SYSTEM 33
sembled from units supplied by the environment: the latter provides the
opportunity for self-assembly, hence for emergence. In sum, the environ-
ment of any system is creative - only, it is selective and exclusive rather
than permissive.
3.4. Evolution
Self-assembly may result in evolution - and this at the molecular, the
organismic, the populational, and other levels. Let us then examine the
general concept of evolution. To this end let us begin by elucidating
the general concept of descent:
t = 0, n = 2°
t =) n = )' t = 2, n = 22
t= 3 • n =1' t = 3, n = 23
4. SYSTEMICITY
where the aij are in general complex numbers. The degree of integration or
cohesion of the system can then be defined as
11' =
la121 + la211
lall + a121 la21 + azzl'
If the aij are not constant but time dependent, w itself will depend on time.
In general, w: T -> [0,1].
If a12 = a21 = 0, the components do not form a system; in every other
case they do. In particular, if component 1 controls component 2, then
° °
a12 = and aZl #- 0, whence < w :-:;; 1. And if there is symmetric interac-
tion, a12 = a21 #- 0. Finally, if all the aij equal unity, w = 1, i.e. the system
is maximally cohesive.
Assume then that it is possible to define in each case a measure w: T ->
[0, 1] of the degree of integration of systems of any given kind. Then by
charting the course of w values we can follow the history of the system
from buildup to breakdown through its stable stage if any. In other words,
we can introduce
POSTULATE 1.7 For every kind of system there is an optimal size, i.e. a
number of components that maximizes the degree of integration (cohesion)
of the system in the given environment. That number is called the critical
size.
An immediate consequence of this assumption is that, for every kind of
system, there is (a) a threshold size, i.e. a number of components below'
SYSTEM 37
oL-------------~--------------~
No N
Fig. 1.9. Degree of integration (cohesion) per number of components vs. total number
of components. No = optimal or critical size.
which the aggregate does not form a system, and (b) a maximum size, i.e. a
number above which the system breaks down.
Remark 1 Unlike most other axioms of our theory, Postulate 1.7 is em-
pirically testable. For one thing it might well be that for systems of certain
kinds there is more than a single critical size. If this were the case we would
have to make a small adjustment in Postulate 1.7. Remark 2 Accretion by
gravitational puIl would seem to refute our postulate. It does not, for the
same process increases the density of matter and radiation energy, and this
eventually sets up nuclear reactions that may lead to either explosion or
collapse.
The preceding axiom concerns the overaIl integrity or cohesion or a
system without regard for the integration of its subsystems. If a system
does have subsystems, not just components, then the cohesion of the sub-
systems competes with that of the overall system. Example 1 A molecule
with all of its valences saturated is stable, hence it coheres poorly with
similar molecules, e.g. its fellow components in a polymer. In general, the
greater the complexity of a molecule the lower its overaIl binding energy.
Example 2 A large human family has weaker interactions with the rest of
society, per family member, than a small family; the members of a large
family spend a larger amount of energy interacting among themselves than
with their social environment. Hence while they fare well in cases of dis-
aster they may not make good citizens.
We compress these observations into
38 CHAPTER 1
POSTULATE 1.8 The more cohesive each subsystem the less cohesive the
total system.
A problem for the systems designer, be he an engineer or an applied
social scientist, is to hit on a structure that will maximize overall integrity.
He cannot maximize the cohesion of each subsystem because then it be-
comes self-sufficient instead of serving the whole. And he cannot minimize
the partial integrities because then the subsystems would become unstable
(unreliable). A compromise solution is to choose medium cohesion sub-
systems and to have more than one perform a given function or role. Such
design enhances the reliability of the system regardless of its nature. See
Figure 1.10.
Finally, another concept relevant to that of systemicity is the notion
of coordination, which must be distinguished from that of integration. If
integration fails the system undergoes structural breakdown. On the other
hand coordination concerns the relation among either components or
functions resulting in functional maintenance. If coordination fails the
system undergoes functional breakdown. There can be integration without
coordination but not conversely. A complex machine out of kilter is inte-
grated but not coordinated. On the other hand organisms are coordinated
and afortiori integrated as long as they live. A possible characterization of
the concept of coordination is given by
Reliability = R = P(a)·P(b)
Probability of failure = 1 - R.
Fig. 1.10. Increasing overall reliability (or degree of integration), by increasing redun-
dancy, i.e. the number of subsystems performing the same function. Pea) is the prob-
ability that component a will perform its regular function(s).
SYSTEM 39
but also a host of hypotheses and data concerning the structure of water
(i.e. the relative configuration of H 20 molecules in the lattice) as well as
hypotheses and data about the dynamics of water bodies - hypotheses
and. data which vary of course according to whether water is in the
gaseous, liquid, or solid phase. In sum, to describe, explain or predict
the properties of water we use both microlaws and macrolaws.
Atomism, an ontological doctrine, is usually, though not necessarily,
allied with reductionism, the epistemological doctrine according to which
the study of a system is reducible to the study of its components. (The con-
verse is false: one can be an epistemological reductionist, yet acknowledge
wholes, emergence, and levels.) The reductionist will claim of course that
we may use macrolaws and, in general, laws of systems, as a convenience,
though in principle we should be able to get along with microlaws (or laws
of components) alone, since the former are reducible to (deducible from)
the latter. This thesis contains a grain of truth but is not the whole truth.
No theory T2 of water as a body follows solely from a microphysical theory
T j of the H 20 molecule - not even by our adjoining what some philoso-
phers call the bridge laws relating macrophysical concepts (e.g. pressure)
to microphysical ones (e.g. molecular impact). Much more than this must
added to the primary or reducing theory T j in order to obtain the second-
ary or reduced theory, namely hypotheses concerning the interactions
among the system components.
The extreme case of reduction is that of straightforward deduction from
a given set of premises, or strong reduction. Examples: reduction of
particle mechanics to continuum mechanics, and of ray optics to wave
optics. (The converse reductions are impossible.) These are rather excep-
tional cases. In general we must resort to a more complex strategy, namely
weak reduction, or deduction from a primary theory in conjuction with a
set of conjectures and data congenial with but alien to the former. The
structure of this inference is:
Tj U Subsidiary hypotheses and data concerning interactions
among components ~ T 2 .
The subsidiary hypotheses constitute a model of the composition and
structure of the system. Since this model, though couched in the language
of Tj, is not included in Tj, we are in the presence not of straight (or
strong) but of partial (or weak) reduction. (Cf. Bunge, 1977f.)
Note that we are not stating that the properties of water, or of any other
macro system, are mysterious. On the contrary, they can be explained at
SYSTEM 43
least in outline. For example, the exceptionally high boiling point and
evaporation heat of water are explainable in terms of the hydrogen bonds
linking together all the H 20 molecules in a water body, bonds which are
in turn explained by the composition and the structure of the H 20 mo-
lecule. But the point is that the intermolecular hydrogen bonds do not
occur in the study of the individual H 20 molecule. In other words,
although water is composed of H 20 molecules it does not reduce to
H 20 - pace the efforts of enlightened philosophers to crush the holistic
monster (e.g. Kemeny and Oppenheim, 1956; Oppenheim and Putnam,
1958, Putnam, 1969).
In short, atomism is almost as false as holism, the difference lying in
that, whereas the former stimulates research, the latter blocks it. Each of
these views has a grain of truth that the systemic view preserves and
expands:
5. CON'eLUDING REMARKS
The idea of a system, as distinct from a loose aggregate, is very old. Yet
it has only recently been elucidated and exploited systematically. The mere
suggestion that the thing we happen to be looking at, manipulating, or in-
vestigating might be a system rather than an unstructured object or a mere
heap, will guide our study and handling of it. Indeed, if we suspect that a
certain thing is a system then we shall endeavor to identify its composition,
its environment, and its structure.
The order in which the three coordinates of the system concept appear
is natural rather than accidental. Indeed, listing a system's components
must precede any question regarding its environment and its structure; and
identification of the environment comes before display of the structure,
because the latter is the collection of relations among the components and
among these and environmental items. True, when coming across certain
systems, such as a plant, a watch, or a galaxy, we often begin our search
with the totality and its environment, ending up by disclosing its composi-
tion and its structure. But when investigating a forest, a social system, and
a fortiori a social supersystem such as a nation, we first meet its compo-
nents (or atomic parts) in their milieu and try to figure out the structure of
the whole by studying the behavior of the individual components. In either
case, i.e. whatever our mode of perception may be, the conceptual analysis
of a system must proceed in the way indicated - identification of compo-
sition, environment, and structure - if only for mathematical reasons.
44 CHAPTER I
CHEMISM
All concrete systems are physical entities, in that they possess energy-
a measure of changeability (Vol. 3, Postulate 5.3). However, some concrete
systems are more than just physical entities, in the sense that the categories
of physics are insufficient to account for them. For example, some are
social systems. And although some of the concepts of physics - in partic-
ular the notion of energy - are necessary to account for any society, most
are unnecessary. On the other hand social science requires new concepts,
such as those of artifact, social group, justice, and culture. The peculiar
concepts of social science are supposed to represent emergent properties
of a society, i.e. properties that are not possessed by its components.
Likewise the study of the components of any society, i.e. social animals,
calls for a number of distinctly biological categories, such as reproduction,
sex, and death. And if the animals happen to have a highly evolved nerv-
ous system, we shall need still other categories, such as the concepts of
altruism and planning, which are usually regarded as psychological.
Finally, organisms are composed of chemical systems, which are physical
entities of a very special kind, namely composed of atoms and molecules
engaged in chemical reactions.
Thus the totality of concrete entities may be grouped into five genera:
the set of physical things (e.g. atoms and fields), the set of chemical sys-
tems (e.g. protein-synthesizing units), the set ofbiosystems (organisms and
their organs), the set of sociosystems (societies and their subsystems and
supersystems), and the set of artificial things. (The set of psychosystems,
or animals endowed with mental abilities, is assumed to be a subset of that
of biosystems.) The first three genera precede the other two in the sense
that they supply the components of the higher order things or take part
in some other way in their assembly. Since the lower a genus the more
populous it is, we may depict the structure of reality as a pyramid: see
Figure 2.1. This suggests the plan of this volume. (More on this structure
in Ch. 6.)
In this chapter we shall hardly touch on physical systems: first, because
they are the best understood of all; second, because we have actually
dealt with them in Vol. 3. We shall focus instead on certain key chemical
categories of crucial importance in the study of life, such as those of che-
45
46 CHAPTER 2
BlOCH EM 0 SYSTEMS
CHEMOSYSTEMS
PHYSICAL THINGS
Fig. 2.1. The five main genera of things constituting the furniture of the world.
mical reaction and chemosystem, protein and nucleic acid. These cate-
gories are certainly more specific (less universal) than those of thing, pro-
perty, system, and others we have encountered so far. On the other hand
our chemical categories are far more universal than some of those that
have dominated ontology for centuries, such as those of mind and purpose.
1. CHEMICAL SYSTEM
1.1. Atom
Since we shall make ample use of the concept of a chemical system, we
must characterize it, the more so since no definition of it in systems theore-
tic terms seems to be available. And, since a chemical system is one whose
components are reacting chemicals, we must start by characterizing the
notions of a chemical and of a reaction.
The components of a chemical system are chemical atoms or molecules.
A detailed characterization of things of these kinds calls for a number of
specific scientific concepts. Luckily atoms have an essential or root pro-
perty from which many another depends, namely atomic number. And
molecules have an essential property of their own, namely atomic com-
position. These two properties will suffice for our limited purposes.
Our starting point is then the notion of chemical atom of a kind, or
elementary chemical species, or element for short. There are 100 odd ele-
ments, among them hydrogen (H), carbon (C), nitrogen (N), and oxygen
(0), the atomic numbers of which are 1, 6, 7, and 8 respectively. Every
member of such a natural kind or species is a material system formed by at
least two things of different genera, namely a system of nucleons (protons
or neutrons), and an aggregate of electrons. In turn the nucleons are prob-
ably composed of subunits (quarks?), and the atoms can join forming
molecules. See Figure 2.2. (This diagram suggests that molecules and their
CHEMISM 47
r
Molecule
Atom Atom
I
~,
~Electmn
C
!
Fig. 2.2. A molecule is composed of two or more atoms. Each of these is in turn com-
posed of a nucleus and one or more electrons. The number of protons in the nucleus
equals the atomic number of the atom. Nucleons are possibly held together in the
nucleus by meson fields. In turn every nucleon (proton or neutron) is presumably com-
posed of smaller units (quarks?). Altogether six different kinds of thing are involved in
a molecule.
components constitute a sup semi lattice the order of which is that of the
part-whole relation c. Thus if al and az are two atoms forming a molecule
m, then sup {ah az} = m. However, we shall not pursue this point here.)
In short, we steal from physics and chemistry
d = AI ~ z = {n A c A \1 S n S N}
n nA =
N
A = UnA, and mA 0 for m =1= 11.
n=l
In other words, the totality of atoms is the disjoint union of all the atomic
kinds. The latter are those displayed in the periodic table. It is not known
whether the number of atoms of each atomic kind is finite.
So much for the (atomic) components of a molecule.
1.2. Molecule
A molecule is a system composed of atoms - or, if preferred, of nuclei
and electrons. Molecules range from hydrogen or Hz (molecular weight
= 2) to deoxyribonucleic acids or DNA (molecular weight up to 1 billion).
There are about 4 million known species of molecules, 96 percent of which
contain carbon. The combinations of atoms to form molecules are brought
about by either ionic bonds or covalent bonds. In turn molecules stick
together by virtue of hydrogen bonds or van der Waals forces (e.g., hydro-
phobic forces). All of these bonds are ultimately electric forces and they
are studied by quantum chemistry, an application of quantum mechanics.
In ontology we take such bonds for granted and let chemists inquire into
their nature. (Remember the fate of Hegel's alleged refutation of Ber-
zelius' brilliant hypothesis that chemical bonds are electrical: cf. Hegel
(1830), Sec. 330). The metaphysician accepts gratefully what the chemist
is willing to teach him about the spontaneous formation (self-assembly)
of molecules out of atoms, or of other molecules, as well as about the
breakdown or dissociation of such systems as a result of their interaction
with their environment (in particular thermal motion). The philosopher's
task is not to compete with the chemist but to try and discover the struc-
ture and peculiarities of chemical processes vis a vis other kinds of process.
If in the course of his work he comes up with results helping to clarify the
foundations of chemistry, so much the better.
The statement that all molecules are composed of atoms can be exact i-
CHEMISM 49
fied in a number of ways. We shall proceed as follows. Consider for ex-
ample the first atomic species or
IA = H = The set of all hydrogen atoms = {x Eel Z(x) = I}.
This is the set of all conceivable molecules, in turn a superset of the set of
all really possible molecules.
The structure of A* is obviously that of a semigroup under the binary
operation . of combination or concatenation. More exactly, the structure
<A*, .) is the free semigroup generated by the set A of atoms, and this
because every element of A* is either a member of A or a string of mem-
bers of A. Now, as noted before, A* contains not only the really possible
molecules but also the unlawful ones. If we shrink A* to the subset A** of
really possible molecules, we face the difficulty that A** is not closed under
the operation· of combination or concatenation (i.e. it does not contain
all the concatenates). Equivalently: . is a partial operation on A**, i.e. one
not defined for every pair of members of A**. We shall therefore say that
the system $'** = <A**, .) is a partial semigroup embedded in the free
semigroup $'* = <A*, .) generated by A. In other words, we make
POSTULATE 2.2 Let .r;f* = <A*, .) be the free semigroup generated by the
set A of all atoms of all kinds (Postulate 2.1), and call
A** = {z E G I Z = xj'" xrz ... xf!JN & Xi E 'A & l s i s N
& 0 s mi < 00 & Z is lawful}.
Then
(i) the partial semigroup $'** = <A**, . ) is embedded in $'*;
(ii) for any two things X and y in A**, if xy is in A**, then xy represents
the combination of x with y (or the xy compound).
Note that we are not postulating that $'** is a commutative semigroup.
Hence the isomers xyz and yxz will be distinct molecules.
The previous axiom supports the following conventions that will come
in handy in the sequel.
CHEMISM 51
DEFINITION 2.2 Let d** be the set defined in Postulate 2.2 Then
(i) the set of compounds is the complement of A** to A, i.e.
M = A** - A = A** n A;
(ii) any subset of M, of the form
M; = {z Eel z = X\"l X~2 ... xF/N & Xz E'A & 1 sis N},
with fixed values of the mz for 0 s mz < 00, is a molecular species;
(iii) every member of a molecular species is a molecule.
(Note the strategy of first defining a set and then naming its members.)
DEFINlTlON 2.5 Let A** be the totality of compounds and I: A** --* N
a function assigning each atom or molecule a natural number. Then the
length of a member t of A** equals the number of atoms in z. More pre-
cisely,
Ifz == xrl.x~12 ... X~N E A**, then I(z) = ml + m2 + ... + mN'
this concept of length applies to molecules of all shapes, not just the
linear ones. Thus the length of the planar molecule CH 4 is, according to
our definition, equal to 5.
So much for structural matters. (For somewhat different treatments see
Rosen (1959) and Chiaraviglio (1965).) Let us now turn to chemical
processes.
i)~liINltION 2.6 Let The the time span (relative to a given reference frame)
and let
F"" (F1;F2' ... ,Fn>: T~ Rn
be a state function for an arbitrary system rJ of a certain kind, satisfying
the laws in a set L. Further, let the first m < ncoordinates of F {at a given
time) equal the number of comp'Onents of each kind of chemical in the
system (at that time). Then
CHEMISM 53
(ii) the value of the composition state function at time t E T, i.e, F cCt),
is called the instantaneous composition of (j;
(iii) the set of all possible instantaneous compositions of fl, i.e.
Se«(j) = {FcCt) E JRn It E T & Fe satisfies L},
is called the composition state space of (j.
Example The composition state function of a system where water is
formed out of hydrogen and oxygen, according to the reaction 2Hz + O2
-" 2H zO, has
the initial value Fe(t,) = (2, 1,0)
and (on completion of the reaction)
thejinalvalueFe(tj) = <0,0,2),
where the integers are the numbers of molecules of the chemicals in the
system. If we wish to represent the entire process from start to end, rather
than only the initial and the final states, we can use the partial concen·
trations, to obtain
Fe = <CH2 , CO2 , CH20 ),
where each coordinate of the triple is a function from T to R. A condition
for being able to use these functions is that the system be rather bulky,
as is tacitly assumed in chemical kinetics.
Finally we can make
course of 'r, and this change is not just a transport from or to environment
of 0'.
According to this definition there can be talk of a chemical system only
where and while chemical reactions are in progress. Thus a dead battery is
no longer a chemical system. Before the reactions start, and after they have
reached completion, the system is a physical one and only potentially a
chemical system. (A chemical system is then an essentially dynamical
system that can be at most in a state of dynamical equilibrium,. or steady
state, which is the one at which every reaction proceeds at the same pace as
its converse, so that there is no net variation of the composition of the
system.) This suggests
DEFINITION 2.8 A concrete system 0' is a physical system during the period
'riff the composition state function of 0' is constant (e.g. zero) throughout 'r
or, if it changes, it does so by transport from or to the environment of 0'.
In sum a chemical system is a system whose chemical composition
changes in the course of time, however slowly, and does so not just by
importation or exportation of chemicals, but by virtue of processes in-
volving its components. These processes are of course the chemical reac-
tions. Let us look into these.
[p]
II ,/11
/>1
I:i
1// [B]
f :,
START
.rAJ
Fig. 2.3. Composition state space of a chemical system in which only reactions of the
form A + B ~ P occur. [X] designates the concentration of X. Each line within the
conoid, from beginning to end, represents one possible complete reaction compatible
with the reaction schema as well as with the formula for the rate of formation of P from
A and B.
(i) the full trajectory of the state point F(t) in SdK) is called the total
history of (j during 7: ~ T:
we get
Cst
~~~:r.A~~---C:
Fig. 2.4. The reactions A ~ Band B ~ C add up to the overall reaction A ~ C. Each
reaction is represented by an arrow in the cartesian three-space of the concentrations
(or the numbers of atoms or molecules).
58 CHAPTER 2
THEOREM 2.1 Let R be the set of all chemical reactions that can occur in
a given chemical system, + the binary operation of reaction addition,
- the unary operation of reaction inversion, and 0 the null reaction. Then
the structure (Ji = <R, +, -, 0) is a group.
Now, whereas some reactions are concurrent, others are successive.
The notion of reaction succession is obvious:
two members of R. Then 'i p,ecedes rj in time iff ' j sets on after 'i' I.e.
CHEMISM 59
where the last two reaction constants are comparable, i.e. k3 ~ k;. We
leave the reader the task of representing the six reactions in our state space
CHEMISM 61
=-E,--;:5--..;>------1 E 5 1------<;>------>
E,P
~---j E P /(-------'
Fig. 2.5. A self-controlled chemical system of the negative feedback type. The reaction
product P combines with the catalyzer E forming the ·ntermediate compound EP and
thus slows down the left-to-right reactions.
2. BIOCHEMICAL SYSTEM
2.1. Biomolecule
We shall call biomolecule any macromolecule that, like a protein or a
nucleic acid, is essential to some typically biological process such as cell
62 CHAPTER 2
The cell
1
Nucleus
Organelles Mitochondria
Chloroplasts
Macromolecules
/
Nucleic
1
Proteins
1
Poly-
~ Lipids
Mol wt 10"-10" acids saccharides
1
Mono-
1 1 1
Building blocks Amino Simple Fatty
Mol wt nucleotides acids sugars acids,
100-350 glycerol
1 1
Intermediates Ribose, a-Keto Phospho- Acetate,
Mol wt carbamyl acids pyruvate. malonate
50-250 phosphate malate
Precursors from
~I/ CO2
the environment H2O
Mol wt 18-44 N2
Fig. 2.6. Build-up processes that start with simple molecules in the environment and
end up in living cells. 'Mol wt' is short for 'molecular weight'. From Lehni~ger (1975).
64 CHAPTER 2
In short, there are two sets of units (monomers) from which the polynu-
cleotides or nucleic acids are formed:
Ribonucleotides R = aUg U c U u
Deoxyribonucleotides D = a' U g' U c' U t'.
We proceed to characterize the nucleic acids and in the first place the
members of the DNA family. We start with the building blocks in the set
D = a' U g' U c' U t' and form the set of all possible strings or sequences
of members of D, with the sole restriction that the composition of each
compound be equal to the union of the compositions of its constitutents or
precursors. I.e. if x and yare in D and combine to form z = xy, then
re A(Z) = re A{X) U re A(y), where're A(U)' stands for the atomic composition
of u. (Cf. Chap. 1, Sec.1.2.) The new set D* formed from members of D'
contains all the conceivable combinations, with repetitions and permuta-
tions, of deoxyribonucleotides. In particular D* contains all the polynu-
cleotides forming the DNA molecules as well as monotonous strings such
as xn = xxx ... x with x E 0'. We shall restrict our attention to those mole-
cules that contain all four nuc1eotides, i.e. members of a', g', c', and t'.
Call Dt the subset of D* formed by the strings containing all four
deoxyribonuc1eotides in all conceivable numbers and locations, such as
wxyz, zxwyx, yywxxyzzz, where w E a', x E g', Y E c', and Z E t'. The set
Dt contains all the components of all the DNA molecules we need to
consider - both the actual and the merely possible ones. The structure of
Dt is that of a semigroup. More precisely, ~4 = (Dt, .) is a subsemi-
group of the free semigroup ~ = (D*, .) generated by D.
DNA molecules are double strands of things of kind Dt. These double
strands are coupled in such a way that every A base in one of the strands
is paired off to a Tbase in the other, and likewise every C base is paired off
to a G base in the other. Hence we make
CHEMISM 65
POSTULATE 2.4 Let!?) = <D*, . >be the free semigroup generated by the
set D = a' U g' U c' U [' of deoxyribonucleotides under the operation
. of chemical combination, subject to the condition that, for any x and
y in D*, if z = xy, then ~D(Z) = ~D(X) U ~D(Y)' Further, call !?)4 =
>
<Dt" the sub semi group formed by those sequences that contain mole-
cules of all four kinds in D. There are systems composed of two chains
of molecules in Dt, whose structure includes the bijection p : D ---> D such
that for any XA E a', XG E g', Xc E c', and XT E [',
POSTULATE >
2.5 Let fJi = <R*, . be the free semigroup generated by the
set R = aUg U c U u of ribonucleotides, and call fJi 4 = <RI, . the >
subsemigroup formed by those sequences that contain molecules of all
four kinds in R. Every member of R4 is a possible RNA molecule.
The union of all kinds of DNA and RNA molecules is the totality of
nucleic acids:
N = DI URI.
The case of proteins is similar. Here the building blocks are the amino
acids or rather some of them, such as glycine or NHzCHzCOOH. More
precisely, the protein amino acids form the genus
AA = Ala U Arg U Asn U Asp U Cys U GIn U Glu UGly U His U lie
U Leu U Lys U Met U Phe U Pro USer U Thr U Trp U Tyr U Val.
66 CHAPTER 2
This set has the semi group property under the operation . of chemical
combination. Also: the structure dd = <AA*, .), where AA* is the set
of all possible strings or sequences of members of AA, is the free semi-
group generated by AA. The members of this group are called polypeptides
and the longer among them proteins. More often than not a protein con-
tains at least one member of every species included in AA. However, there
are important exceptions to this rule. Therefore we cannot restrict our
attention to the sub semi group formed by those sequences of amino acids
that contain representatives of all 20 kinds of protein amino acids.
We summarize the foregoing in
p: Dt x 2E --> Dt x Dt
such that p(d, e) = <d', dl/), with d, d', dl/ EDt, e E 2 E , and ceD(d) ,
ceDed'), ceD(dl/) E 2E.
As mentioned above, DNA molecules control their own reproduction,
i.e. are their own templates. But they are not the only molecular systems
acting as templates: RNA molecules template for the synthesis of proteins.
We must therefore elucidate the notion of a template. See Figure 2.7.
Fig. 2.7. System" is a O'-template, i.e. " provides a sort of mold for the assembly (e.g.
synthesis) of 0' out of environmental entities.
o
CT
OUPlICATION
Dupliclition
DNA
..~i? / \
Transcriptioll
/ / \
4' / \
RNA i? / \
..",'<' / \
1
PROTEIN
Transl ..uion
"
'. .. _;'
RNA
/
/
----::::-:-::::::-=-::-4) pROTEIN
TRANSLATION
\
\.i
la) (b)
Fig.2.S. Two versions of the "central dogma". (a) The standard version (e.g. Watson.
1976). (b) An alternative version (Crick, 1970). For a translation of the metaphorical
expression 'information flow' into the nonmetaphoricallanguage of quantum chemistry,
wait a few decades.
CHEMISM 69
a E AA and r E Rt, is read 'r codes/or a'. However, it has been found ex-
perimentally that different co dons can "code for" one and the same amino
acid. (I.e. the "code" is degenerate or many-one, hence not a proper code
allowing one to break any ciphered molecular message.) In other words,
no such function/from AA to Rt exists, so that the converse "translation"
of protein structure into RNA structures is impossible.
A genuine code is a certain correspondence between two sets of artificial
signs and, more particularly, languages (cf. Birkhoff and Bartee, Ch. 8).
The coding and decoding functions of errorless codes are one to one, not
many to one like 7r:. Therefore the use of the linguistic concept of code in
chemistry is metaphorical. The same holds for the expressions 'the tem-
plates instruct (or code for) protein synthesis' and 'the genetic instruc-
tions contained in a gene'. That these are analogies is shown by the fol-
lowing facts. First, a genuine information system is one composed of a
transmitter, a receiver, and an information channel between them; no such
components are apparent in a chemical system. Second, the computations
of the quantity of information allegedly locked in a chunk of genetic mate-
rial (i.e. a segment of a DNA molecule) are phoney - so much so that they
occur in no biological law statements and everyone gives his own favorite
arbitrary estimate. For these reasons the information-theoretic language
employed in molecular biology is a heuristic prop that has now exhausted
its heuristic power. Moreover it is a stumbling block on the road to trans-
forming the "central dogma" into a hypothetical-deductive system explain-
ing the synthesis and the duplication of nucleic acids, as well as the syn-
thesis of proteins, exclusively in physical and chemical terms instead of
describing them in analogical terms such as 'informational molecule',
'information flow', 'transcription', and 'translation'.
Note finally that, like the previous mappings introduced in this sec-
tion - namely p (DNA replication), SR (RNA synthesis) and Sp (protein
synthesis) - 7r: is not defined by Postulate 2.10, which characterizes a whole
class of such functions. However, unlike p, SR and Sp, 7r: is known in detail:
indeed it is given by the table that summarizes the so-called genetic code.
What is still unknown is the origin of the "code". Nor is it known whether
nucleic acids have always been "used" to synthesize proteins.
it is assumed that the precursors of the first organisms, on the Earth and
probably elsewhere, were extracellular biochemical systems, and that
the first cells emerged by self-assembly from, or the merger of, some such
biochemical systems. And it is sometimes conjectured that the immediate
precursors of cells may have been systems containing both proteins and
nucleic acids. Such systems, though not possessing all the properties of
contemporary cells, would have their main emergent properties, namely
the abilities to metabolize, self-reproduce, and mutate. (Cf. Kiippers,
1975.) However, the problem of the origin of life will be taken up in the
next chapter.
A iI i I I
B ~i ~
E A+B I C I A+B+Ei C+E
~I
~
E C I I
I
I
j
(b) (c )
Fig. 2.10. Three kinds of system: (a) physical system with four non-reacting compo-
nents; (b) chemical system with one nonreactive component (E); (c) biochemical system:
all components react, and one of them (E) is an enzyme or proteinic catalyzer. Whereas
physical and chemical systems may be almost completely closed (full line) or open only
in some respects, biochemical systems must be semiopen (dashed line), for they prey on
their environment.
74 CHAPTER 2
LIFE
Life has long been a mystery on which mystery mongers have thrived. It is
no longer such: biologists are beginning to understand what makes organ-
isms tick, and physicians to manipulate the control systems responsible
for that ticking - which, incidentally, is not like a clock's. This understand-
ing could not be forthcoming as long as organisms were studied either on
their own level exclusively (holistic approach) or as physical systems de-
void of emergent properties (levelism), and in either case apart from their
history. The theory of evolution and molecular biology are to be credited
for having transformed the mystery of life into the problem of life: its
origin and maintenance, its evolution and extinction.
We have come to understand that many of the peculiarities of organisms
have molecular roots. Thus cell division, a typically biological event, is
the outcome of a chemical process. Moreover we understand now that
other biological properties and processes - such as those of emerging by
way of self-assembly of nonliving units, and of being subjected to natural
selection - are properties shared by all systems. (Recall Ch. 1, Sec. 3.2.)
In other words, we have come to realize that organisms are systems com-
posed of chemical systems, and moreover that they have evolved from
lower level things. This suggests the plan of this chapter: we start out with
biochemical systems (studied in Ch. 2), endow them with certain emergent
properties to obtain biosystems, and examine some peculiarities of biol-
ogical adaptation and evolution.
The upshot of our study will be a merger of ontological pluralism (or
emergentism) with moderate epistemological reductionism. Indeed we
shall take the view that ontological monism - whether mechanistic or
spiritualistic - is false because it denies the emergence of novel properties
and laws as one moves from physical systems through chemosystems to
biosystems. On the other hand the program of explaining the higher (more
complex and newer) levels by the lower ones - i.e. epistemological reduc-
tionism - has proved fertile and is in keeping with contemporary biology.
Indeed the latter - particularly molecular biology and evolutionary biol-
ogy - is explaining life without explaining it away, i.e. without leveling
down (let alone up), and is doing it with the help of chemistry, though not
in terms of it exclusively.
75
76 CHAPTER 3
( a) (b)
Fig. 3.1. The self-assembly of colloidal particles around a protein molecule (P) in an
aqueous solution (a), and the coalescence of colloidal particles into a coacervate (b).
78 CHAPTER 3
able solution of nucleotides and nucleic acids could perhaps start synthe-
sizing RNA molecules, which in turn would control the synthesis of pro-
teins as well as of further DNA molecules. Furthermore a protecting
double layer of lipid molecules could self-assemble around such a DNA-
RNA-protein system. Such a semipermeable membrane would increase
the concentration of biomolecules, hence the rate of biochemical reac-
tions, and would keep out disrupting (e.g. dissociating) agents. New reac-
tions would thus be possible, ending up in the formation of new kinds of
biopolymers and even of trans molecular systems such as vesicles and
organelles of various kinds.
Given biomolecules in a suitable medium, the formation of organelles
such as chloroplasts and ribosomes is far from miraculous. Indeed a ribo-
some, for example, is not an impossibly complicated system: it consists
of 58 macromolecules, three of them RNA molecules, and the rest proteins
of various kinds. What the ribosome does - its function - is, chiefly, to
synthesize proteins from precursors in its environment.
The next step is, of course, the self-assembly of a few such transmolec-
ular systems to form a cell, however primitive - hence inefficient, ergo
ill-adapted. The rest is history - evolutionary history driven by mutation
and selection. In short, the process of emergence of life may have looked
like Figure 3.2. .
This diagram condenses what is common to various rival hypotheses
concerning the precise self-assembly of primitive cells. (See Florkin (1960),
Oparin (1968), Fox and Dose (1972), Miller and Orgel, (1974), Oro, Miller,
Ponnamperuma and Young (Eds.) (1974).) There is rapidly accumulating
Euca~tic cells
1
\
Primitive cells
SElF-
ASSEMBLY
BIOCHEMICAL
SYSTEMS
Coacervates (1 J
Biomolecules
CHEMICAL
SYSTEMS Carbon
compounds
evidence that most of the buildup processes depicted in Figure 3.2 have
been going on on our planet, and at least the first few steps elsewhere as
well. In particular, the analysis of Precambrian rocks, of meteorites, and
recently of some celestial objects, has been yielding amino acids, fatty
acids, and even nucleic acids.
We compress and generalize the above into
POSTULATE 3.2 There are systems of a kind B such that, for every member
b of B,
(i) b is composed of chemical and biochemical (hence metabolizing)
subsystems;
(ii) b incorporates some of the biomolecules it synthesizes (rather than
releasing them to its environment);
(iii) the possible activities of b include (a) the rearrangement, assembly,
and dismantling of components; (b) the replacement of nonfunctional
components and the removal or neutralization of some chemicals; (c) the
80 CHAPTER 3
capture and storing of free energy (in ATP molecules, glycogen, and fat)
for future consumption;
(iv) b can adjust to some environmental changes;
(v) b is a component of some supersystem (organism) or else of some
population of systems of the same kind;
(vi) some of the subsystems of b are capable of reproducing parts of b;
(vii) some of the subsystems of b control some of the processes occur-
ring in b in such a way that a fairly constant inner environment (milieu
interieur) is maintained in the system;
(viii) all of the control systems in b are interconnected by signals (dif-
fusion of chemicals, propagating chemical reactions, electrical signals,
etc.) and so constitute a signal or information network;
(ix) one of the control systems of b (its genetic system or genome) is
composed of nucleic acid molecules, is unique to b, and controls the de-
velopment and the reproduction of b;
(x) the descendants of b are similar to b but may possess some new traits
of their own (by mutation or gene recombination);
(xi) b competes in some respects with other systems of the same species
as well as of other species, and cooperates with the same or others in some
other respects;
(xii) b descends proximately or remotely from biochemical systems
devoid of some of the properties listed above, and its nucleic acid mole-
cules keep traces (hence some record) of the ancestry of b;
(xiii) b lasts, as a member of B, for a limited time lapse.
The name of the B's should come as no surprise:
bolize this way do not make the grade: they are either nonliving or sick.)
Likewise clause (iii): a nonliving (e.g. biochemical) system need not be
self-cleaning and self-repairing but could accumulate reaction-inhibiting
chemicals (e.g. toxic substances) that would eventually bring some or even
all of its reactions to a halt.
Note that in clause (xi) we have included not only competition, the heart
of Darwinism, but also cooperation. Of course in the vast majority of cases
cooperation is not deliberate, i.e. it does not consist in solidarity, but it is
not less real for that. Without the unwitting cooperation of plants, animals,
and other organisms, the terrestrial atmosphere might not be very different
from what it was in the beginning, i.e. about five billion years ago: for one
thing it might still be reducing rather than oxydizing. Note also a peculi-
arity of heredity in biosystems (clause (xii)). Like other complex systems,
such as stars, mountains, and artifacts, biosystems betray their past. But,
unlike nonliving systems, biosystems concentrate their archives in a tiny
part of themselves, namely their genome.
Not all of the functions (properties, processes and activities) we attribute
to an organism are actually carried out by it throughout its life history.
Thus metabolism and cellular division may be temporarily suspended, as
is the case with spores and dormant seeds. Some of the properties of bio-
systems are then dispositions, or potential properties that may actualize
under favorable environmental circumstances. (For the notion of poten-
tiality see Vol. 3, Ch. 4.)
We have not included certain properties that are obviously possessed by
biosystems - and therefore often listed as definitory - because they are
also common to all biochemical systems and even to some physical sys-
tems. Among them are (a) having a more or less definite boundary separat-
ing the system from its environment, and (b) being subject to natural selec-
tion. That the former property is shared by many a macro physical non-
living system, is clear. As for the latter property, Postulate 1.6 attributes it
to all systems.
Finally note that, according to Definition 3.1, chromosomes are not
alive because they do not metabolize, and ribosomes are not alive because
they do not replicate. Likewise viri are not alive because they do not func-
tion at all outside some host cell - so much so that independent viri are
often crystals. (Only the host cell-virus system is alive.) Nor do robots,
however sophisticated, qualify as biosystems, if only because they are made
of mechanical and electrical components instead of biochemical ones. A
82 CHAPTER 3
DEFINITION 3.2 (i) A cell is any biosystem such that none of its components
is a biosystem. In obvious symbols,
XEC =dj xEB & \fy (YE<&'(X) =>y¢B);
(ii) the cellular composition of a biosystem is the set of cells that compose
it:
X EB => 'C c(x) = dj 'C(x) n C, where <&'(x) = {y Eel y c:::: X};
(iii) an organism is a cell or a multicellular biosystem which is not a pro-,
per subsystem of a biosystem :
XEO =d/XECVXEB&('Cc(x)c C& l(3y) (YEB&x-<y)
bonds amongst its components, and such bonds are possible only if the
individuals concerned are not far apart. Remark 2 Nor is it necessary to
specify the planet in defining the general notion of a biosphere: the defini-
tion applies to any planet and it does not hold for the set of ecosystems on
all possible planets because such a set is not a concrete system.
1.4. Biolevel
We have distinguished, more or less explicitly, six different biotic levels,
namely those of cell, living component of an organism, organism, popula-
tion, ecosystem, and biosphere. We may as well display them for reference
purposes:
B[ = cell level = the set of all cells
Bz = organ level = the set of all organs
B3 = organismic level = the set of all organisms
B4 = population level = the set of all populations
B5 = ecosystem level = the set of all ecosystems
B6 = biosphere level = the set of all biospheres.
The family (3 = {Bf, B 2 , B 3, B 4, B 5, B 6} illustrates the concept of a level
structure (Definition 1.8). So, the set (3 together with the level precedence
relation deserves a name of its own:
DEFINITION 3.4 The set (3 of biotic levels together with the relation < of
level precedence, i.e. S?J = «(3, <), is the biolevel structure.
S?J has also been called the scala naturae and the hierarchy of life. These
are misnomers: a ladder proper is supposed to lead somewhere, and a
hierarchy proper involves a dominance or subordinat"ion relation, while S?J
does not.
Now, the biotic levels we have included in (3 are not the only logically
possible ones. In particular, one can imagine the existence of subcellular
organisms; in fact it has occasionally been claimed that such biosystems
exist. We shall assume that whatever is alive or composed of living beings
is in (3:
biologists nowadays would claim that biotic levels are so many stages of an
evolutionary process: that every biolevel has emerged spontaneously from
the preceding (prebiotic or biotic) level. Actually we need no separate
axiom to incorporate and refine this assumption, as it is a mere specification
of Postulate 3.1 :
2. BIOFUNCTION
(or sickness) as well as of total health (or health in all respects). One ob-
vious candidate is this: The degree of sickness of a biosystem in a given
environment, in a given respect, at a given time, is its relative imbalance
(or deviation from the physiological norm) in that respect at that time.
Assuming optimistically that such normal values can always be found, we
propose
DEFINITION 3.6 Let F,-: B x ExT --+ lR be the ith component of the
state function of biosystems of kind B in environments of type E, and Fi
the corresponding (physiological) normal value of F i . Further, let W =
{Wi E [0,1] I 1 :s;; i :s;; N} be a set of positive reals that add up to unity,
and such that Wi measures the weight or relative importance (for the whole
biosystem) of the property represented by the function F i . (Cf. Vol.
3, ch. 2, Sec. 4.2, Definition 2.15 of property weight.) Then
(i) the degree of sickness of the biosystem b E B in the environment e E
E, in the ith respect and at time t E T, is
si(b, e, t) = I F,{b, e, t) - Fil / IF,{b, e, t) + P,-I;
(ii) the degree of health of the biosystem b in the environment e, in the
ith respect and at time t, is the complement of s,-(b, e, t) to unity, i.e.
hi(b, e, t) = 1 - si(b, e, t);
(iii) the total degree of sickness and the total degree of health of organism
bin e at tare
N
S(b, e, t) = 1:: wis,.(b, e, t) and H(b, e, t) =1 - S(b, e, t)
i=l
respectively.
The values of s,. and hi are real numbers in the unit interval. And each
is the same whether F,{t) equals nFi or (l/n)Fi , with n an arbitrary integer.
The conceptual interest of the above formula resides in its making sense
only within a theoretical model of biosystems of the species of interest.
Its methodological interest lies in its reminding us that scientific medicine
is just a chapter of biology (including psychology in the case of man).
Finally we come to
DEFINITION 3.7 A thing is dead iff, having been a biosystem, it has ceased
to perform all of the functions listed in Postulate 3.1.
Remark 1 We could not define death as maximal degree of sickness, or
88 CHAPTER 3
----DEATH
Fig. 3.3. The successive collapses of the state space of a dying biosystem: a gradual
loss of properties and of systemicity.
LIFE 89
which are linked to one another, and succeed one another, in various
ways. If eucaryotic (nucleated), a cell consists of a nucleus and a cyto-
plasm. The former is composed of DNA and RNA threads, and the
latter of subsystems such as ribosomes, mitochondria, chloroplasts (in
plants), and cilia - none of which are alive.
Each of the nonliving subsystems of a cell performs certain activities,
usually called 'functions', such as storing energy, synthesizing proteins,
breaking down excess biomolecules, detecting imbalances, or transmitting
signals. Such subsystems of the cell are therefore called 'functional units'.
In multicellular organisms (metazoans) this division of labor often car-
ries over to entire cells or even cell groups such as tissues and organs. In
either case each subsystem of a biosystem, whether or not the former is
alive, discharges certain functions, i.e. does certain things, that are pecu-
liar to it, and others which are done also by subsystems of other kinds.
For example, neurons not only generate and transmit signals but also syn-
thesize ATP, nucleic acids, and proteins,
We elucidate these notions by means of the following definition, which
employs the notion of a process elucidated in Ch. 1, Sec. 1.7:
(ii) a and a' have the same value during the time interval of interest, i.e.
Yea, b, e, t) = V(a', b, e, t') for t and (' in the given interval.
Remark 1 Some biofunctions, such as metabolism and cell division,
are common to all organisms. Others are species-specific. For example
photosynthesis is peculiar to plants, and the synthesis of enzymes that
occurs in the digestion of meat is peculiar to carnivores. The grouping of
organisms by biofunction clusters provides just as natural and fundament-
al a classification as evolutionary taxonomy; only, it is out of fashion.
Remark 2 The more numerous the functions an organism can perform; the
better equipped it is to cope with environmental challenges. Multicellular
organism that are highly differentiated will therefore be selected by rich
and changing environments. Also, because they have spare cells, meta-
zoans are covered against the accidental loss of some cells, so they have
better chances of survival than organisms without such spares. (Recall the
discussion of reliability in Ch. I, Sec. 4.1.) These two factors help explain
the prosperity of metazoans on our planet. In short, the terrestrial en-
vironment has been fit for complex and versatile organisms endowed with
a high degree of redundancy. Remark 3 Alternative definitions of bio-
logical value are of course thinkable. One alternative worth looking into
is to equate biovalue with efficiency. Another is teleological: Valuable to
an organism (or a species) X is whatever contributes to X's attaining its
goals. We cannot adopt this alternative because we are not assigning goal-
seeking activities to all organisms. Goals will be introduced in the next
chapter, with reference to organism endowed with complex nervous
systems. Remark 4 As evolution proceeds, the value of some subsystems in-
creases (preadaptation) and that of others decreases, sometimes to the
point of vanishing altogether. In the latter case the subsystem may dis-
appear (as in the loss of the ability to synthesize certain vitamins), or it
may be kept functionless for hundreds of thousands of years. Thus it is
possible, nay quite likely, that man has a number of fossil (functionless)
proteins, genes, entire cells, and even cellular systems - i.e. subsystems
that have long ceased to serve any biofunction. But this is very hard to
prove. Remark 5 It is just as likely that certain hereditary pathological
conditions were valuable in the past or may yet become valuable. This is
the case with sickle-hemoglobin, which is associated with resistance to
malaria. Likewise certain congenital malformations, such as flat feet and
poor sight, have spared the lives of many men in modern wars. (War
selects the unfit.)
92 CHAPTER 3
2.3. Biocontrol
have mUltiple control systems grouped into tightly integrated signal net-
works. Another peculiarity is that such networks function well- amazingly
well in most cases. 'Well' does not mean either at full blast or else with
maximal economy, but instead in whatever manner is most valuable to
the organism as a whole. (Recall Definition 3.9 ofbiovalue.) For example,
the rate at which any given protein is synthesized is neither minimal (max-
imal economy of resources) nor maximal (waste) but somewhere in be-
tween - above equilibrium and short of waste. This near optimality is
peculiar to biosystems.
Unlike biology, physics and chemistry are dominated by extremal
principles such as Hamilton's, i.e. basic law statements according to
which, of all the conceivable processes of a given kind, only those are
actualized that either minimize or maximize a certain magnitude, e.g. the
action. (Cf. Appendix B, Sec. 1.4.) These principles "operate" of course
for the physical components of a biosystem, but these components inter-
twine in such a manner that certain properties of value to the organism
as a whole are near-optimized rather than either minimized or maximized.
I.e. such near optima are not determined locally by each separate system
but result from the interaction among the various subsystems of the or-
ganism. In fact since the various components of the control network of
the whole organism are interdependent, the near optimization of certain
local variables will hamper that of others. So, what would be best for each
subsystem is unattainable. But the next best is attainable, and this is a
compromise among a number of optima. This compromise may be called
overall (or organismic) near optimality - not full optimality because no
organism is perfect. (A perfect biospecies would not be subject to evolu-
tion.) Only the biological organization of successful species tends to
optimality.
An organism may be said to possess an efficient biocontrol network if
it is capable of redressing whatever imbalances may occur - i.e. if it can
detect and correct them. If not, the organism becomes sick and may die.
More precisely, every malfunction of a specific biocontrol system is a
specific sickness. (The converse is false. For example infectious diseases,
if overcome, are so by the proper functioning of the immune response
system.) And if several biocontrol systems malfunction, or do not function
synergically, a nonspecific (generic) sickness supervenes. Natural death
is probably the final state in a process of generic (nonspecific) sickness.
Biocontrol has given rise to the belief that organisms are goal-seeking:
that they act in everything so as to ensure their own preservation or at
LIFE 95
least that of their species. First the theory of evolution through natural
selection, and in recent times the theory of control, should have corrected
that illusion. In the first place self-control is a property of all biochemical
systems - by virtue of their enzymatic activity - whether or not they are
alive. And if system integrity were defined as equilibrium (steady state),
then every biochemical system would be seen to behave as if it reached
for its own preservation. Secondly, biocontrol is not always efficient: it
falls short of optimality when the organism becomes sick and it fails
utterly at death. Thirdly, biocontrol can fail, even in healthy organisms,
when certain imbalances grow beyond bounds during catastrophes such
as floods and droughts. (After all, the great majority of terrestrial bio-
species have become extinct.) Fourthly, biocontrol is occasionally defec-
tive. For example, only a small part of the RNA synthesized in the cell
nucleus migrates to the surrounding cytoplasm to control protein synthe-
sis: most of it is broken down into nucleotides without ever being "used".
Another example: many genes are never "expressed" or activated. Fifthly,
genetic control is not so strict as to prevent mutations, most of which are
deleterious. In any case, we need not invoke goal-directedness to explain
why biocontrol is normally beneficial for the organism as a whole: the
organisms not endowed with efficient control networks succumb before
reproducing, and that's all.
2.4. Development
The history of any system can be represented as the trajectory of its re-
presentative point in a suitable state space for the system. (Cf. Ch. 1, Sec.
2.2.) In particular, the life history of an organism can be represented as a
line in its state space. The part of the life history comprised between the
beginning (birth) and the attainment of the adult state of the organism is
called its development. (See Bonner (1974).) The development process of
an organism has three main strands, each of which is taken care of by a
bunch of components of the state function representing its possible states.
These strands or aspects of development are:
(i) growth, or multiplication of cells of certain kinds (at uneven rates for
different subsystems of the organism);
(ii) morphogenesis, or formation of new subsystems;
(iii) differentiation, or specialization of cells.
All three aspects of development proceed under a triple control:
enzymatic, genetic, and environmental. The environment supplies or
96 CHAPTER 3
Thorn (1972), with its total disregard for morphogens, physical constraints,
natural selection, chance, and history, does not go beyond classical mor-
phology except for the addition of a mathematics of dubious utility.)
Morphogenesis may be regarded as a particular case of self-assembly
of molecules, organelles, or cells, as the case may be. Like other self-as-
sembly processes, the morphogenetic ones are guided by both internal
forces and environmental resources and circumstances. The formation of
an organelle, a cell, or an organ is controlled by genes and other morpho-
gens in conjunction with the milieu interieur and the environment. The
DNA and RNA molecules, the hormones and other morphogens control
the chemical reactions and the diffusion proceses, as well as the assembly
and dismantling of the various units, but nothing gets done unless the en-
vironment supplies the materials and the energy.
A self-organization process can start spontaneously in an initially
homogeneous system of reactants and morpho gens. Thus a local chance
fluctuation in their relative concentration can diffuse throughout the me-
dium, ending up in a differentiated system (Turing, 1952). However, reac-
tion-diffusion processes are not the only morphogenetic mechanisms, nor
are they typically biological. Alternative mechanisms are: (a) sorting pro-
cesses (grouping of functionally similar units, as in the case of polyribo-
somes), (b) cooperative interactions among molecules or cells of different
kinds, (c) packing, folding, and other mechanical effects of accretion and
external pressure; (d) enzymatic syntheses of morphogens proceeding until
a new entity is formed - and inhibited thereafter.
Let us take a quick look at processes of the first kind, i.e. sorting proces-
ses (Steinberg, 1963; Mostow (Ed.), 1975). Let C be a set of cells of two
kinds, A and B. Assume that all these cells are motile and that like cells
adhere to one another more strongly than do unlike cells. Moreover, sup-
pose that the cells of type B adhere to one another less strongly than those
of type A. In sum, assume that the adhesion works (or energies) compare
like this:
W A > WB ~WAB'
.~. 0 Morphogenesis
00. - - - - - -
.~~O o
Fig. 3.4. The formation of an onion-shaped organ constituted by two concentric
spheres composed of cells of different kinds, A and B. according to Steinberg (1963).
DNA
RillA
G A T C A T TAT TAG A 1 T A T " .......
........
~ ~ ~ ~ ~ ~
0
j 1
OJ IT] ~
I I I Q] CD
Fig. 3.5. Two kinds of mutation: by permutation and by deletion (or insertion) of
nucleotides.
vidual variability". It also explains the large number of biospecies. And the
limited environmental resources, with the accompanying competition for
them, explains why biopopulations do not grow exponentially (except for
limited periods). Finally, changes in environmental conditions - partic~
ularly important changes such as glaciations and continental drifts-
e)(plain why (a) most biospecies that have been on Earth have become
extinct, and (b) the number of species can jump up or down during certain
brief periods. (One such revolutionary period was the beginning of the
Cambrian, when many species of multicellular organisms emerged.)
In view of the fashionable claim that whatever an organism is and does
is programmed in its genome, it may be helpful to insist that what is in-
herited is a bit of genetic material carrying certain potentialities. These
potentialities mayor may not be actualized in the course of the organisms'
life, according to the possibilities offered by the environment. (For the
concepts of potentiality and actuality see Vol. 3, Ch. 4.) We may as well
be more precise and adopt
pens in the case of organs that do not perform their specific function
during the entire life of an organism, is that their nonfunctional states
precede in time the functional ones.
A second point where developmental biology and the theory of heredity
make contact with metaphysics is over the question of preformation vs.
emergence. There have been two classes of solution to the problem of ex-
plaining the development of an organism from egg through adulthood to
final breakdown. One is to assume that all of the properties and patterns
that appear in the course of development are just the unpacking or un-
folding (development) of what was contained in potency (preformed) in the
egg. Or, to put it into modern parlance, the assumption is that the genome
contains a full set of "instructions" steering the organism through the suc-
cessive stages of its development. This is the preformationist hypothesis
held by Aristotle and Leibniz, and resurrected by some contemporary
scientists who believe that biology is fully reducible to molecular biology.
Preformationism denies real novelty, in particular the emergence of new
properties that accompanies self-organization (Ch. 1, Sec. 3.2); and it
minimizes or even ignores the role of the environment.
The partisans of epigenesis hold, on the contrary, that development is
an epigenetic or creative process in which novelty occurs at every stage.
They maintain that the organism starts at a low state of organization and
acquires new components and properties, as well as ioses others as it de-
velops. Epigeneticism is of course the hypothesis underlying all of clas-
sical embryology as well as the contemporary studies on the origin of life.
The molecular biology revolution has all but eclipsed epigeneticism.
What is fashionable nowadays is genetic preformationism. Yet fashion
need not concide with truth. The claim that development is just the deploy-
ment or implementation of a blueprint contained in the genome is no more
illuminating or correct than holding that it is guided by an immaterial
elan vital (Bergson, 1907) - or that it consists ultimately in the embodiment
of Platonic self-existing geometric shapes (Thorn, 1972). Surely develop-
ment can be neither purely preformative nor purely epigenetic, if only be-
cause the environment supplies the raw materials for it and weeds out the
maladapted organisms, as well as offers chances to the preadapted ones.
It seems that each view, either preformationism or epigeneticism, has
alternatively accelerated and blocked the development of biology. Pre-
formationism may have stimulated the purely descriptive phase of em-
bryology and the beginnings of molecular biology, whereas epigeneticism
may have stimulated experimental embryology (particularly investigations
LIFE 103
3. EVOLUTION
3.1. Adaptation
A salient feature of our biosphere is its rich variety. There are at least
three million biospecies (Dobzhansky, 1970), and no two strictly identical
individuals in any of them. (Even so-called identical twins are somewhat
different.) This variety is a result of four billion years of evolution. And
the evolutionary process is in turn the net effect of a combination of genic
changes (mutations and recombinations) with environmental selection.
The process seems to be continuing: there is no reason to believe that it
stopped the moment man became aware of it. In fact the extinction of a
number of species has been certified during our own lifetime, and it is
likely that new species are forming right now. In short, bioevolution is
revolutionary - or creative (Bergson, 1907), or emergent (Lloyd Morgan,
1933).
Anthropomorphically speaking, evolution is a trial and error, or success
and failure, story. Success is called adaptation, and failure maladaptation.
A biopopulation of well adapted individuals thrives or expands, whereas
one of maladapted individuals declines. This looks obvious but ceases to
be so as soon as one stops and examines the very term 'adaptation': he
104 CHAPTER 3
then realizes how tricky it is (Williams, 1966). In fact, the word 'adapta-
tion' is ambiguous, as it designates at least three different concepts,
namely the following:
Al = suitability of a subsystem (organelle, organ, etc.) to a function,
or high biovalue of the subsystem for the entire organism;
A z = adjustment of the organism to its environment;
A3 = fertility of a biopopulation.
That these are indeed different concepts, is realized upon displaying
their structure. Al is a relation between subsystems and organisms, A z
a relation between organisms and environments, and A3 an intrinsic
property of organisms. A methodological examination confirms the
difference. In fact the way to find out whether something has the property
A I is to conduct certain biochemical or physiological experiments or
observations in a constant environment. By varying environmental factors
we can ascertain whether an organism possesses property A z. And the
criterion for A3 is the count of offspring.
In sum, we must be reconciled with the fact that 'adaptation' is an
ambiguous term. This seems obvious, yet it is overlooked by those who
claim that only population genetics uses "the correct" concept, namely
A 3 • And it is equally overlooked by those who claim that the Darwin-
Spencer principle of the survival of the fittest is tautological. Indeed, sur-
vivorship or fertility (A3) is not the same as fitness or adjustment (A z)
but an indicator of the latter. The logical relations among the three con-
cepts are these: A3 => A z => Aj, i.e. subsystem biovalue is necessary for
adjustment, which is in turn necessary for fertility.
It goes without saying that adaptation of any kind is time-dependent.
I.e., A), A z and A3 vary in time, either because of internal changes, or
environmental changes, or both. The case of sudden environmental
changes is particularly interesting, for they are bound to favor some
biopopulations while hurting others. In the first case one speaks of pre-
adaptation, in the second of premaladaptation. In keeping with the above
distinction among three adaptation concepts, we distinguish three pre-
adaptation types:
PAl switch of function(s) to other(s) of greater biovalue;
PAz improved adjustment to the environment;
P A3 greater fertility.
As with the case of adaptation, the first is necessary for the second, which
is in turn necessary for the third, i.e. P A3 => PAz => P A).
Neither of the three kinds of adaptation or preadaptation calls for tele-
LIFE 105
30
<
.2
....
1
POSTULATE 3.4 All populations evolve, and the evolution of each is a pro-
cess whereby spontaneous (genic) changes ensue in unequally adapted
individuals, amongst which the environment selects the best adapted ones.
Selection is called natural if it is spontaneous, and artificial if controlled
by man.
Because biopopulations are not isolated from each other but interact
with other biopopulations forming ecosystems, the evolution of each must
affect that of others. This process has a name of its own:
Since module permutations are far less radical than changes in composi-
tion, we make the further simplifying assumptions that
(3.19)
Similary, we posit that the variety 01 is just as stable as 10, and that 00 and
11 are equally stable:
(3.20)
Our final assumption is that the mutations between the extreme varie-
ties 00 and II proceed via the intermediate species 01 and 10, i.e.
II flnij II ==
:1 (1 - flnu) (l - flnu) (1 - finn)'
(1 - fln22 - fin)
Iflnu (3 - fln22 - fin) (3 - fln22 - fin)
3- fln22 - fin
I-Hl- fln22 - fin) fln22 fin t(1 - fln22 - fin)
I to - fln22 - fin) fin fln22 t(1 - fln22 - fin)
ip - finn)' fin)
It(l - fln22 -
(l - finn) - finn
(3 - fln22 - fin) 3 - fln22 - fin
:3- fln22 - fin
(3.23)
NI = 2(2mll - I) (l-dl)N
Ny = 2(1 - mll) {I - dl)N
3 - m22 -m (3.24)
N~ = Ny
Nt = (l - m22 "--m)NY.
(3.25) & (m22 +m < I) or, equivalently, (3.25) & (m21 + m24 > 0).
3.6. Bioprogress
Evolutionary biologists agree that all modern complex organisms descend
from primitive organisms: recall Figure 3.2. So, there would seem to be
no doubt that evolution has been progressive. Not so fast. First of all, the
following kinds of evolutionary process are usually distinguished:
Some lineages, then, have advanced over certain periods while others have
not. The second problem is that there is no agreement on the very notion
of bioprogress - the more so since progress through increasing specializa-
tion is sure to lead to stagnation or extinction.
A number of alternative definitions of bioprogress have been proposed
(cf. Williams, 1966, Bowler, 1976). Here are some: bioprogress is increas-
ing histological differentiation, morphological complication, survivorship,
adaptability, freedom from the environment, or amount of genetic infor-
mation. All of these proposals presuppose that progress is global rather
than partial (in particular respects), and none of them lends itself easily to
quantitation. In particular, the information-theoretic definition is pseudo-
quantitative, as there is no unambiguous and objective way of measuring
or computing the "quantity of information stored in an organism" (Lwoff,
1962; Apter and Wolpert, 1965). After all, biosystems are not identical
with their genomes, and the latter are more than just information pro-
cessing devices.
We shall take it that bioprogress consists in increasing biovalue, and
shall proceed to refine the distinction (Ayala, 1974) between uniform
(relentless) progress, on the one hand, and overall (net, average) progress
on the other. Our elucidation will make use of the concept of system pro-
geny (Definition 1. 15(v) and Definition 3.9 of biovalue. Since the former
was framed for systems in general, we shall first adapt it to biopopulations;
and this adaptation will involve the notion of a variable collection, i.e. one
with changing membership (Vol. 3, Definition 5.13). Let us then start with
system (e.g. organ) of every member of PI in P(Po), and assume that all the
subsystems of that kind perform (roughly) the same function(s}. Then
(i) the K's have advanced on the whole with regard to l: between Po and
PI' for t :> 0, iff the value of subsystem of kind l: for the members of PI is
significantly greater than its value for the members of Po;
(ii) the K's have advanced uniformly along P(Po} with regard to l: be-
tween Po and Pt. for t :> 0, iff the value of subsystem of kind;; for the
members of every population Pu in P(Po), between Po and Pt, is greater than
its value for the immediate ancestors of the Pu's.
Whereas the first clause elucidates the notion of net or average progress,
the second clarifies that of uniform progress. Uniform progress implies
net progress but not conversely: i.e. there may be temporary setbacks in a
progressive lineage,
Finally we make
the emergence of new species, they have been fatal to others. In any case
most species have become extinct instead of persisting on the road of
improvement.
Another presupposition of the hypothesis of universal and global pro-
gress is that there exists a fixed goal - an ideal or paradigm - towards
which successive generations move. This hypothesis would make sense in a
Platonic biology, where the archetype is perfect, every real organism being
a sort of blurred carbon copy of the former. But of course there is no such
thing as Platonic biology, although there are vestiges of Platonism in
biotaxonomy. The very notion of an immaterial object - such as the Idea
of the Perfect Mosquito - guiding the evolution of concrete things is in-
consistent with the ontology of modern science. (Cf. Vol. 3, Introduction
and Ch. 2.) Besides, since the environment changes rather erratically and
sometimes dramatically, if there were a target then it should be moving
rather than fixed, and so evolutionary processes zig-zag rather than go in
a definite direction.
In sum, the biosphere does not obey the so-called "law of progress".
Progress, when it does occur, is a trend not a law, and one occurring in
some lineages, in some respects, and for some time. In the biosphere, just
as in the sociosphere, there are not only progressives (e.g. mammals) but
also conservatives (e.g. ants) and reactionaries (e.g. lice). Besides, whereas
many progressives have become extinct, some conservatives are still
around: drastic environmental changes and fierce competition can wreck
the most brilliant careers.
For all the above reasons, the most we can assert is
4. CONCLUDING REMARKS
The traditional biometaphysics are vitalism and mechanism. The strength
LIFE 119
of the former lies in its ability to ask questions, some of them pertinent,
that the latter fails to answer. Whereas classical vitalism has hardly
evolved, mechanism has. Two kinds of mechanism are popular today:
(a) physicochemicalism, or the thesis that organisms are just physico-
chemical systems, and (b) machinism, or the thesis that organisms are
machine-like systems, hence fashioned after a design or plan, and acting
in a goal-seeking manner.
Physicochemicalism is the heir to traditional mechanism, and machin-
ism a hybrid of mechanism and vitalism. Whereas the model of the first
is the chemical reactor, that of the second is the computer. Neither of
these biometaphysics is consistent with the whole of biology, although
physicochemicalism has been fruitful where machinism has been sterile.
Physicochemicalism is incompatible with biology because it denies em-
ergence and levels, and machinism because it takes for granted and
posits plans and goals where there are none.
The alternative we propose is biosystemism, or the application of sys-
temism (Ch. 1, Sec. 4.2) to the study of life. Biosystemism recognizes the
bios as an emergent level rooted to the chemical one. Therefore it accepts
the lay difference between living and nonliving - denied, e.g. by Pirie
(1960), Kendrew (1966) and Keosian (1968) - as well as the molecular-
biological account of the self-assembly of biosystems from biochemical
precursors, and the theory of evolution by genic changes and natural
selection.
We are faced then with four different biometaphysics, each with its
accompanying problem circle, method, and goal: see Table 3.1. The oldest
and most persistent of them is vitalism, which is also the most popular
biophilosophy. Actually there are two kinds of vitalism: ontological and
methodological. The former states that organisms are distinguished by
some immaterial principle, the latter that they are so complex and variable,
that they cannot be tackled with the ordinary. methods of science. Ontolo-
gical vitalism has been pronounced dead long ago, even by biologists who
continually lapse into it - such as Monod (1970) and Jacob (1970) when
fantasizing about the project te!eonomique of every organism. On the other
hand the claim that organisms cannot be studied by the ordinary scientific
method but call for a special method (i.e. methodological vitalism) is still
very much alive except among molecular biologists (Bunge, 1973a, Ch.3).
As for machinism, its distinctive thesis is that organisms are like ma-
chines in their having a design or plan and a purpose. However, two
kinds of machinism must be distinguished: verbal and serious. Verbal
machinism restricts itself to repeating uncritically the organism-machine
120 CHAPTER 3
TABLE 3.1
Four approaches to the study of life
Mechanism
Physico-
Vitalism Machinism chemicalism Biosystemism
Ontological Life trans- Organisms Organisms Organisms are sui
Framework cends matter are machine- are physico- generis systems com-
and is guided like: they chemical posed of biochemical
by an imma- have plans systems. systems and possessing
terial principle and goals. emergent properties
and laws.
Problem 'What is it 'What is it 'How is it All biological prob-
Circle for?' ques- for?' and composed?' lems, in particular
tions. 'How does and 'What regarding subsystems,
it work?' are its prop- their self-assembly,
questions. ertiesT and evolution.
Method Description The whole array of
of organism Physico-chemical methods particular methods of
as a whole. plus building of analogs. the various chapters
of natural science.
Goals Intuitive Technology- Physicoche- Biological explanation
understand- wise expla- mical expla- and prediction.
ing of the nation and nation and
whole forecast. prediction.
analogy and to warning that biology may learn more from engineering
that from physics and chemistry (Polanyi, 1968). This kind of machinism
can be discarded by noting that biology does not have a single law state-
ment, let alone theory, containing the concepts of plan (or design) and
purpose. (Even the most serious attempt to incorporate the concept of
goal-directedness into biology, namely that of Sommerhoff (1974), fails
to mathematize correctly that concept as well as to exhibit examples of
laws and theories where the concept occurs.)
Serious contemporary machinism is the heir to Descartes and La Met-
trie and culminated with the theory of self-reproducing automata (von
Neumann, 1966). It is rather popular with the computer science crowd, as
can be seen by scanning the J. Computer and System Sciences. This ap-
proach has stimulated, in particular, biological applications of the theories
of automata and information. However, three decades of work along this
line have yielded precious little; in particular, it has not contributed to
the molecular biology revolution or to evolutionary biology. No wonder,
for machinism skips the chemical level and ignores history: it thinks in
LIFE 121
c H c H
.
.§
Common
ancestor
.§ Camnlon
Cbl
C. I
Fig. 3.7. Humans (H) and chimpanzees (C) are genetically about as different as two
subspecies of mice (a). They differ mainly at the organ level (b), particularly in respect
of the nervous system, hence of behavior. See King and Wilson (1975).
RULE 3.1. Each bioentity must be studied on its own level and should be
explained with the help of its adjoining levels.
Thus an organ should not only be studied on its own level but should
also be analyzed both downward (in terms of its subsystems) and upward
(in terms ofthe entire organism in its environment). This holds particularly
for the human brain, which is not only a neural system but also a sub-
system of a component of social systems. But this is the story to be told
next.
CHAPTER 4
MIND
The present chapter deals with one of the oldest, most intriguing, and most
perverse of all the problems lying in the intersection of philosophy and
science, namely the so-called mind-body problem. This is the system of
questions about the nature of mind and its relations to matter.
Some eminent thinkers - notably Spencer (1862), E. du Bois-Reymond
(1872), and Popper and Eccles (1977) - have pronounced the problem to
be insoluble. I submit that part of the problem is that it is usually formu-
lated in inadequate terms - namely those of ordinary language. These are
inadequate not only because ordinary language is notoriously poor and
inexact, to the point of being of little use in investigating the problem of
matter, but also because it begs the question. Indeed European languages
are loaded with a preconceived solution to the problem of mind, namely
psychophysical dualism, or the doctrine that mind is something apart from
matter. The very expression 'mind-body problem' suggests that mind and
body are distinct entities on the same footing - i.e. things. We must then
look to science for inspiration.
In science we do not speak of the motion-body problem, or of the
reaction-substance problem, or of the digestion-digestive tract problem,
or of the mobility-society problem. We speak instead of the motion of
bodies, the reaction of chemicals, the digestive functions of the digestive
tract, and the mobility of a society. We do not reify properties, states, or
events - except when it comes to the properties, states, and events of the
nervous system. We shall close this gap, that keeps the study of mind a
scientific anomaly in the grip of myth, by replacing ordinary language by
state space language, which is mathematically precise and is shared by
science and scientific ontology.
Another difficulty besetting the so-called mind-body problem is the
extreme youth of neuroscience: recall that Ramon y Cajal's neuronal
hypothesis is less than one century old. Surely there were important spo-
radic discoveries in earlier times, starting with Hippocrates' hypothesis
that the brain - rather than the heart or the liver - is the organ of emotion
and thought. However, it is only in recent years that a concerted attack at
all levels has been launched (Worden et al. (Eds.), 1975).
A third difficulty is that, far from being a strictly scientific problem, ours
124
MIND 125
some entity and processes in some thing, we must ask what minds - i.e.
what is the thing that perceives, feels, remembers, imagines, wills, and
thinks. This is the very core of the so-called mind-body problem, i.e. the
identification of the subject or referent of the mentalist predicates.
Those who hope the mind-body problem to be soluble have proposed
two solutions to it: that what does the minding (perceiving, feeling, think-
ing, etc.) is the mind (or soul or spirit), and that it is the brain. According
to the former the mind is an immaterial entity wherein all mental states
and processes occur: feelings, memories, ideas, and the like would be in or
of the mind, which would in turn be an entity separate from the body.
According to the second answer the mind is not a separate entity but a
set of brain functions: perceiving, imagining, thinking, and the like would
be brain processes. The former thesis is called psychophysical dualism, the
latter materialism or naturalism.
However, there are a number of varieties of dualism and of materialism.
(Cf. Vesey, 1964; Smythies, 1965; Armstrong, 1968; O'Connor, 1969;
Borst, 1970; Popper and Eccles, 1977; Margolis, 1978.). Our main task in
this chapter is not to discuss them but to formulate a version of material-
ism consistent with the principles laid down in the previous chapters and in
Vol. 3 of this work. This is emergentist materialism, or the doctrine ac-
cording to which (a) all mental states, events and processes are states of, or
events and processes in, the brains of vertebrates; (b) these states, events
and processes are emergent relative to those of the cellular components of
the brain, and (c) the so-called psychophysical (or psychosomatic) rela-
tions are interactions between different subsystems of the brain, or among
them and other components of the organism (Bunge, 1977e). This kind of
materialism is monistic with respect to substance and pluralistic with
respect to properties.
Here are some reasons for favoring emergentist materialism:
(i) because it eschews the ghostly mental substance without thereby
denying mental facts, emergentist materialism is compatible with the
scientific approach;
(ii) emergentist materialism - as we hope to show - is free from the
fuzziness that characterizes dualism with its talk of an immaterial mind and
of "correlations" between it and the body;
(iii) unlike dualism, emergentist materialism is consistent with the
general concepts of state and event that can be gleaned from all the sciences.
(On the other hand according to dualism mental states would be the only
states that fail to be states of some thing, and mental events the only
MIND 127
events that fail to be changes of state of some thing - this being why
dualism agrees more with theology than with science.)
(iv) Unlike dualism, emergentist materialism/osters interaction between
psychology and the other sciences, in particular neuroscience, and this
precisely because it regards mental events as special biological events;
(v) unlike dualism, which digs an unbridgeable chasm between man
and beast, emergentist materialism accords with evolutionary biology,
which - by uncovering the gradual development of mental faculties along
certain lineages - refutes the complacent superstition that only Man is
endowed with Mind;
(vi) unlike dualism, which postulates an unchanging mind, emergentist
materialism agrees with developmental psychology and neurophysiology,
which exhibit the gradual maturation of the brain and behavior.
The only drawback to emergentist materialism is its so far not having
been formulated in a precise and detailed way. Indeed, emergentist ma-
terialism is little more than a programmatic hypothesis with enormous
heuristic power. We shall try and remedy this shortcoming in the course
of this chapter, by building a general theory of the mind consistent not
only with our ontological principles but also with current neuroscience
and psychology.
() ()
()
NOW LATER
(b)
Fig. 4.1. (a) Two neural microsystems occupying roughly the same spatial region yet
functionally detached: they 'subserve' different functions. (b) Psychons can be itinerant
even if the components stay in place, because their connections change in time.
DEFINITION 4;5. Let)) be a neural system and '6'i))) the neuronal composi-
tion of)) at time t. Further, call
the function such that Ct(a, b), for a, bE 0'i))), is the strength (intensity)
of the connection (coupling, link) from neuron a to neuron b at time t.
Then the connectivity of)) at t is represented by the matrix formed by all
the connection values, i.e.
DEFINITION 4.6. A connectivity is constant iff it does not change once estab-
lished (i.e. iff C t is independent of time). Otherwise it is variable.
POSTULATE 4.1. All animals with a nervous system have neuronal systems
that are committed, and some animals have also neuronal systems that
are plastic (uncommitted, self-organizable).
Apparently worms, insects and other lowly animals have only totally
(or nearly totally) wired-in neuronal systems. On the other hand the plastic
cortex of man is the largest in all animals and a sort of vast reserve army
ready to meet emergencies, most of which never happen. Hence whereas
MIND 133
POSTULATE 4.2. The neuronal systems that regulate (control) the internal
milieu, as well as all the biofunctions of the newborn animal, are com-
mitted (wired-in).
Example The temperature and acidity of the intercellular fluid are re-
gulated mainly by certain wired-in neuronal systems, and so are the breath-
ing and sucking movements of the newborn animal. But they are not the
only regulators: also the endocrine system plays an important role in in-
ternal control.
2. BRAIN STATES
events: they are signalled by physiological changes outside the brain and
in turn react upon them. Finallycomes the brain's cognition of some of its
own states, i.e. self-consciousness. We shall deal with it in Sec. 5.1.
We proceed now to explain the state space representation of CNS
activity. Consider an arbitrary neuronal system, be it a small neuronal
circuit (fixed or itinerant), a sizable subsystem of the CNS, or the entire
brain. Like every other system, it can be represented by a state function
F = <Flo F2 , ••• , Fn). Ifwe are interested not in the precise localization of
the system and its components, but only in its overall nonspatial properties,
we can assume that F is only a time-dependent function whose values are
n-tuples of real numbers. I.e. we can set F: T -> JRn, with T ~ JR, and JR the
set of real numbers. (We take the host of the neural system, i.e. the whole
animal, as its reference frame.)
Each component Fi of the state function F can be decomposed into a
constant(or nearly constant) part F~ and a variable part F~: see Figure 4.2.
Obviously, either can be zero during the period of interest. However, the
point is that, while the rate of change of F~ vanishes at all times (i.e.
p~ = 0), that of n does not. The latter may therefore be taken to represent
the activity of the neuronal system in the ith respect. For this reason we
make.
DEFINITION 4.10 Let F : T -> JRn be a state function for a neural system ).I,
and let F = Fe + Fv, with Fe = 0 (zero rate of change) for all t E T. Then
(i) ).I is active at time tiff Fv(t) =1= 0;
(ii) the intensity of the activity of ).i over the time lapse 'r c T equals
the fraction of components of).i active during 'r;
(iii) the state of activity of).i at time t is s = Fv(t);
(iv) the (total) process (or function) ).I is engaged in over the time interval
'r c T is the set of states of activity of).i:
F'
DEFINITION 4.11 Let n-(v, 'r) be the total process (or function) of a neural
system v in an animal b during the time interval 'r c T. The corresponding
specific/unction (activity, process) of v during 'r is
n-s(v, 'r) = n-(v, 'r) - U n-(fl., 'r) with fl. 9= v.
p;<.b
Ct(i, j) = cijFY.
(For some mathematical models of the dynamics of cortical systems, see
Marr (1970), Wilson and Cowan (1973) and Nass and Cooper (1975).)
However, here we are interested only in the general principle:
POSTULATE 4.5 For any neural system ).I of an animal, the instantaneous
state of activity of ).I decomposes additively into two functions: Fv =
A + E, where A does not vanish for all t E T, whereas E depends upon
the actions of other subsystems of the animal upon ).I.
COROLLARY 4.1. All and only animals endowed with plastic neural systems
are capable of being in mental states (or undergoing mental processes).
COROLLARY 4.3 Mental functions (processes) cease with the death of the
corresponding neural systems.
Examples Brain injuries, strokes, and of course death bring about the
cessation of the mental functions normally discharged by the affected
neural systems. Dualism does not explain such a "correlation".
'(ii) the K-mind, or mind of species K, during period 'C', is the union of
the minds of its members during 'C':
'-.
M(K, 'C') = U m(x, 'C').
xEK
Since the members of the set called 'mind' are brain functions (pro-
cesses), it makes no sense to say that the brain is the physical "basis" of
the mind. And since the human mind is nothing but the union of all the
individual human minds, it makes no sense to speak of the collective mind
of mankind as if it were an entity or even a functional system. On the other
hand the mind of an individual animal does have functional unity: it is
a functional system (Definition 1.9 in Ch. 1, Sec. 1.7), for every process
(function) in it is lawfully related to at least one other process (function)
in it. The "neural basis" of the unity of the mind, on which many philoso-
phers have insisted, is this: the components of the plastic neural super-
system P, far from being uncoupled, form a system by Postulate 4.3.
Hence
COROLLARY 4.5. The mental functions of (processes in) the plastic neural
supersystem of an animal are coupled to one another, i.e. they form a
140 CHAPTER 4
functional system. (In short, for every animal b and every period .. of its
existence, m(b, ..), if nonvoid, is a functional system.)
Consequently when a neurosurgeon cuts a living primate brain into two,
he splits the plastic system into two, and thus also the corresponding mind
of the animal. (See Corollary 4.7.)
Our last convention in this section is
with the help of physical concepts, such as those of pressure and conduc-
tivity, misses the point. Firstly, although such events as nightmares and
feeling a sour taste are not describable in strictly physical terms, they are,
at least in principle, describable in neurophysiological terms. Secondly,
in ordinary life we may never care to describe mental states in neural
terms, just as we do not care to describe such ordinary physical events as
a car crash or the burning of a light bulb with the paraphernalia of physical
theory. It would not be worth our while to do so; we are satisfied knowing
that, if need be, such events, or some of their features, could be accounted
for in scientific terms. Remark 5 Pace Armstrong (1968), mind and brain
are not identical: there is no more brain-mind identity than there is lung-
respiration identity. In our version of the so-called identity theory, the
set of mental states is a subset of the states of the plastic neural system of
the animal: see Figure 4.3.
Organism states
Brain states
.. --;::.
F,
Fig. 4.3. Projection of the state space of a human being. Both axes represent neuro-
physiological properties; one of them is identical with a psychological property. The
arc of curve represents a mental process, such as composing a melody, which is only
partly conscious. One part of it is abnormal, e.g. consists of auditory halluci'lations.
142 CHAPTER 4
interactions are well defined only for concrete things (Vol. 3, Ch. 5, Def.
5.30). To speak of mind-body interaction is just as mistaken as to speak
of shape-body, or property-thing, or composition-system, or behavior-
animal interactions. However, it does make sense to speak of mental-
bodily interactions provided this expression is taken to abbreviate "inter-
actions among plastic (uncommitted) neural systems, on the one hand, and
committed neural systems or bodily systems that are not part of the CNS,
on the other". Example The interactions between cortical and subcortical
regions of the CNS, between sensory and motor areas, between ideational
neural systems and external receptors, between the visual cortex and the
pituitary, etc.
Paradoxically enough, while the dualist has no right to claim that mental
events influence nonmental bodily events (because he has no clear idea of
such influence), the monist is entitled to this opinion. Indeed, since mental
events are for him neural events of a kind, they can influence and even
cause other events in any subsystem of the same body because of the in-
tegrative action of the CNS. In short, because mental events are neural
events, and the causal relation is defined for pairs of events in concrete
things (Vol. 3, Ch. 6, Def. 6.23), we have
COROLLARY 4.6. Mental events can cause nonmental events in the same
body and conversely.
Consequently disturbances of nonmental (e.g. metabolic) biofunctions
may influence mental states and, conversely, mental events, such as acts
of will, may influence nonmental bodily states. This is what neurology,
neurochemistry, psychiatry, education, and advertising are all about.
Example 1 Yogis learn to regulate their heartbeats, intestinal contractions
and even rate of oxygen intake (cf. Miller, 1969). Example 2 A blow in the
head can "knock out the mind" erasing all memory of events immediately
preceding the accident. Example 3 By implanting electrodes in selected
sites of the brain, and sending electric currents through them, the neuro-
physiologist can control some mental states and, through them, elicit or
block behavior such as rage (cf. Delgado, 1969).
The most dramatic and revealing of all surgical interventions is of course
the sectioning of the bridge between the two cerebral hemispheres, or
corpus callosum (Sperry, 1964, 1966, 1977; Gazzaniga, 1967; Bogen,
1969). The split-brain patient behaves in many situations as if he were in
two minds - as forecast by Fechner in 1860. For example, while his left
hemisphere may wish to read, his right hemisphere may wish to go for a
MIND 143
walk, and so a conflict arises within one and the same skull. This is not
surprising in our framework: indeed, according to Postulate 4.3 and
Corollary 4.4, there is a single mind as long as there is a single plastic
supersystem. But if the latter is cut into two then two plastic systems em-
erge, each with its own mind or system of mental functions:
ing others. In sum, the mental can be contrplled, nay is being controlled,
in various ways precisely because it is not immaterial: if it were, it would
be immune to all attempts at manipulating it.
What holds for the physical control of the mind holds also, mutatis
mutandis, for the mental control of the body, so often cited as proof of the
superiority of Spirit over Matter. Consider three well known phenomena:
hunger, anger, and yoga. The process involving the hunger feelingis nor-
mally this. Sensors in the digestive tract send signals to the hypothalamus,
which in turn alerts the cortex. The motor cortex and the cerebellum con-
trol the muscales intervening in the search and reach for food, and in eat-
ing. The dualist thinks he can afford to skip all these links: he discards the
process ensuing in the cortex's wanting the body to eat, and speaks of the
direct action of the mind on the muscles. His account is far too simple -
and unscientific. Saying that X ate because he meant to eat or had the
intention of eating is just redescribing an ordinary situation in ordinary
language terms, not giving an explanation of the process. And yet such ac-
counts are as popular with ordinary language philosophers (e.g. Alston,
1974) as the account of behavior in terms of instincts was among ethologists
half a century ago.
Consider anger next. The dualist is satisfied with "explanations" like this
one: "She kicked him because she was angered by his words". The phy-
siological psychologist, on the other hand, analyzes this ordinary knowl-
edge account as a causal chain of bodily events: Hearing an utterance ~
Activation of psychons in the forebrain and of the adrenals ~ Action of
forebrain and adrenals on motor center ~ Movement of limb. Likewise,
when confronted with yoga exercises and biofeedback techniques for
moving things without the intervention of any muscles, the physiological
psychologist concludes that the autonomic nervous system is not as fully
autonomous as we used to think, but can be subjected at least in part to
control by cortical systems. There is no evidence for the action of an im-
material mind on the body and there can be no such evidence, for the
ghostly is inaccessible to science. Epicurus knew this twenty-two centuries
ago.
anywhere. So much so, that in the past the heart and the liver were as-
sumed to be the "seats" of the mind. However, as long as we remain on the
ordinary knowledge level we cannot refrain from citing evidence for the
thesis that the mind is coextensive with the body: e.g. when your foot hurts
you feel pain there, not in your brain. So, ordinary experience is ambigu-
ous: it supports now the thesis of the nonspatiality of the mental, now the
thesis that the mind can wander throughout the body. This suggests that
the nonspatiality thesis is confused, and that introspective evidence is
insufficient or even irrelevant. We should examine both.
The thesis of the nonspatiality of the mental contains two different
hypotheses. One is that mental events are unextended, the other that they
occur nowhere - except of course "in" the mind, which is allegedly imma-
terial and therefore unextended. The latter hypothesis implies the former
but not conversely. Indeed, mental events might be unextended yet occur
somewhere - say in the brain. Likewise a physical event that takes no
time occurs at some time or other. In our ontology there are no events in
themselves but only events in some thing or other (Vol. 3, Ch. 5). Hence
the question of the space "occupied" by an event is the problem of the ex-
tension of the changing thing. Thus not the firing of a neuron but the firing
neuron is spatially extended.
Events occur wherever the "eventing" (changing) things may be (Vol. 3,
Ch. 6, Sec. 4.2). In particular, mental events occur in some plastic neural
system or other. So, in principle and often in practice, the neurophysiol-
ogist can locate a thought process, e.g. by eliciting it with electrical stimu-
lation (Penfield and Rasmussen, 1950). Hence the ordinary language
philosopher's claim that "it makes no sense at all" to talk about mental
states and events as being located somewhere in the body has been experi-
mentally refuted. Mental events are just as localized as the corresponding
plastic neural systems. If the latter have a fixed location, so have the mental
events they engage in. And if the neural system in question happens to be
itinerant, the mental event itself will be itinerant - just as much as a light
wave. In short, mental events occur in the brain. However, the mind,
conceived of as the set of all mental events (Definition 4.14), is nowhere.
What about pain in limb, particularly a limb that has been amputated
(phantom limb)? The short answer is that, just as we see stars in the sky,
not in our brain, so we locate pains in several parts of the body, and some-
times either localization is wrong. A long answer is that the cortex of the
adult human being has a map of the entire body (the somatotopic represen-
tation). It is here that the signals coming from various parts of the body are
146 CHAPTER 4
decoded and located. If they come from the stump of an amputated limb,
the patient can feel them as originating in the missing part because he has
formed since childhood a comparatively unalterable map. (There are alter-
native neurophysiological explanations of phantom experience.)
In sum, mental processes are located wherever the corresponding
psychons are located; likewise they occur in time and take some time to
happen. This generalization from contemporary neurophysiology -
foreshadowed by the Hippocratic school- comes as no surprise for the
materialist, for whom the res cogitans is a res extensa, namely the brain.
What about ideas: are they too in spacetime? The answer to this am-
biguous question depends on the construal of 'idea'. If taken as idea-
tion processes, ideas are in the brains that happen to think them up - and
only there and at the time they are being thought. On the other hand the
so called product of any such process, i.e. the idea in itself, is nowhere in
spacetime because it does not exist by itself: we only feign it does. For
example, although thinking of the number 3 is a brain process, hence one
located in spacetime, the number 3 is nowhere because it is a fiction exist-
ing by convention or fiat, and this pretence does not include the property
of spatiotemporality. What holds for the number 3 holds for every other
idea - concept, proposition, or theory. In every case we abstract from the
neurophysiological properties of the concrete ideation process and come
up with a construct that, by convention, has only conceptual or ideal
properties. (Cf. Vol. I, Ch. 1.)
To sum up: Whereas the mental is in some head or other, ideas in them-
selves, having no being of their own, are nowhere.
Rule 4.1 invites either of the following strategies with regard to men-
talist predicates: elimination, definition, or deduction with the help of
neurophysiological concepts. Of these the second, i.e. definition (or iden-
tification), has been commended by the so-called identity theorists (e.g,
Smart, 1959). That is, they have proposed that, for any mentalist predicate
M, there is a neurophysiological predicate N such that M = N, where =
is the relation of ordinary (or strict) identity. Ordinary language philo-
sophers have claimed that this cannot be, for "it makes no sense" (Mal-
colm, 1964). The obvious rejoinder is that no scientific hypothesis
whatsoever makes sense to them, for they insist on remaining within the
confines of ordinary language, hence ordinary knowledge.
Another, related objection to the identity thesis, and one commended
as definitive by Popper (Popper and Eccles, 1977), is as follows (Kripke,
1971). If the identity is to be taken just as seriously as "Heat is the motion
of molecules", then it must be a necessary identity, in the sense that it must
hold in all possible worlds - whatever these may be. However, this cannot
be, because "it seems clearly possible" (to Kripke) that M (e.g. pain) could
exist without the corresponding brain state N, or that the latter could exist
without being felt as pain - whence the identity is contingent and there-
fore flimsy. Rejoinder: (a) scientists and science-oriented philosophers do
not waste their time speculating about (logically) possible worlds: they
want to explore the real world (Bunge, 1977a); (b) the difference between
necessary (or strict) identity and contingent identity does not occur in
ordinary logic or in science. In sum, the sophistic objection to materialism
holds no water.
Surely our Rule 4.1 is not a summary of the state of the art but rather a
programmatic injunction, i.e. one that can guide research. As such it can
be confirmed. And it should be discarded (or simply forgotten) if found
barren. But, far from barren, it is behind many a success in contemporary
psychology and neuroscience.
Note that Rule 4.1 is the weak not the strong reductionist thesis: we
favor the partial reduction (definition or deduction) of the mental to the
neurophysiological, not the elimination of its emergent properties (cf.
Bunge, 1977b, 1977c). Thus we do not state that a frog enjoys listening to
a Beethoven quartet, or that what a human feels when listening to it is no
different from a frog's sighting a fly. We only say that enjoying Beethoven
is a process in a human auditory system educated to face such experience.
The partial reducibility (definability or deducibility) of mentalist predi-
cates does not entail that every ordinary language sentence concerning
MIND 149
3. SENSATION TO VALUATION
o Ideation unit
Senscr
Primary
C(jtcal
II E~citiltion Other
o--..--{ ...._ _ _ _-\( Perceiving \~~~'==:Osensor(s)
Module ~ibition •
Projection
/P s,~ P
5,
s~~
P, 5,
(a) (b)
Fig. 4.5. Sensation and perception. (a) Same sensation elicits different perceptions
according to the plastic neural assembly it activates. (b) Different sensations give rise to
same perception by activating the same plastic neural system.
case the bodily events are caused by external events, and the maps are far
more numerous than the body schemata. In fact there is an entire atlas of
the external world, made up of dozens of maps - one for shapes, another
for colors, another for motion, and so on. Our assumption is
B,{K) = U B,-(x);
xEK
(All of the above definitions are understood to refer to a given stage in the
development of the individual or the evolution of the species. As an animal
develops its behavioral repertoire changes. Something similar holds for
the evolution of behavior.)
Our first assumption is that behavior, far from being primary, is deri-
vative:
THEOREM 4.1. No two animals behave in exactly the same way [even if
they are conspecific].
The assumption that a nervous system may contain a plastic subsystem
(Postulate 4.1), together with Postulate 4.9 and Definition 4.24, entails
158 CHAPTER 4
of the type" X eats (or sleeps, or mates) because X has an urge (or drive)
to eat (or sleep, or mate),' are not illuminating as long as the drive concept
remains an occult hypothetical construct. They become genuine explana-
tions the moment drives are construed in physiological terms, e.g. as
detection of imbalances (deficits or superavits) of certain physiological
variables such as sugar or noradrenalin levels in the blood. In this case
the hypothesis that behavior X is caused by drive Y can be put to the test,
namely by manipulating the imbalance and watching behavior changes.
The entire drive theory can be compressed into one definition and one
assumption:
DEFINITION 4.27. Let S be a set of items and b an animal. Further, let Z:b
be a partial order on S. Then the structure "Yb = (S, Z:b) is a value
system for b at a given time iff .
(i) b can detect any member of S and discriminate it from all other items
in S;
(ii) for any two members x and y of S, b either prefers x to y(x Z:b y)
or conversely (y Z:b x) or both (x "" bY) at the given time.
Remark 1 This is a comparative concept of value. A quantitative concept
is that elucidated by utility theory and occurring in decision theory. Re-
mark 2 Note the difference between the concept of psychovalue elucidated
by Definition 4.27 and that of biovalue introduced in Ch. 3, Sec.2.2. The
biovalue of a for b at t is the objective value that a possesses for b at t.
On the other hand the psychovalue of a for bat t is the value that b assigns
to a at t. These value assignments may be biologically mistaken, in which
case psychovalues conflict with biovalues. The psychophysical dualist
would regard this as a conflict between mind and body, while the monist
would regard it as a conflict between the brain or a subsystem of it, on
the one hand, and the rest of the body or a different brain subsystem, on
the other. In any case, since such conflicts do occur, we had better not
try to reduce one kind of value to the other even though both valuations
are biological processes. Remark 3 Value systems are known to change
for one and the same organism, not only in the course of its development
but also with circumstances. Therefore we incorporate this fact into our
hypothesis:
POSTULATE 4.13. All animals are equipped with a value system, and those
capable of learning can modify their value systems.
Finally we introduce the notion of choice. An animal presented with a
set of alternatives may abstain from choosing among them or may pick
one of them. If it tries out some of the options and finds out the hard way
which lead to the most valuable outcomes, the animal will be said to have
learned to choose. But in all cases there must be some freedom in order
for there to be genuine choice. I.e. the animal must be able to choose any
of the alternatives open to it, and some choices must be (at least initially)
mistaken (i.e. disvaluable) for there to be choice learning. Hence we make
MIND 161
4. RECALL TO KNOWLEDGE
POSTULATE 4.14 All animals have memory of some of their past states, and
none of all of them.
There is probably a large variety of memory systems or mechanisms.
162 CHAPTER 4
We may class them into fixed and itinerant. While invertebrates and lower
vertebrates have probably only fixed memory systems (i.e. systems with
constant neuronal composition), higher vertebrates have also itinerant
memory systems, i.e. systems that are formed just for the occasion. The
latter are particularly interesting in the case of man, where remembering
seems to consist not in the reactivation of some fixed neural circuit but
"in the fresh production or reconstruction of an item" (Bindra, 1976, p.
330). In any case, memorizing is not passive recording but a kind of acti-
vity of certain neural systems possibly involving the strengthening of
synaptic connectivities (Anderson, 1972, 1973). Nor do we "store" past
events in our memory, let alone in our mind: if we store anything at all it is
in our brain.
Some memory systems have the ability to learn. According to Defini-
tion 4.9, learning is the acquisition of new neural functions, whence it
involves plastic (not only wired-in) neural systems. This definition differs
from the usual behavioral definition of learning as modification of be-
havior in response to stimulation. The behaviorist definition is concerned
with the effects rather than the causes, and it could be applied to any thing,
for any thing will behave differently if stimulated differently.
To emphasize the idea that learning is a modification of neural activity,
we make
DEFINITION 4.31 The experience of an animal at a given time is the set of all
it has learned up until that time.
That is, the experience of an animal is its accumulated learning, i.e. the
collection of changes in its plastic neural systems. These include of course
nonperceptuallearning. On the other hand "learning" to avoid harmful
stimuli is not learning proper according to our definition, for it may well
MIND 163
4.3. Thinking
Forming concepts, propositions, problems, and directions are examples
of thinking; so are comparing and transforming concepts, propositions,
MIND 165
DEFINTION 4.36 Two thoughts are equivalent iff they consist in thinking of
the same constructs. I.e.
Oa(C) '" Ob(C') iff C = C' for any animals a and h.
It would have been desirable to define a construct as·a set of equivalent
neural events instead of presupposing the concept of a construct, for it
would root constructs on neural events while preserving their abstract
quality. Unfortunately we have been unable to find the proper equivalence
relation that would allow one to construct such a definition.
4.5. Creativity
All higher vertebrates are creative, and man is so superlatively. This may
be conceded while at the same time it is doubted whether the notion of
168 CHAPTER 4
POSTULATE 4.18 Every creative act is the activity, or an effect of the ac-
tivity, of newly formed neural systems.
Although creativity is maximal in humans, it is likely to be found among
all mammals and birds - if only the theological and empiricist prejudice
against the concept of creativity is overcome and the phenomenon is
investigated scientifically, particularly among young animals placed in
unusual circumstances. We make bold and assume that all higher verte-
brates are creative: .
POSTULATE 4.19 All animals endowed with plastic neural systems are
creative.
However, so far as we know only humans are absolutely creative, i.e.
capable of inventing or discovering certain "things" before any other ani-
mals. We shall use this assumption in defining humanness in Sec. 5.
What is the ontological status of the so-called creations of the human
mind? To obtain a sensible answer we must first distinguish the creative
MIND 169
creations and, far from being idle or merely entertaining, are an essential
part of modern culture. But it would be sheer animism to endow such
fictions with real autonomous existence and causal efficacy - the way
idealist philosophers from Plato on have been doing, or the way Popper
(1968, 1972, 1974, 1977) does. A differential equation does not exist by
itself and therefore is harmless; only thinking of it can have some effects.
In short, ideas in themselves are fictions and as such have no physical
existence.
5. SELF TO SOCIETY
POSTULATE 4.20 All animals are aware of some stimuli and some are also
conscious of some of their own brain processes.
MIND 171
DEFINITION 4.43 The consciousness of an animal b is the set of all the states
of the eNS of b in which b is conscious of some eNS process or other
in b.
In other words, consciousness of brain event x is direct knowledge of x.
(This is what introspection is all about.) Moreover we may hypothesize
that a conscious event is a brain activity consisting in monitoring (record-
ing, analyzing, controlling, or keeping track of) some other brain activity,
much as a voltmeter measures the electromotive force between two points
of an electrical network. I.e. we adopt
It is a well known law that, as learned tasks are repeated, they tend to
become automated, i.e. unconscious:
THEOREM 4.3 All animals capable of being in conscious st-ates are able to
perform free voluntary acts.
If consciousness is not exclusively human, neither is free will. And both
are legitimate subjects of scientific research.
(i) the personality of b is the functional system composed of all the motor
and mental functions of b;
(ii) a person is an animal endowed with a personality.
Remark 1 It follows that not only humans but all animals with mental
abilities are entitled to be called 'persons'. Remark 2 Personality is a
whole-body property: there is no special organ of personality, since it is
defined as the union of behavior and mentation. Remark 3 For the same
reason the destruction or ablation of considerable portions of the CNS,
as well as the amputation or the paralysis of limbs, ensue in personality
changes. Irreversible and massive destruction of the brain destroys per-
sonality altogether, in which case the animal ceases to be a person both
psychologically and legally. Remark 4 Learning shapes and often enriches
personality; it can even produce radical personality changes. Remark 5
People who undergo shattering experiences (war, concentration camp,
prison, or ideological conversion) may acquire whole new personalities.
Remark 6 In deep sleep or in coma we lose much if not all of our person-
ality: we become nonpersons or nearly so. (But of course when reawaken-
ing we recover - or rather reconstruct - our personality.) Remark 7 Since
the behavior and the mentation of animals endowed with plastic neural
systems depend partly on their environment, one and the same animal is
bound to exhibit different personalities in different environments - e.g.
a tyrant at home and meek at work. (For the various determinants of
personality see Gray, (1972b ).) Remark 8 There is no personal identity or
enduring personality any more than there is digestive or cardiovascular
identity: the behavior and the mentation of a higher vertebrate are more
changeable and vulnerable than any other bodily function. (For an early
criticism of the idea of personal identity see Hume, 1739, Part IV, Sec.
VI.) Remark 9 Since mind is not incorporeal stuff but a collection of brain
functions, and since there are no two strictly identical brains, "identical"
twins reared in the same family have different minds, hence are different
persons instead of constituting (as Shaffer, 1977, has claimed) a single mind
with two bodies. Remark 10 Since split brain patients have two minds
(Corollary 4.7), they also have a double personality, i.e. they are two
different persons united anatomically. Remark 11 If an· individual were
to get a new brain by transplant, the personalities of both donor and
recipient would disappear, and a third personality would emerge. Indeed,
the brain in its new encasement would receive new stimuli and fail to
receive some of the customary stimuli. And, because it would be con-
MIND 175
DEFINITION 4.51 If a and b are animals, then a and b cooperate with one
another iff the social behavior of each is valuable to the other or to a third
animal.
When cooperation concerns things (e.g. goods) Qf some kind, it is called
sharing, and participation when it concerns activities. More explicitly:
POSTULATE 4.24 A set of con specific animals forms a social system if and
only if each of them cooperates with some other members of the same set
[i.e. shares in the resources of the society or participates in some of its
activities].
The early evolutionists stressed competition at the expense of coopera-
tion, and both Freud and Lorenz regarded aggression as instinctive and
pervasive. We have learned since that the struggle for life involves both.
For example, flocking and the formation of colonies are effective defenses
against predation (Hamilton, 1971). Group grazing, hunting in packs, and
active defense involve cooperation - not to speak of higher forms of
sociality such as division of labor and games. No cooperation, no socio-
systems.
Only some forms of cooperation are purposeful, even fewer conscious.
Purposeful cooperation deserves a name of its own:
POSTULATE 4.25 In every social system (a) there are deviants in some re-
spects, and (b) some of the fellow members of the deviant components sub-
ject them to a pressure that is a monotonically increasing function of the
deviance.
In most vertebrate societies social structure has a rather weak reaction
on individual behavior. Indeed flocks and troops of various kinds split and
even dissolve altogether under altered environmental conditions. In hier-
archical primate societies, on the other hand, the over-all social structure
seems to be overpowering as long as the social system lasts. This makes for
social stability - as well as for the occasional rebellion. The deviant, under-
180 CHAPTER 4
privileged, or just young, are often outcast. In a few cases they may find a
more favorable habitat offering new opportunities and new challenges. In
other cases the deviants may join forces to attempt a restructuring of their
group. In any case, just as genic variety is a motor of biological evolution,
so social deviance is a motor of social evolution.
5.5. Communication
All social animals, and many which are not, can exchange information of
some kind, and some of them using a great variety of means. We shall
adopt some of the standard concepts used in describing animal communi-
cation (Morris, 1938; Smith, 1965). They are all included in
S. Further, call ff' = (S*, 0> the free semigroup on S [Le. the set of con-
catenates of members of S]. Then ff' is a language iff
(i) S* is a set of learned animal signals;
(ii) there is at least one animal capable of understanding some of the
messages carried by members of S*.
Remark 1 This definition allows for private languages such as those deaf
children can invent (Goldin-Meadow et al., 1977). Incidentally, Wittgen-
stein and his followers have denied that private languages could exist.
Remark 2 Chimpanzees can learn certain sign languages (Gardner and
Gardner, 1969). This finding has weakened the hypothesis that language is
exclusively human - weakened rather than refuted, for the chimpanzees
in question were taught a man-made language, namely the American Sign
Language or ASL, whereas humans create languages. Moreover, although
ASL has as much expressive power as any natural language, chimpanzees
cannot make full use of it because they cannot think all human thoughts.
Remark 3 By using borrowed languages, chimpanzees can express thoughts
of their own and thus refute the dualist's contention that it is "really senSe-
less to conjecture that animals may have thoughts" (Malcolm, 1973). We
are now sure, on the basis of experiments with teaching artificial languages
to chimpanzees, that animals can think. Remark 4 Because chimpanzees
can be trained to communicate not only by means of ASL but also with the
help of plastic chips (Premack, 1971) and via computers (Rumbaugh et al.,
1973), man can no longer be defined as the talking animal. On the other
hand we can evidently postulate that humans (and chimpanzees and
possibly other animals as well) have an innate linguistic ability or poten-
tiality or disposition. This is not to say that man is born with a knowledge
of certain basic structural features common to all natural languages. What
we are born with is the vocal tract, the Wernicke and Broca areas, and
subsidiary neural systems - as well as a social environment, carrier of a
tradition, that stimulates the acquisition and development of verbal
languages.
182 CHAPTER 4
In the next chapter we shall argue that every human society, no matter
how primitive or evolved, is composed of four main subsystems: kinship,
economy, culture, and polity. We submit that all four are already found in
nuce in some prehuman societies.
That prehuman societies have a kinship system, i.e. a social network
based on reproduction, seems obvious. Nor is there any doubt that some
have also an economic system, i.e. a social network based on work. As for
politics in the large sense of management of social activities, it is well
known than in certain animal societies law and order, as well as defense,
are well organized. (Cf. Gray Eaton (1976) for the case of the Japanese
macaque society.) Less well known, but true nonetheless is that some pre-
human societies have a culture of sorts, i.e. a system whose members en-
gage in social activities that are mental rather than biological, productive,
or J:lanagerial. The patterns of such activities can moreover be transmitted
throughout the community as well as handed over to the next generation -
i.e. traditions are formed, spread, and are kept. Such a transmission is ef-
fected by imitation and occasionally also by teaching, hence with the help
of signals. This has been found to be the case among Japanese macaques
(Kawai, 1965) and chimpanzees (Menzel et al., 1972). In both cases the
initiation of new traditions called for some bold individuals - creative
deviants - that sought newness rather than fearing it.
Because economic, cultural and political systems are already found in
prehuman societies, although in a primitive form, we should characterize
the corresponding concepts:
things before any other animals (Definition 4.41). We submit that this is
the case of all humans, from Homo erectus through habilis and sapiens
to sapiens sapiens, i.e. us. That is, we assume
6. CONCLUDING REMARKS
Wind is the motion of air: there is no wind apart from moving air. Like-
wise there is no metabolism over and above metabolizing systems, or
social change over and above social systems. In every factual science,
states are states of concrete entities, and events are changes of state of such
objects. It is only in traditional psychology and philosophy of mind that
talk of behavior independent of behaving animals, and of mentation along-
side or above the brain, is tolerated, nay encouraged by the dualist myth.
This separation between the mind and that which does the minding - be-
tween function and organ - has kept psychology and the philosophy of
mind far apart from neurobiology and has prevented a full utilization of
the scientific approach to the mind-body problem. The philosophy of
psychophysical dualism is responsible for the backward state of research
into that problem. Anyone interested in the progress of the science of
mind must therefore reject the prehistoric myth of the autonomous mind.
The rejection of psychophysical dualism does not force one to adopt
eliminative materialism in either of its versions - i.e. that mind and brain
are identical, that there is no mind, or that all things have mental capaci-
184 CHAPTER 4
SOCIETY
I. HUMAN SOCIETY
1.1. Definitions
Our Postulate 4.26 characterizes a human society as an animal society
188 CHAPTER 5
only occasionally and in any case not fashioned by them or at least not to
standardized patterns. Again, some birds and mammals play, and still
others have evolved coded sign systems (though not languages proper).
However, only humans organize (manage) some of their own activities,
and teach behavior patterns and feeling and thinking patterns, and create
novel objects (artifacts, works of art, organizations, words, etc.). In sum
only humans possess, socially if not individually, all ten properties listed
above.
The preceding characterization of human society, though sketchy,
suffices for the moment. We proceed to express it in terms of concepts
elucidated in the previous chapters:
that are neither individual nor societal, such as shyness and conformity
(or its dual, deviance). Take the latter, analyzed by Definition 4.54. In it
'F' represents a property of an individual, the average F a resultant pro-
perty of a sociosystem, and the standard deviation d of the distribution of
F values in the society represents an emergent property of the latter. (So
much so that if d is small we can say that the sociosystem is conformist.)
Our first hypothesis is that the concept of an emergent societal property
is not idle:
Fig.5.1. Three sectors: (a) school system (connected indirectly via school board) or
retail marketing system (connected indirectly via wholesale distributor); (b) manufactur-
ing system (where the inputs to some subsystems are outputs to others); (c) postal net-
work (controlled by central post office).
SOCIETY 193
1.3. institution
ous cult. We are interested here in yet another acceptation of the word
'institution', namely the one used in speaking of the Family, the Mail, the
Bureaucracy, or the Law.
Social philosophers are divided over the nature of institutions, hence
over the proper way of studying them. The two classical doctrines in this
field are holism and individualism. The holist regards an institution such
as the Law as a disembodied corpus of rules or laws that individuals en-
force, obey, or violate. According to this view the institution hovers above
the persons and moreover is superior to or more valuable than them.
No wonder holism, in some from or other, is inherent in totalitarian
ideologies. The individualist rejects such an abstraction and regards the
Law as the set of judges, lawyers and policemen - with total neglect of the
subjects of the Law, namely the common citizens. When confronted with
a sentence such as 'The Law compelled X to do Y', the individualist trans-
lates it into 'Judge (or policeman) U forced X to do Y'. But of course this
is a poor substitute for the original sentence, as the person U in charge of
enforcing the law behaves as he does because of what he is, namely a
member of a certain legal system with a code of legal procedure. Different
individuals might have behaved in roughly the same way. On the other
hand the same individuals, in a different legal system, might behave dif-
ferently. In conclusion, the legal system in question is neither the set in-
cluding the individuals U and X nor a set of supraindividual rules. Hence
although both holism and individualism have their points, neither is com-
petent to account for the behavior of the individual in society. Holism fails
because it postulates totalities transcending the components, and indi-
vidualism fails because it ignores the totalities.
Our systemist view is as follows. A legal system is a subsystem of some
society, i.e. it is a part of a human community set off by certain peculiar
social relations of the M type, such as those of judging and passing sen-
tence. And what is called 'the Law', with an awesome capital, is the col-
lection of all legal systems. In general, we shall construe an institution
as the set of all the F-sectors of all societies, regardless of their spatial and
temporal propinquity. Thus the collection of all school sectors is called
Education, the set of all trade unions Organized Labor, the set of all
postal systems the Mail, and so on. (Recall Definition 5.4 of an F-sector
and Definition 5.5 of an F-system.) We make then
DEFINITION 5.6 Let Z = {O"b 0"2, •.. , O"m} be a set of human societies, and
call Fik the kth sector of society 0" i' i.e.
SOCIETY 195
Undoubtedly such institutional rules can be rendered explicit for the most
part, just as the rules for doing properly some job or playing some game
can be formulated explicitly up to a point. But the writing down of a rule
does not confer autonomous existence upon it: this may only suggest,
wrongly, that it has one - perhaps because, when inscribed on a clay tablet,
a stone slab, or a papyrus, it can outlast its author. The upright and
merciless judge enforces rules just as the foreman has an operating manual
observed.
The actions abiding by the rules of an institutional game may be called
'rule directed' as long as we do not suggest thereby the idealist thesis that
rules exist independently of the rulers, the rule enforcers, and the ruled -
anymore than the laws of nature are separate from the things satisfying
them. The rules of social behavior are inherent in the very relationships
holding among the members of the system in question. (Much the same
holds for any deviations from that behavior: the delinquent may be an
outlaw in the sense that he does not observe the code of law, but he is as
much a member of the legal system as the judge - so much so that there is
no need for a legal system where there is no occasion for delinquency.) In
short the institutional rules, and also the patterns of breaking such rules
and the rules for punishing these violations, are incorporated into the very
network of social relationships that hold the system together. And it is not
"society" that sanctions them or punishes any deviations from them, as the
holist would have it, but rather some members of society, namely those
in charge of regulating the functioning of the social system. By conceiving
of institutional rules as patterns of the optimal or at least desired function-
ing of social systems, it becomes easier to understand how those rules ori-
ginated, where they fail, and in what respects they could be improved upon
if not repealed altogether.
So much for a preliminary conceptual framework. Let us now proceed
to analyze more closely some of the typical concepts of the social sci-
ences, starting with that of work.
2.1. Work
Work, one of the peculiarities of human behavior, is a kind of activity-
not just an action but a regular one. An activity, whether performed by
animal or by machine, may be defined as a patterned sequence of actions
SOCIETY 197
aiming at some preset goal. Thus, far from being erratic, an activity pro-
ceeds in such a way that, normally, the difference between attainment and
target is gradually decreased (Wiener, 1948, Ch. IV). See Figure 5.2. We
compress these remarks into
every entry of which is the set theoretic union of the partial entries with
the same indices.
Work is a special kind of activity~ namely one thought to be useful. We
define it as follows:
t
o
I
, initial Vilhl<:l
Time
Fig. 5.2. A system attaining asymptotically a goal state in either of two typical ways:
(a)monotonically (full line) or (b) by trial and error (dotted line). Note in the latter case
the phenomenon of overshooting (momentarily overreaching the goal).
198 CHAPTER 5
DEFINITION 5.8 If x is a system, then x works iff there exists a thing y and
an animal z (equal to or different from x) such that
(i) x exerts an activity on y (in a way other than by eating y) and by so
doing produces an appreciable change in the state of y, and
(ii) the thing y in the new state brought about by the activity of x is of
value to z.
We distinguish three kinds of work, namely labor (L), cultural work
(K), and managing (M). Labor aims at transforming nonhuman things
into more valuable things; cultural work aims ultimately at modifying
brain functions; and managing consists in controlling both labor and
cultural work. See Figure 5.3. Let us first examine labor.
2.2. Labor
We start off by making
c
U
A L
T
U
o R
E
Fig. 5.3. The three main kinds of work: labor, culture, and management of both. The
outputs of labor: material goods and services. The outputs of cultural work: cultural
goods (e.g. books) and services (e.g. education).
SOCIETY 199
DEFINITION 5.11 Let 0' be a society with composition 'f?(0') and social rela-
tions in S, and call K c .9"(0') a proper subset of the social structure of 0'.
Then
(i) the element K j of K, for 1 ::;; j ::;; s, is called the jth kind of cultural
(or secondary) work done by members of p iff K j is an activity capable of
evoking feelings or thoughts, or supplying ideas intervening in the primary
production of 0' ;
(ii) the subset 2Pj c 'f?(0') occurring in the domain of the cultural work
relation K j , is called thejth set of cultural (or secondary) producers of 0';
SOCIETY 201
(iii) the union 2p = UlP j is called the cultural (or secondary) manpower
of 0".
Remark 1 Although cultural work aims at changing the way people feel
or think, K is not disjoint from T. Indeed, all cultural work involves trans-
forming some thing or other, if only the state of the surrounding air.
Moreover there are useful works of art, such as artistically designed pieces
of furniture, as well as useful works of science, such as scientifically con-
trolled instruments. Remark 2 According to clause (i), writing a book that
may be published is doing a spot of cultural work; even writing or per-
forming a piece of rock "music" is perpetrating some cultural work. Not so
engaging in praying, dreaming, or transcendental meditation - which is
sleeping anyhow. Likewise drawing the blueprint for a house or a machine
or the pattern for a dress is engaging in cultural work, whereas drawing
exclusively for one's pleasure, or humming a melody, are not. To be cul-
tural an activity must be social or socializable as well as artistic or intel-
lectual.
We shall return to the matter of cultural work in Sec.3.2. Let us now
characterize the relations aimed at influencing social behavior in an im-
mediate way, rather than through the subtle medium of culture:
POSTULATE 5.4 Every human society has both a cultural manpower (i.e.
2p =F 0) and a managerial one (i.e. 3p =F 0).
From this platitude it follows that, under risk of destroying society,
it is impossible to do entirely without either the cultural force or the man-
agerial force. What one should aim for is maximizing the harmonious
cooperation among all three work forces, i.e. easing the tensions and there-
fore eliminating the need for keeping repression forces. In particular,
rather than try quixotically to eliminate management one should strive
to make it work in the interest of material and cultural production rather
than in its own interest. Such a blending of management with productive
forces obtains in a well run modern industrial plant, marketing system,
communications network, hospital, or school- in short wherever some
people work according to the canons of advanced (scientific) technology.
Such sociosystems deserve a name of their own:
Fig. 5.4. The main subsystems of any society. Outer circle: whole society. Inner con-
centric circle: kinship system. Eccentric circles: economy (gathering, hunting, fishing,
manufacturing, trading, financing); culture (artistic, technological, scientific, humanistic
creation and their diffusion); and polity (government of all kinds, nongovernmental
political activity, and the military). The overlaps suggest that some persons participate
in two or even three systems. E.g. a scientist belongs to the cultural system by virtue of
his research and teaching; to the economic system if he writes textbooks; and to the
political system if he acts as consultant to the government.
204 CHAPTER 5
DEFINITION 5.15 Let 0' be a society with labor force IP, cultural force
2p, and managerial force 3P. Further, let the subscripts L, K'and
k
G identify
whatever is associated with primary production, cultural production, and
political administration respectively. Then
(i) the subsystem of 0' represented by
Fig. 5.5. Example ofinternational economic and political relations between two nations
IT, and IT,. Note the two intranational relations besides the four international ones: the
former are indicated by dotted lines.
r r
MANPOWER _ _~ PRODUCTION
GOODS &
RAW SERVICES
MARKETING
MATERIALS
WASTE
ENERGY - - FINANCING PRODUCTS
etc ...
Fig. 5.6. Schema of the economic system. Inputs: natural resources (raw materials and
energy) and human resources. Outputs: goods, services, and waste products. Some of the
outputs are fed back into the system (e.g. as tools). Controls: culture (mainly techniques
and advanced technology), and politics.
210 CHAPTER 5
recall the various items entering the production, servicing, use, and circula-
tion of artifacts. They are chiefly the following: (a) the idea or design of
the artifacts; (b) the idea of the possible use(s) of the artifact; (c) a plan or
design for the manufacture of the artifact; (d) the raw material and the
energy required to produce the artifact, (e) the actual production process;
(f) the marketing of the finished product; (g) the control of the entire
process. Of all these seven ingredients only one, the fourth, is supplied by
nature. Because all the others are ultimately in the hands or brains of
people, the artiphysis is totally at their mercy. Men, or similar social
animals, create it and can destroy it. Artifacts have no will of their own
and, indeed, no existence of their own: even the most sophisticated among
them are but tools. And technology has no autonomy either: it is culti-
vated only in societies that care for it. A band of technoclasts could kill
every technologist and destroy every artifact on Earth in a few days. So, to
claim that technology and the set of artifacts have a life and a momentum
of their own, that they cannot be stopped, is bizarre to say the least.
We summarize the foregoing in
POSTULATE 5.6 (i) Every human society has members able to design, manu-
facture, repair, or use artifacts; (ii) every nonliving artifact [tool or ma-
chine] serves only the purpose(s) of some animal or other.
It follows that no machine, not even a highly evolved robot, has a pur-
pose of its own: whatever the goal seeking mechanisms built into it, they
have been designed and ultimately manufactured by rational beings in
their own interests. Hence machines stand no chance of taking over
unless of course programmed to do so. It follows also that, when a non-
living artifact imitates rational animals in certain regards, it does so by
proxy not spontaneously. A further consequence is that artifacts can nei-
ther evaluate nor think on their own. That they do not evaluate is obvious
since they have no goals other than those built into them by their de-
signers. (This vicarious evaluation capacity may reside in the hardware, as
when a machine is designed to seek or to avoid certain stimuli or else to
rank them. Or in the software, as when a machine is programmed to weigh
apples or alternative routes for mail delivery. In either case delegated eva-
luation may trigger a vicarious decision - e.g. that of discarding an under-
weight apple or an inconvenient route. When speaking of 'the vicarious
abilities of machines' the technologist underlines 'abilities' whereas the
philosopher emphasizes 'vicarious' - unless of course he takes the former's
word.)
SOCIETY 211
As for our claim that machines cannot think, it is quite obvious in the
light of the psychobiological view on thought as an activity of certain
neural systems (Definition 4.35). Here again machines, unlike rational
beings, have no purposes - in particular cognitive goals - of their own, so
they can properly be said to act by delegation. Surely machines can help
us solve certain problems, but they cannot discover any because formulat-
ing a problem is not a rule directed activity and, as long as we do not
master a rule, we cannot design or program a computer to follow such a
rule. (This is also why there can be no self-programming machines: if a
machine were to have problems of its own we would quickly scrap it.) Also,
machines can be used to process the assumptions of certain theories, but
they do not help create them - again for the same reason, namely because
theory construction is not a rule directed activity but a thoroughly creative
one. In short tools, however marvelous, are nothing but tools.
this cultural supersystem is much less tightly knit or integrated. Thus the
chemical engineer working in an oil refinery is more strongly attached to
the economic system than to the cultural system stricto sensu - unless of
course he is a research chemist, in which case he will belong to both the
economic and the cultural systems, and may accordingly be torn by a
conflict of loyalties. In either case the input-output schema of the cul-
tural system is sketched in Figure 5.7.
This schema serves as a reminder that, whether in the narrow or in the
comprehensive construal, a cultural system is a concrete system - not just
"a body of meanings, values and rules". More precisely, a culture is a
concrete system composed of living rational beings engaged in activities
of various kinds, some of which go beyond the call of biological need, and
all of which are social because they draw from tradition and influence
social behavior. I.e. the creation and the diffusion of culture are social
processes, shaped by society and in turn having an influence on the econo-
my and the polity.
Our concept of a culture as a concrete system is at variance with the
idealistic notion of a disembodied system of ideas, values, and rules, which
is the concept favored by most anthropologists. Actually this other notion
is that of a type of culture rather than of a particular culture. Thus where a
cultural system (in our sense) includes certain actions, such as writing and
inventing, a culture in the anthropological sense includes only the patterns
of such actions - e.g. the rules that are supposed to regulate those actions.
Nor does our concept of culture coincide with that of an ideal superstruc-
ture - made up of ideas, values, and norms - mounted on the material
structure (economy and polity). Unlike these dualistic concepts of culture,
ours is monistic.
ECONOMY POLITY
CULTURAL
MANPOWER
ARTIFACTS
ENERG
=1 }
\ j,I
~ENCULTURf;,D
CULT~RAL
WASTE
Pt:OPlE
ARTIFACTS
~-----'
Fig. 5.7. Input-output model of the cultural system. Inputs: manpower (pivotal and
auxiliary), artifacts (paint brushes, paper, pencils, books, archives, etc.), and energy
(light, heat, etc.). Outputs: cultural artifacts (musical scores, sculptures, journals, etc.).
people transformed by the latter. and waste products. Some of the outputs are fed back
into the system. Controls: economy and polity.
SOCIETY 213
Fig. 5.8. The contemporary cultural system and its subsystems. Each box represents
both creation and diffusion. The arrows represent both information flows and controls.
Art and ideology, which have the most numerous memberships in traditional societies,
have the fewest information links with the remaining subcultures. (From Bunge, 1977g.)
214 CHAPTER 5
not just oddity - and is free to develop it without too much interference
from the economic and political systems. A declining culture is one that
ceases to value discovery and invention, and takes refuge in lamaistic
repetition or in withdrawal from reality. The management of a cultural
system - the politics of culture - can encourage creativity, channel it, or
nip every creative bud. A dead culture is still a culture and one that can
remain in that state for a long time. But because of the functional rela-
tionships among the three main artificial subsystems of every society, any
major change in the economy or in the polity is bound to have cultural
repercussions.
The cultural system, then, coevolves with the economic and political
systems. However, the modes of evolution of these subsystems may be
quite different, and neither is merely an image of the other. Thus cultural
growth is compatible with zero economic growth and political stagnation,
provided the political system does not interfere destructively with cultural
evolution. Moreover, while economic development is limited by natural
resources, there are no such limits to cultural growth: the more we know
the more fresh problems we can pose and solve.
Yet there are of course economic and political constraints on cultural
evolution. For one thing any given society can support only so many pro-
fessional composers, mathematicians, philosophers, and so on. But at
least here, unlike the case of the economy, a solution is in sight, namely
automation and the amateur pursuit of cultural interests. In fact the form-
er is freeing many hands and brains from routine work, and the good so-
ciety should turn that added freedom into leisure and creation instead of
unemployment. A revaluation of culture should encourage people to en-
gage in cultural work for the fun of it rather than primarily as a source of
income. Everyone should be able to master whatever cultural skill he or
she likes, and to practice it during the increasingly long leisure hours
rendered possible by automation. This need not encourage substandard
cultural production. On the contrary, it is the increasing substitution of
extrinsic motivation (pay, promotion, fame) for intrinsic motivation that
is lowering the quality of cultural work. (For the negative effects of ex-
trinsic rewards on children see Lepper et at., 1973.)
To conclude. Our view of culture is at variance with idealism ("The
spirit is the driving force") as well as with economism ("Culture is but a
byproduct of the economy"). If every culture is a subsystem of a society
then it has its own dynamics, hence some degree of autonomy, and it also
interacts with the other subsystems, namely the economy and the polity.
SOCIETY 215
Therefore it is neither totally free and omnipotent nor totally enslaved and
powerless. Just as some members of the economic system exercise econo-
mic power, and some members of the political system wield political
power, so some members of the cultural system exercise cultural power,
particularly if they are entrenched in certain cultural subsystems, whether
private or governmental. For example, the school system of a community
exerts some cultural influence on all the inhabitants of the community-
often as strong an influence as that exerted in the past by organized religion.
This influence is not restricted to purely cultural matters. A creative cul-
tural organization may study, discuss, and even propose and circulate
blueprints for economic and political action. Such proposals do not
amount to action but may arouse and guide it. After all people are being
moved by thinking individuals all the time. Why let obsolete ideologies do
all the pushing when it is possible to draw rational and feasible blueprints -
i.e. to ground social action on the social sciences?
difference between the objective and the subjective conditions for a rev-
olution. Let us take a quick look at two key concepts in political psychol-
ogy: those of adhesiveness, and feeling of unfairness.
Whether or not a society is highly cohesive is of less consequence to its
"survival" than the way that cohesiveness is "perceived". A social fabric
with a low actual cohesiveness may be "seen" as highly cohesive (because
e.g. of the repression of dissent); and conversely a highly cohesive society
may be "seen" not to be so (because e.g. of a vocal nonconformist minority).
The key concept here is that of attachment of an individual to a group or
system. We may say that an individual feels attached to a social group or
a social systerri if he believes that the benefits (material, cultural, emotion-
al, etc.) he derives from sticking to it are greater than the price he has to
pay for doing so. In other words, the adherence or attachment of individual
x to a given group or system Gi is A,{x) = B,{x) - C,{x), where Band C
are the benefit and cost functions respectively. There is positive attachment
to Gi if the benefits outweigh the costs (i.e. Ai(X) > 0). Otherwise there is
either rejection (A,{x) < 0) by x of Gi or indifference (A,.(x) = 0) of x to
Gi . The total adherence or attachment of individual x to a society a formed
by n social groups Gi is
n
A(x) = 1: A,{x),
i=l
the equity or inequity of his station in society and his evaluation of the
equity (or inequity) ideal. The imbalance between the two will contribute
to his attachment or disaffection to the prevailing social structure. There-
fore the social scientist is interested not only in actual equity but also in
perceived equity, desired equity, and desirable or feasible equity. Let us
take a quick look at these concepts.
Anything that can be distributed or allocated, from goods and services
to leisure hours to political participation, can be assigned with various
degrees of fairness or equity. Now, what is fair in one society may be un-
fair,in another. Thus whereas in certain societies an average of 10 school-
ing years would be fair for all, in others this would constitute too heavy a
burden, hence be unfair to the majority, and in still others it would be in-
sufficient, hence just as unfair to the majority. This notion of an objectively
desirable value of a property (such as the number of schooling years)
must be distinguished from that of a subjectively desired value (desider-
atum). Whereas the latter may be unrealistic and incompatible with fair-
ness, the former should in principle be determinable by studying the state
of development and resources of the society. Such an objectively deter-
minable value (or distribution, as the case may be) we shall call an
equity goal, in contrast with the multiple equity desiderata.
We may assume that for each feature or variable there is one measure of
equity or, dually, inequity. Whatever the feature may be, the objectivy
inequity is some function f of the difference between the actual value X
of the variable and its goal value G, i.e. i(X) = f(X - G). Example 1 The
variable has a single value at a given time, as is the case with the mean
number of schooling years. In this case a suitable inequity indicator is
just the difference between G and X divided by the greater of the two.
Example 2 The feature in question is a distribution, such as that of arable
land among the families of a society. If the goal is a straight line (e.g. the
area ofland is proportional to the number of families), then the Gini index
is a suitable measure of inequity: see Figure 5.10. Example 3 The goal
distribution is symmetric - as is the case of the distribution of working
hours vs. age. In this case inequity equals skewness, of which there are
vanous measures.
The acuteness of the perception of inequity, as well as the desire for
equity, depend on a number of factors in addition to actual inequity.
One of them is social mobility. In a primitive society equity is taken for
granted, whereas in a highly stratified society inequity may go just as un-
SOCIETY 219
Fig. 5.10. Inequality in the distribution of wealth. The Gini index is the difference
between the area under the straight line (equality) and the area under the actual curve.
questioned for being sanctified by ideology. On the other hand where social
mobility is appreciable many persons feel status differences as iniquitous
and also avoidable either by individual or group action. If the former
proves ineffectual, the time is ripe for dissent and even sedition. We may
then adopt the following version of the Aristotle-de Tocqueville-Marx
law: Extreme inequity ends up by being perceived and negatively evalu-
ated, that diminishes the attachment to the social structure, and this in
turn leads to either indifference or active rebellion - in either case to a
deterioration of the political system.
Needless to say, inequity may be economic, cultural, or political. In-
equity felt in just one respect is neither necessary nor sufficient for a group
to rebel against the social fabric. Any group that fits fairly well in the
prevailing social (economic, cultural, and political) order - i.e. that is
not marginal - is likely to be conservative even though that group may
not be the principal beneficiary of that order. What do seem to be neces-
sary and sufficient for members of a social group to rebel against the pre-
vailing social order are (a) to feel being discriminated against in all three
respects (economically, culturally, and politically), (b) to feel free from the
traditions attached to the given social order (i.e. to be cultural not just
economic or political outcasts), and (c) to be able to organize themselves
into a system.
Finally we come to the notion of freedom. There are as many kinds of
freedom as there are kinds of social behavior. Since human social activi-
ties occur in either of the three main artificial subsystems of society, we
may group the species of freedom into three genera: economic, cultural,
and political. Freedom to work (or to loaf) is economic freedom; freedom
to learn is cultural freedom - and so is the freedom to remain ignorant;
and freedom to participate in making decisions affecting the entire com-
220 CHAPTER 5
POSTULATE 5.7 Every person and every society are free in some respects,
none in all.
A completely free society - i.e. one in which every person were totally
free - would be impossible, because the exercise of one person's freedoms
is limited by that of his neighbors'. And a thoroughly enslaved society
would be just as impossible, because every person needs to exercise some
SOCIETY 221
~ SOCIAL STRUCTURE
those on which all the others depend - such as the kinship relations and the
relations of economic, cultural, and political power. We may characterize
them as follows:
DEFINITION 5.17 A root (or basic) social relation is a social relation that is
(i) bonding,
(ii) person to person (rather than system to system), and
(iii) an interaction (rather than a one-sided action).
We assume that all social relations are either basic or generated by
relations of this kind:
such that, for any components a and b of (1, if Xab then there are compo-
nents c and d of (1 such that Ycd, where c and d are possibly equal to a
and b respectively;
(ii) for any (bonding or nonbonding) system-to-system relation U in the
family of subsystems of (1, there is a bonding person-to-person relation V
in the structure of (1 such that, for any subsystems x and y of (1, if Uxy,
then there is at least one component a of x, and another b of y, such that
Vab.
Example 1 Slave owners a and b, living in different places and times
(hence not bonded), are socially equivalent ( '" ) by virtue of being masters
(M) of slaves (S): a '" b = dj (3x) (3y) (Sx & Sy & Max & Mby). Example
2 Factory a supplies concern b with merchandise (or goods) c, or Sabc for
short. This boils down to the following: members of a produce c, and
members of a (the same or different) deliver c to members of b. In obvious
symbols: Sabc = dj (3x) (3y) (3z) (x E <c(a) & y E <C(a) & Z E <C(b) & Pxc
& Dycz).
Social bonds may in turn be classed into voluntary and involuntary.
Voluntary relations are those the individual can enter by choice - e.g.
friendship, participation in community activities, and even - in industrial
societies in times of full employment - job relations. Examples of involun-
tary social ties are the filial relation, the relation of being trained at a
nursery school, the prisoner-captor relation, and the bondages of slavery
and serfdom. We assume that, just as bonds determine nonbonding rela-
tions, so involuntary ties determine some of the voluntary ones. For ex-
ample, being a prisoner, a serf, or a slave, determines most of the other
SOCIETY 223
relations a person can bear to other individuals. The mere existence and
the basic character of involuntary relations suffices to brush aside the
philosophy of voluntarism, according to which the individual will is the
source of all behavior. The truth lies between the extremes of voluntarism
and total determinism: man can act of his own free will (Theorem 4.3) but
not always (Postulate 5.7).
Of all the involuntary social bonds family relations are the most basic
because they are determined by reproduction, which is a biofunction. The
family is thus situated at the biosocial interface. The root of the family is
then the div-ision of the sexes. In other words, the same-sex relation ~ s
splits the total membership '?f(o-) of a society 0- into two disjoint sets:
males (M) and females (F). In short, '?f(o-)/ ~ s = {M, F}. A human family
of the nuclear or restricted kind may be defined as an aggregate of human
coresidents among whom the relations of mate (or spouse) or of parent-
hood obtain. The relation f.-l of mate (or rather its graph) is included in the
cartesian product M x F. The relation 'fr of being parent of somebody is a
binary relation in '?f(o-). And the relation p of coresidence too may be
construed as a binary relation in the same set. This is all we need to form
DEFINITION 5.18 Let 0- be a society with sexual division '?f(o-)/ ~., mate
relationship f.-l, parenthood bond 'fr, and coresidence tie p. The nuclear
family F(b) of the component b of 0- is the sociosystem constituted by the
mate(s) and the children of b who live with b:
'?f(F(b» = {x E '?f(o-) I (f.-lxb V 'frbx) & pxb},
rff(F(b» :::J rff(o-) & rff(F(b» n '?f(o-) =l= 0,
f/'(F(b» = S U T, with S = {f.-l, 'fr, pl·
A member of a nuclear family may have more than one spouse (as in the
cases of bands, polygamy, and polyandry) provided all his or her mates
are coresident. But each person belongs to exactly one nuclear family. (A
childless widow constitutes a one person nuclear family but may belong to
an extended family.) The set of nuclear families of a society is the compo-
sition of the nuclear kinship system of the society. That set is split into
disjoint sets by the relation of belonging to the same family. The import-
ance of this relation and of the partition it induces justifies making
224 CHAPTER 5
For any x, y E 'I&'(U), X'" f Y =df (3z) (z E ~(U) &X E 'I&'(F(z» &
Y E 'I&'(F(z»};
(ii) the nuclear kinship system of u is the subsystem whose members are
in some family or other, so that its membership is
I
'l&'iu) = {x E 'I&'(u) (3y) (y E 'I&'(u) & x E 'I&'(F(y»};
(iii) the family structure of u is the set of all the nuclear families of u,
i.e. the partition of'l&'iO') by the relation'" f of family equivalence:
X EXECUTIVES (Managers)
OVERSEER
WORKINGMEN
(iv) the class structure of (J is the family of all partial class structures of
(J, i.e.
I
C«(J) = {,qlIA«(J) A = economy, culture, or polity}.
If the quotient of ~«(J) by '" A fails to partition the composition of (J
into disjoint subsets, then there is no A-control in the society, and thus
the society is A-classless. In other words, (J is classless in the respect A
iff ~«(J)/ '" A = {~«(J)}. It follows that an A-classless society is possible
only if each of its components participates extensively in activities of
kind A. The problem of participation will be addressed in the next sub-
section.
(ii) the degree of marginality (or alienation) of group Spq in 0" is the
complement to unity of its degree of participation, i.e.
p.(Spq, 0", "') = 1 - n(Spq,O", "');
(iii) the degree of overall participation in 0" is
From the findings of social scientists we infer that all societies are dif-
ferentiated, though none maximally, and that there is some measure of
participation in all. That is, we propose
POSTULATE 5.9 For every society a, I"" I> 1, and 0 < o(a, "") < 1,
o< n(a, "") < 1.
Every society is more or less cohesive: were it not, it would not be a
system. Social cohesion is of course one of the emergent societal proper-
ties: it is a systemic property not possessed by any of the individuals
composing a society. Cohesion can be spontaneous, forced, or a com-
bination of the two. Spontaneous cohesion, the one that does not result
from coercion, is undoubtedly the more important of the three. When
strong it ensures the stability of the social fabric; when weak it calls for
either structural change or the violent repression of any attempt to effect
structural changes.
In Ch. 4 we assumed that cohesion derives from sharing (of things and
services) and participation (in activities): recall Postulate 4.24. We shall
presently discuss a measure of spontaneous cohesion based on that as-
sumption (Bunge and Garcia-Sucre, 1976). No participation at all, i.e.
extreme segregation, results in instability, which ends up in fragmentation.
Yet maximal participation has a similar dissolving effect, for a society of
equals in every respect is a collection of competitors in no need of mutual
assistance, i.e. in no need of belonging to the society. In other words, if
every individual is either nothing or everything in a society, i.e. utterly
marginal or ubiquitous, then cohesion is minimal. Maximal cohesion must
then occur somewhere in between, when each individual shares several,
though by no means all, cells with fellow members. In sum, cohesion must
result from moderate participation. We proceed to formalize this intui-
tive idea.
The more members two social groups share the stronger they are
bonded. However, these bonds operate only if the groups are different,
in particular functionally complementary to one another. More precisely,
the effective bonding between two social groups A and B is the numerosity
of their intersection times that of their symmetric difference, i.e. IA n BI·
IA 6. BI· A low degree of differentiation (i.e. a smalllA 6. B Ivalue) may
be compensated for by a high degree of participation (i.e. a large overlap
value IA n BD and conversely. Certainly differentiation by itself does not
232 CHAPTER 5
and fictions, such as ideas in themselves (i.e. apart from ideating brains).
Here we shall note only how convenient it would be for all oppressors to
be able to persuade their subjects that their bondage belongs in the world
offairy tales and mathematical formulas, so they could not possibly suffer
from it, let alone rebel against it.
Sociologists, who study social relations and social structures, and poli-
ticians, who attempt to control them, cannot take seriously the claim that
social bonds and the social organizations they constitute are fictions. They
regard them as just as real as physical or chemical bonds. Hence, when
intent on changing them they resort to social action rather than incanta-
tion.
5. SOCIAL CHANGE
DEFINITION 5.29 Let a be a society with social structure matrix IISikll and
corresponding population density matrix
D(a, t) = (1IN(a, t»·IINik(a, t)ll.
The net structural change in a over the time interval [tl> t21 is
Example At tl> a is divided into two social classes: masters (M) and
slaves (S). At time t2 > II> a is leveled to a society of free men (F) with
conservation of the total population. (I.e. either the revolution has been
bloodless or the losses in lives have been compensated for by births or
immigration.) The net class change (or change in class structure) is then
POSTULATE 5.10 Let D(a, t1) and D(a, t2) represent the social structure
of a society a at times t1 and t2 respectively. Then there exists an operator
E(a; tb t2) such that
D(a, t2) = E(a; tl> t2)D(a, t1)
representing the mode of change of a over the period [t1' t2] and called the
social evolution operator for that period.
This is a substantive assumption that may be found empirically want-
ing, particularly over a long period: the actual social evolution may well
be nonlinear for all we know. However, we may assume that the postulate
holds at least for short periods.
From the above postulate and Definition 5.29 it follows that the net
structural change is
SOCIETY 237
where I is a suitable unit matrix. Given the initial state D(a, t 1) and the
mode of evolution E(a; tl> t2), the final structure, i.e. D(a, t2), remains
uniquely determined. This is the direct or forecast problem. The inverse
or explanation problem is this: Given the values of the density matrix at
two different times, e.g. now and one decade ago, find the mode of
evolution, i.e. compute the evolution operator. Being an inverse problem,
in general it lacks a unique solution and is much harder than the forecast
problem. In fact, consider again the example studied a moment ago. The
following evolution operators, among others, account for the same struc-
tural change:
-1 o 0 -1 o
E= 0 -1 0 and E = 0 -1
1 0
It can be shown in general that the following holds (Bunge, I 974d) :
THEOREM 5.2 The relative growth of any cell in a column of the social
structure matrix occurs at the expense of the relative decay of some other
cell(s) in the same column.
For example, if a column consists of just two cells and one of them grows
exponentially, then the other must decay exponentially. Thus all growth
has to be paid for by some decline.
Finally, note that social change is the likelier the more heterogeneous
or varied a society is, and this if only because the more social cells there
are, the more numerous the possible transitions among them. For the same
reason, the more homogeneous a society the less changeable it is.
5.2. History
DEFINITION 5.30 Let a be a society with social matrix IISikl1 and popula-
tion matrix IINikll. Then
238 CHAPTER 5
(i) (7 remains stationary (or in a state of static equilibrium) in the kth re-
spect (where 1 ~ k ~ n) over the period [0, 'r] iff Nik(t) = constant for all
1 ~ i ~ n and all t E [0, 'r];
(ii) (7 is stable (or in a state of dynamical equilibrium) in the kth respect
(1 ~ k ~ n) over the period [0, 'r] iff all the cells in the kth column of
IISikl1 expand or shrink at the same rate:
Nik(t) = at with a E R+ and a =!= I for all I ~ i ~ m;
11
0
Iz
0
0 0
0 0
Nl
Nz
=
j,N,;
hN, I'
0 0 13 14 N3 13 N 3 + 14 N 41
0 0 0 0 IN4 o i
V = 1 - (1IN(O)) ~i.t;N;(O).
This index equals 0 for peaceful revolutions, to for those that decimate
each social stratum, and t for those that take the lives of half the popula-
tion in each group.
So far we have drawn no distinction between spontaneous and planned
social change - hence between descriptive models and prescriptive ones.
But all contemporary societies, even those with economic free enterprise,
are sUbjected to planning to some degree or other, and it may be argued
that they need more (and better) of the same if they are to survive. Take
the rather obvious case of population growth. In the absence of any social
or cultural constraints, and pretending that there are no environmental
strictures either - in particular no interpopulational competition - the
expected population size satisfies the rate law
6. CONCLUDING REMARKS
In this, the last chapter, we shall generalize some of the results of the
previous chapters. We shall thus come up with the core of a systemic
world view. This Weltanschauung is a continuation of certain philosophical
traditions and is also in tune with contemporary science. But it should
not be mistaken for the popular "systems philosophy," a new version of
holism according to which every thing is a system (false) and the patterns
of being and becoming are basically the same at all levels (false). Our
systemist philosophy is neither holistic nor atomistic: it acknowledges
the variety of properties, kinds and patterns found in the world and, by
using certain elementary formal tools, it avoids the obscurities of tra-
ditional philosophy. Ours is, in short, a kind of scientific ontology. (Cf.
Vol. 3, Introduction.) Let us review some of its assumptions.
POSTULATE 6.3 The universe is a system - namely the system such that
every other thing is a component of it.
245
246 CHAPTER 6
POSTULATE 6.4 At the present stage of the evolution of the universe there
are five system genera: physical, chemical, biological, social, and technical.
I.e. the family of system genera is g = {S!> S2, S3, S4, Ss}.
The cautious clause about the present is a reminder that the higher
kinds have not always existed (so far as we know), and may either become
extinct or give rise to further system genera. Thus the exploration of
outer space might reveal the existence of further system genera, and a
nuclear holocaust would wipe out S3, S4 and Ss on our planet.
kinds, such as protons, there is little doubt that things of new kinds come
into being every so often. I.e. at least part of the marvelous variety of the
world is a product of change. Now, there are two main kinds of change:
quantitative (e.g. motion) and qualitative (e.g. chemical combination).
In the former case all of the axes of the state space of the thing are used;
in the latter some new axes pop up and others drop out as change pro-
ceeds. However, even quantitative change may end up in qualitative
newness, as in the case of high speed collisions of atomic nuclei, the
reshuffling of the components of a system, and growth by accretion or
multiplication. Let us take a look at reshuffling and accretion, which
would seem to be unlikely emergence mechanisms.
A reorganization or restructuring is of course a change in the structure
of a system, i.e. a change modifying some of the links among the system
components even though it may not alter appreciably the intrinsic pro-
perties of the said components. Two simple examples are isomeric transi-
tions (in an atomic nucleus or in a molecule) and institutional reorganiza-
tions without changes in p"ersonnel. In a closed system, repeated shuffling
ends up in a state of equilibrium and disorder where no novelty can occur.
On the other hand in an open system certain reorganizations accompanied
by qualitative novelty and increase of order can occur. Examples The
cooling of a liquid until it solidifies forming a crystal; the cooling of a
metallic ring until it becomes superconducting; the magnetization of an
iron bar; and the formation of patterns in fluids and in chemical systems.
(Cf. Turing, 1952; Glansdorff and Prigogine, 1971; Haken, 1975.)
Another mechanism that may ensue in qualitative novelty is accretion
or assembly. Thus stars and planets are believed to have formed by gravi-
tational accretion of cosmic gas and dust. However, novelty is more likely
to emerge along processes where heterogeneous things come together to
form systems. This may occur even if "mistakes" - i.e. departures from the
common pattern - are possible, that jeopardize the completion of the
assembly process. Now, in principle assembly can proceed either in one
stage - whether instantly or gradually - or in several. For example, a one
stage assembly process may consist in the addition of one component per
unit time; and a multiple stage assembly process may start with the forma-
tion of decamers out of monomers, proceeding with the formation of
hectomers out of decamers, and so on. If there is a nonvanishing probabil-
ity that each item (whether a monomer, a decamer, etc.) joins another in
the "wrong" way - e.g. forming an unstable system - then the direct or
one stage process is unlikely to yield systems of high complexity. As the
A SYSTEMIC WORLD VIEW 249
POSTULATE 6.5 The more complex a system, the more numerous the stages
in the process of its assembly.
The dual of this hypothesis concerns the breakdown or dismantling of
systems: it states that complex systems can decay in different ways, each
way corresponding to the weakening of one set of bonds among the com-
ponents. (Even comparatively simple things such as mu mesons can have
multiple disintegration schemata.) In short we lay down
POSTULATE 6.6 The more complex a system, the more numerous its pos-
sible breakdown modes.
6.4. Emergence
Our next assumption is that every concrete system has assembled from, or
with the help of, things in the same or lower order genera. In order to
formulate this assumption with some precision we need
DEFINITION 6.1 Let S = {S1' S2, S3, S4, S5} be the family of system genera.
Then Si precedes Sj' or Si < Sj for short, iff things of genus Si take part
(either as components or as agents) in the assembly of every system in Sj.
Our hypothesis is
POSTULATE 6.7 For any x E Sj there is at least one Y E Si' where Si < Sj'
such that y has taken part in the assembly of x.
Thus far we have only asserted that every system has been preceded by
pre-existing things - which is the reciprocal of the holistic thesis that the
whole precedes its parts. But we have not said exactly what precedes what.
This we do now:
POSTULATE 6.8 The precedence relations among the system genera are
250 CHAPTER 6
SOC TECH
BID
CHEM
I
<
I!
PHYS I
taken for naive reductionism, i.e. the belief that all it takes to account for
a whole is a knowledge of its parts. This belief is mistaken because systems
possess properties absent from their components, hence not to be dis-
covered by studying the latter in isolation from one another. Analysis is
necessary but insufficient to account for the formation and persistence of
systems: we must also study systems on their own level and interaction
with things of different genera if we wish to understand them. This is what
the theories of systems of various kinds - physical, chemical, biological,
social, and technical - are all about.
The philosophy that combines an acknowledgment of emergence with
the thesis that emergence is explainable and predictable within bounds
may be called rational emergentism. This philosophy, foreshadowed by
Sellars (1922), supersedes both atomism (though not its allegiance to
science) and holism (though not its insistence on emergence), and it in-
corporates a critical realist theory of knowledge - of which more in the
next two volumes of this Treatise.
6.6. Synopsis
It is closing time. We close by characterizing the ontology presented in
this volume and its companion, The Furniture of the World, in terms of a
few traditional isms. (A single ism is bad because it is one-sides and rigid.
Many-sidedness and flexibility can be achieved only by a system of match-
ing isms.)
Our ontology endorses
(i) naturalism or materialism, for it countenances only material existents
and discounts autonomous ideas, ghosts, and the like; but not physicalism
(or mechanism), as this denies that all things are physical entities;
(ii) systemism, for it holds that every thing is either a system or a com-
ponent of one; but not holism, as it rejects the myths that the whole is
incomprehensible, prior to its components, and superior to them;
(iii) pluralism as regards the variety of things and processes, hence the
plurality of kinds of thing and laws; and also monism as regards the sub-
stance that possesses properties and undergoes change (namely matter) as
well as the number of worlds (just one);
(iv) emergentism with regard to novelty, for it holds that, while some
bulk properties of systems are resultant, others are emergent; but not
irrationalism with regard to the possibility of explaining and predicting
emergence;
252 CHAPTER 6
(v) dynamicism, for it assumes that every thing is in flux in some respect
or other; but not dialectics, for it rejects the tenets that every thing is a
unity of opposites, and that every change consists in or is caused by some
strife orontic contradiction;
(vi) evolutionism with regard to the formation of systems of new kinds,
for it maintains that new systems pop up all the time and are selected by
their environment; but neither gradualism nor saltafionism, for it recog-
nizes both smooth changes and leaps;
(vii) determinism with regard to events and processes, by holding all of
them to be lawful and none of them to come out of the blue or disappear
without leaving traces; but not causalism, for it recognizes randomness
and goal striving as types of process alongside causal ones;
(viii) biosystemism with regard to life, for it regards organisms as ma-
terial systems that, though composed of chemosystems, have properties
not found on other levels; but neither vitalism nor machinism nor mechan-
ism;
(ix) psychosysfemism with regard to mind, for it holds that mental func-
tions are emergent activities (processes) of complex neural systems; but
neither eliminative nor reductive materialism, for it affirms that the mental,
though explainable with the help of physical, chemical, biological, and
social premises, is emergent;
(x) sociosystemism with regard to society, for it claims that society is a
system composed of subsystems (economy, culture, polity, etc.), and pos-
sessing properties (such as stratification and political stability) that no
individual has; hence neither individualism nor collectivism, neither idealism
nor vulgar materialism.
The reader accustomed to dwell in a single ism or in none is likely to
throw up his hands in despair at the multiplicity of isms embraced by our
ontology. Let this be said in defense of such mUltiplicity. First, it is pos-
sible to synthesize a variety of philosophical isms provided they are not
mutually inconsistent - i.e. provided the result is a coherent conceptual
system rather than an eclectic bag. (We have tried to secure consistency by
adopting the axiomatic format.) Second, it is necessary to adopt (and
elaborate) a number of philosophical isms to account for the variety and
mutability of reality - provided the various theses harmonize with science.
Third, tradition can be avoided only at the risk of unfairness and ignor-
ance: rather than dismiss our philosophical legacy altogether, we should
try and enrich it.
APPENDIX A
SYSTEM MODELS
1. INPUT-OUTPUT MODELS
ment and in turn reacting on (some features of) the latter. We call this the
complete black box, or black box for short. The simplest complete black
box is sensitive to stimuli of just one kind and accepts just one stimulus at
a time - i.e. has a single input terminal; and it also has a single output
terminal, i.e. acts just in one fashion on its environment and performs one
action at a time. This, the simplest black box, is called a dipole. Any more
complex black box is called a multipole: it is equipped with a certain num-
ber of input terminals of the same or different kinds. See Figure AI. We
shall concentrate our attention on the dipole, which is the building block
of every other input-output model of a system.
From a strictly formal point of view a dipole is just a double entry table
of items (e.g. numbers), or set of ordered pairs (u, v), where u E U is an
input value and v E V the corresponding output value. In other words, a
black box of this kind boils down to a binary relation f (perhaps a func-
tion) with graph or extension ,ff(f) c: U x V, where U is the collection of
possible inputs and V that of possible outputs. (This is what Mesarovic
(1968) calls a general system. But it is a model of a system and moreover
not completely general, for it neglects the internal structure of the system.)
The simplest of all such model systems, i.e. the dipole, has inputs and
outputs of the yes or no type: i.e. the terminals are either on (value 1) or
off (value 0). That is, U = V = 2 = {O, I}, andf: 2 -+ 2 is the transfer
function. Example 1 The conformist box:f(O) = 0,/(1) = 1. Example 2
The nonconformist box: f(O) = 1, f(l) = O. Example 3 The capricious
(yet not chaotic) or stochastic box: Pr(O I 0) = p > 0, Pr(11 0) =
1 - p, Pr(O 11) = q > 0, Pr(111) = 1 - q, where p is the probability
of response 0 to stimulus 0, and similarly for q. (Note the difference be-
tween the stochastic box, characterized by constant values of p and q,
hence statistically reliable, and a totally chaotic box, i.e. one with no
transfer function.) All three boxes are coarse models of a specialist; the
third models the creative type.
Other black boxes accept continuous streams of stimuli and deliver
o-u
(a) (b) (C) (d) (8)
Fig. Al. (a) The bare black box. (b) The passive monopole, or system without outputs.
(c) The spontaneous monopole, or system without inputs. (d) The dipole, or simplest
complete black box. (e) The reactive multipole.
SYSTEM MODELS 255
continuous streams of outputs. That is, their transfer functions are smooth
functions of time. Example 1 Differentiator. The state of the system at
time t E Tis the ordered pair <u(t), u(t), where u : T --+ Rand u is the rate
of change of u. The state space of the system is S = U x V, where U
is the image of u and V that of v = U. Example 2 Delay line. Here v(t) =
u(t - 8) with 8 > O. I.e. the current output equals the input at the earlier
time t - 8. The state space is S = U x {u(t - 8) It E T}.
In the last two examples the input and output values were parametrized
by time - i.e. U and V were construed as the images of time-dependent
functions. In general the output value v(t) at time t will depend not only
on the current value of the input (as in Example 1) but also on some or
even all of the inputs accepted earlier by the system (as in Example 2).
I.e. in general the system will have memory (or be hereditary). A fairly
general input-output relation covering all such cases for the dipole, and
easily generalizable to the multi pole, is
where M summarizes the structure of the box (Bunge, 1963a). The domain
of integration, which stops at t, has been chosen to conform with the
antecedence principle (Postulate 6.12 in Vol. 3, Ch. 6, Sec. 5.3). If M is
a delta function, then the box is memoriless. If F is a linear function then
the above represents a linear dipole. In a great many cases M does not
depend on t but only on the past (if at all), and F( u(7:)) = u(t - 8), where
8 is a time lag. In such cases the integral is the convolution of the two func-
tions, and the mathematics is greatly simplified.
Strictly speaking a black box is a suitable model not of a system but rather
of a system component. In many cases a system can be adequately modeled
as an assembly of black boxes connected in a suitable way. The basic
modes of coupling are as follows.
(i) Series coupling: see Figure A2a. If the two component boxes are
characterized by the tran'sfer functions f: U --+ V and g : V --+ W, then
the resulting box is characterized by the composition h = fog with values
w = g(f(u)). Note that, in general, function composition is not commuta-
tive; thus g might filter out all stimuli except those in a certain set, while
256 APPENDIX A
~W=f(9lU))
(a)
v,
(b)
(c)
Fig. A2. (a) Series combination of two dipoles. Example: cloth manufacturer g supplies
clothing manufacturer f (b) Parallel connection of two dipoles. Example: two branches
of a cloth manufacturing concern. (c) Feedback coupling (or self interconnection). Ex-
ample: blood pressure control system.
the list of all the transfer functions characterizing its components when
inter-connected, i.e.
ENVIRONMENT
CONTROLLING System
Controlled
Output
System
Fig. A3. Feedback control system. The detector compares the output (actual state) of
the controlled system and, if it differs from the set point, drives the response or corrective
device, which acts on the controlled system to correct the imbalance.
The study of specific control systems belongs to the special sciences, such
as mechanical and electrical engineering, biochemistry, physiology, eco-
nomics, and politology. And the scientific study of the features common to
all control systems, regardless of their composition, is of course cyberne-
tics (Wiener, 1948). Cybernetics was born in advanced engineering but,
since it is stuff-free (i.e. independent of the kind of material), it assumes no
special physical, chemical, biological, or social laws, so it covers all levels.
It is therefore a chapter of the general theory of systems as well as of
scientific ontology.
The supersystem composed by the controlled and the controlling sys-
tems can be described as follows. The total state function of the super-
system is
F = (Flo F 2, ... , Fp, Fp+l' Fp+2' ... , F n >: T -----> lR n,
where the first p coordinates are the so-called state variables and the re-
maining n - p components the control variables. These components are
subjected to three sets of conditions:
(i) constraints on the state variables (e.g. their values must lie within a
certain region of the state space);
(ii) control or steering conditions specifying the manner in which the
control variables affect the state variables, e.g.
Pi = gi(F, t) for 1 =s; i =s; p;
(iii) optimization conditions on certain state variables (e.g. minimization
of energy consumption, or maximization of the rate of production of some
substance).
SYSTEM MODELS 259
( ) ( ) a E R+ for t ~ 0
ut=Ht= {
o for t < 0
is converted into v(t) = at. Obviously this output, unless checked, gets
eventually out of hand. This can be prevented by coupling the integrator
to a controller that diverts part of the output and sends it back to the input
terminal, so that a closed loop results: see Figure A4.
In the simplest case the output fraction kv(t), where Ikl < 1, is fed back
diminished but without distortion, so that the input-output equation reads
now
u v
c
Fig. A4. Integrator with feedback control. The effect of the con'trol subsystem C is to
convert a constant input into either an exponentially growing response (for k > 0) or an
exponentially decaying one (for k < 0).
260 APPENDIX A
a
v(t) = ~Tkr(l-exp( -Ikl t» decelerated growth approaching
the target value allkl
Note that both the controlled and the controlling subsystems have been
treated as black boxes: their mechanisms have not been specified. Nor
could they without roping in definite law statements (physical, biological,
or what have you).
The above example illustrates the black box model of a control system.
In most cases such a model is insufficient and a grey box model is required,
i.e. one involving not only an input vector u and an output vector v but
also an (internal) state vector x. The simplest example of such a model of
a control system is, of course, the linear system
x= Ax + Bu, v = Cx,
where the state matrix A, the input matrix B, and the output matrix Care
constant. The corresponding block diagram is shown in Figure A5.
~
f, f,
0
F, F,
(al (b)
Fig. A6. Two kinds of stable orbits in state space. (a) Spiral approach to static equili-
brium. (b) Stationary orbit.
the stability limits. Far from being a passive toy of its environment, such
a system adapts itself actively to it and thus manages to endure. Because
such systems are of paramount importance in the life sciences and in the
social sciences, not to mention technology, they deserve a name of their
own:
2.1. Generalities
Take a black box and observe or conjecture its internal states without
however disclosing the mechanism that makes it tick: you have a grey box.
Psychologists say that an intervening variable has been added to the
stimUlus-response pairs, and that the new variable is not a "hypothetical
construct" but a mere formal link between inputs and outputs. But if the in-
tervening variable is interpreted as representing the mechanism that trans-
forms inputs into outputs, then instead of a grey box we have a dynamical
box (Appendix B, Sec. 1). A grey box is then midway in depth between a
black box and a translucid box or mechanismic model. We may picture it
SYSTEM MODELS 263
U v
8"
8m
Fig. A7. Transition of system from state Sm to state S. under the influence of input u.
264 APPENDIX A
kinds of system. (Perhaps for this reason von Neumann, 1951, mistook it
for a chapter of formal logic.)
All information processing systems may be regarded, at least to a first
approximation, as automata: nervous systems, servomechanisms, com-
puters, TV networks, and so on. Therefore the referent of automata theory
is not a narrow species of system but the entire genus of information pro-
cessing systems, whether physical, chemical, living or social. And because
the theory is concerned with certain traits of concrete systems, it is not a
formal theory but a factual one. Indeed it provides an exact and simple
(hence also superficial) model of a system interacting with its environment
regardless of any specific features of interest to the special sciences - such
as the kinds of material it is built from and the way it is energized. For
this reason automata theory belongs not only in advanced technology but
also in ontology. Let us therefore take a quick look at it.
TABLE Al
Basic concepts of the theory of deterministic automata
Symbol Mathematical nature Factual interpretation
Z Set Set of distinct unit inputs
Q Set Set of distinct unit outputs
Individual Null input or pause
Binary operation Concatenation
S Set State space
M Function Transition (next state) function
N Function Output function
So Individual Initial state
F Set Set of final states
(i) ~ = {aO' 0"1, ... , O"k}, called the input alphabet, is a nonempty set with
k + 1 elements called letters, everyone of which represents a unit input on
A from its environment;
(ii) Q, called the output alphabet, is a set with two elements: 0 and 1,
where 0 represents the absence of an output and 1 a unitary output;
(iii) let ~* be the set of finite concatenations, called tapes, of elements
of ~ (i. e. letters), and let 0"0 E ~ be such that, for every tape x E ~* , 0"0 X =
XO"o = x. Likewise let Q* be the set of finite concatenations of elements of
Q, with 0 E Q such that, for every y E Q*, Oy = yO = y. The structures
<~*, 0,0"0) and <Q*, 0,0) are free monoids; ° represents the combination
or concatenation of successive inputs to or outputs of A;
(iv) S, called the state space of A, is a finite nonempty set with n elements,
where n > 1; every member of S represents an internal state of A;
(v) the function M: S x Z* ---> S is called the transition (or next state)
function; if s E S and x E Z*, then M(s, x) E S represents the state A goes
into when A, while in state s, admits input x;
(vi) the function N: S x Z* ---> Q* is called the output function and is
such that N(J, x) = 1 E Q* for every f E F c S and every x E ~*, where
N(s, x) represents the output of A when acted on by the input x E ~*
while in state s;
(vii) So is a distinguished element of S representing the initial internal
state of A;
(viii) F is a nonempty set included in the state space S, and every mem-
ber f of F represents a final state of A;
(ix) A makes no spontaneous transitions; i.e. the null input has no ef-
fect: for every s in S, M(s, 0"0) = s;
(x) the internal states form sequences, i.e. the effect of a compound in-
put x ° y, where x, y E Z*, equals the effect of the second input acting on
the automaton in the state to which the first input carried it:
M(s, x y) = M(M(s, x), y).
0
acterize two different automata with the same internal states reacting dif-
ferently to the same unit inputs (71 and (72' They may thus represent some
trait of the behavior of two different insects confronted with the same
circumstances.
5,
5,
~_~~J~ SI s2
~---~~~
a; a;
M1(S, (71) SI S2 SI 5,
M1(S, (7z) Sz So Sz
5, 5,
a; a;
M2(S, (71) So Sz Sz
5, 5"
Mz(s, (7z) SI So So a; a;
5,
a; a;
5"
Fig. A8. Two different three state automata, each characterized by a different transi-
tion function (M, and M, respectively), subjected to the same stimuli.
maton A. Then
(i) the restriction of the transition function M to So is called the response
function: rp(x) = M(so, x), for x E Z*;
(ii) a state s E S of A is accessible iff there is some input x E 2* such that
s = rp(x) - otherwise s is inaccessible;
(iii) a tape x E Z* is accepted (or recognized) by A iff rp(x) E F (i.e. if x
has an overt effect when acting on A in so);
(iv) the behavior of A is the set of all the tapes accepted (or recognized)
by A: peA) = df{XEZ*lrp(X)EF};
(v) two automata A and A' are behaviorally identical iff peA) = peA').
Note the difference between the above notion of behavior and that
used in the so-called behavioral sciences.
And now a few typical consequences. Firstly
THEOREM Al If two inputs x, y E Z* have the same effect, then the effects
SYSTEM MODELS 267
of the compound inputs xz and yz are the same for an arbitrary tape
Z E Z*. I.e. for all states s E S and all inputs x, y, Z E Z*,
M(s, x) = M(s, y) => M(s, x ° z) = M(s, yo z).
Example Let ~ = {00, 01> 02}, S = {so, S1}, F = {S1}, and assume the
stochastic matrices
s,
So
270 APPENDIX A
EN V IRONMENT
RECEIVER
Fig. A9. The components of an information system. The effects of the environment on
the channel, when random. are lumped under the category of noise.
272 APPENDIX A
CHANGE MODELS
In this Appendix the accent will be on change. We shall exhibit both kine-
matical and dynamical models, first of quantitative, then of qualitative
change. And we shall study only extremely general change models, i.e.
such that are portable from one field of inquiry to another. As in the case
of Appendix A, most of the material to be presented is standard or has at
least appeared before. But the interpretation and the philosophical discus-
sion is often novel.
I. KINEMATICAL MODELS
The equilibrium property can be found by going over to the state space,
namely by dividing one equation by the other to obtain
dF1 _ G1
-dF~- - -G2 '
which represents the possible processes in the state space. The system is
stable iff dF2/dF1 approaches a definite number as Fl increases. This
number is the slope of a straight line in the state space, namely the equi-
librium trajectory. If on the other hand the system diverges from the (ima-
ginary) equilibrium line, it is unstable. See Figure BI. (The equilibrium
trajectory may be a finite segment or just a point.)
For the sake of definiteness think of p as the mass density and of $ as the
total mass of the system.
We now assume the following factual hypothesis:
i = J v'
dVs + Ss dS jm
wherejn is the projection ofj onto the interior normal to S. See Figure B2.
Using now Green's theorem, a purely mathematical result, we derive
Fig. B2. Transactions of a system with its environment as reflected in the rate of change
of the former's extensive property e. This rate consists of the (positive or negative) con-
tribution of the source (or sink s) and the (positive or negative) contribution of the flow j.
CHANGE MODELS 277
II
fields" (Jost, 1965, p. xi). In addition, one hopes that this approach may
avoid the inconsistencies that mar the quantum field theories. Its role in
metaphysics is yet to be investigated.
We close this subsection with an historical remark. The Lagrangian
framework was foreshadowed by Maupertuis as a synthesis of dynamics
and metaphysics. It was brought to fruition by Lagrange, who freed it
from metaphysics, and it was generalized beyond dynamics by Maxwell.
The axiom contained in Definition B2 (iv) is one of the most secure and yet
misinterpreted principles of science. Maupertuis took the (action) integral
I to be always a minimum and interpreted this as a sure sign of divine
wisdom: Nature always chose the paths that minimized I. It was later
discovered that in some cases (notably in optics) I is not a minimum but
a maximum, so the principle was rechristened from that of minimal to
that of extremal action. Nevertheless a number of physicists, among them
Planck and Margenau, continued to regard Maupertuis' principle as an
expression of purpose and design. This is mistaken if only because the
principle, together with some lemmas borrowed from the variational cal-
culus, entails the equations of motion - in which no traces of purpose or
design have ever been detected. It is not that the system behaves in order
to minimize or maximize its action: it is just a law, and an extremely per-
vasive one, that all systems behave that way. Because of this generality the
Lagrangian framework should be recognized as belonging to the intersec-
tion of science and metaphysics.
playa heuristic role and save time. But they have hardly any ontological
import except insofar as they show that different mechanisms may result
in similar overall behavior patterns, whence the study of behavior does
not exhaust that of reality. For example brownian particles, domestic
flies, and drunkards move randomly, but the underlying mechanisms are
very different, so that the observation that all three are kinematically
analogous has no metaphysical significance. What does have ontological
import is the finding of dynamical analogies among systems of different
kinds, i.e. the discovery that certain mechanisms - such as transport, dif-
fusion, merging, cooperation, competition, random shuffling, and casual
mutation - are conspicuous on all levels. So, let us turn our attention to
mechanisms.
2. DYNAMICAL MODELS
2.l. Generalities
A dynamical model of a system is a kinematical model (Sec. I) some vari-
ables of which are interpretable as "driving" others. In other words, a
dynamical model involves a change mechanism, i.e. either a subsytem or a
process whereby the changes in some properties of the system effect
changes in some other properties of the same system. Morever all such
changes are spatiotemporal or reducible to changes in time (such as rates)
or in space (such as gradients). .
Example 1 The geographic distribution of a biospecies can be explained
as the outcome of the action of three separate mechanisms: random dif-
fusion, drift, and environmental resistance (or selection pressure). A
dynamical model of the spatial pattern of a species (or rather a family of
populations of the given species) is this. The rate of change of the popula-
tion density p of the given collection of populations at some point x on
the surface of the planet is
2.2. Formalities
We repeat in a slightly mor~ exact way the definition of a mechanism given
at the beginning of the last subsection:
If these rates are nil, no processes occur. If they are positive, each in-
dividual eventually attains and surpasses its optimum level. Furthermore
if cooperation sets in from the start, a will help b in proportion to the
286 APPENDIX B
latter's needs, i.e. in such a way that the imbalance f3 - B will be offset;
and b in turn contributes in proportion to the imbalance a - A, though
not necessarily with the same intensity. This mutual help, we have as-
sumed, is positive as long as there is a deficit (negative balance); it be-
comes nil on attaining satiety, and negative as soon as there is superavit
(positive balance). The simplest formulas representing such a cooperative
process are the nonlinear equations
A = au + ada - A)B,
with aij :2: 0 for i,.i = 1, 2.
The coefficient a12 measures the strength of the help a gets from b, while
a21 measures the intensity of the help a lends b. At the optimum point <a,
(3) there is neither help nor hindrance: each component fends for itself, so
the system breaks down momentarily.
We spare the reader the treatment of the general case in the vicinity of
the optimum point, showing that, in effect, no sooner does a attain the a
level than its partner b turns against it forcing it to slow down its rate of
consumption or of production. (Cf. Bunge, 1976.) Intead, we shall study
the particular case when the two partners behave in exactly the same way.
In this case, i.e. for Cl = C2 = C (no initial advantage), all = a22 (same
individual rates), a12 = a21 (symmetric help) and a = f3 (same require-
ments), we are left with a single equation of evolution, namely
x= A,B.
The general solution to this equation is found to be
x(t) = allt + . ac .
c + (a - c) exp (-a12 at)
If each partner is self-sufficient (all = a22 > 0), cooperation only speeds
up the process up to the optimum point, and curbs it after that point.
Otherwise (i.e. if all = a22 = 0), mutual help makes the process possible.
Indeed, the second term of the last equation represents the growth of A
(which is equal to B) from the initial value c towards the asymptote A
(00) = a > c. Rivalry from this point on prevents unlimited growth in
either consumption or production. See Figure B3. This limitation takes
place whether or not there is an external limitation such as scarcity of
resources: the control is strictly internal. Not that there are no external
limitations; only, our model does not take them into account.
General dynamical models of this kind have at least three possible uses:
CHANGE MODELS 287
,
A 8 Runa~ay
I growth
I Cooperatm-competition :
I controfktd gf"O'Nth
I
/
/
Fig. B3. Full line of a system composed of partners that cooperate and compete (a 12 #-
0) but are powerless in isolation (i.e. when all = a22 = 0). Dotted line: stagnation in the
absence of interaction (i.e. when a 12 = 0).
(a) they may suggest the search for real systems (e.g. in morphogenesis,
neurology, ecology, or sociology) that approach a zero growth state in
some variable as a result of their cooperative-competitive structure; (b)
they may stimulate the building of specific theories covering real systems
behaving in the way outlined by the model; (c) they may help in planning
social controls. In particular, our model suggests redesigning society in
such a way that each member helps his/her neighbor in proportion to his/
her own ability and the other party's needs, and also controls his/her
partner's growth and is in turn controlled by it, thus preventing scarcity
as well as waste, and thus combining the good features of competition with
those of cooperation.
has no consequences on the other targets of the system - i.e. that the ef-
fects are mutually independent, so that only their cumulative effect is of
interest. Since the probability that a bullet fails to hit a target is q = 1 -
p, the probability that the target will fail to mutate under the action of n
random bullets is
qn = (1 - p)n ~ r pn for large n.
And the probability that all m sensitive components will remain unaltered
after a round of n firings will be the m-fold product of the above, i.e.
Q ~ (e-pn)m.
Note that we are not assuming that all mutagenic agents are external to
the system concerned: it suffices that they be external to its sensitive com-
ponents. For example, chromosomes are subjected not only to external
radiation but also to thermal and chemical actions of the surrounding
protoplasm. And every cultural system is the beneficiary or victim not only
of economic and political actions but also of its own creative and destruc-
tive members.
BIBLIOGRAPHY
292
BIBLIOGRAPHY 293
Bunge, Mario, and Maximo Garcia Sucre (1976). Differentiation, participation and
cohesion. Quality and Quantity 10: 171-178.
Bunge, Mario, Johann Galtung, and Mircea Malitza, Eds. (1977) Mathematical Ap-
proaches to International Relations. Bucharest: Romanian Academy of Social and
Political Sciences.
Bunge, Mario, and Rodolfo Llinas (1978). The mind-body problem in the light of
neuroscience. Proc. 16th World Congress of Philosophy.
Calvin, Melvin (1969). Chemical Evolution. New York & Oxford: Oxford University
Press.
Chiaraviglio, Lucio (1965). Sequential machines and the coding of polypeptides. J.
Theoret. BioI. 8: 130-140.
Chomsky, Noam (1968). Language and Mind. New York: Harcourt, Brace & World.
Cowan, Jack (1976). Are there modifiable synapses in the visual cortex? In Rosenzweig
and Bennett, 1976, pp. 133-143.
Craik, Kenneth J. W. (1966), The Nature of Psychology. Stephen L. Sherwood, Ed.
Cambridge: University Press.
Crick, Francis (1970), Central dogma of molecular biology. Nature 227: 561-563.
Delgado, Jose M. R. (1969), Physical Control of the Mind. New York: Harper & Row.
Deutsch, Karl W. (1968), The Analysis of International Relations. Englewoods Cliffs,
N. J.: Prentice-Hall.
Dobzhansky, Theososius (1970), Genetics of the Evolutionary Process. New York &
London: Columbia University Press.
Dorf, Richard C. (1974), Modern Control Systems, 2nd ed. Reading, Mass.: Addison-
Wesley Publ. Co.
Du Bois-Reymond, Emil (1872), Ueber die Grenzen des Naturerkennens. In Reden,
2nd ed. Leipzig: Veit und Co.
Dunbar, Max (1960). The evolution of stability in marine ecosystems: natural selection
at the level of the ecosystem. Am. Natur. 94: 129-136.
Edelen, Dominic G. B. (1962). The Structure of Field Space. Berkeley: University of
California.
Eigen, Manfred (1971). Selforganization of matter and the evolution of biological
macromolecules. Naturwissenschaften 58: 465-523.
Feigl, Herbert (1958), The 'mental' and the 'physical'. In H. Feigl, M. Scriven, and G.
Maxwell, Eds., Minnesota Studies in the Philosophy of Science II, pp. 370-497. Min-
neapolis: University of Minnesota Press.
Feinberg, Martin (1972). On chemical kinetics of a certain class. Arch. Rat. Mech.
Anal. 46: 1-41.
Florkin, M. ,Ed. (1960). Aspects of the Origin of Life. Oxford: Pergamon Press.
Fox, Sidney W., and K. Dose (1972). Molecular Evolution and the Origin of Life. San
Francisco: Freeman & Co.
Freeman, Walter J. (1973). A model of the olfactory system. In M. A. M. Brazier, D. O.
Walter, and D. Schneider, Eds., Neural Modeling, pp. 41-62. Los Angeles: UCLA
Brain Research Institute.
Fritsch, Bruno (1974), Wachstum Begrenzung als Machtinstrument. Stuttgart: Deutsche
Verlags-Anstalt.
Frondizi, Risieri (1953). The Nature of the Self. A Functional Interpretation. New Haven,
Con.: Yale University Press. Repr.: Carbondale, Ill.: Southern Illinois Press, 1971.
BIBLIOGRAPHY 295
Gall, John (1977) Systemantics: How Systems Work and Especially How They Fail.
New York: Quadrangle.
Gardner, R. A., and B. T. Gardner (1969). Teaching sign language to a chimpanzee.
Science 165: 664-672.
Gazzaniga, Michael S. (1967). The split brain in man. Sci. Am. 217, 2: 24-29.
Gierer, H., and H. Meinhardt (1972). A theory of biological pattern formation. Kybern-
etik 12: 30-39.
Ginsburg, S. (1962). An Introduction to Mathematical Machine Theory. Reading, Mass.:
Addison-Wesley Pub!. Co.
Glansdorff, P., and I. Prigogine (1971). Thermodynamic Theory of Structure, Stability
and Fluctuations. London: Wiley-Interscience.
Glass, Leon (1973). Stochastic generation of regular distributions. Science 180: 1061-
1063.
Goldin· Meadow, Susan, and Heidi Feldman (1977). The development of language-like
communication without a language model Science 197: 401-403.
Gray, J. A. (1972a). The psychophysiological nature of introversion-extroversion: A
modification of Eysenck's theory. In Nebylitsyn and Gray, 1972, pp. 182-205.
Gray, J. A. (1972b). Learning theory, the conceptual nervous system, and personality.
In Nebylitsyn and Gray, 1972, pp. 372-399.
Gray Eaton, G. (1976). The social order of Japanese macaques. Sci. Am. 235, 4: 96--106.
Griffin, Donald R. (1976). The Question of Animal Awareness. New York: Rockefeller
University Press.
Gross, Charles G., and H. Philip Zeigler, Eds. (1969). Readings in Physiological Psy-
chology. Motivation. New York: Harper & Row.
Guinzburg, A. (1968). Algebraic Theory of Automata. New York: Academic Press.
Haken, H. (1975). Cooperative phenomena in systems far from equilibrium and in ~on
physical systems. Rev. Mod. Phys. 47: 67-121.
Hamilton, W. D. (1964). The genetical theory of social behavior. J. Theoret. BioI. 7:
1-16, 17-32.
Hamilton, W. D. (1971), Geometry of the selfish herd. J. Theoret. BioI. 31: 295-311.
Harris, Marvin (1968). The Rise of Anthropological Theory. New York: Thomas Y.
Crowell Co.
Harrison, M. A. (1965). Introduction to Switching and Automata Theory. New York:
McGraw-Hill.
Hebb, Donald O. (1949). The Organization of Behavior. New York: Wiley.
Hebb, Donald O. (1968). A Textbood of Psychology. Philadelphia: W. B. Saunders Co.
Hebb, Donald O. (1974).What psychology is about. Am. Psychol. 29: 71-79.
Hegel, Georg Wilhelm Friedrich (1830). Encyklopadie der philosophischen Wissen-
schaften im Grundrisse, 3rd ed. Heidelberg: Verwaltung des Oswaldschen Verlags.
Henderson, Lawrence J. (1913). The Fitness of the Environment. Boston: Beacon Press,
1958.
Hinde, Robert A. (1974). Biological Bases of Human Social Behaviour. New York:
McGraw-Hill.
Holland, John H. (1975). Adaptation in Natural and Artificial Systems. Ann Arbor: Uni-
versity of Michigan Press.
Homans, George C. (1974). Social Behavior: Its Elementary Forms. New York: Har-
court, Brace, Jovanovich.
296 BIBLIOGRAPHY
Hubel, D. H., and T. N. Wiesel (1963), Receptive fields and functional architecture of
monkey striate cortex. J. Physiol. 165: 559-568.
Hull, David (1974), Philosophy of Biological Science. Englewood Cliffs, N.J.: Prentice-
Hall.
Hume, David (1739). A Treatise of Human Nature. L. A. Selby-Bigge, Ed. Oxford:
Clarendon Press, 1888.
Jacob, Franc;:ois (1970). La logique du vivant. Paris, Gallimard.
Jacob, Franc;:ois (1977). Evolution and tinkering. Science 196: 1161-1166.
Jost, R. (1965). The General Theory of Quantized Fields. Providence: American Mathe-
matical Society.
Kalman, R. E., P. L. Falb, and M. A. Arbib (1969). Topics in Mathematical System
Theory. New York: McGraw-Hill.
Kawai, Masao (1965). Newly acquired pre-cultural behavior of the natural troop of
Japanese monkeys on Koshima Islet. Primates 6: 1-30.
Kemeny, John, and Paul Oppenheim (1956). On reduction. Phi/os. Stud. 7: 6-19.
Kendrew, John C. (1966). The Thread of Life. London: G. Bell and Sons Ltd.
Keosian, John (1968). The Origin of Life, 2nd ed. New York: Van Nostrand Reinhold
Co.
King, M. -C., and A. C. Wilson (1975). Evolution at two levels in humans and chimpan-
zees. Science 188: 107-116.
Klir, George (1969). An Approach to General System Theory. New York: Van Nostrand
Reinhold Co.
Klir, Jiri, and Miroslav Valach (1967). Cybernetic Modelling. London: Iliffe Books
Ltd. ; Prague: SNTL.
Klir, George, and Gary Rogers (1977). Basic and Applied General Systems Research: A
Bibliography. Binghampton, N.Y.: SUNY-Binghampton.
Kripke, Saul (1971). Identity and necessity. In Milton K. Munitz, Ed., Identity and Indi-
viduation, pp 135-164 New York: New York University Press.
Krohn, Kenneth, Rudolph Langer and John Rhodes (1967). Algebraic principles for
the analysis ofa biochemical system. J. Compo Syst. Sci. 1 : 119-136.
Kiippers, Bernd (1975). The general principles of selection and evolution at the mole-
cular level. Prog. Biophys. Molec. BioI. 30: 1-22.
Lange, Oscar (1965). Wholes and Parts: A General Theory of Systems Behavior. Oxford:
Pergamon Press; Warsaw: PWN.
Lehninger, Albert L. (1975). Biochemistry, 2nd ed. New York: Worth Publishers Inc.
Lepper, M. R., D. Greene, and R. E. Nisbett (1973). Undermining children's intrinsic
interest with extrinsic rewards. J. Pers. Soc. Psychol. 28: 129-137.
Levins, Richard (1968). Evolution in Changing Environments. Princeton: Princeton Uni-
versi ty Press.
Levi-Strauss, Claude (1963). Structural Anthropology. New York: Basic Books.
Lewontin, Richard C. (1974). The Genetic Basis of Evolutionary Change. New York &
London: Columbia University Press.
Llinas, Rodolfo, and Mario Bunge (1978). Restricted applicability of the concept of
command in neuroscience. Behav. Brain Sci. 1: 30-31.
Lloyd Morgan, C. (1933). The Emergence of Novelty. London: Williams and Norgate.
Luria, A. R. (1966) Human Brain and Psychological Processes. New York: Harper &
Row.
BIBLIOGRAPHY 297
Luria, A. R. (1969) Speech development and the formation of mental processes. In Mi-
chael Cole and Irving Maltzman, Eds., A Handbook of Contemporary Soviet Psycho-
logy, pp. 121-162. New York: Basic Books.
Lwoff, Andre (1962). Biological Order. Cambridge, Mass. : MIT Press.
MacGregor, Ronald J., and Edwin R. Lewis (1977). Neural Modelling. New York &
London: Plenum Press.
Malcolm, Norman (1973). Thoughtless brutes. Proc. Add. Am. Phi/os. Assoc. 46: 5-
20.
Malcolm, Norman (1964). Scientific materialism and the identity theory. Dialogue 3:
115-125.
Mandl, F. (1959). Introduction to Quantum Field Theory. New York: Interscience.
Malsburg, C. von der (1973). Self-organization of orientation sensitive cells in the striate
cortex. Kybernetik 14: 85-100.
Margolis, Joseph (1978) Persons and Minds: The Prospects of Nonreductive Materialism.
Dordrecht & Boston: D. Reidel Publ. Co.
Marr, David (1970). A theory for cerebral cortex. Proc. Roy. Soc. (London.) B176:
161-234.
Maynard Smith, J. (1968). Mathematical Ideas in Biology. Cambridge: University
Press.
McNaughton, Robert (1961). The theory of automata: a survey. Adv. Compo 2: 379-
421.
Menzel, E.W., Jr., R. K. Davenport, and C. M. Rogers (1972). Protocultural aspects of
chimpanzee's responsiveness to novel objects. Folia primatol. 17: 161-170.
Mesarovic, M. D., Ed. (1964). Views on General Systems Theory. New York: Wiley.
Miller, Stanley, and Leslie Orgel (1974). The Origins of Life on the Earth. Englewood
Cliffs, N.J.: Prentice-Hall.
Miller, Neal E. (1969). Learning of visceral and glandular responses. Science 163:
434-445.
Milsum, John H., Ed. (1968). Positive Feedback. Oxford: Pergamon Press.
Monod, Jacques (1970). Le hasard et la nlicessite. Paris: Ed. du seuil. Engl. transl.:
Chance and Necessity. New York: Knopf, 1971.
Moore, Robert Y. (1976). Synaptogenesis and the morphology of learning and memory.
In Rosenzweig and Bennet, Eds., pp. 340-347.
Morris, Charles W. (1938) Foundations of the theory of signs. International Encyclopae-
dia of Unified Science, vol. I., No.2. Chicago: University of Chicago Press.
Mostow, G. D., Ed. (1975). Mathematical Models for Cell Rearrangement. New
Haven, Conn.: Yale University Press.
Mountcastle, Vernon (1957). Modality and topographic properties of single neurons of
eat's somatic sensory cortex. J. Neurophys. 20: 408-434.
Nass, Menasche M., and Leon N. Cooper (1975). A theory for the development of fea-
ture detecting cells in visual cortex. Biolog. Cybernet. 19: 1-18.
Nebylitsyn, V. D., and A. J. Gray, Eds. (1972). Biological Bases of Individual Behaviour.
New York & London: Academic Press.
Neumann, John von (1951). The general and logical theory of automata. In L. A.
Jeffress, Ed., Cerebral Mechanisms in Behavior. New York': Wiley.
Neumann, John von (1966). Theory of Self-Reproducing Automata. Ed. and completed
by A. W. Burks. Urbana: University of Illinois Press.
298 BIBLIOGRAPHY
Rabin, Michael 0., and Dana Scott (1959). Finite automata and their decision
problems. IBM J. Res. & Devel. 3: 114-125.
Rashevsky, Nicholas (1972). Some remarks on the central nervous system. Bull. Math.
Biophys. 34: 231-242.
Ratcliffe-Brown, A. R. (1935). On the concept of function in social science. Am. Anthr.
37: 397-402.
Reynolds, Vernon (1976). The Biology of Human Action. Reading & San Francisco: W.
H. Freeman and Co. Ltd.
Rosen, Robert (1959). The DNA-protein coding problem. Bull. Math. Biophys. 21: 71-
95.
Rosen, Robert (1970). Subunits and subassembly processes. J. Theoret. Bioi. 28: 415-
422.
Rosenblueth, Arturo, Norbert Wiener, and Julian Bigelow (1943). Behavior, purpose
and teleology. Phi/os. Sci. 10: 18-24.
. Rosenblueth, Arturo, and Norbert Wiener (1950). Purposeful and non-purposeful
behavior. Phi/os. Sci. 17: 318-326.
Rosenzweig, Mark R., and Edward L. Bennett, Eds. (1976). Neural Mechanisms of
Learning and Memory. Cambridge, Mass.: MIT Press.
Rumbaugh, Duane M., Timothy V. Gill, and Ernst von Glasersfeld (1973). Reading and
sentence completion by a chimpanzee (Pan). Science 182: 731-733.
Ruse, Michael (1979). Sociobiology: Sense or Nonsense? Dordrecht & Boston: D. Reidel
Pub!. Co.
Rutledge, L. T. (1976). Synaptogenesis: effects of synaptic use. In Rosenzweig and
Bennett, Eds., 1976, pp. 329-339.
Schlick, Moritz (1925). Allgemeine Erkenntnislehre, rev. ed. Trans!. A. E. Blumberg,
General Theory of Knowledge. Wien-New York: Springer-Verlag, 1974.
Sellars, Roy Wood (1922). Evolutionary Naturalism. Chicago: Open Court.
Sellers, Peter H. (1967). Algebraic complexes which characterize chemical networks.
SIAM J. Appl. Math. 15: 13-68.
Sellers, Peter H. (1970). Combinatorial analysis of a chemical network. J. Franklin Inst.
290: 113-130.
Sellers, Peter H. (1971). An introduction to a mathematical theory of chemical reaction
networks 1. Arch. Rational Mech. Anal. 44: 22-40.
Shaffer, Jerome A. (1977). Personal identity: the implications of brain bisection and
brain transplant. J. Medicine and Philosophy 2: 147-161.
Smart, J. J. C. (1959). Sensations and brain processes. Phi/os. Rev. 68: 141-156.
Smith, Emil (1970). Evolution of enzymes. In Paul D. Boyer, Ed., The Enzymes, Vo!.
I, pp. 267-339. New York: Academic Press.
Smith, W. John (1965). Message, meaning, and context in ethology. Am. Natur. 99:
405-409.
Smythies, J. R., Ed. (1965). Brain and Mind. London: Routledge & Kegan Paul.
Sommerhoff, Gerd (1974). Logic of the Living Brain. New York: Wiley.
Spencer, Herbert (1862). First Principles. Final ed.: London: Watts & Co., 1937.
Sourkes, Theodore (1962). Biochemistry of Mental Disease. New York: Harper &
Row.
Sperry, Roger W. (1964). Neurology and the mind-body problem. In Robert Isaacson,
300 BIBLIOGRAPHY
Ed., Basic Readings in Neuropsychology, pp. 403--429. New York: Harper & Row.
Sperry, Roger W. (1966). Brain bisection and mechanisms of consciousness. In John
C. Eccles, Ed., Brain and Conscious Experience. Berlin, Heidelberg & New York:
Springer-Verlag, 1966.
Sperry, Roger W. (1977). Forebrain commissurotomy and conscious awareness. J.
Med. Phi/os. 2: 101-126.
Spinoza, Benedict (1677). Ethics. Oxford: University Press, 1927.
Steinberg, Malcolm S. (1963). Reconstruction of tissues by dissociated cells. Science
141: 401--408.
Thom, Rene (1972). Stabilite structurelle et morphogenese. Reading, Mass.: W. A.
Benjamin Inc.
Toulmin, Stephen (1972). The mentality of man's brain. In A.G. Karczmar and John
C. Eccles, Eds., Brain and Human Behavior, pp. 409--422. New York, Heidelberg &
Berlin: Springer-Verlag.
Turing, Alan (1950). Computing machinery and intelligence. Mind NS59: 433--460.
Turing, Alan (1952). The chemical basis of morphogenesis. Phi/os. Trans. Roy. Soc.
(Lond.) B237: 7-94.
Tyson, John J. (1976). The Belousov-Zhabotinskii Reaction. Berlin, Heidelberg, New
York: Springer-Verlag.
Vaucouleurs, Gerard de (1970). The case for a hierarchical cosmology. Science 167:
1203-1213.
Vesey, G. N. A., Ed. (1964. Body and Mind. London: George Allen & Unwin.
Vol'kenshtein, Mikhail V. (1970). Molecules and Life. New York: Plenum Press.
Vygotskii, L. S. (1962). Thought and Language. Transl. E. Hanfman and G. Vakar.
Cambridge, Mass.: MIT Press.
Waddington, Conrad Hall (1959). Evolutionary systems - animal and human. Nature
183: 1654-1658.
Wallerstein, Immanuel (1974). The Modern World-System: Capitalist Agriculture and
the Origins of the European World Economy in the Sixteenth Century. New York:
Academic Press.
Watson, James (1976). Molecular Biology of the Gene, 3rd ed. Menlo Park, Cal.: W. A.
Benjamin, Inc.
Wiener, Norbert (1948). Cybernetics. New York: John Wiley & ~ons.
Williams, George C. (1966). Adaptation and Natural Selection. Princeton: Princeton
University Press.
Wilson, Edward O. (1975). Sociobiology: The New Synthesis. Cambridge, Mass.: Har-
vard University Press.
Wilson, H. R., and J. D. Cowan (1973). A mathematical theory of the functional dynam-
ics of cortical and thalamic nervous tissue. Kybernetic 13: 55-80.
Winch, Peter (1958). The Idea of a Social Science. London: Routledge & Kegan Paul.
Worden, Frederic G., Judith P. Swazey, and George Adelman, Eds. (1975). The
Neurosciences: Paths of Discovery. Cambridge, Mass.: MIT Press.
Zadeh, Lofti A., and Charles A. Desoer (1963). Linear System Theory: The State Space
Approach. New York: McGraw-Hili.
INDEX OF NAMES
301
302 INDEX OF NAMES
Young, R. S. 76, 78
INDEX OF SUBJECTS
305
306 INDEX OF SUBJECTS
rationality 167
template 67 -69
reaction 54-60
thing 14
red uction 42, 75
thinking 164-166
reproduction 98-99
RNA 62, 64-70
rules 196 universe 9-10, 245-246
ACKNOWLEDGEMENTS XV
INTRODUCTION 1
1. Ontological Problems 1
2. The Business of Ontology 3
3. Is Ontology Possible? 6
4. The Method of Scientific Ontology 7
5. The Goals of Scientific Ontology 9
6. Ontology and Formal Science 13
7. The Ontology of Science 15
8. Ontological Inputs and Outputs of Science and Technology 19
9. Uses of Ontology 22
to. Concluding Remarks 24
1. SUBSTANCE 26
1. Association 26
1.1. Concatenation and its Ontological Interpretation 26
1.2. Axiomatic Foundation of Association Theory 28
1.3. Consequences 31
1.4. Atom Aggregates 36
1.5. Clustering 37
1.6. Historical Remark 38
2. Assembly 39
2.1. Juxtaposition and Superposition: Intuitive Idea 39
2.2. Formalization 40
2.3. Definitions 42
2.4. Some Consequences 45
309
310 CONTENTS OF 'ONTOLOGY I'
4. Concluding Remarks 55
2. FORM 57
1. Property and Attribute 58
1.1. Difference between Property and Attribute 58
1.2. Attribute-Property Correspondence 59
2. Analysis 62
2.1. Property in General and Property of a Particular 62
2.2. Intrinsic and Mutual, Primary and Secondary 65
3. Theory 69
3.1. Unarization and Dichotomization 69
3.2. Basic Assumptions and Conventions 72
3.3. Laws as Properties 77
3.4. Precedence and Conjunction of Properties 80
3.5. Similarity 85
3.6. Indiscernibility 90
4. Properties of Properties 92
4.1. Identity and Difference of Properties 92
4.2. Property Weight 94
4.3. Resultants and Emergents 97
4.4. Properties of Properties 98
5. Status of Properties 99
5.1. The Reality of Properties 99
5.2. A Critique of Platonism 102
5.3. The Problem of Universals 104
3. THING 110
1. Thing and Model Thing 110
1.1. Thing: Definition 110
CONTENTS OF 'ONTOLOGY I' 311
2. State 123
2.1. Centrality of the State Concept 123
2.2. State Function 125
2.3. Law Statements as Restrictions on State Functions 128
2.4. State Space: Preliminaries 131
2.5. Definition of a State Space 133
2.6. Equivalent Representations of States 136
2.7. State and State Preparation 138
2.8. Concluding Remarks 139
4. POSSIBILITY 164
1. Conceptual Possibility 165
1.1. Possibility Concepts 165
1.2. Four Concepts of Conceptual Possibility 165
1.3. Conceptual Possibility: Relative 168
3. Disposition 179
3.1. Intuitive Idea 179
312 CONTENTS OF 'ONTOLOGY l'
4. Probability 185
4.1. Abstract Concept 185
4.2. Probability State Space 187
4.3. Propensity Interpretation 190
6. Marginalia 199
6.1. Modal Logic and Real Possibility 199
6.2. Possible Worlds Metaphysics 202
6.3. Modality and Probability 204
6.4. Randomness 208
6.5. Probability and Causality 210
6.6. The Many Worlds Interpretation of Quantum Mechanics 211
5. CHANGE 215
1. Changeability 216
1.1. Preliminaries 216
1.2. Changeability 218
2. Event 221
2.1. The Ordered Pair Representation of Events 221
2.2. The Event Space 222
2.3. The Representation of Processes 226
2.4. The Space of Lawful Events 229
2.5. Keeping Track of Changing States 232
2.6. Rate, Extent, and Change Potential 237
3. Process 243
3.1. Serial Change: Types 243
3.2. General Concepts and Principles 251
6. SPACETIME 276
1. Conflicting Views 278
1.1. The Three Main Views 278
1.2. Approaches to Chronotopics Building 281
2. Space 283
2.1. Interposition 283
2.2. A Philosopher's Space 285
2.3. The Physicist's Space 287
2.4. Bulk and Shape 293
2.5. Concluding Remarks 294
3. Duration 296
3.1. Intuitive Idea 296
3.2. Before and After 297
3.3. Duration 301
4. Spacetime 305
4.1. Spacetime, the Basic Network of Events 305
4.2. Position in Spacetime 309
4.3. Change in Spacetime 312
BIBLIOGRAPHY 334