Self-Organization and Agency: In Chemistry and In Process Philosophy

by Joseph E. Earley

Joseph E. Earley is Professor of Chemistry at Georgetown University, Washington, D.C.

The following article appeared in Process Studies, pp. 242-258, Vol. 11, Number 4, Winter, 1981. Process Studies is published quarterly by the Center for Process Studies, 1325 N. College Ave., Claremont, CA 91711. Used by permission. This material was prepared for Religion Online by Ted and Winnie Brock.


Dr. Earley compares process philosophy with the science of chemistry — both open-system structures which exist in an antecedent world.

Nature abounds in compound individuals. Discrete, functioning entities are made up of components which are, in some sense, also individuals. Scientists sometimes need to be concerned with whether aggregates (e.g., species of plantsl) or components (e.g., quarks2) are ‘real’, but such questions are not generally regarded as having great importance for science. It has often happened, however, that scientific developments have had major significance for subsequent philosophical discussion of problems of the one and the many. Recently, there has been considerable increase in scientific understanding of the spontaneous development of spatial and temporal organization (structure) in physical, chemical, and biological systems.3 In an earlier note (PS 11:35), I suggested that this progress in science raises points that may be helpful in dealing with a question of current importance for process philosophy. This paper provides support for that suggestion. The first section introduces the philosophical problem. The middle sections provide brief nontechnical introduction to scientific concepts. The final section combines both topics.

I. The Problem of the Compound Individual

In rejecting Plato’s doctrine, and also the atomistic cosmology of Democritus, Aristotle had assigned fundamental importance to ousia (‘substances’ or ‘supposits’), the discrete persistent entities of ordinary experience (this ox, that tree) (NPE 45, 204). Whitehead’s rejection ofthat position was emphatic: "It is fundamental to the metaphysical doctrine of the philosophy of organism that the notion of an actual entity as the unchanging subject of change is completely abandoned" (PR 29/ 44).

Paul Weiss disagrees with his teacher’s doctrine, stating:

Each actuality is a substance. It maintains a hold on whatever it contains, produces, and intrudes upon. It persists and it acts. It has an irreducible, independent core, and receives determinations from insistent, intrusive forces. . . . If an actuality were not a substance, its parts would not belong to it, and it would disperse itself in the very act of making its presence evident. The very items which it dominates, it would not control; nor would it continue to be despite an involvement in change and motion. It would be inert and solely in itself, or it would be a mere event. In either case, it would not be a source of action. (FC 107)

Most interpreters agree that Whitehead, in his final synthesis at least, held that "the final real things of which the world is made up" (PR 18/ 27) are both microscopic and noncomposite. For instance, William Christian states "now actual entities and nexus belong to different categories of existence" (IMW 408). When John Cobb suggested that a chemical molecule might properly be regarded as an "enduring object" (CNT 89), Donald Sherburne rejected the idea, maintaining that "electrons and protons are structured societies" and stating that it was Whitehead’s opinion that "at the level of electrons and protons we have not yet gotten down to personally-ordered strands of actual occasions" (PPCT 89). Ivor Leclerc has summarized (NPE 286) the orthodox interpretation: "Fundamental in this [Whitehead’s] doctrine is that it is the ultimate constituents into which all compounds are analyzable that are the tine physical existents."

It is clear that Whitehead was keenly aware of contemporary scientific developments as he constructed his system. His final synthesis was completed during the period in which chemistry and nuclear physics were being systematized in terms of a small number of "elementary" (noncomposite) particles. The particle physics of our day offers no similar confidence that specific entities of any sort may correctly be regarded as simple. "Elementary" is used in an operational sense that varies with the circumstances.4 Entities which are properly regarded as elementary in low-energy experiments turn out to have definite and significant internal structure when higher energies are employed. To the extent that Whitehead’s actual entities must be both microscopic and noncomposite (as the orthodox interpreters hold), then to the same extent there is a problem as to how the category of actual entity could be applicable, as Whitehead required all his categories to be (PR 3/4), at least in the universe accessible to physical science.

Whitehead considered each enduring individual human person to be a society of actual entities and that at any moment in the life of a person there exists a "regnant" occasion which somehow "presides" (PR 109/ 167) in the human body. Rem Edwards (PS 5:195) discussed severe problems that arise in applying Whitehead’s categorial scheme to human individuals. Edward Pols agrees that Whiteheadian actual occasions are "miniscule," but finds a different interpretation necessary for "high-level" entities. In a subtle analysis of rational action and the rational agent Pols5 explores the consequences of considering both the agent and the action as being "primary in an ontological sense. From the standpoint of psychology, as well as from that of microphysics, Whitehead’s conceptual system seems to have difficulties that may not have been apparent some years ago.6

Justus Buchler has developed a system7 which seems to be the polar opposite of Whitehead’s position, as described above. "Whatever is discriminated in any way (whether it is ‘encountered’ or produced or otherwise related to) is a natural complex, and no complex is more ‘real’ or more ‘natural’ or more ‘genuine’ or more ‘ultimate’ than any other." In contrast to the doctrine, ascribed to Whitehead by Leclerc, that "the truly active entities must be identified with the ultimate constituents, those which are not themselves composite" (MN 104), Buchler teaches that there is no level of complexity which has any sort of priority, that any one sort of aggregate is, ontologically, just as good as any other sort.

Recently F. Bradford Wallack has claimed that the orthodox interpreters have all been mistaken; that Whitehead intended the category of actual entity (actual occasion) to include such macroscopic, long-lived, and diverse entities as Julius Caesar, the Castle Rock at Edinburgh, a system of nebulae. She defends the proposition that "the actual entity is any concrete existent whatever" (ENPWM 16ff.). It seems to me that Wallack’s interpretation of Whitehead is close to the doctrine of Buchler. It might seem to be a scandal to the faithful that a thesis so at variance with the received interpretation of Whitehead’s system should be given any hearing.

I agree with James Felt, S.J., (PS 10:57) that it is necessary to reject Wallack’s contention that Whitehead, in his final synthesis, propounded the doctrine that she advocates. However there are many passages in Whitehead’s writings (e.g., SMW 93f.) which indicate that the orthodox interpretation catches only a fraction of the richness of Whitehead’s thought. Some admixture of the Buchler-Wallack tinge might well be appropriate for a neo-Whiteheadian system which would be consistent with main currents in contemporary thought. Considerations that are connected with recent progress in understanding self-organizing chemical and physical systems lead me to propose a criterion for application of the category of actual entity that retains many of the distinctive features of the philosophy of organism while dealing with some of the objections raised against that system, and including aspects of the Buchler-Wallack approach.

II. Structure and Stability in Chemistry

A new kind of organization, called dissipative structure has recently been discovered3 in physical and chemical systems (SNS). Before dealing with this new development, we briefly review well-established concepts.

The diiodine molecule, I2, is composed of two iodine atoms joined by a covalent bond. Each of the two iodine atoms is composed of a positively-charged nucleus, a number of core electrons which comprise filled electronic shells, and an unfilled (valence) shell of electrons. The covalent bond arises from delocalization of the valence-shell electrons of both atoms, so that all these electrons may be regarded as pertaining to the molecule as a whole. The result of this bond-formation is that there is a single value of internuclear distance which is a defining characteristic of the molecule of diiodine. If, for any reason, the distance between the two iodine-atom nuclei should be less than that distance, there will be an unbalanced force (mainly arising from interactions of the filled shells) which will tend to increase the distance between the atomic centers. Conversely, if the atoms should be further apart than this equilibrium internuclear distance, then there would be an unbalanced force (primarily due to valence-shell interactions) which would tend to decrease the interatomic distance. At ordinary temperatures diiodine molecules vibrate continuously around the equilibrium internuclear distance, alteratomic distance. At ordinary temperatures diiodine molecules vibrate continuously around the equilibrium internuclear distance, alternately becoming stretched and compressed. The diiodine property of self-restoration after a disturbance.8

The flat, pungently-scented, purple crystals which are the most common form of elemental iodine consist of diiodine molecules packed in layers. Fairly weak forces within and between these layers arise from cooperative motions of the electrons in the outer electronic shells of the diiodine molecules. The structure of this regular aggregate of molecules cannot be described in terms of any one variable (in this respect it differs from the diiodine molecule, for which the single variable of internuclear distance sufficed). It is possible, however, completely to describe this structure in terms of a small number of independent parameters. Examples of such variables would be: the distance between layers, the distance between adjacent molecules, and soon. For a simple aggregate such as solid iodine, a half-dozen or so parameters would suffice; for more complex structures, a larger but still finite number of parameters would be needed. There is one geometric structure, corresponding to particular values for each of the defining parameters, which has the same property of self-restoration after disturbance that was noted for the molecule of diiodine. At any real temperature (one that is above absolute zero) all the components of the crystal are in motion, but all remain near their equilibrium positions.9

A purple haze can be seen in closed vials containing iodine crystals. The forces holding diiodine molecules in the crystal are modest, and liquid iodine is not stable at ordinary temperatures and pressures, so that solid iodine sublimes to give a rather large concentration of diatomic molecules in the vapor phase. At any given temperature, there is a particular concentration of gaseous iodine that is in equilibrium with solid iodine. If the vapor is pumped away, more solid will sublime; if additional gas is introduced, new crystals will form. The combination of solid iodine and one particular concentration of gaseous iodine is similar to the spatial structures previously discussed, in that this arrangement has an intrinsic capacity for regeneration after a modest perturbation.10 Both of these classes of arrangements are called equilibrium structures. They have the capacity to persist in closed systems indefinitely.

Just as the diiodine molecule vibrates about its equilibrium internuclear distance, and molecules in solid iodine jiggle around in three dimensions, still keeping the overall structure of the crystal intact, so, too, there are fluctuations in the concentrations of all chemicals which participate in chemical equilibria. In each small region of space the concentration of a given reagent is not quite constant, but undergoes increases and decreases, usually small.

III. Chemical Dissipative Structures

Ordinary chemical reactions are messy affairs. Customarily, chemists hold that any real reaction may be regarded as being composed of a number of elementary steps. The defining characteristic of an elementary step is that its rate is proportional to the product of the concentrations of the reactants involved in that step. An elementary step that will be important for our purposes involves the reaction of one molecule, called Y, with two molecules of another kind, denoted X. To the extent that this reaction is elementary, the reaction velocity doubles as the concentration of Y doubles, but increases by a factor of four if the concentration of X doubles. All real reactions are composed of many elementary steps, and the equations which tell how rate varies with concentration for real reactions are more or less complicated algebraic combinations of the simple equations which apply to the several elementary steps.

When many chemicals interact, the rate of production (or of destruction) of each molecular species is influenced by the concentrations of all the others, as well as by such environmental factors as temperature and whether or not the system is illuminated. Increasing the concentration of X, say, may speed production of molecule P, but, at the same time, it may retard production (or increase the rate of destruction) of species Y. In the vast majority of cases that have been studied, mixtures of chemicals change in composition, as time proceeds, in such a way as to approach a condition of chemical equilibrium.

Novel chemical principles are illustrated by the rather simple reaction sequence:


B + X Y + Q


Y + 2X 3X

This sequence of reactions describes the conversion of A and B (the reactants) to P and Q (the products) while X and Y are both formed and destroyed (intermediates). The first three reactions are not remarkable, but the fourth step (used above as an example) has the unusual feature of being autocatalytic. That is, because this reaction leads to net production of species X, and the rate of reaction increases as the concentration of X increases, this fourth step keeps getting faster and faster.

In the particular reaction sequence shown, the second and third reactions increase in rate as the concentration of species X increases, but these two reactions use X up, rather than producing it. Differential equations can be used to state the same relationships that the chemical equations express. The concentrations of A, B, P, and Q are held at constant values throughout the system for a single experiment, but these values can be varied from one experiment to another. For certain values of these adjustable parameters, a remarkable result is found. Concentrations of X and Y do not smoothly approach their equilibrium values (as do concentrations of all chemicals involved in ordinary reactions) but rather these concentrations oscillate around the condition of equilibrium. That is, the concentration of X repeatedly increases above the equilibrium value, then falls below it. Meanwhile, the concentration of Y is varying in the converse sense, being lower than its equilibrium value when X is high, and higher than its equilibrium value when X is low. This reaction sequence describes a chemical oscillator. In recognition of the extensive study of this abstract model by Ilya Prigogine and his associates in Belgium (SNS), this mechanism is called "the Brusselator."

A real chemical reaction which gives rise to repeated oscillations of solution color (red to blue, and back again) has been discovered by Russian workers. (It is called the Belousov-Zhabotinskii, or B-Z, reaction.) The detailed sequence of elementary steps (about twenty) which must be involved in the B-Z reaction has been worked out by Richard Noyes.11 This mechanism (called "the Oregonator" in honor of the location of Noyes’s laboratory) is well understood and involves nothing but ordinary chemistry, but it is too involved to discuss here.

Both the Oregonator (the model of the mechanism of an actual chemical system) and also the Brusselator (an abstract scheme which lends itself to study and simulation) have the important feature that they give rise to true limit cycles. What this means is that the values (for various times) of the variables (concentrations of the intermediates X and Y for the Brusselator) that are solutions of the differential equations all lie on a single closed trajectory. That is to say, there is one unique sequence of states of X concentration (and a single definite sequence of states of Y concentration which is associated with the X sequence) which will be followed by the oscillating system. For given values of system parameters (temperature, reactant concentration, etc.), each oscillation will be a duplicate of the previous one, once an initial ‘warm-up’ period is over.

A system that is on a limit cycle may be said to be stable in a sense that is related to, but differs in an important way from, the kinds of stability that have previously been described. Any perturbation that causes the system to be in a state that lies off the limit cycle will cause chemical reactions to occur that will bring the system closer to that cycle. (This must be the case if the limit cycle is the only set of states that satisfies the defining differential equations, under the environmental conditions that prevail.) Stability is the defining property of structure. The self-restoring characteristic of the limit cycle indicates that some novel sort of structure is present.12

The important difference between this new kind of structure and the kinds previously discussed is that the new sort exists in open rather than in closed systems. In the Brusselator, for instance, reactants (A and B) are continually supplied and converted to the products (P and Q). Somehow, these products must leave the system, if the concentrations of P and Q are to remain constant. In order for the network of reactions to operate, and to maintain the system on the limit cycle, part of the energy-content of the reactants is dissipated in formation of lower-energy products. To maintain constant temperature, this energy must be removed. Heat must flow out of the system. The system must be capable of exchanging both mass and energy with the surroundings: it must be open. The new sort of structure which forms under such conditions is called "dissipative structure." Such structures should clearly be distinguished from "equilibrium structures" which persist only in closed systems.

The stability of the diiodine molecule results from interplay of mutually opposing forces. The repulsion of inner-shell electrons is balanced by attraction resulting from delocalization of valence-shell electrons. The stability of the crystal of elemental iodine is due to an analagous balance, but one involving a larger number of factors. The stability of the chemical dissipative structure is due to a balancing of processes. The autocatalytic production of intermediate X is balanced by two processes using X up, and supplemented by formation of X from A. There is a balance struck, but one that leads to oscillation, rather than to stasis.

In describing chemical oscillators, I said that the concentrations of the intermediates X and Y alternately went above and below their equilibrium values. Strictly speaking, that was a misstatement. Chemical dissipative structures occur only under conditions that lead to what is called ‘chemical instability’. That is, a situation for which there is no ‘equilibrium’ Solution of the defining differential equations, but rather at least three steady-state solutions, some of which are not stable. For instance, in the case of the Brusselator, there is a single steady-state solution if the concentrations of reactants (A and B) is not large,13 relative to the concentrations of products (P and Q). However, if products are removed and/or reactant concentrations are increased, we come to a critical concentration ratio (SNS 98). Once that ratio is exceeded, there turn out to be not one but three steady-state solutions to the defining equations. Two of these steady-state solutions are stable, in the sense explained previously (small perturbations engender forces which tend to restore the system to the steady state). The third solution (the one that corresponds to the equilibrium state that had existed before the critical concentration-ratio was exceeded) is an unstable steady-state solution. It is possible (at least in principle) to arrange matters so that the system is in this unstable steady state. Barring fluctuations, the system will remain in this condition indefinitely. But should the system suffer any perturbation, however infinitesimal, the state of the system will start to change, and that change will continue until one or the other of the two steady states is reached.

An interesting question is: how does the system decide which of the two possible nonequilibrium steady states to approach? The answer to this contemporary version of the problem of ‘Buridan’s Ass’ turns out to be that the detailed nature of the initial perturbation makes all the difference. From a given initial state, a system will change in one sense if stimulated by certain fluctuations and will change in entirely other ways if set in motion by different infinitesimal disturbances (SNS 289).

One of the most striking features of the B-Z reaction is its propensity to generate spectacular patterns ofcolor14 that retain their definition while undergoing repeated color changes. Using computer simulations based on the simpler Brusselator model, it is possible to understand the development of such spatial order. The diffusion of chemical species through the reaction medium requires time. In cases for which reactant and product (A,B,P,Q) concentrations are not held constant but are allowed to follow the usual (nonlinear) equations which govern ordinary diffusion, development of spatial order is to be expected, if the chemical equations of the Brusselator apply.

Pretty examples of this feature of chemical dissipative structure are given in computer-simulations of the diffusion-dependent Brusselator in a one-dimensional medium (SNS 124). The four reactions shown above are considered to occur in a tubular container. The concentrations of A, B, P, and Q are held fixed at the ends of the tube, but, in the body of the tube, these concentrations vary in accordance with the kinetic relationships that result from chemical reactions going on, and also in accordance with the usual laws of diffusion. There is a steady-state solution which corresponds to a monotonic variation of concentrations between the ends and middle of the tube. Once the reactant concentrations have been increased (or the product concentrations have been decreased) so that the critical ratio has been exceeded, that steady state becomes unstable, and two other steady-state solutions come into existence.

This experiment illustrates some features of chemical self-organization that seem to be especially important from the point of view of process philosophy, so it seems worthwhile to present additional heuristic discussion. Suppose we start with a tube containing an orange jell, and we provide reservoirs containing a green-colored reagent at both ends of the tube. In the presence of a high concentration of the reagent the tube is green, for an intermediate concentration it is white, and for a low concentration, orange. Inside the tube, the Brusselator reactions (given above) occur. The subcritical steady state would have the end regions similar and the center different: like the Nigerian flag, green-white-green. As the concentration of reagent is increased, the system would be taken beyond the chemical instability, and there would be two equally probable nonequilibrium steady states. To continue our simile, one of these would resemble the flag of The Ivory Coast, orange-white-green, while the other would resemble the flag of Ireland, green-white-orange. In the computer simulations, the concentrations at the ends of the tubes can be varied in a regular manner. Under that simulated change the initial steady state, once set up, persists far into the regime of instability, providing no adventitious variations of concentrations away from that steady state are allowed. However, if fluctuations (that exceed certain minimum values in magnitude and extent) are included, then the instability of the initial (Nigerian) steady state is demonstrated, and the system approaches one (Irish) or the other (Ivory Coastian) of the two stable nonequilibrium steady states. The interesting point is that which of the two steady states is approached depends on the details of the initiating fluctuation.

In the computer simulations, the start-off perturbations can be specified to any desired degree of accuracy. In real chemical systems, the B-Z reaction for instance, quite similar bifurcation of spatial solutions occurs. Decision-points of the same sort arise repeatedly in the natural development of the embryo of every biological organism. There are many situations in biology15 and economics16 in which opposing nonlinear relationships give rise to instabilities that are quite analogous.

For real systems, things are so involved that it makes little sense to talk of the details of the initiating fluctuation. Any measuring device which was adjusted so as to gather information about the macroscopic changes which follow upon decision would not be able to register information about parameters which distinguish one sort of fluctuation (Irish) from the other sort of fluctuation (Ivory Coastian). In Whitehead’s words, these transitions must be regarded as "internally determined, but externally free" (PR 27/ 41). That is, the system is governed by deterministic laws, and the nature of the possible steady states are fixed by these laws (the differential equations corresponding to the chemical equations which have been given previously). But these deterministic laws allow for several steady-state solutions. (There are several ‘real potentials’ [PR 65/102] accessible.) The choice between these possible futures is made, in a condition of chemical instability, by stochastic (nondeterministic) processes, such as random internal fluctuations of the system.

One additional feature should be mentioned before we conclude this discussion (regrettably technical) of computer simulations of development of spatial structure by the Brusselator. For some combinations of external conditions and internal fluctuations, regions of high order can occur imbedded in larger regions of a lower degree of order (SNS 136). (At the risk of straining an analogy beyond usefulness, I call attention to the flag of the Republic of South Africa.) These computer simulations of an abstract model (closely related to the features which characterize self-organizing chemical and biological systems) clearly show that the origin of intrinsic capacity for self-definition can be understood in detail, at least in relatively simple cases.

IV. Process and Agency

In the passage quoted previously, Weiss contends that neither a process nor an event can be a source of action. After a study of earlier opinions on such questions (NPE 1-253), Leclerc comes to a conclusion that seems to be close to the position that Weiss advocates. Leclerc proposes that, in certain cases, compound individuals can function as unitary, persistent sources of agency. When the acting of component entities on each other" is "fully reciprocal," then "the compound entity acts as a whole, that is, as one, with reference to, and on, other wholes." This, he suggests," is how many substances are to be conceived as constituting one substance" (NPE 113). Leclerc does not further specify what he means by fully reciprocal action.17

Considering that the world is, at bottom, composed of ‘elementary particles’ amounts to using bricks or building-stones as models of actual entities. Using ‘fully reciprocal action’ as a criterion would qualify a diatomic molecule like diiodine or, less naively, a set of molecules in chemical equilibrium, as a model of an actual entity. Regrettably, both of these classes of models run into the objection that Weiss puts so trenchantly. A molecule, or a chemical equilibrium, is a closed-system structure. No such structure can persist if it is engaged in any significant relationship with the rest of the world. As Weiss correctly observes, "it would disperse itself in the very act of making its presence evident" (FC 107). If the container in which solid and gaseous iodine are at equilibrium be opened, the odor of iodine will fill the room, but the equilibrium will be destroyed, and the vessel will, soon enough, be empty.

Whitehead’s principle of relativity (PR 148/ 224) requires that the interaction of an entity with the rest of the world be intrinsic to that entity’s existence. Any criterion of applicability for the first category of existence (the category of actual entity or actual occasion) must involve consideration of at least some of the details of interactions between entities. Dissipative structures, in contrast to equilibrium structures, exist in open-systems and require interaction with the rest of the world in order to maintain themselves in being. In this respect, at least, dissipative structures are more widely-applicable models of reality than are equilibrium structures. Before proposing that there may be a relationship between dissipative structures and actual entities, we give a brief ‘genetic analysis’ of chemical self-organizing systems.

A chemical dissipative structure comes into existence when environmental conditions of a complex open system change, so that an equilibrium state becomes unstable, and a network of relationships (among components of the system) gives rise to several nonequilibrium steady states. In nature, changes of environmental conditions arise from such sources as the melting of polar ice-caps, explosion of dwarf stars, the fall of night. Highly complicated chemical mixtures can be found in any mud-hole. (Simple chemical systems are products of high art.) Autocatalysis is required for existence of dissipative structure, but natural systems at many levels have this feature. The kinds of nonlinear relationships needed to give multiple steady states are common in physical, chemical, biological, and social interactions. The type of organization modeled by the B-Z reaction and by the Brusselator is quite widely distributed.

Once change in system parameters has led to a condition of chemical instability, a stochastic internal fluctuation (or a perturbation from the outside) sets off an evolution (change) of the composition of the system. For instance, in the B-Z reaction the ‘decision’ as to which of several possible spatial structures will actually form may (in the absence of external stimulation) be made by random fluctuation, a fortuitous local excess of one reactant. Subsequently, the system may reach a new steady state, or alternatively, a limit cycle involving several quasi-steady states. In these two cases, a dissipative structure, perhaps with temporal and spatial inhomogeneity, will have come into existence. If neither a steady state nor a limit cycle is reached, the system will continue to evolve. Eventually, whatever features had defined it will be changed, so that the system will become indistinguishable from the surroundings.

In a new dissipative structure (either a steady state or a limit cycle), concentrations of chemicals and rates of mass-transfer between the system and surroundings will not be those characteristic of the previous equilibrium state (or the nonequilibrium steady state that corresponds to it under conditions of instability) but rather they will be the (perhaps quite different) average values which pertain to the new structure. What this means is that the closure of the network of relationships (such as the set of four equations that defines the Brusselator) has been the occasion of (caused) a change in the relationship between a system and its surroundings. Such closures of relationships are not restricted to any particular spatial or temporal scale. Microscopic, short-lived structures of this sort are possible, as are large ones with long characteristic time-scales.

Lewis S. Ford has pointed out that various stages in the development of Whitehead’s doctrine can be discerned by textual analysis (PS 8:145). In the stage that Ford (1:251) has called his ‘first metaphysical synthesis’, Whitehead held that ‘value’ ("the intrinsic reality of an event") could exist on a multiplicity of levels, not just a microscopic one (SMW 93). Ford (PS 3:109) has suggested the characterization ‘divisible but undivided’ as being more appropriate for Whiteheadian actual entities than descriptions such as microscopic or noncomposite. During the years between the completion of SMW and the publication of PR, Whitehead tried out an idea that he does not seem to have included in his final synthesis.18 "Whenever the ‘all or none’ principle holds, we are in some way dealing with one actual entity, and not with a society of such entities, nor with the analysis of components contributory to one such entity" (S 28). I suggest that a coherent neo-Whiteheadian synthesis might result from development of this point of view. There is an ‘all or none’ character about the closure of the networks of relationships that define dissipative structures. Closures of this type provide a basis for understanding how entities, in interaction with others, can be ‘divisible but undivided’.

Each actuality is related to other entities by prehensions, which are components of the concrescences of the others. Prehensions are ‘simple’ if they arise from objects which are single actual entities, but are ‘transmuted’ if they arise from a nexus which functions as a unit, insofar as a particular concrescence is concerned (PR 232). The notion of a simple prehension is related to the assumption that there do exist ultimate components which are not composite. Philosophically significant problems involve highly transmuted interactions. It is important to show how effective transmutation of a multiplicity into a unitary source of action can be achieved. Consideration of some details of physical interactions involving chemical dissipative structures clarifies how effective transmutation may occur for actual entities in general.

Any interaction that can be studied by physical or chemical techniques can be understood in terms of one or more of four classes of fundamental forces (gravitational, electromagnetic, strong, weak). All of these are held to operate by exchange of quanta. Every observing device can be understood on this basis, as functioning by quanta exchange. Sufficiently sensitive detectors may, perhaps, register single such events, but most observing devices, human eyes for instance, are sensitive only to resultants of a myriad of such transactions. Whether or nor particular quanta considered in this kind of an explanation are capable of further analysis, or yet other sorts of forces (and quanta) are subsequently discovered, is not, after all, the most significant point. Since important observing devices function by transmutation, the unification, rather than the generally unobservable ultimate unit, is of principal interest. In this connection, it seems that insufficient attention has been paid to variation in response-times of percipients.

No observer functions instantaneously. There is a characteristic resolution-time for every percipient. Whether a subject finds an object to be a multiplicity or a unit depends on the relationship between the time-parameter appropriate to the object and those that pertain to the percipient.

In the case of chemical dissipative structures, there is a time-parameter which is characteristic of the structure as a whole19 (roughly, the time it takes for the system to traverse the limit cycle). Any observing device that has a time-scale short with respect to the time-parameter of the dissipative structure (as a whole) will register a regular and repeated change of the properties of the system. (The system will be perceived as a multiplicity.) On the other hand, any percipient with a slow response-time, relative to the cycle-time of the structure, will register a steady effect due to the system (now perceived as a unit). This sort of transmutation occurs quite widely. It is difficult (impossible, I would say) to think of a specific perception or interaction which does not involve this kind of transmutation.

The thesis I want to propose is that a compound individual should be considered to be one ‘actual entity’ or one ‘actual occasion’ if, and to the extent that, particular percipients interact with that entity as a unified source of effective action. Patterns of relationships among component parts of the compound individual, exchange of components between the entity and its surroundings, adventitious fluctuations, all are important in concrescence, but characteristics of percipients are also important. It would be high abstraction to inquire whether a certain thing is or is not properly regarded as an actual entity, apart from consideration of the interaction of that entity with others.

It might be objected that all that has been done here is to amplify Whitehead’s dictum, "For physics, the thing itself is what it does" (AI 157). A critic might claim that we have ignored what Neville calls ‘the ontological perspective’ (2:79).

The ‘reformed subjectivist principle’ (PR 167/ 254) requires that satisfaction of ‘subjective aim’ be an intrinsic feature of all actualities. Achievement of subjective aim belongs to the ontological aspect of an actual entity. The relationship, if any, between achievement of subjective aim and interaction with the others is not obvious. It is clear, though, that Whitehead teaches that the efficacy of an actual entity in the ‘creative advance into novelty’ (PR 349/ 530) is intimately related to satisfaction of subjective aim. As Felt (PS 10:57) has pointed out, what needs to be done is to elucidate how the ontological aspect (achievement of subjective aim) of an actual entity might be connected with significant interaction of that entity with others (the epistemological aspect).

In the case of chemical dissipative structures, closure of a defining network of relationships makes a difference for the rest of the world. For certain percipients at least, that difference is such that the system, as a whole, is perceived as, and therefore is, a unified center of agency. This set of relationships is capable of being expressed in mathematical (abstract) terms. If one considers the structure, as a whole, as one actual entity then this set of relationships would be the corresponding subjective aim.

Considering, as I suggest, certain compound individuals as single actual entities does not entail a major change in the interpretation of the ontological principle ("actual entities are the only reasons," PR 24/37). It does require that we attend to the point that: "The ultimate metaphysical principle is the advance from disjunction to conjunction, creating a novel entity other than the entities given in disjunction. The novel entity is at once the togetherness of the ‘many’ which it finds, and also it is one among the disjunctive many which it leaves; it is a novel entity, disjunctively among the many entities which it synthesizes. The many become one, and are increased by one" (PR 21/32, emphasis added). There is no single level of size or time that has unique status. ‘All or none’ principles operate on many levels and give rise to effective actualities of many sorts. Each one of these entities is, by reason of specific closures of relationships, ontologically unified. Every entity defined by such a closure is one actual entity, as well as being a nexus of components.

Main concerns of process philosophy can be dealt with using the approach advocated above. It is not possible to treat these extensively here, but brief notes may indicate the applicability of the point of view that I favor.

The existence of a steady-state solution or of a limit cycle depends on the system parameters (including the initial concentrations of all the chemicals present) and also on what Whitehead terms ‘the primordial nature of God’. Whether or not a new structure emerges is influenced by the conditions that prevail, but the existence (or nonexistence) of high-level interaction patterns, of the sort that mathematicians might discover, is also involved. There is a sense in which these patterns of relationships are entailed in the characteristics of lower-level components, but in a more significant sense, these relationships exist at the level of the emergent structure rather than at the level of the components. The property of X that is significant for origin of the dissipative structure is how-it-reacts-with-Y. It is impossible to discuss such properties for X apart from discussion of X and Y together in the presence of A, B, P. and Q (in short, the system).20

An important task for process philosophy is to clarify the sense in which an entity may be said to retain self-identity while being engaged in action. The time-period characteristic of a dissipative structure is the time required to traverse the limit cycle. The system functions as a society with personal order. Each circuit (oscillation) may be considered a single occasion. (It would be egregious misplacement of concreteness to be concerned about at which part of the cycle one occasion leaves off and the next begins.) Successive members of this personally-ordered society share the same defining network of relationships (subjective aim) and are, thus far, the same. They do differ numerically, as do the many loops that constitute a coil spring. Dissipative structures require continual interchange of mass and energy between system and surroundings. The system must therefore be engaged in continuous, significant interaction with the rest of the world. The world cannot remain exactly the same over the time that the dissipative structure exists. No two circuits of the limit cycle (occasions) are precisely alike in all particulars.

It is possible to deal with ‘upper-level’ or hierarchical structures without resorting to ‘regnant’ occasions. Since percipients encounter relatively persistent sources of agency, patterns of interactions of these agents with each other can give rise to compound individuals of indefinitely-high degrees of complexity.21 Whitehead often wrote, particularly in his less-formal work, as if entities of various sizes could properly be classed as actual entities. Wallack (ENPWM) has developed this idea, but, in my opinion, she has leaned too far in the direction of the ontological universalism that Buchler7 advocates. To consider all heaps, aggregates, and mélanges as equally actual misses the insight that "the ultimate metaphysical truth is atomism" (PR 35/ 53). This seems to be a serious mistake. There do exist, on many levels, patterns of interaction that give rise to structures, that provide the basis for well-founded discriminations of unity. Not every nexus or ‘natural complex’ should be considered to be an actual entity. Some, but not all, compound entities have the property of retaining self-identity while engaged in action, and therefore have an ontological status that is different from the status of those nexus which lack that property.

Formation of open-system structures (whether considered from the point of view of chemistry, or from that of process philosophy) requires structures (of some stability) to exist in an antecedent world. But novel concrescences emerge, due to networks of relationships involving these antecedent structures and also due to specific, contingent fluctuations. The point that needs to be emphasized is that when relationships between components give rise to open-system structures then novel concrescences attain ontological parity with antecedent and component structures. I do not claim that this was Whitehead’s final doctrine, but I do maintain that it accords well with many of the distinctive features of ‘the philosophy of organism’ and is better suited to further development, and to application, than is the orthodox interpretation of Whitehead’s final synthesis.



CNT -- John B. Cobb, Jr. A Christian Natural Theology. Philadelphia: Westminster Press, 1965.

ENPWM -- F. Bradford Wallack. The Epochal Nature of Process in Whitehead’s Metaphysics. Albany: State University of New York Press, 1980.

FC -- Paul Weiss. First Considerations. Carbondale: Southern Illinois University Press, 1959.

IWM -- William Christian. An Interpretation of Whitehead’s Metaphysics. New York: Yale University Press, 1959.

MN -- Ivor Leclerc. "Some Main Philosophical Issues Involved in Contemporary Scientific Thought," in John B. Cobb, Jr., and David Ray Griffin, eds., Mind in Nature. Washington: University Press of America, 1977.

NPE -- Ivor Leclerc. The Nature of Physical Existence. New York: Humanities Press, 1972.

PPCT -- Donald W. Sherburne. "Whitehead Without God," in Delwin Brown, Ralph E. James, Jr., and Gene Reeves, eds., Process Philosophy and Christian Thought. Indianapolis: Bobbs-Merrill, 1971.

SNS -- Gregoire Nicolis and Ilya Prigogine. Self-Organization in Non-Equilibrium Systems. New York: John Wiley and Sons, 1977.

1. Lewis S. Ford. "Whitehead’s First Metaphysical Synthesis," International Philosophical Quarterly 17 (1977), 251-64.

2. Robert Neville. "Authority and Experience in Religious Ethics," Logos 1 (1980), 79-92.



1Donald A. Levin, "The Nature of Plant Species," Science 204 (1979), 381-84; R. R. Sokal, "Population Differentiation: Something New or More of the Same?" in Ecological Genetics, Peter F. Brussard, ed., (New York: Springer, 1980).

2 Kenneth A. Johnson, "The Bag Model of Quark Confinement," Scientific American, 241/1 (1979), 112-21; Yoichiro Nambu, "The Confinement of Quarks," Scientific American 235/5 (1976), 48-71. After reviewing the evidence that neutrons, protons, and some other subatomic particles consist of yet smaller bits (the quarks), Nambu presents arguments to establish that it will never be possible to observe an uncombined quark. He then asks: "If a particle cannot be isolated, or observed, even in theory, how will we ever be able to know that it exists?"

3Advances related to those discussed here have been made in several areas of science. This paper is most closely related to the work of Ilya Prigogine in far-from-equilibrium thermodynamics. Professor Prigogine reviewed his work on the occasion of receiving the Nobel Prize for Chemistry in 1977. This review was published as "Time, Structure and Fluctuations" in Science 201 (1978), 777-85. He has pointed out some of the philosophical implications of his work in a book for persons literate in physical chemistry: From Being To Becoming (San Francisco: W. H. Freeman, 1980). In collaboration with a professional philosopher, he has published a work for the general reader (I. Prigogine and I. Stengers, La Nouvelle Alliance [Paris: Gallimard, 1978]). A revised version of that work, in English, is in preparation. I am grateful to Professor Prigogine for hospitality at the Free University of Brussels during the fall semester of 1976.

4 E.g., R. Rajaraman, "Elementarity of Baryon Resonances in Nuclear Matter in Mesons in Nuclei, volume 1, Monnique Rho and Denys Wilkinson, eds., (Amsterdam: North Holland. 1979).

5Edward Pols, "The Ontology of the Rational Agent," Review of Metaphysics 33 (1980), 689-710. See also Edward Pols, "Human Agents as Actual Beings, PS 8:103-13 (1978).

6Douglas H. Hofsteader has summarized current topics in the study of hierarchically-organized systems in an excellent hook for general readers: Goedel, Escher, Bach (New York: Basic Books, 1979). Herbert L. Simon has made remarkable progress in related areas. Some of these developments are significant for the problems Whitehead treated, but appear not to he fully consistent with his final synthesis, as usually interpreted.

7Justus Buchler, Metaphysics of Natural Complexes (New York: Columbia University Press, 1966), p. 31. See also P. Hare and J. Ryder, PS 10:120.

8 The relationships that define molecules like diiodine are frequently discussed in terms of potential energy. The configuration corresponding to the equilibrium internuclear distance has lower potential energy than any other configuration. Stretching the molecule, or compressing it, increases the potential energy of the molecule. The diiodine molecule is said to occupy a potential energy ‘well’, an inferred depression in an imagined surface in a hypothetical potential energy-configuration ‘space’. Closely related statements can be made about any entity that can be said to ‘exist’, in the sense in which that word is used by chemists.

9 For any such structure it is possible to imagine (if not actually to construct) a surface in a multi-dimensional space which represents how potential energy varies with all of the parameters that define the structure. The structure of the diiodine molecule corresponds to a minimum in potential energy as the single parameter, internuclear clear distance, varies. The structure of the crystal of elemental iodine also corresponds to a minimum in potential energy, but now more than one parameter is important. In both cases the imagery of a patio a potential energy surface is used. Every chemical entity, properly so called, can be regarded as corresponding to such a pit (whatever its dimensionality might be). The defining characteristic of chemical structure is that there exists a configuration such that deviations from that configuration entail forces (broadly used) such that there is a tendency to restore the reference configuration. (Chemists sometimes discuss entities, such as exiplexes and activated complexes, that do not correspond to potential energy minima, but these are not regarded as being ‘real chemical species’.)

10 J. Willard Gibbs showed that each chemical equilibrium can be considered to correspond to a minimum in a free-energy-composition space, closely analogous to the potential energy-configuration pit which corresponds to every spatial structure.

11 Richard M. Noyes and Richard J. Field, "Mechanisms of Chemical Oscillators: Experimental Examples," Accounts of Chemical Research 10 (1977), 273-80; Joel Keizer, "Nonequilibrium Thermodynamics and the Stability of States Far from Equilibrium," Accounts of Chemical Research 12 (1979), 243-49; Richard M. Noyes, "Oscillations in Homogeneous Systems," Ber. Bunsenes Gesell. (Physical Chemistry) 84, (1980), 295-303. (This issue is devoted to articles on various aspects of chemical oscillators.)

12 No analog for potential energy or reaction free energy is now known to be associated with this sort of structure. Discussion of whether a criterion of relative stability of nonequilibrium steady states exists is currently being carried out in the literature.

13The ‘chemical affinity’ of a system is related to the ratio of concentrations of reactants to that of products, both compared to equilibrium values, and is a measure of distance from equilibrium.

14Arthur T. Winfree, "Rotating Chemical Reactions." Scientific American 230/6 (1974), 82-95.

15 (a) Armin Weiss, "Replication and Evolution in Inorganic Systems," Angew. Chem. Intl. Ed. Eng. 20 (1981), 850-60. (b) Manfried Eigen and Peter Schuster, The Hypercycle, A Principle of Natural Self-Organization (New York: Springer, 1979). (c) Manfried Eigen and Ruthild Winkler, The Law’s of the Game: How the Principles of Nature Govern Chance (New York: Knopf, 1981). (d) Hans Kuhn and Juerg Wasser, "Molecular Self-Organization and the Origin of Life," Angew. Chem. Intl. Ed. Engl. 20 (1981), 500-19. (e) H. Fredrik Nijhout, "Color Patterns of Butterflies and Moths" Scientific American 245/5 (1981), 140-53. (1) Gunther S. Stent and David A. Weisblat, "The Development of a Simple Nervous System" Scientific American 246/1 (1982), 136-47. (g) Charles J. Lumsden and Edward O. Wilson, Genes, Mind and Culture (Cambridge: Harvard University Press, 1981).

16 Thomas C. Schelling, Micromotives and Macrobehavior (New York: Norton, 1978).

17 For application and some further specification of this position, see Ivor Leclerc, Motion, Action and Physical Being," International Philosophical Quarterly 21 (1981), 17-26, and Ivor Leclerc, "The Metaphysics of the Good," Review of Metaphysics 35 (1981), 3-26, and references therein.

18 The question of why Whitehead chose to develop the position he did develop, rather than this one, cannot be discussed adequately here. Whitehead’s attack on ‘scientific materialism’ in SMW and his eventual adoption of a position which is interpreted by many, to teach that there are two sorts of actualities, microscopic physical entities and human persons (e.g., Wallack, ENPWM 6), calls to mind Pope’s lines (An Essay on Man, IV):

Vice is a monster of so frightful mien,

As to be hated needs but to be seen;

Yet seen too oft, familiar with her face,

We first endure, then pity, then embrace.

19 The existence of a limit cycle requires that there be two time-scales in the underlying chemical mechanism (e.g., John W. Avitabile, Jr., and Andrew C. DeRocco, "Time Delay, Epigenetic Dynamics and Circadian Oscillations," Biophys. Journal 15 [1975], 180), but the time-scale characteristic of the dissipative structure is not simply related to any underlying process; rather, it arises by the interaction of these component reactions.

20 It has been argued that subatomic particles require a different logic than the one that applies to macroscopic entities. Different levels involve diverse principles. It is inappropriate to ground a scheme of thought on any single level. R. I. C. Hughes, "Quantum Logic," Scientific American 245/4 (1981), 202-13. See also Douglas R. Hofstadter, "Strange Attractors: Mathematical Patterns Delicately Poised Between Order and Chaos," Scientific American 245/5 (1981), 22-43.

21 Lynn Margulis, Symbiosis in Cell Evolution (San Francisco: Freeman, 1981).