Process as an analytical unit suited to historiography of the twenty-first century

By Haines Brown, 13 September 2003

The most obvious thing about historical phenomena is that they represent change in time. We may wish to describe an arbitrary “initial state” of a system in terms of a set of characteristics, and we might define its final outcome as the set of features we observe at a time subsequent to the initial state and functionally related to it. However, we are fully aware that these are artificial constructs imposed on a fluid process that continually changes and cannot be adequately described in terms of fixed properties. An initial state and an outcome are arbitrary demarcations in the continually flowing river of history.

While we have a clear subjective sense of change in the present and no doubt of its reality in history, it is very difficult to convey it in empirical terms—that is, in terms of a set of static qualities. As soon as we associate some empirical quality with a process, it ceases to be a process to become a static structure in thought—a kind of Heisenberg indeterminancy principle, if you will.

But let me step back a bit and make sure I've not outreached shared assumptions. Most things are defined in the dictionary in terms of their essential qualities. What is it that makes these particular qualities “essential,” while others are mere passing accidents? It turns out that it is those qualities that seem to persist that are thought to be essential. We then proceed to use these essential qualities to construct conceptual categories in terms of which we are able to communicate with one another other about classes of things. Accidental qualities, such as the color of an automobile, serve only to distinguish the members of a class.

There's no denying that these quite artificial procedures work quite well in practice because persistent qualities are by definition the same ones perceived by the others with whom we come into contact. The problem is that this utility is limited to a relatively short range. It is meaningful for individuals, whose experience is limited to mere decades (or centuries in cultural terms), and geographically constrained (although we are becoming globalized today, language still divides us). As we expand our view beyond that of the individual or specific culture, its content becomes more complex and change more evident.

That is why historiography, except for very local accounts, is not biography writ large, for it finds complexity and change to be the normal state of affairs, and consistency the exception. In history, phenomena are in principle fluid and cannot be frozen in time by static categories. In fact, there is no reason whatsoever to assume that the useful conventions of daily life are at all appropriate in historiography.

There are situations in which we do try to express change in empirical terms. For example, a movie film represents movement by displaying a succession of still frames. We derive from it the impression of motion only because our sensory apparatus cannot distinguish such a rapid succession of images and blurs them all together. A movie conveys an illusion of motion, but we still lack the ability to represent it in thought, to make dynamic processes an object of study.

Since change cannot really be represented in empirical terms, we must try a different approach for defining an analytical unit if we expect it to be appropriate for historiography.

In a way, this is not difficult. Change, after all, results from a causal relation. A change in one thing is said to cause a change in another to which it is causally connected. Since we are not at this point trying to explain a particular change, but only how to represent change in general in thought, the specifics of the cause are unimportant, and merely its causal relation needs to be posited.

Of course, an appeal to a causal agent and its relation is quite conventional in historical discourse. An event (empirical change within a limited time and space) takes place, and the historian explains it by reference to another event understood to be its cause. Since the causal relation is “abstract” (refers to a relation of qualities rather than the qualities themselves) rather than empirical, it is not treated as primary data, but merely secondary, as derivative. That is, in conventional historiography, a causal relation is inferred between two events if they are sufficiently proximate in time or place and if doing so satisfies “common sense” (which makes up for the deficiencies of the method).

There's no real (“tangible”) evidence that things are causally interconnected, and so there is not the certainty enjoyed when viewing the hard facts associated with the events themselves. Therefore, in conventional historiography, empirical qualities are definitive and primary, while their causal relations are a posteriori and only inferred.

Useful as this may be in daily life, I would like to suggest that prioritizing the empirical is a rather arbitrary and unrealistic choice. I would like to suggest here that we could as well make the causal relation a priori and explain change as its empirical effect. However, we need to keep in mind that in reality causal relations give rise to empirical change, and empirical change supports causal relations, and so the issue is not which is really primary, but which is best represented as primary in thought.

A possible objection to making causal relations primary is that empirical data is felt to be the beginning and end of historical investigation because the historian starts out with the certainty of “hard facts” and uses them to arrive at an explanation of a manifest empirical change. However, this is today no longer quite so obvious.

We now know that there are no “pure” facts uncontaminated by observational hypotheses. We now realize, without any necessary implication of subjectivism, that facts are socially constructed. What is evidence today was mere trash not long ago; facts important in relation to one social group are irrelevant to another; our axiom set, definitions, values and aims tend to shape profoundly what are the facts, their character and their significance.

For example, we have all seen those spectacular images of the cosmos produced by the Hubble telescope. Sometimes people have objected that the skies don’t really appear in such extraordinary color, and so the images must in some way be artificial. However, these images are no more artificial than the impression of viewing the stars with the naked eye through a backyard telescope, for the eye sees only a small part of the spectrum. The colors in the Hubble images convey a much richer truth than the eye can possibly see, and is no more artificial than the image seen through our home telescope, and actually there is more truth value contained in the Hubble images. In either case, the images are mere analogues of the reality being viewed, and are a function of the truth, not the raw truth itself.

If facts no longer seem to us quite so “hard,” and if causal relations can only be inferred, that would seem to put the historiographical enterprise on very thin ice indeed. Consequently, we often end up with a skepticism greater than actually warranted. We know that events took place in the past; we know that these were all processes; and we expect knowledge of the past to be meaningful because we find present processes to be so. A realist approach says: start with what we know to be true, and then build the conceptual equipment necessary to represent it in thought.

If all that we know is that there were processes, we must start with causal relations. One might well suppose there must first be entities to be causally connected, but that is not necessarily so. We can define things as being what is at both ends of a causal relation and go from there, and momentarily we will suggest that motion is an intrinsic property of all things in the universe, and so we don’t need to have proximate events for the source of changes.

Let me offer an example, for this shift in perspective might seem a little bewildering. The term “social class” is usually defined as a set of empirical qualities shared by its members. A member of the aristocracy, then, is a person who acquires, one way or the other, a set of legal privileges that supports his political dominance in society. This is an example of an empiricist definition. We have a group the members of which share a trait, and it gives them a power advantage over other groups. Having established a characteristic of the aristocracy, we might then infer from it a causal relation with the peasantry that we probably can characterize as exploitation.

However, there is an entirely different way to define social class, which is to define it as those people who happen to share a common causal relation. By convention, this causal relation (termed a “relation of production”) is the relation of a person with the means of production. Some people own or possess means of production, and they are called the bourgeoisie (because people whose potentials in life derived from owning means of production tended to live in town). Some means of production can be very small, in which case a member of the bourgeois class might work very hard for little return, perhaps even less than that of a skilled member of the working class. That is, income level is not here immediately definitive of class position.

In our example, we start out with some idea of what we are looking for, which is that members of the bourgeoisie develop through their relation of production, but members of the modern working class, who lack such means, must instead seek to develop through collective struggles. Presuppositions would also be present if we instead start out with empirical facts (infer exploitation from the fact of power rations rather than infer classes from the fact of exploitation). In both cases, the investigation is entered with certain axioms, values, definitions and objectives. This is true of any scientific endeavor. It is impossible and certainly undesirable to set out to solve a problem with a tabula that is entirely rasa.

One should not underestimate the implications of making a causal relation definitive and from it infer an empirical change. For one thing, it makes change primary, and stability only a possible and somewhat exceptional effect of change. For example, the stability represented by a “civilization” is an exception in prevailing flux of history. While the narrower the scope of historical investigation, the more one is able to employ empiricist categories with some success, it can be argued that this is not a sufficient reason for doing so. The issue is usually put in terms such as this: while local history can make sense in narrowly local terms, to what extent are explanations really sound until the local events are placed in relation to long-range changes and broader influences?

However, starting with a causal relation does raise some difficulties. For example, where does the empirical dimension then come from? If we are walking along a woodland path and stumble on a rock, its empirical existence intrudes on our daydreams in no uncertain terms, but causal relations don’t do that. I’ll come back to this momentarily, but in very general terms, new empirical features arise from the constraints upon a causal relation imposed by old empirical qualities. So now we end up obscuring the origin of matter through infinite regress rather than obscuring the ultimate engine of change. Nevertheless, we are better off because matter is directly perceived, and it is not difficult for our short-range outlook to grasp it in thought.

The ultimate basis of this general proposition lies in thermodynamic cosmology, but let me offer here instead a very pedestrian example that will be less troublesome. Suppose I am pouring through a sieve some ungraded sand. The result is a structure that did not exist before: a more or less regular cone-shaped heap consisting of grains only under a certain size. How did this order arise? First, gravity, which refers to the causal relation between the earth and the grains, causes the sand to attempt to pass through the sieve, and those that manage to do so form a cone-shaped heap also because of gravity. However, the physical characteristics of the sieve play a role, for it sorts the sand grains by size. The sieve is a structure that existed initially and so is inherited from the past; it constrains the degrees of freedom of the grains by preventing larger ones from passing through. The resulting heap is a structure that emerges in the future (in relation to the initial state).

So we see here a structure inherited from the past that constrains a present causal relation so as to reduce the degrees of freedom of the process and (of necessity) creates a new empirical structure in the future.

This way of representing a process has three great advantages over the usual empiricist policy of starting with the supposed “hard facts:”

I would like to end with two additional points that might help sharpen the distinction between the conventional empiricist analytical unit and one that represents it as process.

The laboratory model was a key notion in modern natural science, not because of its method, hardly novel, but because of its ideological implications. The laboratory walls served to isolate an experiment from outside influences so that what takes place within has to be solely the result of forces inherent in the objects under investigation. Through repeated trials there arises a general law which is understood to represent an essential truth of things. You can put objects of a certain mass in motion and infer from their behavior such a law as f=ma (force = mass x acceleration). Every time you try the experiment, you obtained nearly the same result, and that tends to validate the formula. It represents a universal law because, if the behavior is intrinsic to all things, it must be is universally present. The formula is thought to represent an essential truth of objects, and we “explain” behavior by subsuming a particular event under it.

The problem with this is that the law discovered in the laboratory is to some extent merely an effect of the laboratory walls. That is, deprived of a causal relation with the environment outside the laboratory, any process taking place within it must be regular and predictable (positively entropic).

Because the laws do seem to apply to isolated phenomena, meaning becomes intrinsic to them, and practical success serves to validate this atomistic view of reality in which the truth of things is self-contained. The lonely individual therefore becomes the appropriate historical agent as he makes optimal decisions for himself. However, this makes a mystery of why the outcome of a historical process is not entirely predictable and mechanically determined. The laboratory model contradicts what we know of history, which is that it is a creative or emergent process in which human struggle in the present frees the future to some extent from the burden of the past.

The other point that needs to be mentioned is that the presumption that all things are always in motion has historically been a stumbling block. Henri Bergson, for example, has been ridiculed for his elan vitale—a mysterious inner force which accounts for things’ emergence. However, empiricism merely sweeps the problem under the rug. It represents change as an infinite regress as each event is caused by another event, which in turn is the effect of yet another. Ultimately there is no explanation of change at all. Newton, for example, invested his atom with a metaphysical and thus profoundly mysterious inner “sympathy,” which eventually become a force of attraction. The shift only seems more naturalistic but in fact hid a supernatural force behind its surface appearance.

However, in the last half century we have come to understood very well why all things are in fact necessarily in motion. It is simply an effect of the original state of the universe. As a result of the Big Bang, the universe started out in a highly improbable state, and according to the Second Law of Thermodynamics, must return as quickly as possible to a more probable state. That is, all things, in themselves must “dissipate,” must experience an increase in entropy, must move toward a more probable and less ordered state. Everything is subject to this rule, not because of some mysterious inner impulse, but because everything inherits an improbable state from the original Big Bang.

That there are almost as many exceptions to this as there are manifestations of it need not detain us here beyond pointing out that dissipation can represent a “thermodynamic engine” that gives rise to new structures in the universe. The emergence of novelty almost equals its dissolution as the universe heads toward eventual “heat death,” which is the most probable form of its energy.

These emergent structures constrain the universe’s dissipation to give rise to time, which is why the universe did not dissipate instantly as the case before the Big Bang “when” there was a “perfect vacuum.” Time, then, is not mere chronological sequence, but more importantly is an effect of a constraint upon dissipation—the famous “arrow of time” which distinguishes the past from the future.

Cosmic dissipation is the ultimate engine of change, but causal relations are its form. That is, it is the dissipation of an improbable state that drives change, but the vehicle of change is the dissipation of specific things, which affect the other things with which they are causally connected.

In short, there are a variety of reasons why the most obvious basic analytical unit to be employed in historiography should be the process. That realization arose in the late nineteenth century, but was quickly buried by the prevailing positivist paradigm. However, at the same time there began to arise a number of developments in the natural sciences which revealed the limitations of the paradigm: thermodynamics, relativity, quantum mechanics, general systems theory, etc., which can now support a viable alternative to an empiricist analytical unit in historiography.

There is no reason whatsoever why historians of the twenty-first century should continue with an antiquated ideology that arose under very specific political and economic circumstances. Whatever ideological forms arise in the historiographic conceptions of the present century, there seems no question but that the basic unit of analysis must be dynamic rather than static. The process as we have defined it here offers one approach to that end.