History as science

A discussion on the H-History-and-theory list, May 2006

Bob Rosen's post had me asking a rather basic question about the nature and defining characteristics of science—beyond its predictive accuracy. My understanding of scientific doctrine is that the results of an experiment must be both independently verifiable and independently reproducible. In the Russian front example, Rosen suggests that the accurate prediction of defeat by the German General Staff is indicative of the social scientific nature of the quantifiable data used to make the prediction. However, because the General Staff happened to be right says as much about the intuitive components of important military analysis and decision-making as it does about its scientific aspects. While it is true that certain components can be quantitatively assessed (fuel availability, production capability, geographic realities, etc.), can it be said that given the exact same conditions regarding the quantifiable data (which can only be applied to the material needs of war) would result in exactly the same outcome? Under what circumstances can such a situation be reproduced exactly? While there may indeed be scientific elements in judging a military situation, there are an equal number of factors that simply cannot be scientifically quantified—the will to survive, a strong sense of nationalism, the terrorization of one's own countrymen, ideological drives, and just plain luck, for example. While these things are predictable to some degree, the exactitude of when and how are the unknown variables. More importantly from a scientific perspective, they are not reproducible for the purpose of finding independent results that would be the same because each conflict has its own qualitative idiosyncrasies. Military analysis of the sort Rosen describes is therefore not scientific. Perhaps our insistence on the exclusively scientific analysis of military activity helps explain why we have such difficulty combating terrorism—a practice based in ideological motives that cannot be reduced to scientific analysis.

Moreover, the push to make history or politics a pure science has always been a mystery to me. Why is it so difficult to reconcile the scientific elements with the non-scientific ones when studying history? Is it an effort to maintain some kind of disciplinary essentialism? Why is it so difficult to admit that there are unquantifiable factors that contribute to both the unfolding of history and in the textual interpretation of that unfolding? Are we so two-dimensionally materialistic that we can’t allow for knowledge that isn’t scientifically determined and based strictly in the material realities of existence? Does anything “not science” equate necessarily with “not knowledge”? Is evidence only legitimate if it is “data”? Certainly our intellectual ancestors would have found such self-imposed, literal limitations on knowledge very strange.

David Tietge

> Bob Rosen's post had me asking a rather basic question about the
> nature and defining characteristics of science—beyond its
> predictive accuracy. My understanding of scientific doctrine is
> that the results of an experiment must be both independently
> verifiable and independently reproducible.

I must apologize for not having followed this thread closely, but perhaps I can inject a brief comment just at this point.

I fear we may be under the burden of a late 19th-century (positivist) conception of natural science. Much could be said here, but I limit myself to the observation that many natural sciences do not attempt to predict, but rather seek to explain how a particular situation came about (”retrodiction”). These sciences are usually characterized as “evolutionary” sciences, and they are distinguished (arguably) by the fact that the outcomes of the object of study have lower entropy than their initial state.

Examples of such evolutionary sciences are cosmology, geology, meteorology, and, of course, evolutionary biology. The accurate prediction of an outcome by launching an experiment is out of the question because outcomes by definition (decreased entropy) are to a degree novel. What is interesting about them is not what has been predicted, but what is unpredictable.

If this be true, then I see no tension between the human sciences, such as historiography, and other evolutionary sciences. Another aspect of this issue is that theories, even in the non-evolutionary natural sciences, don’t simply stand or fall on the basis of experimental outcomes, and so the ideal of accuracy, verifiability and independent reproducibility, while desirable, are not the litmus test of the validty of a theory (a classic discussion is that of Imre Lakatos and Alan Musgrave, eds., Criticism and the Growth of Knowledge [Cambridge, 1970]).

I assume here that the existence of evolutionary natural sciences for situations in which outcomes are less probable than an initial state is a commonplace among natural scientists. If the case is otherwise, or my definition of evolutionary sciences is remiss, that certainly would appear to be an important point to resolve.

Haines Brown

I could not say anything about evolutionary biology. I do not know anything about biology. Three of four evolutionary natural sciences ( cosmology, geology and meteorology) are actually do make predictions.

Cosmology. As it is known from history of astronomy, Clyde W. Tombaugh did make discovery of Pluto using prediction based on observation of Neptune ( by the way calculations used for this prediction did have an error).

Meteorology. Everybody could watch weather forecast on TV or check NOAA web site for hurricanes forecast for 2006.

Geology. Geologists are also doing some forecasts. For example, all diamond exploration around the world is going on in pretty much similiar stuctures ( kimberlite pipes in proterozoic crystall shields). Why? Because all known so far deposites are located in that structures. Another example of attempts of prediction in geology is usage of geophysical data in modeling structure of the Earth ( does not work very well actually, but that is attept of prediction based on data).

All mentioned above were not experiments, but attempts to obtain outcome by analyzing obtained previously data.

Again I do not see this anything like that in history. And I do not beleive that this types of historical forecast of future possible.

More, I am now more than ever do not understand why history should be science. Why history could not be just area of knowledge and/or attempt of explanation of the past of humans? Is something which is not called “science” or “scientific” less valauble?

I hesitated to follow up on Iaroslav's reply, for a thoughtful reading might show we really do not disagree. However, since the point is such an important one, allow me to explore it a little more deeply.

With a broad sweep I distinguished those sciences that validate hypotheses by making predictions, from those which do not. In more technical terms, the former represent a study of (”dissipative”) systems that increase in entropy, while the latter study (”emergent”) systems in which entropy decreases. This distinction, I believe, is a fairly conventional one in the sciences.

However, I need to introduce greater precision here lest there be misunderstanding.

In “dissipative” systems, outcomes are in principle unequivocally determined by their initial state. Knowledge of the initial state allows one to perform an experiment or to view outcomes with foreknowledge of what they will be. Of course, as any science student will attest, outcomes are never exactly what are predicted, but if the results of an experiment fall within a standard deviation, the general law is considered to be valid, and it is assumed the general law is what determined the outcome, given the initial state. Subsuming events under the operation of these general laws is taken to represent an “explanation” of behavior and the source of meaning.

In “emergent” systems in which entropy decreases, the initial state determines only the probability distribution of possible outcomes. Prediction is limited to ascertaining this probability distribution, and not an actual outcome, which is not predictable within the range of possible outcomes.

Given this, I have no objection to Iaroslav's specific points. For example, the weatherman does indeed make predictions, but the probability of their being correct rapidly falls off, and are generally good for only up to about five days, and even then are chancy. The weather predictions I receive assert, for example, that there is a certain percentage likelihood that there will be rain tomorrow. One's ability to predict the outcome of an emergent system is a function of time and system complexity, and the point is not that outcomes are entirely unpredictable, but that to some degree they are unpredictable.

Another nuance is that in any emergent process, there are subprocesses that may in fact be mechanically deterministic. While a probabilistic causality applies to the system as a whole, there are aspects of it that may well be unequivocally deterministic. I believe this addresses one or two of Iaroslav's examples.

But now let us consider human history:

> Again I do not see this anything like that in history. And I do not
> beleive that this types of historical forecast of future possible.

Iaroslav may here be overstating the case a bit, just as I did when speaking of emergent processes in the natural sciences. As a practical matter, prediction in history is certainly possible. For example, if a situation is relatively simple and the prediction is short range, we do it all the time. We are almost certain that next week the Republic of Bolivia will still exist. This element of predictability is probably what we mean by “common sense” in historical explanation. Furthermore, there are certain “auxiliary sciences” such as economics, psychology, anthropology or sociology that come to the aid of the historian. They are used because of their predictive power, but are generally applied to only aspects of an historical system. They also seem to contribute to what we think of as “common sense” in historical explanation.

It seems to me that “common sense” is based on the experience of daily life which is so constrained in time and place that outcomes are highly predictable. That is why common sense lets us down when it is a question of long-range history covering many centuries or quite large geographic areas.

If my distinction here concerning probabilistic determinism in emergent systems and unequivocal determinism in dissipative systems carries any weight, then there seems to me no absolute distinction between the “evolutionary” natural sciences and human evolution.

We have a natural and empirically well-founded presumption that humans are essentially free agents, and it might seem (given the past hegemony of positivism) that natural processes in contrast are unequivocally deterministic and therefore alien to human affairs. And yet, humans are not absolutely free agents. “Men make history, but not in circumstances of their own choosing” (Introduction to _The Eighteenth Brumaire of Louis Bonaparte_). We always act in terms of specific circumstances that constrain the probability distribution of possible outcomes. To admit this does not, in my view, diminish the dignity of the human being, but moves our notion of human liberty from an empty metaphysical bias to become realistically operational in the historical process.

Haines Brown

I am in complete agreement with you. It has for some time seemed to me that this whole issue is about ego and vanity. Being a science confers a stature, an importance, a gravity which some feel History is lacking. That is, as long as it cannot lay claim to being a science History will not receive (from who?) the respect that it deserves, at least in certain epistemological circles. What I wonder about, however, is what would be the practical consequences of deciding that History was or was not, or even could be or could not be, a science? How would such a decision affect historical methodology, explanation, interpretation, criticism? I doubt if coming down on one side or another of this debate will greatly impact how history is done or written about.

’Ta meta ta historika’—when history becomes philosophy.

The word metaphysics derives from the ancient Greek phrase ‘ta meta ta physika’ an ancient editor used to entitle one of Aristotle's treatises. Aristotelian scholars interpret the phrase to mean “beyond the physics”; i.e. continuing to develop ideas [beyond] those presented in Aristotle's Physics. Aristotle said: “We must, with a view to the science we are seeking, proceed first to the aporiae…for in so far as the intellect has aporiae, it is…bound; it is impossible to go forward.” According to Aristotelian scholar Joseph Owens, aporiae refers to “conceptions hindering the intellect from [completely] understanding the four causes empirically established in the Physics”. Aristotle's Metaphysics then is the attempt to overcome those obstacles (”aporiae”) “hindering the intellect” from reaching a complete state of causal knowledge (”The Doctrine of Being in the Aristotelian Metaphysics”, Toronto 1957 pp. 28, 114, 144).

Similarly, Toynbee writes, regarding Aristotle's Metaphysics: “…the intellectual conquest of the field of physics opened up a further field of inquiry beyond it.” Further, Toynbee suggests an analogous field of metahistory, i.e. “ta meta ta historika”. “If one does not go on from physics (or history) to ‘metaphysics' (or metahistory), one will not be able to understand physics (or history) itself, or, in other words, ‘metaphysics' (or metahistory’) is the field in which physics (or history) finds its explanation.” (”A Study of History” vol. 12 p. 228)

To my mind, a good example of a rigorously empirical historian reaching the ’going beyond’ transitional point from history to metahistory is Frenand Braudel asking “What is Society?” (”The Wheels of Commerce” 1979 p. 458) However, Bruaudel does not cross over to metahistory, but his consideration of the question indicates his sense of something “beyond” the empirical. Historians study the empirical material manifestations of societies (political, economic, religious, etc), but generally don’t go “beyond” the history of a society and consider the essence of social being. What is the nature of the bond between great masses of individuals separated in space and time that creates a sense of sameness differentiating them from other groups and causes them to act in concert?

Sartre's “Critique of Dialectal Reasoning” is, to my mind, an example of a comprehensive metahistorical work. (Note: Sartre did not refer to his work as metahistory as Aristotle never used the word metaphysics. Rather these are terms used to characterize the nature of the respective works.) Sartre is attempting to solve intellectual problems that arise in Marxist empirically derived causal theory of social change. For example, he writes: “We established our fundamental agreement with historical materialism… [However,] it must be proved … that conflicts—within a person or a group - are the motive force of History…” As with Aristotle, after establishing an empirical theory of causality and finding it necessary to go “beyond” for a complete understanding, Sartre goes “beyond” Marxist empirical theory seeking completeness. (”Critique of Dialectical Reasoning” 1982 p15)

Similarly, Toynbee's vol. 12 of his “A Study of History” can be viewed as a metahistorical work—albeit it not as systematic and comprehensive as Sartre's. The first eleven volumes are the empirical work and development of causal theories. In the twelfth, which is subtitled “Reconsiderations,” he seeks more completeness. Like Sartre he seeks more complete definitions of terms like: “progress”, “institutions”, “society”, “culture”, “civilization”, etc.

Tom Verso

After reading postings of Haines Brown and Bruce Sanchez, I feel that I need to clarify my point as well.

In regards to Bruce questions. I do not know answers to Your questions. I agree with Your point that this recognition of history is not science but special area of knowledge will not probably affect how is history done now. On other side, when ideology is changing ( science is ideology : “ We beleive in science”), methods of exploring world or humans are changing as well. Sorry that I could not provide with clear answers, but that's as far as I could go now.

In regards to Haines comments. Thanks for clarifications.

First, I do not really want to discuss all this “entropy” and “standard deviation” staff. Why? Because for me such extensive usage of terms originated and used in physics and mathematics is clear sign of positivistics attitude toward history.

Second, I actually understated my position. My position is : predictions in history is impossible. History is not about future for sure. You wrote

“We are almost certain that next week the Republic of Bolivia will still exist.”

It is not certainty—it is wild guess. How much certainty did have Poland on August 31,1939? How much certainty did have Soviets on June 21, 1941? How much certainty did have USA on September 10, 2001? So when historian starts talking about future, he is futurologist but not historian. I want to repeat myself : “ History is not about future”.

And last, if all this is just “ empty metaphysical bias” than what is all this discussion is about?

Regards,

Iaroslav

<snip>
“( science is ideology : “ We beleive in science”),…”

NO,NO,NO,NO. That way madness lies. The Creationists try it out: “you believe in evolution, we believe in Creation”. That muddles two meanings of the word “believe”. Creationists take Creation on faith. Science precisely rejects faith, and accepts (”believes”) only propositions that are not “falsified” by any of the available evidence.

I knew a man who believed (under the influence of LSD) that he could fly, and jumped off a high tower. His belief was falsified at the moment when his body met the ground.

The Boxers believed that they would be invulnerable to bullets if their faith was strong enough. So they charged British troops. Naturally, those who survived had just proved to themselves that their faith was indeed strong enough, while those who died weren’t around to argue otherwise. The British on the other hand “believed” in the science of ballistics, which the Boxers' experiment verified yet again.

Ideologies are closed systems which don’t admit evidence that contradicts them; science seeks such evidence out.

There are those who confuse science with “what scientists do”. Scientists are human and susceptible to belief systems like anyone else. For example, Linus Pauling, despite four Nobel prizes in Chemistry, convinced himself on the basis of no evidence whatsoever that Vitamin C could cure cancer, and treated his wife with it. She died, of course. What Pauling did to win his Nobel prizes was science. His proselytizing on behalf of Vitamin C was ideology.

Naturally every ideology out there would like to have science on its side, and so tries to get scientists to do pseudo-science. Lysenko is the best- known example, but there are also tame scientists working for tobacco companies or drug companies or whatever: wearing white coats, but doing something other than science.

As to history and prediction. History does make predictions, but not about the future. For example: among Holocaust historians there is a long-standing dispute between “intentionalists”, who think that the Final Solution was planned and ordered by Hitler, and “functionalists”, who think that it grew out of local decisions without a central plan. Both sides agree that if (among the billions of captured German documents) an authenticated “smoking gun” document were to surface, the dispute would be settled. The functionalists predict (on the basis of the availbale evidence) that no such document will ever be found.

Christian Gerlach on the other hand has turned up a notation in Goebbels's diary that he thinks is the “smoking gun”. When Germany declared war on the US on 21 December 1941, Hitler convened a meeting of Gauleiters, and one of Goebbels's hats was as Gauleiter of Berlin. At this meeting, Hitler declared (according to Goebbels) that now that they were at war with the US,

it was time to settle scores with the Jews. Hitler had threatened that “a new world war will lead … to the annihilation of the Jewish race in Europe, and in Hitler's mind, Gerlach argues, the entry of the US into the war made it into a world war.

However, the hypothesis that the Fuehrerbefehl came on December 21 1941 leads to the conclusion that nothing done before that date was intended to accomplish the annihilation of the Jews, i.e., to the prediction that evidence will turn up to show that each prior event will have some other explanation. Gerlach will have to explain why Chelmno was already in operation, the Wannsee conference had been scheduled months earlier and the Einsatzgruppen had been massacring Jews since June, among other things.

In other words, scientific prediction is not about peering into the future; it is about setting out the precise conditions that a hypothesis wil have to meet.

In a message dated 6/16/06 9:28:09 AM, Iaroslav Gorokhovski writes:

It is not certainty—it is wild guess.�� How much certainty did have Poland on August 31,1939?� How much certainty did have Soviets on June 21, 1941? How much certainty did have USA on September 10, 2001?� So when historian starts talking about future, he is futurologist but not historian.� I want to repeat myself : “ History is not about future”.

And last, if all this is just “ empty metaphysical bias” than what is all this discussion is about? —Iaroslav

Iaroslav,

The definition of history occurs in time. It's an overview. If you compare the events of the last years of the Roman Empire with actions of the Bush administration, I think history can make certain predictions. History can be practiced in predictive mode only after long stretches in time. It is only then that it responds in a scientific sense. But the gathering and collating of information is also scientific in nature.

Best Wishes,
Bob Rosen

No substantial contribution, but just an effort to avoid misunderstanding.

Iaroslav Gorokhovski wrote:

“In regards to Bruce questions. … I agree with Your point that this recognition of history is not science but special area of knowledge will not probably affect how is history done now.”

Brown:

As this thread has made clear, “science” is an ambivalent word. For example, the wikipedia article on science offers three general meanings:

a) the literal meaning, which simply means a defined object of study. I doubt that this could be source of difficulty, although by demarcating an object of study, we may bring in a bias—a subjective component (a big issue, but not relevant here).

b) It can be the system of acquiring knowledge based on empiricism, experimentation, and methodological naturalism. This meaning is more of a problem for the historian, for while evidence may be empirical, and the method one of naturalism, the experiment has no significant role. Actually, within the philosophy of science, there is strong objection to making scientific knowledge depend on experiment, for many sciences are concerned with processes having unique outcomes. Where would the study of the Big Bang be if it depended on experiment?

c) The organized body of knowledge humans have gained by such research. I suppose there is no objection to this meaning, either, in that historiography represents an organized body of knowledge.

In short, the only element here to which the historian would object is the presumption of the positivist laboratory experiment. If we employ the word “science” as it is often understood as implying the hypothetical-deductive method, then historians have good reason to keep their distance. If, on the other hand, there were agreement with the point often made that there is no one scientific method, and scientists actually often use an abductive method rather than the hypothetical-deductive, then in this case, the natural scientist and the historian are using the same method.

It seems to me that if the historian has some reason to embrace the term “science”, he must at the same time make clear he is using an abductive method rather than the deductive, for most people assume science implies the latter. Is there any advantage in using the term “science”? I believe so. For example, historians are rather open to questionable metaphysical constructs, while scientific naturalism should offer a tool to subject them to criticism.

Iaroslav Gorokhovski:

“In regards to Haines comments. Thanks for clarifications.

Brown:

Let me start by agreeing with you. Attempts, for example, to apply thermodynamics in anthropology and economics have not been impressive. That is, while it can be done, the amount of insight gained appears small, and it is certainly not what most historians are up to.

But I was not suggesting any such simple importation of concepts from the natural sciences into historiography, which your comment seems to imply. Let me offer a context for my previous remarks that might shed a little more light on them.

Much has been written on hierarchy theory and more specifically the hierarchy of sciences. Certain scientific domains are considered more universal than others in that each level of reality at some point emerged from them. Each emergent level has its own peculiar rules of behavior, but it remains within the constraints of the domain from which it emerged. Human history, as one such level of existence, has its own peculiar way of manifesting itself, but we nevertheless remain bound by the constraints, biological and mechanical, of the more universal levels from which human life emerged.

We may justly treat these biological or mechanical constraint as trivialities. Mankind is still bound by gravity, and no feature of human history defies that constraint. I believe what this means is that knowledge of these constraints yields knowledge of the human condition that is rather marginal to what really interests us, not that it is not real. That the Egyptian pyramids depend on gravity is true, but trivial.

I don’t wish to go back and elaborate my point, but only note that awareness of some features of the more universal levels of reality from which human history emerged nevertheless prove to be useful. For example, we cannot advance, innovate, act constructively in history without necessarily dissipating our environment (i.e., the “economy”), and this dissipation imposes real limits on what we can do.

Iaroslav Gorokhovski:

“Second, I actually understated my position. My position is : predictions in history is impossible. History is not about future for sure. You wrote

“We are almost certain that next week the Republic of Bolivia will still exist.”

It is not certainty”

Brown:

You may have misunderstood. I tried to offer an example where common sense suggests that we can predict with a high degree of certainty. One might raise philosophical objections, but it is a fact that we have a very high degree of confidence that Bolivia will exist tomorrow. That probability is a fact.

It is, of course, possible it may not, but we _know_ that this is very remote possiblity, and so in fact we can justifiably predict that it will be in existence tomorrow as long as our predictions are probabilistic. Given this, historians often do predict, usually not about the future, but about the most likely outcome of some past situation. Historians call this “retrodiction.” Particular outcomes such as you bring up, were not unequivocally predictable of course, but if we knew enough about the prior situations, we hopefully would be able to know that such an outcome was possible and even assign a certain probability to it. Without using this probabilistic prediction, the course of history would necessarily appear chaotic. Except for the chronicler, don’t all historians explain a particular outcome in the past as a result of its prior situation including the intentions of the actors?

Iaroslav Gorokhovski:

“So when historian starts talking about future, he is futurologist but not historian. I want to repeat myself : “ History is not about future”.”

Brown:

Here I happen to disagree, but I’m on thinner ice. If historical outcomes are explained in terms of their prior situation, then there is no reason the historian can’t (probabilistically) predict what lies in the future in much the same manner as a meteorologist. Weren’t historians often called upon to offer advice concerning state policies? However, there is good reason why historians are generally not futurologists, for when comfortably studying the past, the outcome of a past process is known, and so our estimate of the probability distribution of the initial state's possible outcomes finds a test. If we fail such a test (we fail to account for the outcome as at least possible), then we know our understanding of the initial state was inadequate. The better we understand the situation in which an event occurs, the more we are able to arrive at the probability distribution of its possible outcomes.

I believe historians have good reason both to embrace and to avoid association with the word science. Howvever, I rather prefer the former because it engages historians in a universal language that helps them avoid becoming more marginalized than they already are. Another reason is that it brings with it a committment to materialism and naturalism that should help historians avoid metaphysical presumptions such as employing a fixed definition of “human nature”.

Haines Brown