Content

HHC - titling added

Philosophy of Science, 8 (4)

Oct. 1941, 567-579.

THE question as to the existence of laws in history has frequently been discussed. A new discussion may yet be useful, since some misconceptions based on incorrect comparisons with the natural sciences have been brought forward by both advocates and opponents of historical laws. We shall try to clarify the problem by applying a few ideas familiar to physicists and astronomers to the conditions peculiar to history. Physics is the most mature of all empirical sciences as to method. In physics the law-concept has been used for three hundred years. It may be assumed, therefore, that most of the difficulties in its application to other fields have their physical counterpart and can be clarified most easily with the help of physical concepts. A few preliminary examples of historical laws will be given towards the end of the article.

The relationship between historical laws and historical prophecies has sometimes been misrepresented. Astronomers can not predict from Newton’s law what the position of the planet Mars will be on next New Year’s Eve. In addition to the law they

*The article is part of a study undertaken with the help of a grant from the Social Science Research Council.

need the knowledge of the positions, velocities, and masses of a few celestial bodies at some given time: they need knowledge of “initial conditions” as the physicist puts it. Knowledge of a law, therefore, is not a sufficient but only a necessary condition of prediction. Evidently the same holds for history. Even if laws according to which wars between industrialized countries proceed were known, it might still be impossible to predict the outcome of the present war. Among other more intricate things we do not know is e.g. the number of airplanes on both sides. This knowledge will not be achieved before the war is ended, when it will be too late for prediction. As far as the past is concerned many analogous deficiencies, probably, never will be removed. Probably, we shall never know how large the population was in ancient Babylonia and how great the percentage of priests, noblemen, merchants, farmers, and slaves. This ignorance not only impairs our knowledge of the initial conditions but, unfortunately, also of the laws. Every scientific law asserts subsistence of recurrent association or regular connection of certain conditions and events. When and where we are completely ignorant of essential initial conditions, we shall never find recurrent associations. Or to put it more exactly: the more incomplete our knowledge of the initial conditions is, the more difficult is discovery of laws. We shall be very modest, therefore, in our expectations regarding historical laws.

In a few natural sciences laws appear in formulations which, beyond mere statements of recurrent associations, seem to include logical necessity. This necessity, however, springs from the deductive form only attained by those sciences. In physics e.g. the three laws of Kepler can be deduced from Newton’s law; likewise virtually all laws of electromagnetism are deducible from Maxwell’s equations. Yet this is a mere matter of logical form that does not affect the empirical content. When Galileo discovered the law of falling bodies by experimentation and measurement, he did not deduce it from more general mechanical laws, since such were not yet known. Deductive connection with other laws was not achieved before Newton, for deductive theories are almost always constructed a considerable time after

discovery of single empirical laws. This is relevant to our problem. Investigation of historical laws still is in an embryonic stage. For a long time to come these laws must not be compared to the laws of nineteenth century mechanics or electromagnetics but to the laws of young and still undeveloped sciences such as stellar physics. Based on numerous observations e.g. the law of Leavitt-Shapley asserts the existence of a functional relation between period and luminosity of variable stars of a certain type. This and many similar isolated empirical regularities in fixed star astronomy are well verified laws ¹without any regard to the hypothetical attempts of deducing them theoretically from more general physical principles. In history where investigation of laws has hardly begun construction of deductive theories would only impair empirical research. At any rate, however, there is no basic difference between isolated empirical laws and laws connected by deduction.

There is one more reason why historical laws on the one hand and mechanical and electromagnetic laws on the other do not compare. The latter are found in laboratories. Laboratories contain artificial apparatus built for specific ends. All physical apparatus are carefully safeguarded against mechanical shock; conforming to their objectives they are isolated electrically or thermally, they are constructed airtight or protected by leads against X-rays etc. Therefore creation of ‘isolated systems’, i.e. of systems not affected by undesirable interference from without, is among the chief aims of laboratory physics. In nature on the other hand there are, except for astronomy, practically no isolated systems. Particularly in history all systems are very incompletely isolated: cultures, countries, states always interfere with each other spiritually, politically, economically. Ancient

Greece

was influenced by the Orient; the Roman Empire was invaded by Germanic tribes, the civilization of these tribes was changed by classical culture and Christianity; modern

China

is influenced by Western capitalism, modern Western philosophy

1. E.g. Kohlschuetter-Adams: intensity of certain spectral lines and luminosity of the star; Adams-Joy: precision of spectral lines and luminosity; Lindblad: intensity of the continuous spectrum and luminosity.

by Chinese and Indian ideas etc. There are no airtight compartments, no isolated systems in history. History, therefore, must never be compared to laboratory physics. It compares only to geophysics, i.e. to physics of earthquakes and sea-currents, to volcanology and meteorology. This is a triviality but, strangely enough, has been forgotten in most of the analyses. Moreover, the aid of laboratory physics which forms the background of meteorology and geophysics and supplies laws from which deductions can be attempted is lacking in history. All this taken into account, historical phenomena are scarcely more difficult to predict than the weather and certainly not more difficult than volcanic eruptions and earthquakes. What would scientists think of a geophysicist who gives up the search for geophysical laws because of their inexactness?

Psychological laws deal with the behavior of human individuals, historical laws with large groups of individuals, namely with cultures, states, nations, occupations, classes etc. They do not correspond, therefore, to the laws of impact by which in classical gas-theory the behavior of the single molecules is regulated (‘micro-laws’), but to the gas-laws (‘macro-laws’). Historical and sociological groups, however, compare to gases considerably differing from those studied in our laboratories. They are contained in vessels with permeable walls; they consist of comparatively few molecules that do not move at random but in a partially orderly way. Moreover, the impulse of some of them is considerably greater than that of others. Little wonder that under such conditions the ‘gas-laws’ do not hold very exactly. These difficulties must be analyzed one by one. ²

1. The ‘permeable walls’, i.e. the incomplete isolation of the historical systems has already been discussed.

2. 0n the other hand the problems of modern quantum-mechanics have no bearing on our problem. Historical laws are macro-laws. Heisenberg’s principle of indeterminacy questions the existence of physical micro-laws; the validity of macro-laws is not affected by quantum-mechanics. The same holds for the so often discussed problem of determinism. Even individuals with “free” will could follow statistical macro-laws.

2. An ordinary gas vessel in a laboratory contains about 10 ²³ molecules. Since at present the whole population of the earth amounts to about 10 ⁹individuals, all historical groups are by many orders of magnitude smaller than statistical systems in physics. Although the accuracy of macro-laws is dependent not only on the numbers of individuals involved, in general statistical laws of smaller groups are less exact. Certain historical groups are particularly small. E.g. a law of some intellectual development might be concerned with not more than a hundred or even as few as ten philosophers or authors. Provided all authors concerned are taken into account, investigation of macro-laws is still justified, since the authors have been singled out from considerably larger groups. Several years ago the statistician Bortkiewicz obtained his ‘law of small numbers’ by studying soldiers of the German army who had died from having been kicked by a horse. We do not insist on the correctness of this special law of Bortkiewicz, ³but the method in general is justified. E.g. a law of intellectual history would be fairly well founded, though it is based on the observation of only fifty French philosophers, these having been singled out from forty million Frenchmen. Astronomers also investigate laws of the behavior of supernovae and ‘white dwarfs’ though only exceedingly few of such stars are known at present. Astronomers, however, do not pick out or leave out objects according to value concepts and personal predilections.

3. Boltzmann based his theory of gases on the assumption that the molecules move ‘at random’ (‘hypothesis of disorder’), a hypothesis formulated more exactly in modern statistical physics (Von Mises). For historico-sociological groups analogous hypotheses are valid to a limited degree only. E.g. states are ‘organized’, i.e. they consist of orderly hierarchies of subgroups. Yet, even in states a residuum of ‘disorderly’ behavior remains. There is, however, a statistical theory of crystals too in modern physics. In crystals the atoms are arranged in three dimensional ‘lattices’. Since over this order random oscillations

3. It asserts that ‘dispersion’ is nearly “normal” in such exceedingly small groups. Cf. L. v. Bortkiewicz: Das Gesetz der kleinen Zahien, Leipzig 1898.

are superimposed, a statistical theory of e.g. electric and thermal conductivity in crystals is possible. The same holds for historico-sociological groups: modern statistical investigation of public opinion is not prevented by the fact that states are ‘organized’.

4. The historical influence of human individuals shows a much greater variety than the physical effectiveness of atoms. Here two kinds of human inequalities must be distinguished. The effectiveness of persons can differ because they hold different positions in an organization. If in an army a general commits treason the effects are incomparably greater than in the case of a private doing the same. Therefore, counting generals and privates, heads of states or public officials and plain civilians with the same weight can become absurd in certain statistical investigations. In states, armies, and similar organizations inequality of position is of decisive importance. For this reason the validity of historical laws certainly is smallest in political history. In this field ‘chance’, i.e. psychological, biological, and other individual circumstances, might greatly impair macro-regularities. Macro-laws might play a greater part in the history of civilization and ideas, of art, science, philosophy, and religion where differentiated organizations are less important.

On the other hand human individuals influence history to different degrees because they differ in their personal gifts and abilities: there are good and bad artists, good and bad scientists and philosophers. Whether these differences are great enough to make statistical investigation and macro-laws impossible can not be decided by a priori arguments, but only by the results of empirical search after such laws. At any rate a ‘statistical’ history, aiming at laws of intellectual, artistic, and religious developments, would greatly differ from traditional historiography. It can neither dwell on masterpieces nor disregard the mass of mediocrities. The question as to whether an artistic or theoretical work is the product of a genius or a bungler would not even enter its investigation. Certain components of these value concepts, however, would reappear in other and more objective shapes. The distinction between long- and short-lived ideas must appear also in ‘statistical’ history, if it should

represent the facts. And the same holds for both the difference between new creations and mere repetitions and the stratification of literary and artistic taste according to social ranks and various levels of education and sophistication.

Time is contained as variable in some physical laws but not in others. E.g. the second law of thermodynamics asserts on a temporal process, stating that the entropy of an isolated system increases with time. The law of Wiedemann-Franz, on the other hand, asserts that, without regard to time, electric and thermal conductivity of metals are proportional to each other, good electric conductors being good thermal ones and vice versa. These two kinds of regularity may be called temporal laws and simultaneity-laws. ⁴

Both types of laws may be expected in history. We shall give a few examples of the first type (in preliminary and vague formulations) and indicate the empirical evidence in brackets.

1. In isolated historical systems tribal organization precedes the beginnings of the state (empirical evidence: ancient

China

, ancient Jews, ancient

Greece

, Germanic states of the early Middle Ages et al. Under external influence, i.e. in non-isolated systems, the development can proceed in a different way: British dominions,

U. S. A.

2. Individualized art and poetry are preceded by anonymous folk-art and poetry, signed paintings and sculptures by non-signed works. Or to put it more generally: the collective mindedness of the period of self sufficient domestic economy and barter economy precedes the individualistic spirit of the period of money economy and economic competition (rise of individualism in Greece in the sixth and fifth, in Rome in the third and second centuries B.C.; the Middle Ages and the Renaissance). This

4. In physics the terms ‘dynamic’ and ‘static’ laws are often used. This terminology has been avoided here, since it is too narrow and since, in philosophy and the social sciences, use of the word ‘dynamic’ often covers deficiencies in scientific analysis. The magic and animistic connotations of this term, discarded in physics three hundred years ago, have not quite disappeared from the social sciences.

law goes back to Hegel and Jacob Burckhardt and has been generalized by Carl Lamprecht.

3. Free artists e.g. sculptors, painters, and architects, gradually develop from craftsmen, such as stonedressers, whitewashers, and masons (

China

, ancient

Greece

, the Renaissance).

4. Worship of eminent individuals first takes sides and is partial, i.e. persons with opposite or divergent aims are not admired or worshipped by the same people; impartial admiration and worship of divergent great personalities develop later: biased hero-worship precedes impartial genius-worship (classical antiquity, the pre-romantic and the post-romantic period in Europe; cf. the study of the genesis of the concept of genius mentioned in footnote 5).

All these ‘laws’ are yet incomplete in so far as only necessary but not sufficient conditions are given. They describe temporal processes in yet rather vague formulations and may be called ‘historical’ laws in a narrower sense. On the other hand there are in history also simultaneity-laws. Though they are usually called ‘sociological’ laws, there is no basic difference between sociological and historical regular connections. The separation is entirely artificial, since sociology by no means disregards investigation of temporal processes. We give a few examples of simultaneity-laws.

5. Wherever learned priests are entrusted with the task of teaching priest-candidates, they systematize the vague and contradictory mythological traditions of the past and develop rational distinction, classification, and enumeration as scientific methods. Even if they turn to worldly subject-matters they develop causal investigation in a very small degree and never investigation of physical laws (Medieval Arabic and Catholic Scholasticism; Jewish Talmudism; the five orthodox Indian “philosophical” systems, especially Sankhya; Buddhistic Scholasticism in

Japan

6. If under favorable circumstances a past culture is revived after centuries, the initiators and bearers of this intellectual movement are characterized by the following traits: they do not belong to the clergy; they are scribes and secretaries in political services and develop under favorable circumstances to free

literati; exceedingly proud of their ability of writing and reading they disdain the illiterate; in their idea the educated is chiefly distinguished by perfection of his literary style; as patterns of style serve literary documents of the past; the authors of these documents are being looked upon as authorities superior to all representatives of subsequent periods (the European Renaissance; the two Renaissances of Confucianism under the Han and Tang dynasties in China; probably also the Renaissance movement in the New Empire of ancient Egypt and the Neo-Parthian Renaissance).

All these historical ‘laws’ have to be considered as preliminary and more or less probable assertions only. They are meant as examples to illustrate what form, approximately, historical laws would take and also on what kind of empirical evidence they have to be based. ⁵Since experiments are not feasible in history, comparison of various countries and cultures is the only way of finding historical laws. ⁶ The cultures compared ought to be as independent from each other and as numerous as possible. E.g. Europe and

America

, the modern era and antiquity, the civilizations of

India

China

Egypt

, and

Persia

must be compared and investigated from identical points of view if historical laws are to be established. In this program the only real difficulty in the application of the law-concept to history is revealed. It originates in the comparatively small number of independent historical systems. If in the natural sciences hypothetical laws are to be tested, it is in most cases easy to find suitable instances

5 After collection of the historical material about twenty other hypothetical laws have been given by the author of this article at the end of his Die Entstehung des Geniebegriffes. Ein Beitrag zur Ideengeschichte der Antike und des Fruehkapitalismus, Tuebingen I926, pp. 324-326. With the necessary scientific accuracy a historical law has been given for the first time in Frederick J. Teggart: Rome and

China

. A Study of Correlations in Historical Events, Berkeley 1939. Professor Teggart verifies statistically that there is a correlation between political disturbances in Western China and the Asiatic frontier regions of the Roman empire on the one hand and barbaric invasions in the Danube and Rhine region on the other. Teggart’s law, however, belongs to a more special type than the “laws” indicated above. It holds, although for many events, for late antiquity only. Evidently this does not impair its scientific value. Mechanics of rigid bodies also is valid in a limited period only. There were no rigid bodies 10¹⁰ years ago. In empirical science all laws hold as long only as the objects and systems they are concerned with exist.

that do not influence each other. In history it is only among the primitive cultures that the number of independent instances is considerable. Higher civilizations are in most cases connected by spatial interaction or the temporal links of tradition. This is a serious obstacle to verification of laws. E.g. analogous intellectual developments in eighteenth century

France

and

England

may be due to mutual influence and consequently must not be extrapolated and generalized. Practically the same holds for classical antiquity and modern Europe. Yet analogies e.g. in medieval Catholic and in Japanese Buddhistic Scholasticism may offer a sufficiently reliable basis for extrapolation. Far remote cultures, therefore, are particularly illuminating and convincing in investigation of historical laws. But also in the other cases painstaking scientific criticism may often succeed in separating direct influence from actual recurrence of the same phenomenon under analogous conditions. However, prediction will always stand as the ultimate test for the correctness of a law. Evidently prediction must not be taken in a temporal meaning only. If a historical regularity, obtained by comparing certain historical systems, is confirmed in other and independent systems, it may be considered, at least for the present, as verified.

On the other hand lack of perfect analogies neither speaks against the possibility of historical laws nor does it form a basic difference between history and the natural sciences. This had been assumed erroneously by the Windelband-Rickert-school and has often been repeated since. True, no two historical individuals are completely alike and history never repeats itself. However, the repetitions in natural science are overestimated by those only who are rather remote from this field of research. He who has ever worked in a laboratory knows that even every apparatus, if it is somewhat more complex, has its individual characteristics and has to be handled with its own special tricks. No two reflectors of the same brand are perfectly alike and even less two planets. In natural science the variety of objects is mastered by the method of gradual approximation: objects may be handled as analogous in the first approximation; their differences are taken into account later in second and third approxima-

tions, when they are put together and compared in new groups. Variety of historico-sociological phenomena surpasses variety of other objects in degree only. There is no reason, therefore, why the method of gradual approximation should not be applied in comparative history too. The conformities in the cultural ideals of Renaissance humanists and Chinese literati-officials can be established first; the differences may be taken into account later.

On the other hand the method of “understanding” (“insight”) which has often been recommended for social science is not sufficient when investigating historical laws. “Understanding” means psychological empathy: psychologically a historical process is “understood” if it is evident or plausible. The main objection to this criterion of the correctness of a historical assertion is that virtually always opposite historical processes are equally plausible. ⁷When a city is bombed it is plausible that intimidation and defeatism of the population result. But it is plausible as well that the determination to resist increases. It would not be plausible, on the other hand, if the bombing changed the pronunciation of consonants in the bombed city. Which process actually takes place can not be decided by psychological empathy but by statistical observation only. In the final analysis the method of understanding is equivalent to the attempt to deduce historico-sociological laws from laws of introspective psychology. However, before regularities are established it is premature to attempt to deduce them. In construction of new empirical sciences the predeductive stage can not be skipped.

There is even the possibility that certain historical laws may never become psychologically evident. Certainly history is based on the behavior of human individuals reacting psychologically. Yet there are in the natural sciences macro-laws that are similar to micro-laws and others that are not. Maxwell’s equations hold for the macro-phenomena caused by electric currents as well as for the behavior of (free) electrons: in this case macro- and micro-laws not only assert the same functional

7. This originates in the fact that the patterns consent-refusal and affirmation-negation predominate in the province of emotional and intellectual processes.

relations but contain also the same variables (electric and magnetic field etc.). On the other hand the gas laws entirely differ from the laws on elastic impact regulating the behavior of molecules in classical gas-theory. In this case the macro-laws connect variables, such as pressure, volume, and temperature of the gas, which have no application at all to single molecules. ⁸ Likewise there may be historical macro-laws connecting parameters only that are meaningless when applied to human individuals; no term that fits emotional or intellectual processes might enter them. In this case the historical law would be deprived of the possibility of psychological understanding and empathy. Yet it might be well verified empirically, since observable facts can be put into a correspondence also with very complex and abstract logical constructs. The law of the shifting of consonants in the Indo-European languages approaches to this type of “non-understandable” historical regularities. The number of such laws may be considerable e.g. in economic history.

History has grown out of other roots than the natural sciences. A man who, desiring to lift a load, is interested in the principle of the lever is the prototype of a natural scientist. A father who takes pictures of his son every year and collects and keeps them may serve as an analogy to the origin of historiography. He is not interested in regular connections and predictions but in the gradual development of his object because he likes it. Nobody will argue against this kind of historiography. It will always persist as long as men love their countries, their communities, and their culture. On the other hand, however, investigation of historical laws should not be obstructed by methodological objections. The greatest danger in this field is the danger of dilettantism and superficiality. The investigator of historical laws must collect, interpret, and compare an immense and highly complex material. Which scholar is an expert on modern and classical, Egyptian and Chinese history simultaneously? Astronomers have mastered analogous difficulties

8. The correspondence of one macro-parameter to a set of quite different micro-parameters is an instance of the relationship that, with a rather vague term, has been called emergence.

by division of labor and cooperation. The observatories have divided and distributed the problems, have collected, each in its field, the immense material according to identical principles and have thus produced the star catalogues and maps which form the basis of their laws. There is yet no analogous cooperation and division of labor in comparative history. Yet its problems are extremely interesting and may become important in a practical way in the future. Many scientists must establish a common program of research and cooperate according to it. By collecting and comparing the material with philological accuracy historical laws will be discovered at last not by general methodological discussions like ours.