Problems in the Logic of Historical Inquiry
Although it will be convenient to employ the term 'history' in the comprehensive sense of signifying the study of sequential changes that have occurred in any subject matter and not only in human affairs, the present chapter nevertheless deals with questions that relate primarily to explanations of past human actions. After a brief discussion of the general character of historical inquiry, we will analyze some of the forms historical explanations usually take; we will subsequently consider some problems recurrent in historical study, and finally examine a number of issues raised by the doctrine known as "historical inevitability."
- The Central Focus of Historical Study
According to Aristotle, poetry, like theoretical science, is "more philosophic and of graver import" than history, because poetry is concerned with what is pervasive and universal but history is addressed to what is special and singular. Aristotle's remark is a possible source of a widely accepted distinction between two allegedly different types of sciences: the nomothetic, which seek to establish abstract general laws for indefinitely repeatable events and processes; and the ideographic, which aim to understand the unique and the nonrecurrent. It is often maintained that the natural sciences and some of the social ones are nomothetic, whereas history (in the sense of an account of human events, as distinct
[547]
from the events themselves) is pre-eminently ideographic.[1] In consequence, it is frequently claimed that the logical structure of the concepts and explanations required in human history is fundamentally different from the logical structure of concepts and explanations in the natural (and other "generalizing") sciences. Let us examine the basis for these contrasts.
Even a hasty inspection of treatises in theoretical natural and social science on the one hand (such as optics and economics), and of books on history on the other hand, suffices to reveal a striking difference between them. For by and large the statements occurring in the former are general in form and contain few if any references to specific objects, dates, or places, whereas the statements in the latter are almost without exception singular in form and are replete with proper names, designations for particular times or periods, and geographic specifications. To this extent, at any rate, the contrast between the natural and some social sciences as nomothetic, and human history as ideographic, appears to be well founded.
It would be a gross error, however, to conclude that singular statements play no role in the theoretical sciences or that historical inquiry makes no use of universal ones. As previous chapters have repeatedly noted, no conclusions concerning the actual character of specific things and processes can be derived from general statements alone; for theories and laws must be supplemented by initial conditions (i.e., by statements singular or instantial in form) if those general assumptions are to serve for explaining or predicting any particular occurrence. Nor does the familiar and often useful distinction between "pure" and "applied" natural science impair the relevance of this point—on the supposed ground that the pure sciences (such as theoretical electrodynamics or genetics) are concerned with establishing only general statements, and that only applied sciences (such as electrical engineering or agronomy) need to concern themselves with particular cases. For even the pure natural sciences can assert their general statements as empirically warranted only on the basis of concrete factual evidence, and therefore only by making use of singular statements. Moreover, many statements commonly acknowledged to be laws of "pure" science have a generality that is at least geographically restricted—for example, the familiar law that a freely falling body at sea level in latitudes between 38° and 39° on the earth's surface undergoes an acceleration of 980 centimeters per second. If laws of this kind, which are specializations of laws not similarly restricted, are excluded from theoretical treatises, the exclusion is at best only a
[548]
question of convenience rather than of principle. Furthermore, some branches of natural science, such as geophysics or animal ecology, are primarily concerned with spatiotemporal distributions and the development of specific individual systems, and are therefore engaged to a large extent in establishing statements singular in form. In short, neither the natural sciences in their entirety nor any of their purely theoretical subdivisions are exclusively nomothetic.
But neither can historical study dispense with at least a tacit acceptance of general statements of the kind cited in theoretical treatises. Thus, although the historian may be concerned with the nonrecurrent and the unique, he must obviously select and abstract from the concrete occurrences he is engaged in studying, and his discourse about what is unquestionably individual requires the use of common names or general descriptive terms. Accordingly, the historian's characterizations of individual things assume that there are various kinds of occurrences, and in consequence that there are more or less determinate empirical regularities which are associated with each kind and which differentiate one kind from other kinds. For example, Greek colonial expansion during the sixth century b.c. has been attributed by one historian to the needs of Greek commercial interests combined with the adventurous spirit of the Greeks;[2]and he obviously took for granted that human beings possess several types of needs, that each type is generally manifested in certain characteristic modes of behavior, that some of these modes frequently result in the founding of colonies, and so on. Furthermore, one phase of historical inquiry consists of so-called "external and internal criticism"— of efforts directed to ascertaining the authenticity of documents or other remains from the past, the precise meanings of recorded statements, and the reliability of testimony concerning past events. But to accomplish these tasks, historians must be armed with a wide assortment of general laws, some of which are undoubtedly accepted tacitly as "common-sense knowledge" while others are adopted because they are endorsed by some natural or social science.
Moreover, historians are rarely mere chroniclers of the past; and they do not always terminate their investigations of some special group of events, even when they have settled the sequential order in which those events actually happened—for example, when they have established that Antony fell in love with Cleopatra before he fled from the battle of Actium. On the contrary, historians usually seek to understand and explain the events they record in terms of causes and consequences, and to find relations of causal dependence between some of the sequentially ordered events—for example, by showing that Antony fled from the
[549]
battle of Actiumbecause of his love for Cleopatra. To be sure, a historian's claim that two such events are causally related can be mistaken; but the historian making the claim presumably believes he has competent grounds for doing so. However, historians do not as a rule profess an ability to apprehend causal connections between individual occurrences by way of some direct, infallible intuition of such connections; and in any case, a given pair of past events can be shown to be causally related only with the help of causal generalizations (whether strictly universal or statistical in form) which are the products of inquiries designated as "controlled investigations" in a preceding chapter. Accordingly, the causal imputations historians make in explanations of human actions in the past are based on assumed laws of causal dependence. In brief, history is therefore not a purely ideographic discipline.
Nonetheless, there is an important asymmetry between theoretical (or "generalizing") science and history. A theoretical discipline like physics seeks to establish both general and singular statements, and in order to do so physicists employ previously assumed statements of both types. Historians, on the other hand, aim to assert warranted singular statements about the occurrence and the interrelations of specific actions and other particular occurrences. However, although this task can be achieved only by assuming and using general laws, historians do not regard it as part of their aim to establish such laws. It is unlikely that anyone would find something radically wrong with a treatise on theoretical thermodynamics which did not contain a single proper name or a single reference to any particular date. But it is even more unlikely that anyone using the word 'history' in its customary meaning would classify a book as a history if it mentioned no particular individuals, times, and places but stated only generalizations about human behavior. The distinction between history and theoretical science is thus fairly analogous to the difference between geology and physics, or between medical diagnosis and physiology. A geologist seeks to ascertain, for example, the sequential order of geologic formations, and he is able to do so in part by applying various physical laws to his materials of study; but it is not the geologist's task, qua geologist, to establish the laws of mechanics or of radioactive disintegration which he employs in his investigations.
However, this discussion must not be construed as an attempt to rule out by a priori reasoning the possibility of "historical laws" of developmental change. There have been many attempts, in recent years by Oswald Spengler and Arnold Toynbee among others, to show that every society or civilization manifests a uniform pattern of successive change, so that, for example, each society allegedly passes through a fixed series of evolutionary stages, in a manner comparable to the birth, adolescence, maturation, and decay of individual biological organisms. Although none
[550]
of these purported "laws" has found acceptance among competent students, their validity can be assessed only in the light of actual historical evidence and cannot be settled by examining only the formal structure of statements contained in the writings of historians. It is nevertheless pertinent to note that, irrespective of the factual worth professional historians ascribe to these alleged "historical laws," they tend to regard attempts at discovering such laws as contributions to sociology (or to some other branch of the "generalizing" or theoretical sciences) rather than to "history proper."[3] Accordingly, despite the fact that some historians undoubtedly use the evidence of the human past to establish laws of developmental change, they do not do so qua historians, in the opinion of most of their professional colleagues as well as on the evidence of the great bulk of historical writing.
- Probabilistic and Genetic Explanations
In any event, we shall assume that historical inquiry is concerned primarily with particular occurrences, and we shall therefore be concerned with the explanations that are typically offered for them. However, historians sometimes undertake to account for some action of just one human being, sometimes for an aggregative occurrence involving the actions of many men. Since there are important differences between explanations for these two kinds of happenings, it will be convenient to discuss separately the two corresponding major types of historical explanations. Moreover, as was suggested by the discussion in Chapter 10 of various generic differences between explanatory variables, historical explanations also differ in the temporal magnitudes of the events they mention. In particular, individual actions as well as the circumstances under which they occur are sometimes described as if they had no temporal dimensions, while the circumstances under which another action occurs, but not the action itself, are characterized in terms of their temporal spread. In consequence, some historical accounts of individual actions in effect assume that the conditions which supposedly explain the actions can be viewed for the purposes at hand as practically instantaneous; but other accounts of individual actions provide developmental or genetic explanations for them. We will therefore begin the discussion with an example of a historical explanation belonging to a subdivision of the first major type, i.e., one that attempts to account for some action of a single individual by stating for its occurrence a condition whose duration
[551]
is ignored. On the supposition that the explanation is representative of the type, we will then (a) discuss the question whether in explanations of this sort the premises contain general laws, (b) give some reasons for maintaining that the logical structure of such explanations is probabilistic rather than strictly deductive, and (c) consider briefly some senses in which this characterization of the structure is to be understood.
1. The example that will serve to illustrate historical explanations of the first type deals with a circumstance connected with Elizabeth's ascension to the British throne in the sixteenth century. As a result of Henry VIII's quarrel with the Roman Catholic Church, his official title at the time of his death read essentially as follows: "By the Grace of God, King of England, Ireland, and France, Defender of the Faith and Only Supreme Head on Earth of the Church of England called Anglicana Ecclesia." It will be recalled, however, that when his daughter Mary ("Bloody Mary") became queen in 1553 after the death of her brother Edward VI, she repealed the acts establishing the ecclesiastical sovereignty of the British monarch and reaffirmed the supremacy of the Pope in Rome. But when Elizabeth succeeded to the throne five years later, she proclaimed herself "Elizabeth, by the Grace of God Queene of Englande Fraunce and Ireland defendour of the fayth. &c."; and in doing so she became the first British sovereign to "etceterate" herself in an unabbreviated official title. Why did she do so? The legal historian F. W. Maitland proposed the following explanation. He first showed that the '&c.' in the proclamation was not there by inadvertence but had been introduced deliberately. He also pointed out that Elizabeth was confronted with the alternatives either of acknowledging with the late Mary the ecclesiastical supremacy of the Pope or of voiding the Marian statutes and breaking with Rome as her father had done—a decision for either alternative being fraught with grave perils, because the alignment of political and military forces both at home and abroad which favored each alternative was unsettled. Maitland therefore argued that in order to avoid committing herself to either alternative for the moment, Elizabeth employed an ambiguous formulation in the proclamation of her title—a formulation which could be made compatible with any decision she might eventually make. In consequence, according to his own succinct summary statement of the explanation, "So we might expand the symbol thus:— '&c.' = and (if future events shall so decide, but no further or otherwise) of the Church of England and also of Ireland upon earth the Supreme Head." [4]
[552]
a. We shall suppose that Maitland's explanation is fully warranted, since our aim is to analyze its structure rather than to discuss its factual validity. However, such an analysis cannot be performed unless the assumptions underlying the explanation are fully articulated; and, as is normal practice in historical writings, Maitland did not make explicit all the assumptions implicit in his account, whether or not he was aware that he was making them. For example, he took for granted, though without formal mention, the fact (crucial for his argument) that Elizabeth did not believe a patently ambiguous proclamation of her stand on the Roman question would by itself provoke the Pope to initiate armed hostilities against England. But in addition to such singular statements, Maitland also took for granted without mention generalizations about human behavior no less crucial for his argument. For example, in maintaining that Elizabeth adopted a noncommittal wording of her ecclesiastical claims in order to permit herself to postpone a major decision until it became safer to make it, Maitland tacitly assumed some sort of generalization about human conduct—a generalization whose precise form is unclear but which in any case must assert a relation between (a) public statements men are expected to issue concerning their ostensible commitment to some policy at a time when definitive commitment is hazardous, and (b) the use of ambiguous language in such statements for the sake of avoiding a premature commitment. Without some such assumption there would be no ground for holding that the formulation Elizabeth adopted in announcing her sovereign claims had anything to do with the quandary with which she was faced.
However, such generalizations are indispensable in all historical explanations of individual actions (even if, as is usually the case, the generalizations are not stated explicitly). This large claim cannot be established beyond every shadow of doubt, except perhaps by a systematic survey of historical writings; nevertheless, the contention can be shown to be highly credible. For historical explanations of this type aim to state the reason (or a reason) why some given individual x decided more or less deliberately to act in the manner y under circumstances z. But the possible reasons for individual actions can be subsumed under a small number of broad categories, where each category can in turn be divided into suitable subordinate classes; and by considering how the possible reasons in each category could be shown to be actually determining (or "causal") conditions for an individual's action, we can persuade ourselves that the above contention is sound. Indeed, three such major categories of reasons seem to be sufficient, and we shall briefly describe them.