Historical method.

The how of historical inquiry.


On this page I gather an assortment of quotations from historians writing about the very business of history, the historical method.

Refer also to my page on the argument from silence.

Martha Howell and Walter Prevenier, From Reliable Sources, page 141:

Although work like [Thomas] Carlyle's [On Heroes, Hero-Worship and the Heroic in History] is surely naive, it is also a mistake to underestimate the effect an individual can have. Imagine what today's world would be like had Mohammed, Confucius, or Christ not lived, if Marx had not written, if there had been no Hitler!

Martha Howell and Walter Prevenier, From Reliable Sources, pages 43-44:

In order for a source to be used as evidence in a historical argument, certain basic matters about its form and content must be settled.

First, it must be (or must be made) comprehensible at the most basic level of language, handwriting, and vocabulary. ....

Second, the source must be carefully located in place and time: when was it composed, where, in what country or city, in what social setting, by which individual? Are these apparent "facts" of composition correct?—that is, is the date indicated, let us say, in a letter written from the front by Dwight Eisenhower to his wife Mamie the date it was actually written? Is the place indicated within the source the actual place of composition? If the document does not itself provide such evidence—or if there is any reason to doubt the ostensible evidence—is there internal evidence that can be used to determine a probable date, or a time period within which the document was created? Can we tell from the content of the document itself or its relationship to other similar documents where it was composed?

Third, the source must be checked for authenticity. Is it what it purports to be, let us say an agreement for the transfer of land from a secular lord to the church or—to mention one of the famous cases of forgery from recent history—the personal diary of Adolph Hitler? Can we tell from the handwriting, the rhetoric, anachronisms of content, from the ink or the watermark or the quality of the parchment—or from the typeface or the electronic coding of the tape—that the document was not composed where it presents itself as having been composed? Is it, perhaps, a forgery from the period, a forgery from a later period, or simply a case of mislabeling by archivists?

At this point Howell and Prevenier list the principal tools that historians use in order to authenticate the sources:

  1. Paleography, the study of handwriting (pages 44-46).
  2. Diplomatics, the study of charters (page 46).
  3. Archaeology, the study of artifacts (pages 46-50).
  4. Statistics, the study of numerical data (pages 50-55).
  5. Miscellaneous tools (page 56):
    • Sigillography, the study of seals.
    • Chronology, the study of timekeeping.
    • Codicology, the study of handwritten books.
    • Papyrology, the study of papyrus texts.
    • Epigraphy, the study of inscriptions.
    • Heraldry, the study of coats of arms.
    • Numismatics, the study of coinage.
    • Linguistics, the study of language.
    • Genealogy, the study of family relationships.
    • Prosopography, the study of names and careers, or the use of biographical data to construct group portraits.

Martha Howell and Walter Prevenier, From Reliable Sources, page 60:

Sources must be evaluated not only in terms of those external characteristics on which we have been focusing, the questions of where, when, and by whom a source was created and whether it is "genuine" or not. Traditionally, they have also been evaluated in terms of what historians have thought of as internal criteria.

Howell and Prevenier now list the chief elements of source criticism:

  1. The genealogy of the document (pages 61-62), whether it is the original, a copy, or a copy of a copy.
  2. The genesis of the document (pages 62-63), the circumstances, authority, and events in or under which it was produced.
  3. The originality of the document (pages 63-64), whether it is innovating or merely passing on already current information.
  4. The interpretation of the document (pages 64-65), the extraction of some kind of meaning from it.
  5. The authorial authority of the document (pages 65-66), the relation of its author to the subject matter, whether eyewitness, earwitness, or even further removed.
  6. The competence of the observer (pages 66-68); is the author qualified to report and capable of reporting critically and with comprehension?
  7. The trustworthiness of the observer (page 68); is the author lying or telling what he or she believes is the truth?

Martha Howell and Walter Prevenier, From Reliable Sources, pages 69-71, on comparisons of sources (which section covers pages 69-79):

Typically, historians do not rely on just one source to study an event or a historical process, but on many, and they construct their own interpretations about the past by means of comparison among sources—by sifting information contained in many sources, by listening to many voices. ....

The essential problem here is distinguishing among the useful, less useful, and useless sources. Generally, historians consider sources to be useless (for reporting purposes) if they derive from other, usually older, sources. Although it is sometimes hard to decide if a source is in some way derived from another, once that assessment is made, eliminating the dependent source is usually easy. It is much harder, however, to rank sources that all seem to be "original" in that each provides an independent account of the particular events in question.

Nineteenth-century historians developed systematic rules for making such comparisons. Two of the best-known rule books of the age, that of E. Bernheim, published in 1889 (Lehrbuch der historischen Methode und der Geschichtsphilosophie [Guidebook for Historical Method and the Philosophy of History]), and Charles Langlois and Charles Seignobos, from 1898 (Introduction aux études historiques [Introduction to the Study of History]), provide a seven-step process, which we have summarized below. As we shall see, the procedure hardly guarantees the kind of scientific proof these scholars and their contemporaries imagined as the historians' goal (only numbers (2) and (6) seem uncontroversial), but it can, nevertheless, provide entry into the challenging world of source comparison.

This sevenfold list Howell and Prevenier give as follows:

  1. If the sources all agree about an event, historians can consider the event proved.
  2. However, majority does not rule; even if most sources relate events in one way, that version will not prevail unless it passes the tests of critical textual analysis....
  3. The source whose account can be confirmed by reference to outside authorities in some of its parts can be trusted in its entirety if it is impossible similarly to confirm the entire text.
  4. When two sources disagree on a particular point, the historian will prefer the souce with the most "authority"—i.e., the source created by the expert or the eyewitness....
  5. Eyewitnesses are, in general, to be preferred, especially in circumstances where the ordinary observer could have accurately reported what transpired and, more specifically, when they deal with facts known by most contemporaries.
  6. If two independently created sources agree on a matter, the reliability of each is measurably enhanced.
  7. When two sources disagree (and there is no other means of evaluation), then historians take the source which seems to accord best with common sense.

Martha Howell and Walter Prevenier, From Reliable Sources, pages 74-75, after using an argument from silence in an example involving Israeli involvement in the 1982 attack on refugee camps in Beirut:

Of course, an argument from silence can serve as presumptive evidence of the "silenced" event only if, as in this case, the person suppressing the information was in a position to have the information, and was purposing to give a full account of the story from which he omitted the crucial information, and if there were no compelling reasons why he should have omitted the information (other than the wish to conceal). ....

Another difficulty with an "argument from silence" is that historians cannot assume—as nineteenth-century scholars such as Seignobos would have assumed—that an observer of a particular "fact" would have automatically recorded the fact. .... In addition, it is clear, silences can be inadvertently created when texts are partly obliterated, lost, or changed in unexpected ways. And, conversely, it is naive to assume that everything that a text reports was actually observed—much less that it occurred!

Martha Howell and Walter Prevenier, From Reliable Sources, page 75:

Although historians must often reason from silences, they more commonly reason from positive evidence, and in their accounts they employ a number of logical processes. Very often historians reason by interpolation or by analogy, as though inserting missing pieces in a puzzle whose overall pattern they can discern by comparison with other, analogous situations. ....

Comparison of this kind can be a useful technique, but it is also a treacherous one. Comparisons are never perfect. Historical actors are creative; they learn from former events.

Martha Howell and Walter Prevenier, From Reliable Sources, page 76:

Reasoning by analogy, although useful and often necessary, is thus often inadequate. Hence, historians employ other kinds of logical processes as well, often turning to what is labeled "the scientific method." In these instances, historians construct testable hypotheses and marshal evidence to test them, following the principles of the physical sciences. Claude Bernard, a nineteenth-century positivist scientist, was one of the first, in 1865, to lay out these steps systematically: (1) observation, (2) hypothesis, (3) fit between the hypothesis and the given facts, (4) verification of the hypothesis with new facts. For historians who would follow this method, "observation" consists of critical analysis of the sources using the methods we have considered in chapter 2 [on pages 43-68, entitled Technical Analysis of Sources]. The "hypothesis" is an effort at explanation—an attempt to make causal connections between the observed "facts." The process is dialectic, so that the resultant hypothesis is then tested by new facts, revised if necessary, and retested. And so on.

Martha Howell and Walter Prevenier, From Reliable Sources, page 77:

Although it is a simple process to think up hypotheses, it is no simple task to formulate hypotheses that actually link the observed pieces of evidence—that can explain the facts available, not those that the scholar might wish to have. Often, it takes many tries before the scholar can formulate a hypothesis that really works—one that satisfactorily accounts for the known evidence. There is no formula for success in this difficult venture.

Martha Howell and Walter Prevenier, From Reliable Sources, page 78:

The difficulties of applying the so-called scientific method to historical research means that historians must often satisfy themselves with rules of logic that appear less watertight, making statements that seem probable, not "proved" in any "scientific" sense.

Martha Howell and Walter Prevenier, From Reliable Sources, page 80, on establishing evidentiary satisfaction (which section covers pages 79-84):

To a large extent, the amount or quality of evidence required depends on the kind of event being studied.

Martha Howell and Walter Prevenier, From Reliable Sources, page 81:

But historians never have just what they want or need. At one extreme is the historian limited to one source. Einhard's Life of Charlemagne is, for example, the only source scholars have about the private life of Europe's first emperor. Like many of the political biographies written today, this one is more hagiography than critical biography, and in the best of worlds historians might well refuse to use it as evidence about Charlemagne's life and his character. But historians, although conscious that they are prisoners of the unique source and bear all the risks that this involves, use it because it is all they have. At the other extreme are historians studying the recent past. They have a great many sources, and in many ways their problems are thus fewer. But even here there is no certainty.

Martha Howell and Walter Prevenier, From Reliable Sources, page 84, on the facts that matter (which section covers pages 84-87):

[T]he status of any "fact" available to the historian is always insecure. Nevertheless, however self-conscious they are about the limits of their knowledge, about its particularity..., historians must construct their interpretations about the past out of information that they deem to be of factlike status—information that is available to them for the purposes of their inquiry.

Martha Howell and Walter Prevenier, From Reliable Sources, page 84:

Often, historians will privilege evidence that seems to point to a recurring picture, to add to a story that seems familiar or repetitive. Always, however, this is a risky choice. In some sense, all events are unique, and every fact about an event is unique. ....

This is not to say, however, that there are not patterns in history, similarities of circumstance that allow the historian fruitfully to compare one place and time with another, to look for patterns of recurrence and thus patterns of causality. Only when one considers how similar people behave in similar situations can one begin to make generalizations about the relationships between events that we call cause and effect. Only then does history become more than the banal repetition of events.

Louis Gottschalk, Understanding History, pages 115-117:

The historian often has to depend upon secondary works (i.e., other historians' second-hand narratives and expositions) for his knowledge of the background into which to fit the contemporary documents, but he frequently also finds that, just as a good secondary account will enable him better to understand a contemporary document, so the correct understanding of the contemporary document will enable him to correct the secondary account. In the end, his knowledge is best tested by a critical analysis of the testimony of contemporaries.

Hence, as a general rule, the careful historian will be suspicious of secondary works in history, even the best ones. He should use them only for four purposes: (1) to derive the setting into which to fit the contemporary evidence upon his problem, being always prepared, however, to doubt and to rectify the secondary account wherever a critical analysis of contemporary witnesses makes it necessary to do so; (2) to get leads to other bibliographical data; (3) to acquire quotations or citations from contemporary or other sources, but only if they are not more fully available elsewhere and always with skepticism about their accuracy, especially if they are translated from another tongue; and (4) to derive interpretations of and hypotheses regarding his problem, but only with a view to testing or improving upon them, never with the intention of accepting them outright.

In general, the rule regarding time-lapse as applied to secondary sources is the reverse of that rule as applied to primary sources. The further away secondary sources are in time from the events of which they tell, the more reliable they are likely to be. That is true not only because impartiality and detachment are easier for remote periods of history, but also because as time elapses, more materials are likely to become available. In addition, the last writer has the help of the materials and interpretations contained in the earlier studies of his subject. Unfortunately, later historians are not always as competent as earlier ones. All too frequently they are just hack-writers, content merely with "re-hashing" the earlier works without presenting new evidence or points of view.

Louis Gottschalk, Understanding History, pages 90-91:

Four general rules will suffice here to indicate why one group of documents may be given precedence over another. (1) As we have seen, incomplete observation and faulty memory are often responsible for the inadequacy of testimony. Because a witness's reliability is, in general, inversely proportional to the time-lapse between the observation of the event and the witness's recollection, the closer the time of making a document was to the event it records, the better it is likely to be for historical purposes. (2) Some documents were originally intended purely as records or aids to one's memory, some as reports to other persons, some as apologia, some as propaganda, and so on. Because documents differ in this way in purpose, the more serious the author's intention to make a mere record, the more dependable his document as a historical source. (3) Because the effort, on the one hand, to palliate the truth or, on the other, to decorate it with literary, rhetorical, or dramatic flourishes tends to increase as the expected audience increases, in general the fewer the number for whose eyes the document was meant (i.e., the greater its confidential nature), the more "naked" its contents are likely to be. (4) Because the testimony of a schooled or experienced observer and reporter (e.g., a professional soldier reporting a battle, an experienced correspondent describing an interview, a veteran policeman reporting an accident, etc.) is generally superior to that of the untrained and casual observer and reporter, the greater the expertness of the author in the matter he is reporting, the more reliable his report.

Louis Gottschalk, Understanding History, pages 100-101:

The historian or psychologist interested in the inner springs of consciousness may, however, sometimes find the idealized personality of an autobiography more meaningful than the more realistic character revealed by better sources. It is also true that for the correct understanding of personal influences, cults, and legends, the idealization by disciples often is a more meaningful historical fact than the actual personality (see Chapter XI).

Louis Gottschalk, Understanding History, pages 139-140:

In the process of analysis the historian should constantly keep in mind the relevant particulars within the document rather than the document as a whole. Regarding each particular he asks: Is it credible? It might be well to point out again that what is meant by calling a particular credible is not that it is actually what happened, but that it is as close to what actually happened as we can learn from a critical examination of the best available sources. This means verisimilar at a high level. It connotes something more than merely not being preposterous in itself or even than plausible and yet is short of meaning accurately descriptive of past actuality. In other words, the historian establishes verisimilitude rather than objective truth. ....

A historical "fact" thus may be defined as a particular derived directly or indirectly from historical documents and regarded as credible after careful testing in accordance with the canons of historical method (see below, p. 150). An infinity and a multiple variety of facts of this kind are accepted by all historians: e.g., that Socrates really existed; that Alexander invaded India; that the Romans built the Pantheon....

Louis Gottschalk, Understanding History, pages 143-144:

Having accumulated his notes, the investigator must now separate the credible from the incredible. .... In detailed investigations few documents are significant as a whole; they serve most often only as mines from which to extract historical ore. Each bit of ore, however, may contain flaws of its own. The general reliability of an author, in other words, has significance only as establishing the probably credibility of his particular statements. From that process of scupulous analysis emerges an important general rule: for each particular of a document the process of establishing credibility should be separately undertaken regardless of the general credibility of the author.

As has already been pointed out (p. 138), some identification of the author is necessary to test a document's authenticity. In the subsequent process of determining the credibility of its particulars, even the most genuine of documents should be regarded as guilty of deceipt until proven innocent. ....

The historian, however, is frequently obliged to use documents written by persons about whom nothing or relatively little is known. Even the hundreds of biographical dictionaries and encyclopedias already in existence may be of no help because the author's name is unknown or, if known, not to be found in the reference works. The historian must therefore depend upon the document itself to teach him what it can about the author. A single brief document may teach him much if he asks the right questions.

Louis Gottschalk, Understanding History, page 147:

It would be relatively easy, even if the Gettysburg Address were a totally strange document, to establish its approximate date. It was obviously composed "four-score and seven years" after the Declaration of Independence, hence in 1863. But few strange documents are so easily dated. One has frequently to resort to the conjectures known to the historian as the terminus non ante quem ("the point not before which") and the terminus non post quem ("the point not after which"). These termini, or points, have to be established by internal evidence — by clues given within the document itself.

Louis Gottschalk, Understanding History, page 149:

In a law court it is frequently assumed that all testimony of a witness, though under oath, is suspect if the opposing lawyers can impugn his general character or by examination and cross-examination create doubt of his veracity in some regard. Even in modern law courts the old maxim falsus in uno, falsus in omnibus tends to be overemphasized. In addition, hearsay evidence is as a general rule excluded; certain kinds of witnesses are "privileged" or "unqualified" and therefore are not obliged to testify or are kept from testifying....

Louis Gottschalk, Understanding History, page 150:

The historian, however, is prosecutor, attorney for the defense, judge, and jury all in one. But as judge he rules out no evidence whatever if it is relevant. To him any single detail of testimony is credible — even if it is contained in a document obtained by force or fraud, or is otherwise impeachable, or is based on hearsay evidence, or is from an interested witness — provided it can pass four tests:

(1) Was the ultimate source of the detail (the primary witness) able to tell the truth?

(2) Was the primary witness willing to tell the truth?

(3) Is the primary witness accurately reported with regard to the detail under examination?

(4) Is there any independent corroboration of the detail under examination?

Any detail (regardless of what the source or who the author) that passes all four tests is credible historical evidence. It will bear repetition that the primary witness and the detail are now the subjects of examination, not the source as a whole.

Louis Gottschalk, Understanding History, pages 188-189:

A large part of historical composition is argumentative. It attempts to answer questions such as "Why did Caesar cross the Rubicon?" or "Did Washington understand the terms of his capitulation at Fort Necessity?" or "Did Nasser really believe that American and English airplanes aided the Israeli Army in 1967?" The argument in such cases usually consists of a mustering of the evidence on one and the other side and of a conclusion either that one side seems more plausible than the other or that, neither side of the argument being wholly convincing, the historian has to suspend judgment. In such cases the historian is dealing with evidence as concrete things as well as abstract thought; the evidence he is dealing with, whether words or artifacts or whatever, still exists — a record of recollection by Caesar, a capitulation signed by Washington, a transcription of an intercepted telephone conversation with Nasser, and the like — even if only in a translated and re-edited version rather than as the original document itself. When presenting his argument in cases like these, the historian therefore is dealing with evidence that is still extant in some form, good or bad, and he is likely to speak of it in the present tense: "Caesar's Commentaries say..."; "Washington's signature is..."; "Nasser's voice sounds...." Such a use of present tenses is not only permissible; it may well be the best way to use them.

Louis Gottschalk, Understanding History, page 163:

Even when the fact in question may not be well-known, certain kinds of statements are both incidental and probable to such a degree that error or falsehood seems unlikely. If an ancient inscription on a road tells us that a certain proconsul built that road while Augustus was princeps, it may be doubted without further corroboration that that proconsul really built the road, but would be harder to doubt that the road was built during the principate of Augustus. If an advertisement informs readers that "A and B Coffee may be bought at any reliable grocer's at the unusual price of fifty cents a pound," all the inferences of the advertisement may well be doubted without corroboration except that there is a brand of coffee on the market called "A and B Coffee."

Gilbert J. Garraghan, A Guide to Historical Method, page 304:

Proof by circumstances (indirect, presumptive, circumstantial evidence), as differentiated from proof by testimony (testimonial evidence), is used to establish the reality of an alleged fact, or to render a doubtful fact certain. Very often circumstances or indications of varying number and significance for each particular case point to one and the same conclusion. Taken individually, they yield as a rule only probability; taken collectively, they issue in certainty when their concurrence is such that it cannot be explained except by the reality of the alleged fact or facts to which they point.

Gilbert J. Garraghan, A Guide to Historical Method, page 305:

Cumulative or converging evidence is virtually circumstantial. It is "a heaping up" (L. cumulus) of bits of evidence, individually never more than probable, and often only slightly so, until they form a mass of evidence, the net result of which is certainty. But, as already noted, the resulting certainty does not issue directly from the mass or cumulus of probabilities, since no number of mere probabilities added together can logically produce certainty. To produce such effect, one must invoke the "principle of sufficient reason," by arguing that the only possible explanation why so many bits of evidence point to the same alleged fact, is that the fact is objectively true.

Gilbert J. Garraghan, A Guide to Historical Method, page 260:

For the reliability of the popular tradition of a historical fact, certain conditions must be fulfilled.

(a) Broad conditions: (1) Unbroken series of witnesses; (2) several parallel and independent series of witnesses.

(b) Particular conditions: (1) Content a public event of importance; (2) general belief for a definite period; (3) absence of protest during that period; (4) relatively limited duration; (5) influence of the critical spirit, and application of critical investigation; (6) absence of denial by the critically minded.

Gilbert J. Garraghan, A Guide to Historical Method, page 273:

Lanzoni [in Genesi], following De Smedt, distinguishes two types of legend, mere legends, and historical legends. The former have no direct or explicit historical content whatever; the latter have content of this kind in varying degree. Both types can be of use to the historian by preserving data of value, whether implicitly or explicitly. The legend itself may be pure fiction, and at the same time incidentally (or, as the philosophers say, praeter intentionem) may picture vividly and even accurately various phases of a vanished culture or civilization.

Gilbert J. Garraghan, A Guide to Historical Method, pages 143-144:

Analogy argues from the resemblance of two things in one or more respects to their necessary or supposedly necessary resemblance also in other respects. The basis of an argument of this kind is the double principle that every being shows certain attributes or traits corresponding to its nature, that every efficient cause has a corresponding effect; consequently, similar beings show similar attributes or traits, while similar causes have similar effects, and vice versa. Historical analogy applies the principle of analogy to historical data as a method of logical proof.

Garraghan goes on to list a few kinds of faulty analogy, one of which is, on page 145, the interpretation of ancient texts according to contemporary ideas and customs. Surely a good caveat to keep in mind when constructing an argument from analogy.

Gilbert J. Garraghan, A Guide to Historical Method, page 146:

History is concerned immediately with single, individual facts; mediately with such generalized truths as can be derived from the individual facts. Generalizations in history may be applicable only to the past, or they may be of universal application, and as such, independent of time and place. Examples of the two types are respectively the statements, "the Athenians were an art-loving people"; "a strongly centralized government is the best in war time." It is only in the case of the latter type that we can speak with consistency of "historical laws." The logical process employed in arriving at either kind of generalization is known as induction or, more specifically, incomplete induction, which may be defined as "the legitimate derivation of general laws or truths from a limited number of individual cases."

Gilbert J. Garraghan, A Guide to Historical Method, page 151:

As a method of investigation in history, the use of statistics may be taken to mean the collection, tabulation, and analysis of numerical facts of a given category, with a view to deducing therefrom averages, proportions, and other uniformities or laws, useful to the historian.

Gilbert J. Garraghan, A Guide to Historical Method, page 153:

An intelligent use of hypothesis conditions all progress in scientific research. As a rule it is only by thinking out various likely explanations of a phenomenon, by testing them one after the other, and rejecting such as are unsatisfactory, that the true explanation is finally brought to light. This is the course pursued by the physicist and other specialists in the natural sciences; it is a necessary procedure in the social sciences as well. Historical hypothesis may be applied not merely to the data supplied from sources, but to the sources themselves in the whole range of problems which they present, such as authorship, textual integrity, interpretation, trustworthiness.

Garraghan necessarily reminds the reader on page 154 that certain dangers attend argumentation by hypothesis, chief among which is the fitting of facts into the hypothesis rather than the fitting of the hypothesis into the facts.

Gilbert J. Garraghan, A Guide to Historical Method, page 160:

Conjecture does not greatly differ from hypothesis. Both terms are often used as synonyms in every day speech; technically, however, they differ in meaning. Conjecture generally regards individual facts or phenomena, while hypothesis, being of wider range and significance, deals typically with bodies of facts, general situations.

Garraghan now divides conjecture into three kinds. The first is conjectural emendation of a text; the second is conjectural restoration of longer passages of a text or even entire documents; the third is conjectural detail, used to fill out the background of a text.

Gilbert J. Garraghan, A Guide to Historical Method, page 162:

The argument from silence aims to prove the non-reality of an alleged fact from the circumstance that contemporary or later sources of information fail to say anything about it. It is sometimes misleadingly called the negative argument; but this can easily be taken to mean something false, namely, that the argument rests on an explicit denial of some fact.

Garraghan offers only two conditions that an argument from silence must fulfill. First, that the writer whose silence is invoked would have certainly been in a position to know about the alleged fact; second, that the writer would have certainly made mention of it under the circumstances.

Gilbert J. Garraghan, A Guide to Historical Method, page 166:

The argument a priori is based on antecedent probability or improbability. Direct evidence may be lacking that a man is guilty of a crime imputed to him; but his known character, antecedents, habits, make it likely or unlikely that he is guilty. Here the reasoning concerns facts or circumstances prior in time to the occurrence of the event in question.