2.1 The Limitation of Testimonial Reliance in Modern History

In modern historical research, the limitation of testimonial reliance is accelerated by the scientization of historiography with the advent of nineteenth-century historicism.Footnote ¹³ Naturally, this process is shaped by general views of science in the eighteenth and nineteenth centuries. Following Aristotle’s lead, all major early modern philosophers excluded history from the realm of science – some even ranked history as the lowest grade of knowledge. The criteria of science proper, so to speak, were (logically) demonstrable knowledge by way of the necessary and universal operations of a syllogism, as Aristotle had thought. The subject matter of history, however, was the particular and the contingent, this or that person, matters of fact that could all, logically speaking, be otherwise. Even worse, historians rarely had first-hand acquaintance with the particulars they wrote about but had to rely on the word of others. The word of others, in turn, was the helpless prey of Pyrrhonic skepticism concerning the reliability of all judgement and observation. In other words, history was about as far from science and certainty as one could get, just barely above mere opinion. Famous philosophers, such as Hobbes and Descartes, derided historical knowledge as the variable product of experience and authority, never reason.Footnote ¹⁴ The status of historical propositions as inherently unscientific was mirrored in the view that the historian was himself a kind of eyewitness to the past – a “man of affairs” whose reliability increased proportionally with his proximity to the events at issue.Footnote ¹⁵

Against this background, history could become a science only by rejecting the supposition that historical knowledge was completely at the mercy of testimonial reliance. This is precisely what historicists did. As Frederick Beiser has argued, nineteenth-century historicists vindicated the scientific status of history by questioning the widespread thesis that historical propositions must be accepted on faith alone.Footnote ¹⁶ The fundamental claim was that history was indeed a science because it had its own method of acquiring knowledge and understanding about the past, and that this method did not itself rely on mere faith in testimony. The historian’s method is perhaps best described as a critical logic of questions and answers concerning the reliability and meaning of all historical propositions. Such questions are, for example: What were the sources of a particular claim about the past? Are those sources reliable? What can one conjecture from comparing different testimonies? Most importantly, against Pyrrhonism, nineteenth-century historicists emphasized that modern historical research does not rely on testimonial sources alone. The modern historian may use relics from the past, such as coins, medals, inscriptions, and monuments and so on, as physical sources for criticizing testimony.Footnote ¹⁷ The very fact that the historian’s questions can be raised, and sometimes answered, showed that historical propositions were indeed a form of knowledge. Consequently, by having a method of its own, history could claim the status of a science that produces not logically demonstrable truths but probable propositions about the past.Footnote ¹⁸

The source criticism of the historical-critical method brought profound changes to the role of testimony in history. The methods themselves have their roots in biblical criticism, classical philology and comparative linguistics. It was from these disciplines that a critical methodology was imported and further developed in the nineteenth-century version of scientific historiography.Footnote ¹⁹ This critical methodology consisted chiefly of theories for the evaluation of the fidelity of the information chains that produced the historian’s sources and a concomitant reevaluation of cognitive values: trust in the authority of testimony and tradition was replaced by critique and suspicion of the truth-value of material from the past. The most well-known example of this development is undoubtedly Leopold von Ranke’s (1795–1886) school of critical history. Ranke’s role in this process was not primarily that of an innovator – for the critical methodology that Ranke endorsed was mostly borrowed from others – but as the prime organizer of an influential paradigm for professional historical research. The most fundamental feature of this new Rankean paradigm was the idea that the historian’s source material must be placed at the heart of historical study.

For Ranke’s school, one form of source material was valued above all others: the archive document. As Kasper Risbjerg Eskildsen has argued, it was with Ranke that the archive became the main site for producing historical knowledge.Footnote ²⁰ The historian must, according to the Rankean paradigm, study archive material based on rigorous source criticism and a clear distinction between primary and secondary sources. Among all the material that historians may use to support their accounts, the primary sources are of the highest value. Primary sources are documents either produced by or contemporary with the very events that the historian studies – either as descriptions or expressions – and will thereby allow the historian to study the past directly; without the subsequent distortions of traditional accounts of the events produced at a later date (secondary sources). As Georg Iggers has argued, Ranke’s most famous and discussed dictum – to study the past “wie es eigentlich gewesen” (as it really happened) – states that the historian’s task is to go beyond traditional accounts and refrain from moralizing judgements.Footnote ²¹ Thus, Ranke’s seminal distinctions as well as his famous dictum limit the historian’s dependence on testimony. This limitation comes in two different forms: (i) reliance on testimony from historical agents was relativized in relation to other testimony and non-testimonial source material, and (ii) the testimony of traditional accounts, essentially previous historians, was critiqued, or by-passed, with the focus on primary sources.

The Rankean revolution centres on the empirical concept of primary sources. This means that the key question is: What material from the past is the best kind of historical source material? Ranke himself favoured primary sources in the form of written archive documents, typically confidential diplomatic reports.Footnote ²² Staying true to his quest for authentic primary sources, it was only natural that Ranke himself was one of the chief instigators of expanding the amount of available historical source material by vast archive explorations in Prussia, Austria, Venice and the Papacy.Footnote ²³ Still, the sources relevant for Ranke were circumscribed by his focus on the history of politics and the state, which meant that the primary sources were, accordingly, material relating to the political bodies under investigation. Subsequently, historians have expanded the range of source material to include all kinds of material remains from the past. This development starts on a more comprehensive scale already with Johan Huizinga’s use of visual evidence, and, later, greatly influenced by the Annales School and Marc Bloch’s arguments for inferring the past from our present landscape. The rise of “history from below” during the twentieth century – investigating the past from the perspective of class, gender and everyday life – has naturally meant that altogether new kinds of primary sources became relevant for historical research.

The starting point for Ranke’s scientific history was not the historian’s questions, concepts, ideas or hypotheses. On the contrary, objective research can only start by freeing oneself from such notions since they are, of course, shaped by the historian’s own preconceptions. Scientific history begins, instead, with the very archive material upon which everyone can fix their attention, that is, the primary sources themselves.Footnote ²⁴ For the Rankean historian, the value of the source material is tantamount to the authenticity and accuracy it shows itself to have by way of external and internal source criticism.Footnote ²⁵ If only the historian freed himself from bias, the application of the historical-critical method would allow the primary sources to automatically reveal the past as it really happened. Past reality was not veiled for the Rankean historian but visible to the naked eye – were it not for the blindness caused by preconceptions and bias.Footnote ²⁶ This attitude was epitomized in a famous quote in which Ranke exclaimed, “[I] wish that I could as it were dissolve myself and only let the things speak, the mighty powers appear.”Footnote ²⁷ Ranke did not, as Chris Lorenz puts it, acknowledge that all observation is theory-laden, which is common in present-day philosophy of science.Footnote ²⁸

Tellingly, the very word for “source” (Quelle), which gains currency within the German historicist tradition in the early nineteenth century, embodies the Rankean idea of accessing the past directly through primary sources. Suggesting images of pure water flowing from a well spring, the term betrays romantic notions of the historian’s untainted access to past reality through authentic, primary source material. Thus, the metaphoric sense of “source” tends to obscure the fundamental distinction between the data conveyed by the sources and historical knowledge. In the grips of this confusion, Grafton argues, Ranke tended to view specific kinds of archive documents as “transparent windows to past states and events rather than colorful reconstructions of them, whose authors […] often wished to convince their own audience of a personal theory rather than simply to tell what happened.”Footnote ²⁹ This confusion was not, of course, shared by all historicists at the time. As Johann Gustav Droysen (1808–1884) already argued, source material does not speak for itself but stands in need of the historian’s analytical and interpretive skills to become historical knowledge.Footnote ³⁰

It is important to emphasize the restricted, empiricist extension of Ranke’s methodological revolution in historiography.Footnote ³¹ The Rankean revolution limits but does not reject reliance on the testimony of historical agents from the domain of scientific history. What it does reject is the authority of traditional accounts, in favour of attending to primary sources, as well as the acceptance of statements by historical agents at face value without corroboration. Scientific history does not depend on faith in testimony alone. But this is not a movement against testimonial reliance as such. Instead, it is a movement in which the quest for the most authentic and reliable witnesses becomes the chief task of the scientific historian. Rankean historiography is, therefore, keenly interested in separating direct from indirect witness, as well as with evaluating the epistemic virtues that witnesses must possess to deserve the historian’s belief.Footnote ³² As a matter of fact, the appeal to reliable eyewitnesses is crucial for Ranke’s identification of primary sources with direct access to the past. As Ranke wrote:

In other words, although historians could not themselves observe the past first-hand, they could gaze into the past directly by way of the authentic archival documents and relics it left behind in the present.Footnote ³⁴ Gazing into the past directly, however, was premised on the possibility of fulfilling the “founding myth” of scientific historiography.Footnote ³⁵ This was that describing the past objectively is indeed a real possibility for historians. As Lorenz has argued, this myth relies on three connected requirements which today seem completely unfulfillable.Footnote ³⁶ First, historians must dissolve themselves at their desks, thereby allowing the primary sources to speak for themselves. Second, there must be a clean break between past and present, requiring perhaps 50–100 years of distance between historians and the object of their study. Without separating “historical time” from the present, historians cannot view the past objectively due to possible personal interest and involvement.Footnote ³⁷ Third, the archival documents and relics themselves must be products of objective and impartial social processes. Ranke assumed that such processes were guaranteed by the state archive as a neutral repository of raw material from the past.

The conceptual product of the Rankean paradigm is the idea of a historical past that directly mirrors the past-as-actuality. This is an idea of a historical past for which the elimination of forgery, bias, partisanship, value judgements, preconceptions, and so on, automatically allows the truth of the past-as-actuality to be revealed in the historian’s primary sources. On this conception, source criticism, as Quellenforschung, is the method of scientific history. The proper application of that method enables the historian, ideally, to see the past itself exposed 1:1 in its remains. If authenticated, testimony is one possible direct source to the past. As a result, this view of historical research narrows the historian’s knowledge claims to information that withstands the trial of source criticism. It cares not for concerns about narration, interpretation or the silences of the archive, issues that were not to be fully engaged until the 1960s in mainstream historical theory. It was also this narrow conception of method that was canonized as scientific history in a series of manuals during the late nineteenth century, such as Ernst Bernheim’s Lehrbuch (1889) and Charles-Victor Langlois and Charles Seignobos Introduction (1897). Both provide detailed explorations of best practices for allowing the past to arise as it originally was from the historian’s authenticated source material.

2.2 The Conceptual Rejection of Testimony in Modern History

One finds a philosophically far more penetrating argument against testimonial reliance in Collingwood’s work. This is not to say that Collingwood is unique, even if he undoubtedly was a great innovator in the philosophy of history. Collingwood was influenced by other philosophers of history, such as Giambattista Vico and Benedetto Croce, and ideas similar to Collingwood’s have been presented by hermeneuticians such as Wilhelm Dilthey, Heinrich Rickert and Hans-Georg Gadamer. But the main reason Collingwood deserves space paralleling the Rankean paradigm is his unrivalled influence in English-speaking philosophy and theory of history. This influence stretches far beyond the confines of academia – The Idea of History has been included on the Times Literary Supplement’s list of the most influential post-Second World War books.Footnote ³⁸ Importantly, in contrast to most classical authors on the historical-critical method, Collingwood engaged the question of testimony in detail.Footnote ³⁹ The reason for this is that Collingwood’s very idea of history was developed in opposition to testimonial reliance. He argued that this opposition was a key characteristic of modern historical research. As we shall see, Collingwood’s rejection of testimony relies on conceptual rather than empirical grounds.

Collingwood was not impressed by the philosophical underpinning of the Rankean model of scientific historiography. For Collingwood, Langlois’s and Seignobos’s methodology manual was “about as useful to the modern reader as would be a discussion of physics in which no mention was made of relativity.”Footnote ⁴⁰ This stark judgement is motivated by the fact that Collingwood disagreed with the most basic presupposition of the Rankean model of scientific history. Contrary to Ranke and his followers, Collingwood did not think that historical research starts with the source material. He explicitly derided “modern anti-scientific epistemologists” for their idea that “when we have made our minds a perfect blank we shall ‘apprehend the facts’.”Footnote ⁴¹ Scientific history did not, Collingwood argued, start with the source material but with questions posed by the historian. Collingwood was very clear on this point. He wanted to show that the “questioning-activity” is the “dominant factor in history, as it is in all scientific work.”Footnote ⁴² Collingwood thereby argued that the Rankean historian got the wrong end of the stick: “you can’t collect your evidence before you start thinking: because thinking means asking questions …nothing is evidence except in relation to some definite question.”Footnote ⁴³ His fundamental argument was based on the insight that question and evidence are correlative concepts. Evidence is that which allows you to answer the question you are asking. As Collingwood argued:

Collingwood’s argument has far-reaching consequences for the historian’s relation to material from the past. Contrary to the Rankean model for scientific history, Collingwood’s discussion is not about what kind of material the historian should favour, nor about the correct critical apparatus for assessing the authenticity, credibility and accuracy of the source material. In fact, Collingwood dissolves the very idea of there being such a thing as the best material or practical method for studying the past to be established independently of the historian’s questioning-activity. Fundamentally, Collingwood’s argument shows that the historian’s relation to past material cannot be sufficiently explored empirically. Rather, what is needed is a philosophical investigation of the ways in which the questioning-activity conceptually disposes historians in relation to the material that they study.

For understanding the paradigmatic role of the questioning-activity in modern historical research, Collingwood provided a canonical story of how historiography freed itself from testimonial reliance.Footnote ⁴⁵ According to Collingwood, history evolved immensely in the nineteenth and early twentieth centuries – it had, in his rather grandiose terms, gone through a “Copernican revolution.”Footnote ⁴⁶ Collingwood pictured this as a process in which historical thought became autonomous as historians rejected the authority of testimony. By autonomy, Collingwood means: “the condition of being one’s own authority, making statements or taking action on one’s own initiative and not because those statements or actions are authorized or prescribed by anyone else.”Footnote ⁴⁷ In essence, prerevolutionary history consisted of believing authorities upon testimony. Postrevolutionary history, on the contrary, is based on autonomous reasoning from evidence. Even if the metaphor of a “Copernican revolution” is extravagant, it does rightly single out a decisive feature of the change in the idea of history that Collingwood wanted to describe. The centre of gravity in history had moved from the authority of witnesses of the past to the historian as a self-authorizing scientist who constructs accounts of the past based on evidence.

This epistemological revolution in history was styled by Collingwood as a three-phase development. First came “scissors-and-paste” history, which was prerevolutionary, and after the revolution came “critical history” and finally “scientific history,” or history proper, which is based on Collingwood’s seminal notion of reenactment.Footnote ⁴⁸ By scissors-and-paste history, Collingwood denotes a form of writing about the past that he dates to antiquity and the Middle Ages. Such writing about the past, considered pseudo-history by Collingwood, is conducted by excerpting and combining the testimonies of different authorities. By critical history Collingwood denotes a practice that he dates to the seventeenth century, in which historians engage in a systematic critique of the authorities to be included in their own narrative. Both scissors-and-paste and critical history, however, focus only on questions about the truth and falsity of statements by historical agents. In contrast, scientific history, or history proper, focuses on the meaning of past statements in relation to the historian’s questioning-activity.

At the core of Collingwood’s reasoning was a sharp conceptual distinction between testimony and evidence. Belief in testimony, Collingwood claimed, “stops … where history begins.”Footnote ⁴⁹ In history proper, one does not merely believe the testimony of authorities. Rather, these “authorities” become only evidence from which historians, upon their own authority, infer answers to their own questions about the past. Collingwood goes so far as to say that not only is history independent of testimony; it has “no relation to testimony at all.”Footnote ⁵⁰ History is a wholly reasoned form of knowledge proceeding through questions, evidence and criticism. The criterion, therefore, of what is to be accepted as a historical fact is not the trustworthiness of a witness, but the historian’s imaginative reconstruction and reenactment based on evidence. According to Collingwood, scientific history is practiced when the historian is “twisting a passage [by a historical agent] ostensibly about something different into an answer to the question he has decided to ask.”Footnote ⁵¹

For Collingwood, the epistemological foundation of history cannot be the authenticity of witness rapports. The autonomy of history, derived from the primacy of questions and imaginative reconstruction, implies that there is “nothing other than historical thought itself, by appeal to which its conclusion may be verified.”Footnote ⁵² This argument entails that the criterion for what counts as a historical fact is not the trustworthiness of historical witnesses, but the historian’s imaginative reconstruction and critical assessment of past actions and events based on a wide range of sources used as evidence. Given the primacy of questions and imaginative reconstruction, it will ultimately be the historian, in critical dialogue with their community of peers, who determines which account of past actions and events that is most supported by the evidence and therefore to be labelled historical fact.

On Collingwood’s view, historical fact is a dialogical concept in which the significance of actions and events of the past is seen in relation to the understanding and knowledge we have in our present historical context. If the historian simply received information about the past through testimony, in the manner of receiving a letter in the mail, then this information does not in itself constitute historical knowledge. History is not just a method for transferring information from the past to the present. If that were the case, historical knowledge would always be second rate to being contemporaneous with the past events themselves. The perfect experts on the causes of the Thirty Years’ War would be people who lived at that time, while historians are at the mercy of the crumbs of information that they happened to leave behind. Naturally, this presupposes that people in the past already possessed the knowledge that historians pursue, which is precisely what Collingwood denies.Footnote ⁵³

On this score, there are similarities with Hans-Georg Gadamer’s notion of a “fusion of horizons,” and Gadamer was indeed explicit about his debts to Collingwood.Footnote ⁵⁴ Historical knowledge and understanding do not proceed by somehow comparing (past) statements with the reality that they are about. Rather, it proceeds on the logic of question and answer – only by first understanding the question to which a statement was meant to be the answer can we understand that statement itself. As Gadamer points out, however, in this process it is always necessary to go beyond mere reconstruction. The meaning of a past statement is, in history, dependent not merely on the question it was meant to answer in the past but also on the question it is used to answer for the historian. In this process, historians necessarily bring in their own preunderstanding and cannot “avoid to think of what the author accepted unquestioningly.”Footnote ⁵⁵ This means that testimony can never be sufficient for acquiring what is properly called historical knowledge. What goes by that name is necessarily dependent on the historian’s own horizon of understanding.

In sum, modern, scientific historians study past events that are not available to observation. The method of their study is to infer accounts of past events from something that is accessible to observation; namely, material traces in the present – the historian’s evidence. Consequently, knowledge of the past is properly called “historical” only to the extent that this knowledge is based on what the evidence at our disposal obliges us to believe about events in the past. Even if one may have arrived at the very same propositions about past actions and events by other means – say memory accounts, time travel or hearsay – then this would not qualify as historical knowledge. The reason being that knowledge does not count as historical simply for being about this or that event or action in the past, but only if the knowledge in question has been produced by way of the historical method of inferring claims about the past from evidence in the present. Thus, the data about the past that testimony may carry are, by definition, not historical knowledge of the past.

2.3 The Historical Past of Method

The conceptual rejection of testimony has – often without apparent influence from Collingwood – been endorsed by many seminal authors in historical theory.Footnote ⁵⁶ The historian Carlo Ginzburg labelled it the arrival of an “evidential paradigm” in historical research.Footnote ⁵⁷ In this context, the concept of paradigm is not used in the Kuhnian sense of methodological agreement and uniformity. Rather, when Ginzburg speaks of an “evidential paradigm” he refers to a large-scale transformation of the historian’s basic understanding of their source material. According to Ginzburg, modern historical knowledge is conjectural by nature: historians observe certain traces; they pose a hypothesis about what caused them; the traces serve as evidence to confirm or falsify their hypotheses. All source material of the past – from the farmer’s pots and ploughs to the philosopher’s treatises and testimony from both worldly and religious leaders – should be treated not as authorities to be believed based on the pedigree of their origin but merely as evidence.Footnote ⁵⁸

The key feature of the evidential paradigm is not the idea of understanding source material as relics or remains. This was already acknowledged by nineteenth century advocates of scientific history such as Ranke and Bernheim. The idea of remains is, as such, fully compatible with the ambition to gaze at the past directly. What distinguishes the evidential paradigm is, instead, a distinct epistemic relation to the object of investigation – it stipulates that all kinds of source material should be subsumed under the category of evidence. If the source material is understood as evidence, the relevance of that material is dependent on the specific questioning–activity of the historian. As Collingwood showed, evidence is never relevant in and of itself. Rather, it is relevant only in its capacity for answering the question that the historian has decided to ask. In other words, there is an internal connection between the evidential use of source material and the autonomy of history as a discipline.

It is difficult to overestimate the significance of the evidential paradigm for the idea of modern historical research. It is by treating sources as evidence that history qualifies in the general category of science. Contrary to poetry and prose fiction, historical accounts of the past must be constrained, on the one hand by the available source material, and on the other, by the ability of historians to prove the reliability and validity of their inferences from sources used as evidence to a community of peers. Tellingly, Marc Bloch endorsed the arrival of the evidential paradigm as “a glorious victory of mind over its material.”Footnote ⁵⁹ This paradigmatic change came to reshape the historian’s self-image as one in which a radical break had occurred in contrast with their premodern colleagues.

The evidential paradigm freed historians from epistemological subordination. As Bloch wrote, the memoirs of Saint-SimonFootnote ⁶⁰ may frequently give us nothing but fictitious news about the events of the reign of Louis XIV, but they cannot, when used as evidence in spite of themselves, fail to express the mentality of a great noble at the court of the Sun King. By treating the material from the past as evidence, historians could know “far more of the past than the past itself thought good to tell.”Footnote ⁶¹ Inspired by the glorious victory, it was not uncommon that testimony was disqualified as an altogether unreliable epistemic source among proponents of the evidential paradigm.Footnote ⁶² As Bloch argued, for the historian, there is no such thing as a “good eyewitness”; all witness accounts are limited in their perspective, subjective, and dependent on the variant faculty of memory.Footnote ⁶³

A distinct idea of the historical past is born with the evidential paradigm. As I argued, the paradigmatic Rankean view of the historical past is one in which the historian apprehends the past directly by objectively gazing its authentic remains – if historians have been able to make their minds a perfect blank. In contrast, the historical past of the evidential paradigm is the outcome of the historian’s questioning-activity. While the Rankean historian observes the past directly in the sources, the historical past for the evidential paradigm is a product of method.Footnote ⁶⁴ The latter is not an entity that the historian somehow observes in the remains, but a past constructed to explain the evidence in relation to the question historians have decided to ask. This very attitude to material from the past was by Collingwood defined as historical thinking:

Consequently, the questioning-activity of the historian introduces a break between the past-as-actuality and the past as constructed from evidence in historical research.Footnote ⁶⁶ The latter is, as Michael Oakeshott argued, a past that can only be found in history books.Footnote ⁶⁷ For the Rankean historian, this would be a major drawback. The goal of scientific history was to know the past-as-actuality, revealed in the most authentic primary sources, freed from the historian’s preconceptions. Refuting this aim is a fundamental part of the conceptual rejection of testimony in the philosophy of history. The truth historians are looking for, Collingwood wrote, was not “possessed, ready-made, by the writer whom we are studying.”Footnote ⁶⁸ Similarly, Arthur Danto argued “the whole point of history is not to know about actions as witnesses might, but as historians do, in connection with later events and as part of temporal wholes.”Footnote ⁶⁹ In other words, testimony is rejected because historiography involves a distinct form of understanding that goes beyond the knowledge and understanding of events that historical witnesses could possibly have had.

Seminal historians and historical theorists were certainly very keen on separating historical knowledge from testimony. Less certain is whether this separation amounts to a coherent position. Would a wholesale rejection of testimony not also undermine the historian’s task of establishing basic factual claims about “what happened” in the past? For example, to what extent was the Winter Palace plundered when it was seized by the Bolsheviks on 26 October 1917? Eyewitness accounts of those present at the time would be vital sources of information for every attempt to establish what happened at a factual level. That knowledge may, in turn, serve as material for historical research concerning, say, moral norms and discipline during the Russian Revolution. Consequently, how would the historian’s account of the past ever get off the ground if all testimony is rejected as not only unreliable but also irrelevant for a distinctly historical way of knowing the past? In the following section, I will address this fundamental objection and clarify the conceptual rejection of testimony for establishing historical facts. Thereafter, in the fourth section, I proceed to the relation between testimony and historical understanding.

3.1 Material Dependence

One of the most influential recent proponents of the ontological argument is Paul Ricoeur. In his last major work, Memory, History and Forgetting, Ricoeur develops a position that seems to go directly against the basic claims of the evidential paradigm. According to Ricoeur, testimony constitutes the very bedrock of historiography – as the “womb of history.”Footnote ⁷² He contends that the practice of historical research would be unimaginable without testimony. Ricoeur’s argument is framed in opposition to the archival turn of modern historiography. He writes:

This passage highlights several important features of the ontological case for testimony. First, it is certainly true that, given the lack of other kinds of documentation, eyewitness testimony must have a fundamental role for the historian’s account of past events. As Ricoeur rightly emphasizes, how would knowledge in such cases be possible at all without relying on the attestation of those present at the events themselves?Footnote ⁷⁴ Second, Ricoeur is very much to the point when he claims that historiography does not start in the archive. As he writes with reference to Michel de Certeau’s work, the archive is “not just a physical or spatial place, it is also a social one.”Footnote ⁷⁵ In contrast to what Ranke and his followers assumed, the archive is not a neutral site of knowledge production that would somehow have the power of “assuring the objectivity of historical knowledge.”Footnote ⁷⁶ Rather, the archive material is always already a product of power relations that determine what is considered worth preserving.Footnote ⁷⁷

Historiography is not only materially dependent on the testimony of historical agents, but also on the decisions of archivists for the very existence of the material itself. Not only is the archive selective on a social and political basis, but the very idea of a complete archive is also incoherent. “[A]rchive as much as you like: something will always be left out,”Footnote ⁷⁸ as Pierre Nora exclaimed. In other words, the historian’s source material seems to depend on other human beings all the way down.

Ricoeur’s argument about the necessary social relations of source material offers a well-needed remedy for the fetishization of the archive in scientific history. Nevertheless, the argument is of limited relevance for appreciating the specific role of testimony in relation to historical facts.

Ricoeur clearly relies on reasoning from what W.H. Dray called “vacuous contrasts.”Footnote ⁷⁹ To claim that source material in general is always, in one way or another, the product of either human observation or human relations is a truism. It is equally true and unproblematic to say that historical research is always only based on a selection of source material. But as Dray pointed out in relation to the latter, this does not yet say anything specific about history. The contrast is vacuous – if it is meant to somehow set history apart from other sciences – since the feature one has identified is shared by all the sciences. Consequently, one ends up with the platitude that historical knowledge, like knowledge in general, is dependent on human observations and relations. This line of reasoning will also inflate the concept of testimony colossally by including all material dependent on human beings, that is, everything from train tickets to assertions of an eyewitness. Obviously, this broad concept cannot do justice to the central historical research practice of comparing the assertions of a witness (testimony) not only with other testimonies but also with non-intentional material produced by the events under investigation. The ontological argument reveals the basic, material dependence on testimony in history, but it does not get us very far.

Ricoeur’s womb-metaphor highlights our fundamental dependence on other human beings for information about the historical past. In general, we tend to underestimate the fact that “our observation” comes mostly from the observation of others, and this is all the truer for historians who may observe the evidence but never the past itself. Historiography does share this typically unrecognized and widespread dependence on the word of others. As Avishai Margalit writes “This [dependence] is true for all our walks of life: science, religion, history, court, and of course for our collective memory. … I am caught in a network of witnesses.”Footnote ⁸⁰ Important as these points are, this fundamental dependence was not denied by classical proponents of the evidential paradigm. For example, both Collingwood and Bloch readily acknowledged this basic form of reliance.Footnote ⁸¹ As Bloch famously wrote:

Bloch continues by emphasizing that “[a] good half of all we see, is seen through the eyes of others,” and that “all knowledge of mankind, to whatever time it applies, will always derive a large part of its evidence from others. In this respect, the student of the present is scarcely any better off than the historian of the past.”Footnote ⁸³ For Bloch, however, it is important to qualify this dependence by pointing out the relativizing input from material evidence. The “tracks” and “residues” of the past are things that the historian can indeed “see with his own eyes.”Footnote ⁸⁴

It should now be clear that the challenge of testimony requires further specification. The key issue is not whether history does depend on information gained from relying on statements by other human beings. As we have seen, advocates of the evidential paradigm were not denying this basic dependence. The question at issue is, instead, what kind of dependence the historian’s testimonial reliance amounts to. What does reliance on testimony mean in historical research? For this one must turn from the ontological domain towards questions about the epistemic relation between testimony and historical knowledge. In recent work, this epistemic relation has been articulated in efforts to align historical knowledge with the flourishing field of the epistemology of testimony in analytic philosophy.Footnote ⁸⁵ The basic argument is that historiography is akin to knowing by testimony since historians too must rely on the word of others as epistemic justification.

3.2 Epistemic Dependence

It is impossible to assess the epistemic relevance of testimony for historiography without first defining the relevant meaning of testimony. While the ontological argument tends to broaden the concept of testimony, in (analytic) epistemology discussions testimony often denotes the specific speech acts of telling and/or asserting.Footnote ⁸⁶ From such speech acts, it is argued, one may acquire testimonial knowledge. Given this focus on the nature of specific speech acts, testimony becomes a philosophical term of art for knowledge and belief acquisition via tellings. How to understand testimonial knowledge is, however, at the very heart of the debate. Roughly, the general idea is that we have testimonial knowledge when we believe something because someone tells us. The focus of the philosophical debate has been about the justification of testimonial knowledge, that is, whether it is acceptable to call testimony knowledge, and if so under what conditions.

There are two traditional camps in the epistemology of testimony: (i) reductionism or evidentialism, and (ii) non-reductionism or, later, “the assurance view.” David Hume (1711–1776) is often construed as the father of reductionism, while Thomas Reid (1710–1796) is styled as the progenitor of non-reductionism. Reductionism has historically been dominant and is often considered the traditional position, but today there is a plethora of reductionist, non-reductionist and intermediate accounts.Footnote ⁸⁷ The bone of contention between reductionists and non-reductionists is the source and character of epistemic justification. In brief, non-reductionists claim that testimony is a sui generis source of justification of knowledge, and reductionists claim that testimony can amount to knowledge only if supported by independent reasons for belief provided by the individual’s own perception, memory and reasoning. However, the dispute is not about whether it is, in general, possible to transmit knowledge and belief via testimony. Both sides accept the commonsense fact that we do constantly learn about the world from what other people tell us. Rather, what reductionists deny, and non-reductionists confirm, is that beliefs based on testimony can be epistemically justified merely in virtue of the fact that another person has asserted the claim in question.

The most interesting question for historiography is the possibility of non-reductive testimonial knowledge. As we have seen, the role of reductive testimonial knowledge is, of course, commonplace in the ubiquitous practice of using testimony as historical evidence. In fact, this reductive practice is exemplified every time a historian compares the assertions of a witness with non-intentional evidence material. The case for non-reductive testimonial knowledge in historiography, however, was pioneered by C. A. J. Coady in his landmark book, Testimony: A Philosophical Study (1992). Coady’s seminal book, which shaped subsequent discussions in the epistemology of testimony, argued that – contrary to the argument for autonomy by Collingwood and others – historiography is inherently dependent on non-reductive testimonial knowledge. Coady does not, however, discuss the ways in which historiography includes narration and analytical concepts, such as the Unification of Germany or the Renaissance. His claim is not that such notional totalities, beyond the perception of individuals, can be known by way of testimony. Coady is interested only in what he calls historical facts. Specifically, he is interested in facts that can be expressed in singular existential propositions, such as “In September 1830, there were three days of street-fighting in Brussels.”Footnote ⁸⁸ In other words, Coady emphasizes that there are indeed cases of facts about the past, contrary to Collingwood’s claim, for which the truth of the matter was possessed ready-made by the historical agents themselves. As Coady argues,

As a well-known proponent of non-reductionism, Coady’s motive for discussing historiography was to show the soundness of that position in relation to historical facts. Coady’s main philosophical argument is that the possibility of linguistic communication already guarantees some positive correlations between statements and facts. Understanding a language in use requires that one relates to some of the reports expressed in it as true. For how else would the hearers ever be able to assign content to the speaker’s assertions – if they cannot presuppose that there is any reliable correlation between the speaker’s assertions and the truth? The reductionist thesis is therefore found to be inherently flawed. Rejecting all testimony as inherently unreliable is not a coherent position. One cannot “understand what testimony is independently of knowing that it is, in any degree, a reliable form of evidence about the way the world is.”Footnote ⁹⁰ For Coady, epistemic reliance on testimony in historiography, as well as in everyday conversations, simply means that one must (logically) presuppose that the word of others is a reliable, but defeasible, source of information or data.

The main strength of Coady’s argument is that it articulates features of testimonial reliance that historiography shares with everyday knowledge and belief transmission. The (a priori) reliability of testimony, as a logical presupposition of linguistic communication, holds, of course, for historiography in just the same way as it does in everyday situations. Thus, Coady’s argument exposes the fallacy of rejecting all testimony as inherently unreliable. This argument, in turn, shows that the Enlightenment ideal of epistemological individualism is untenable also in historiography.Footnote ⁹¹ The very practice of comparing testimonies relies on giving credence to some testimony over others, which, ipso facto, prohibits the rejection of all testimonial reliance.

Contrary to what Coady believes, however, the fact that linguistic communication presupposes a norm of truthfulness does not warrant anti-reductionism about testimony.Footnote ⁹² The reason being, of course, that it is the very same norm that also makes untruthful testimony possible.Footnote ⁹³ One cannot justify testimonial reliance in any individual case by pointing to the fact that global error is philosophically unthinkable.

In relation to historiography, the main weakness of Coady’s argument is his failure to understand what philosophers of history, especially Collingwood, were rejecting in rejecting testimony from history. For Coady, the main aim was to show that testimony is a reliable form of evidence “about the way the world is.”Footnote ⁹⁴ From the start, he describes testimony as an attestation that “is evidence towards the settling of the matter.”Footnote ⁹⁵ In Coady’s view, epistemic reliance upon testimony is a species of evidential inference in which the evidence happens to be the word of another person. No significant conceptual distinction can therefore be made between evidence and testimony for Coady. As I showed in the previous section, however, the rejection of testimony among classical philosophers of history is the result of stressing the very conceptual distinction that Coady neglects. The former argued that there is an important conceptual distinction between relating to source material as testimony or as evidence. Consequently, when classical philosophers of history rejected testimony, they were not, as Coady assumed, rejecting testimony as a species of evidence. On the contrary, they would all have agreed with Coady: in historical research, testimony is treated as evidence. Instead, what they rejected was testimony as authority.

As we saw, for Collingwood reliance upon testimony is specified as the distinctive act of believing another person on their say-so. This interpretation is strongly supported by the fact that Collingwood systematically connects the concept of testimony with relying on the authority of ready-made statements, and not with the mere act of using another person’s statement as evidential input in one’s own reasoning.Footnote ⁹⁶ What one receives in testimony are assertions to be accepted simply upon the other person’s authority. The main worry was that accepting a proposition on the mere say-so of another person seems – precisely as contemporary epistemologists argue – to involve a deferral of responsibility for the belief acquired from historians to the witnesses in the past.Footnote ⁹⁷ If a historian simply accepted the say-so of an authority, Collingwood argued, then the latter would be responsible for what went into the historian’s narrative.Footnote ⁹⁸ That, in turn, is in conflict with the very job description of the modern historian who should be “relying on his own powers and constituting himself as his own authority.”Footnote ⁹⁹ Furthermore, as the next section shows, Collingwood’s rejection was very much to the point since the social fact of deferral of responsibility, which non-reductive testimony involves, does not amount to epistemic justification.

3.3 Different Logics of Justification

As we have seen, the conceptual rejection of testimony from historiography should not be confused with rejecting the use of testimony as evidence. Classical philosophers of history did not reject testimony as evidence nor as a source of data, but rather testimony as a ready-made authoritative account of past events. Nevertheless, further specification of the distinctiveness of historical knowledge is needed. One may still ask: If it is sometimes legitimate to rely on testimony as ready-made, authoritative answers in everyday life, as it arguably is, then why are we not allowed to do so also in historical research? In other words, can testimony be something more than just evidence for the historian? These questions connect directly with ongoing discussions in the epistemology of testimony in which much attention has been given to questions about the categorical difference between relating to the word of others either as evidence or as assurance.Footnote ¹⁰⁰ Richard Moran, one of the main advocates of this distinction, has argued that to accept a speaker’s testimony is not to take it as good evidence for belief, but to believe what is said on their say-so, their assurance. When testimony is understood as assurance it is “the speaker who is believed, and belief in the proposition asserted follows from this.”Footnote ¹⁰¹

In Moran’s work, assurance and evidence are distinct epistemic attitudes with different logics of justification. He explicates the difference with the following example: I look outside the window on a sunny day and see people bundled up in thick clothes. First, I consider their behaviour as evidence and draw the conclusion that it is colder outside than it otherwise looks. Second, my friend arrives and tells me that it is cold (or not). In the latter case, my friend’s verbal behaviour is not evidence from which I make my own conclusions about the weather. My friend tells me something and I believe him. This means that my entitlement to believe is based not on my own assessment of what is said, but on the speaker’s standing by his word. My friend gives me assurance that it is cold outside. As Moran writes, the spirit in which he presents his words is not, “‘Now I have spoken; make of it what you will’ but rather ‘Take it from me’.”Footnote ¹⁰² Consequently, relating to all tellings as evidence misconstrues the fact that the distinctive feature of testimony is that the speaker’s telling is not data waiting for interpretation, but assurance for the truth of the statement. Moran therefore describes the epistemic phenomenon of telling as an aspect of a “basic relationship between people.”Footnote ¹⁰³

Importantly, assurance and evidential relations involve distinctly different logics of justification. Relating to something as a piece of evidence, I do not think of it as a message intentionally aimed at me. Moran illustrates this relation with a characteristic detective story scenario. The murderer did not leave a piece of evidence at the crime scene in order to make me believe something or other. Or rather, if he did, it constitutes “tainted” evidence:

This explicates a qualitative difference between two epistemic attitudes: one can relate to testimony either as assurance or as evidence. When articulating this contrast, Moran points out that if I see someone’s testimony under the aspect of the concept of evidence, I should focus on unintended messages, that is, the speaker blushing, stammering or otherwise revealing that she has something to hide. Should she attempt to influence my opinion, then this would come out as manipulative. In contrast, if I relate to something in the spirit of assurance, then precisely the fact that the speaker wants to tell me something, and wants me to believe it, is a reason for me to believe him: “What this provides me with is different in kind, though not necessarily in degree of certainty.”Footnote ¹⁰⁵ Consequently, the two epistemic attitudes involve different logics of justification. What takes the appearance of support in the case of assurance, amounts to disproof in the case of evidence.

3.4 The Limits of the Evidential Paradigm

How does historical research relate to the difference between the epistemic attitudes of assurance and evidence? Facing this question, several cracks emerge in the evidential paradigm. First, it should be obvious that historians are not always only interested in the unintended evidence of testimony but also in the intended content. In such cases, Moran’s distinction between different logics of justification holds also for historiography. The fact that a historical agent intentionally tells something, and wants his interlocutors to believe him, does not automatically make his statements fraudulent evidence. This implies that one central claim of the evidential paradigm in historiography – that testimony is evidence simpliciter – is misleading. Contrary to non-intentional source material, such as minutes, train tickets and legal codes, the fact that testimony is produced with the explicit intent to convey certain information about the past does not diminish its credibility as proof. Of course, testimony is still treated as evidence for answering the historian’s questions, which involves a demand for corroboration from other testimony and non-intentional material, but it is nonetheless treated distinctly as a candidate for truth (assurance) and not as mere data awaiting interpretation. In this sense, testimony is an epistemic source like no other in historiography.

The solidity of the evidential paradigm is challenged also by the fact that historians inevitably rely on tellings as assurance in their work. No researcher, historian or other, can possibly verify all the sources they rely on first-hand. Furthermore, the idea that historians could possibly relate to all tellings from historical agents as merely unintended evidence is incoherent. As Bloch and Collingwood emphasized, a central part of modern historical research is to interpret testimony in spite of itself. Especially Collingwood, who tended to be much more assertive on this issue than Bloch, claimed that such interpretations belong to the very essence of history. The scientific historian, Collingwood wrote, is “twisting a passage ostensibly about something different into an answer to the question he has decided to ask.”Footnote ¹⁰⁶ In other words, to do history is to draw conclusions about what a particular statement betrays about, say, cultural norms or the (hidden) motives of the agents. Still, it is not even logically possible that historians would always only read testimony against the grain. The impossibility of relating to statements by historical agents only as the evidence of unintended messages is obvious also in Bloch’s classical example of reading Saint-Simon’s statements as evidence of court mentality.Footnote ¹⁰⁷

If historians want to use statements in spite of themselves, they must first identify what the speakers are doing with their words. This identification is, as such, not independent from questions about the truth of the agent’s statements. The question of truth is, however, not answered by Rankean gazing at the past. Rather, one evaluates the agent’s statements as intentional messages in relation to other statements and our general knowledge of the context. Only this assessment will allow one to identify what Saint-Simon is, plausibly, doing with his words: is he telling, joking, guessing or reciting? Without this basic identification of the speech acts, one will not be able to make claims about the behavioural evidence of Saint-Simon’s statements either. Clearly, any judgement about what kind of mentality Saint-Simon is inadvertently expressing will be dependent on whether one can identify when Saint-Simon is speaking sincerely and when he is not. One should, of course, not deny that these different aspects of Saint-Simon’s statements are often mutually intertwined. Thus, understanding the mentality expressed in Saint-Simon’s statements may also be crucial for correctly understanding his speech acts.

If all material of the past were treated merely as behavioural evidence, then historians could not get any foothold in their material to begin with. Held strictly within an (behavioural) evidential paradigm, the historian would be limited to making claims about the possible material causes of the remains from the past. Accordingly, historians would be limited to suppositions about whether the author of the source material was left-handed, perhaps a quill pen was used, and that one can see signs of stress in the handwriting and so forth. It is only by not relating to source material as behavioural evidence, but also as tellings offering assurance, that the credibility of the historical agent’s original tellings can be examined. Often enough it may also be precisely the truth of the statements, rather than their status as cultural evidence, that interests the historian.

The evidential paradigm is misleading to the extent that it implies that testimony qua assurance has no role to play at all in historiography. Assurances are a central epistemic source also in historical research. In everything from everyday conversations to scientific research, one may relate to the tellings of others either as assurances, that is, as candidates for truth, or as (behavioural) evidence from which one arrives at one’s own conclusions. As such, this distinction does show that Collingwood’s exclamation that history has “no relation to testimony at all”Footnote ¹⁰⁸ is an overstatement – if read as a denial of testimony as assurance in historical research. Acknowledging the role of assurance implies that this Element advances a mitigated form of evidentialism, in contrast with the austere evidentialism that some of Collingwood’s and Bloch’s claims seem to imply.Footnote ¹⁰⁹

3.5 Assurance and Epistemic Justification

It is not obvious that classical philosophers of history, such as Collingwood and Bloch, were indeed rejecting testimony as assurance from historical research. What they certainly rejected, as I will argue next, was, instead, that a historical agent’s assurance alone can serve as epistemic justification for historical knowledge. In this respect, the question of testimony in historical research connects, again, with a much-debated issue in the recent epistemology of testimony: Is testimony an independent form of knowledge, or does it, in the end, require validation from evidence? Answering this question, it will be pertinent to situate the case of historiography in relation to one of the most debated arguments for non-reductive testimonial knowledge; namely, the so-called Interpersonal View of Testimony (IVT).Footnote ¹¹⁰

Several influential exponents of the IVT have argued that the justification of testimonial knowledge comes from the relationship of trust between the interlocutors. According to Benjamin McMyler, testimony provides justification by way of “the particular kind of reason that it does in virtue of the interpersonal relationship between us.”Footnote ¹¹¹ Or, as Edward Hinchman states, telling involves an invitation to trust.Footnote ¹¹² Every time you tell me something, you want me to believe it because you said it, and my belief follows from our relation of trust. According to Paul Faulkner, the relationship of trust between the interlocutors is in itself “a reason providing state. … [O]ne where an audience (as truster) expects a speaker (as trusted) to try to say what is true because the audience is dependent on the speaker doing so.”Footnote ¹¹³ The relationship of trust, therefore, is a kind of social fact that pushes people towards truthfulness in their communication with others. Thus, trusting someone for the truth implies that the listener has the right to resent the speaker in the event of cheating.Footnote ¹¹⁴ Conversely, the speaker who invites the audience to trust what she says is entitled to feel slighted if her words are not received as trust-based reasons for belief.Footnote ¹¹⁵ Thereby, telling creates obligations. In the paradigm case, I ought to expect my interlocutor to tell me the truth and I have the right to resent her if she does not.

The IVT is much debated in the contemporary epistemology of testimony. The central problem is: Why would the ethical and social dimensions of interpersonal relationships validate, or even make it probable, that my informant is not misleading me? How can the social and ethical facts of testimonial reliance be translated into epistemic warrant? Naturally, the proponents of the IVT do not want to suggest that we should always trust people when they tell us something. Considering this unclarity about the epistemic dimension, Jennifer Lackey argues that the IVT does not succeed in showing that knowledge from testimony is a separate, independent form of knowledge. Testimony is indeed a source of knowledge, but as data, and it is subject to all the tests that otherwise apply to data. As Lackey writes:

In other words, we trust testimony, not on the mere authority of the speaker, but because we have independent grounds. Trust does not solve the problem, because we must explain why we should trust the speaker. This means that, in the end, the epistemic justification of testimonial reliance is explained only by evidentialism.Footnote ¹¹⁷ This Element is not, however, the place for a detailed account of the debate about the IVT in the epistemology of testimony. Instead, I will use the problems identified for the IVT to explicate the role of justification from testimony in historical research. As I will show, the problems concerning the independent status of testimony as epistemic justification are further accentuated in historical research.

3.6 Communal Historiographic Argumentation

For the IVT, the concept of interpersonal trust explains why testimony can function as an independent form of knowledge. Typically, there is no such interpersonal relation of trust between the historian and historical agents. Already this fundamental difference implies that testimony cannot have the same justificatory role of providing assurance in the case of history. This difference is, however, not only because people in the past have not given historians their word. Even if a historical witness did give the historian his word, this does not mean that the historian, therefore, is related to the statement of the witness in the same way as in everyday conversations. The crucial difference is, of course, that historiography is a distinct research practice, while everyday conversations are not. Consequently, when the historical agent’s assurances enter historiography, they become, by definition, subject to a space of contestation and argumentation. This fundamental space of historiographic argumentation was described by Ludwig Wittgenstein as “the historical proof game.”Footnote ¹¹⁸ Collingwood, in turn, used a similar metaphor:

Today, Collingwood’s view about the rules of the game – which he later calls the definition of historical thinking – is very much in line with a broad explanationist consensus about proof in historiography. Explanationism stresses that all knowledge is relative to specific epistemic contexts, fallible, and assumes a hypothetical character – there is no foundation for knowledge claims beyond the ways in which the game of proof is played within the community of professional historical research. In this spirit, Lorenz has argued that “The problem of justification in philosophy of history boils down to the question of what kinds of argumentation historians use to argue their claims to knowledge.”Footnote ¹²⁰ A similar focus on justification through argumentation is found among many theorists writing about historical evidence and proof today.Footnote ¹²¹

The metaphor of games and players is very appropriate since it turns our attention towards the communal nature of historiography.Footnote ¹²² Because history is communal, the historian must appeal to independently ascertainable grounds when arguing about whether a testimonial account should be accepted. Appealing to independent grounds entails a conceptual shift in one’s relation to the word of others. As Collingwood phrases the change: “As soon as there are such grounds, the case is no longer one of testimony.”Footnote ¹²³ If one is appealing to grounds available to everyone, then one is, in fact, no longer accepting something merely based on one’s own relation of trust to the person offering testimony.

The crucial role of communal reasoning also explains why the model of inheriting justification, sometimes called epistemic buck passing, cannot serve as an argument for testimonial reliance in historiography.Footnote ¹²⁴ This argument, endorsed by Mark Day, claims that testimony is to be viewed not as a transmission of knowledge but as a transmission of justification. As Day writes:

Day is right concerning justification for the content of the belief, but he underestimates the difference the communal nature of historiography makes for the transmission of justification. To engage in historical research is to engage in a critical practice that invites further reasoning about why some testimony should be accepted. As we saw, this does not mean testimony reduces to behavioural evidence in historiography, but it means that the historian’s reason for accepting a statement is necessarily related to grounds in a way that acceptance of the word of a friend in everyday conversations is not. In the case of friendship, my personal relationship to my friend is my grounds. Not only that I may accept my friend’s word because of our friendship, but also that in relation to the word of my friend there is normally no doubt that needs grounds to be defeated. The default position in history, however, is different. When questions are raised about reliability, testimony is always-already in need of independent grounds to convince other historians of the reasonability of its acceptance.

Contrary to everyday conversations, historians must always appeal to grounds beyond their own relationship to the events or persons involved. For instance, historians are already appealing to something independently available by claiming that someone is a trustworthy witness. Such claims are inevitably connected with producing arguments, available to the community of historians, for why one should consider someone reliable. But to say that I take someone’s word because they have been shown to be reliable is no longer simply to accept something on another person’s say-so. By appealing to someone’s trustworthiness, I am, in essence, relying on independent grounds, available to everyone, and not merely accepting that person’s word. I have, in fact, reasoned independently and come to the conclusion that, on this subject, it is reasonable to accept that person’s statement. In addition, historians are only justified in such reasoning if they can convince their colleagues that they too would have come to the same conclusion. The leap to something independently acceptable, which historical research necessarily involves, entails that testimony cannot have the same justificatory role as it may have in everyday conversations. For the historian, contrary to everyday conversations, all transmission of justification is subject to the communal process of historiographic argumentation.

In sum, it is important to explicate and sometimes moderate the conceptual rejection of testimony in modern history. Historical research does indeed share the basic dependence on testimony that fundamentally shapes knowledge acquisition in all sciences and everyday life. But there are also significant differences. The impossibility of deferring responsibility derives from the communal nature of historical research, which implies that reasons for belief are subject to a collegial argumentative process. Nevertheless, historians do not relate to testimony only as evidence in the sense of data to be interpreted, but also as assurances with intentional content, offering candidates for truth. Concerning the epistemic dimension of testimony in historiography, it is very important to avoid dualism: historiography involves relating to the words of witnesses both as evidence and as assurance, and these distinct epistemic attitudes are not mutually exclusive.

The entire discussion has so far focused on the relation between testimony and historical facts. The main point of comparison was knowledge and belief acquisition in the epistemology of testimony in general, and particularly the role of non-reductive testimonial knowledge in historiography. This framework is, however, useful only for the limited yet important aspect of establishing singular propositional facts in historiography. Furthermore, (analytic) epistemology of testimony is often labouring under a fundamental presupposition that historiography does not always share. One assumes that the speaker knows what the hearer wants to know, and that the task is to investigate how the transmission of knowledge between the interlocutors is possible.Footnote ¹²⁶ As we already saw, however, seminal philosophers of history have argued that historiography involves a distinct form of understanding that goes beyond the knowledge and understanding of events that historical witnesses could possibly have had. In fact, Danto considered going beyond the understanding of historical witnesses as the distinctive feature of historiography. This contention – which became formative for narrativist philosophy of history – was articulated in Danto’s account of the historian’s narrative sentences. Historiography, Danto argued, necessarily involves describing past actions and events in retrospective terms that were (logically) unavailable to contemporaries. His by now famous example was “The Thirty Years’ War started in 1618.”Footnote ¹²⁷ It is to the relation between historical understanding and testimony that this Element must turn next.

4.1 The Rebirth of Testimony in Historiography

The 1970s brought revolutionary changes for the role of testimony in historical research. Naturally, the use of testimony as evidence for establishing facts carried on as before, but there was undoubtedly also a distinctly new kind of interest in testimony. The subjectivity of witness accounts, which was considered a central weakness in the nineteenth-century turn to scientific history, rose up as the main strength of testimonial material. This turn involves focusing on the how rather than on the what of witness accounts – the truth of the telling rather than the telling of the truth. Using Christopher Browning’s distinction, historians became interested in the “authenticity” of testimony rather than merely their “factual accuracy,” or lack thereof.Footnote ¹³² As Assmann writes, this new contribution of testimony is framed not as a pathway to objective facts, but rather as the royal road to authentic experience: “Their point is less to tell us what happened than what it felt like to be in the center of those events; they provide very personal views from within.”Footnote ¹³³ With this shift of interest, it was not uncommon for testimony to be assigned the role of an immediate link to the past, transmitting directly the emotional and experiential dimension of historical events, especially concerning the so-called limit events of the twentieth century.Footnote ¹³⁴

Today, testimony is a major topic in many fields of the humanities – from gender studies to analytic epistemology and phenomenology – and the interdisciplinary field of testimony studies has grown exponentially since the 1980s.Footnote ¹³⁵ One overarching development is, of course, the rise of interest in experience, memory and subjectivity within all the humanities. For historiography, this turn was originally fuelled primarily by democratic aspirations of creating counternarratives “from below,” but memory has moved from being an alternative, minority movement to becoming the hegemonic form of popular historical representation. Some scholars even claim that professional historians have lost their positions as the go-to interpreters of the collective past. At least in public domains, historians are increasingly replaced by eyewitness accounts of those “who were there when it actually happened,” as stylishly framed by journalists, authors, documentary producers, game designers and museum curators.Footnote ¹³⁶ As Sara Jones has argued, the appeal of testimony is closely connected with the idea that eyewitnesses can offer historical understanding something unique (a [perceived] authentic account of the past), thus relying on “the sense that [witnesses] must know what it was like because they were there and the promise that this experience can be transmitted to the listener, reader, visitor or viewer.”Footnote ¹³⁷

In historiography, the new interest in testimony connects with two concurrent scholarly and popular trends: the rise of victim testimony and the politics of identity. For the former, it is widely acknowledged that the rise of victim testimony is inextricably linked to the process of relating to witness accounts from Holocaust survivors, which was set in motion in the wake of the Eichmann trials in the 1960s. This process inspired a massive upsurge of both public and scholarly interest in memory accounts, a turn that the French historian Annette Wieviorka has labelled the advent of an “Era of the Witness.”Footnote ¹³⁸ This characterization is not peculiar to Wieviorka; many influential scholars have acknowledged the turn toward memory accounts for understanding events in the collective past. For instance, Jay Winter has written that memory is “the historical signature of our own generation,” Pierre Nora has claimed that our time is an “era of commemoration,” and Kerwin Lee Klein has argued that memory has become the dominating metaphor for understanding how the present should come to terms with the collective past.Footnote ¹³⁹

The turn toward memory accounts is closely connected with the politics of identity. As Nora argued, there are as many advocates of memory accounts of past events as there are collective groups in society.Footnote ¹⁴⁰ This claim is well-supported by the fact that memory accounts typically serve as instruments for advancing the status of group identities. In many cases, articulating the experiences of oppression through testimony is tantamount to a claim for the acknowledgement of a particular group identity. The American historian Allan Megill proposed a formula for highlighting this connection: “Where identity is problematized, memory is valorized.”Footnote ¹⁴¹ Furthermore, this development is often linked with an ethically motivated idea of multiperspectivity. Memory and testimony give researchers access to the counter-stories of historically disadvantaged communities, stories that can be used to criticize hegemonic narratives and highlight subalternity.Footnote ¹⁴² Consequently, accounts from memory have become a central tool for writing histories about experiences that had previously been either simply avoided or downright silenced.Footnote ¹⁴³

4.2 The Culturalist Challenge to History

The comprehensive turn towards testimony and memory has been labelled “one of the main challenges in the theory and practice of history in the recent few decades.”Footnote ¹⁴⁴ The rebirth of testimony, it is argued, entails that previously upheld oppositions between historical understanding and memory accounts should be dismantled. Both history and testimony offer ways to understand historical events, and both are equally conditioned by linguistic and cultural mediation as well as present concerns of politics, power and identity.Footnote ¹⁴⁵ Thus, historical writing can be subsumed under the metaphor of memory and characterized as but one of many cultural vehicles for collectively remembering the past in the present. As the debates about history and memory can fill a small library, the following will focus only on features especially relevant for the specificity of historical understanding in contrast with understanding the past via testimony.Footnote ¹⁴⁶ There are, of course, also several notable critics of the ongoing general rapprochement between history and memory.Footnote ¹⁴⁷

For the case of testimony, the most popular method of rapprochement is to subsume history under the category of culture. This leads to the conclusion that academic history is just one among many cultural practices for relating to the past. As we have seen, memory and history have traditionally been viewed as opposites ever since the advent of the Rankean paradigm. Memory was downgraded as subjective, unreliable and partial, whereas history has been placed on an epistemic pedestal as objective, reliable and universal. Drawing on work in memory studies, especially Peter Burke, Aleida and Jan Assmann, and Patrick Hutton, Marek Tamm has argued for overcoming this opposition by viewing historiography itself as a culturally specific memory practice in its own time. As Burke proposed, one may treat historiography “much as Halbwachs treated memory, as the product of social groups such as Roman senators, Chinese mandarins, Benedictine monks, university professors, and so on.”Footnote ¹⁴⁸

In brief, Tamm proposes a “solution” to the opposition between memory and history by arguing that “history is a cultural form exactly like, for instance, religion, literature, art, or myth, all of which contribute to the production of cultural memory.”Footnote ¹⁴⁹ According to Tamm, however, the reduction of historical writing to cultural memory does not spell the end of history’s “scientific pretensions.”Footnote ¹⁵⁰ Historiography is still a very specific medium of cultural memory with its own rules and traditions, including documentary evidence and professional critique.Footnote ¹⁵¹ The consequence of the reduction is, instead, that one must reject “the illusion” of a “determined, fixed historical past, lying beyond the pale of living memory and clearly distinct from our present concerns.”Footnote ¹⁵² What one must accept is that history writing is one among many “memorial activities” in which individuals and groups “recollect and construe the past selectively through various media.”Footnote ¹⁵³

This argument is as correct as it is irrelevant. Of course, historiography can be viewed as a specific cultural form. The professional historian was certainly well represented as a producer of cultural memory during the heyday of European nationalism.Footnote ¹⁵⁴ Still, this culturalist argument is completely beside the point for understanding the conceptual distinctiveness of historiography in relation to testimony. Without pausing, Tamm contends that “in terms of cultural memory, history is a cultural form.”Footnote ¹⁵⁵ This is obviously true, but Tamm provides no argument whatsoever for why it would be sufficient to consider historiography only in terms of cultural memory in the first place.

For instance, one may consider Ludwig Wittgenstein’s Tractatus Logico-Philosophicus (1922) in cultural terms as an expression of late Viennese Fin-de-siècle literature. This is, however, only one way to understand the work. In contrast, one may understand the Tractatus as philosophy and engage directly with the conceptual questions that the work poses about logic, language and reality. Certainly, the same goes for historiography. One may relate to historiography either as a cultural practice or as a philosophical idea with distinct presuppositions, principles and a priori concepts. While it is true that qua cultural practice historiography may serve the same function of cultural memory as witness literature or myth, this says nothing yet about the conceptual relation between testimony and historiography as forms of thought about the past. In other words, the only thing Tamm’s solution to the opposition between memory and history offers is a description of what historiography looks like from the viewpoint of a cultural historian. One can, and should, proceed to consider the relation from the perspective of philosophy and the theory of history.

4.3 The Historicity of History

Philosophically more astute, although similar in content, is the argument for rapprochement considering that historiography is, like testimony, always-already framed by the historian’s own, present historical situation. This argument for the historicity of history, hailing at least from Walter Benjamin, takes many different forms but all of them reject a simple binary between history and memory.Footnote ¹⁵⁶ For example, Dominick LaCapra and Paul Ricoeur reject the binary because history grows out of cultural memory and relations of transference, María-Inés Mudrovcic because “historical time” can no longer be separated from the present, and Mark Day because historiography depends on the historian’s multiple relations to the past.Footnote ¹⁵⁷ The postmodern revelation that history is constructed in the present, Assmann argues, persuades one to accept that “history is itself a form of memory.”Footnote ¹⁵⁸ In essence, these scholars pose the question: Can the historian’s relation to the past ever be strictly a historical one? In other words, is the ideal of question-driven rational reasoning from evidence, which is the cornerstone of the evidential paradigm in historical research, really an apt description of what an individual historian’s relation to the past looks like in real life? And where does the motive for engaging in historical research come from?

This reality check, so to speak, is especially relevant considering recent work in historical theory. Influential theorists have shown how past actions and events refuse to become the dead historical past of evidence, and instead returns in experience and claims for justice and recognition in the present.Footnote ¹⁵⁹ As a result, the evidential paradigm, along with the focus on narrative form, has recently been problematized as a narrow and one-dimensional understanding of the historian’s relation to the past. Instead, many historical theorists have argued for approaches that highlight the multifarious relations between historians and the past that they investigate.Footnote ¹⁶⁰ According to Mark Day, who invented the “relations with the past” notion, the main concern for twenty-first century historical theory should be to assess how epistemic issues in history are entangled with relations to the past that he calls “evaluative, preservative, dialogic, material and practical.”Footnote ¹⁶¹ These relations are inseparably connected with the epistemology of history, and the latter cannot be properly understood without taking the former into account.Footnote ¹⁶²

The relations Day speaks about are not principally a matter of choice but relations every historian enters into by the mere fact of inhabiting a world inherited from previous generations. This feature is most clear with the relations he calls material and practical: we begin our lives with artefacts from the past already in place and conventional (practical) ways of relating to events in the past. The “evaluative,” “preservative” and “dialogical” relations are similarly not a matter of choice for the historian but part of the conditions for acquiring historical knowledge at all. The evaluative concerns the necessity of selection and thick descriptions in history, the preservative is about the historian’s basic reliance on testimony and the dialogical is the possibility of being challenged by voices from the past in historical interpretation. As LaCapra has argued, such challenges are vital for articulating the historian’s subject-position.Footnote ¹⁶³ Without the opportunity to be challenged by testimony, it would be impossible for the historian to attain a reflective relation to the ways in which their own preconceptions may lead to epistemic and hermeneutic injustice.Footnote ¹⁶⁴

The argument for multiple relations with the past in historiography is important. Every historiographical work will inevitably start from the individual historian’s historically determined, multifarious relations with the past. For example, the practical dimension is evident already in the historian’s choice of what testimony to use as evidence. As is well-known from the history of historical research, much testimony was not even considered relevant source material for doing history until the turn toward history “from below” in the 1960s. The fact that women, workers and subaltern subjects are indeed also historical agents became obvious for historians only gradually. In this respect, every historiographical work will be shaped by the individual historian’s practical relations to the past. For example, the concern for neglected and silenced subjects may serve as crucial input concerning what testimony historians analyze in their professional work. Undoubtedly, historians may even consider it as part of their ethical responsibility to also use previously neglected testimony as evidence in their research. Consequently, the ethical aspect of the historian’s choices prior to their research work implies an important connection between historical research and the practical past.

4.4 The Ideal Past of History

Does the historicity of history imply that the conceptual distinction between the historical and practical past collapses? No, but it highlights that precision is needed for appreciating the import of the distinction. The purpose of the distinction, at least for Oakeshott, was not to provide empirical categories within which actual historiography can be placed into boxes as either historical or practical. If this were the case, then the distinction would indeed collapse since no historian can escape practical relations to the past completely.Footnote ¹⁶⁵ Instead, the historical and the practical is an ideal-type distinction similar in kind to, for example, Aristotle’s distinction between six forms of political constitution.Footnote ¹⁶⁶ For the latter, the types of constitution are not descriptive of existing ancient Greek states but denote rather the logical possibilities of political organization. Consequently, Aristotle’s principles of classification between poleis ruled by one, a few, or by many, do not become useless simply from the fact they do not correspond with Greek political experience, and arguably no state could ever be ruled by one person alone. The purpose of the distinction was, instead, to philosophically explicate the internal tendencies towards stability or instability, justice or injustice, within logically distinct ideal types of political constitution. Recently, Martyn P. Thompson has argued lucidly for an ideal-type interpretation of Oakeshott’s distinction. He writes:

Consequently, the fact that individual historians always have multiple relations to the past is not a problem for the conceptual distinction between historical and practical past. Just like every existing monarchy, all historiographical work will inevitably be the product of many different historical and social circumstances – historians may even understand their own work as interventions in such circumstances. Regardless of this fact, however, it will still be possible to isolate and discern the ways in which ideal-type elements are present within the real-life existing empirical cases of historiography. In other words, the fact that historiography will always harbour several different relations to the past does not spell the end for the (logical) possibility of autonomy and critical knowledge about the past.

In Oakeshott’s philosophy of history, the practical and the historical denote different (ideal) conceptual spaces. This means that “the practical” and “the historical” do not refer to different past realms but to different kinds of constructions of present evidence. The practical and the historical past are, respectively, produced by attending to material from the past via either a practical or a historical mode of understanding. In this context, the term “history” is used as a term of art. For Oakeshott (and Collingwood), history is a specific form of understanding and not, as in everyday usage, a word synonymous with a specific temporal location, the past. Consequently, to say that testimony is not historical, as both Oakeshott and Collingwood would, is not to deny that testimony may be from the past, nor that testimony is conditioned by past circumstances. Rather, not historical means not the product of historical forms of understanding.Footnote ¹⁶⁸

If past material becomes the focus of our practical understanding, then a practical past is created. In such cases, and testimony is a good example, we are concerned with “the past” for the sake of the present and the future. The practical past is, by definition, a past that is not thought of as worth knowing for its own sake but only “in relation to ourselves and our current activities.”Footnote ¹⁶⁹ We use this kind of past “to make valid practical beliefs about the present and the future.”Footnote ¹⁷⁰ This practical past is everywhere available and consists of the “accumulation of symbolic persons, actions, utterances, situations and artefacts,” which Oakeshott regarded as an “indispensable ingredient of an articulate civilized life.”Footnote ¹⁷¹ By contrast, the historical past is a past that is thought to exist independently of our own concerns. The past of history is studied for its own sake and deserves investigation in its own right.Footnote ¹⁷²

4.5 Autonomy Is Not Impartiality

Does the seeming commitment to own sake-ism expose “the historical past” as simply a sophisticated, covert argument for the old Rankean ideal of impartial and objective history? In other words, is the argument for autonomy an argument for impartiality?Footnote ¹⁷³ For clarifying this issue, it is worth-while considering how Collingwood – who considered Oakeshott’s philosophy of history a “high-water mark” – responded to the question of impartiality.Footnote ¹⁷⁴ In his lecture to the Stubbs Historical Society in 1936, titled “Can Historians be Impartial?”, Collingwood probably surprised his audience when he denied not only that impartiality is possible but also that it is even desirable. Collingwood argued:

In this quote, Collingwood sees partiality as the pervasive preconception and worldview we can never avoid. Our attempts to avoid partiality by disavowal or positioning are a futile struggle against a logical impossibility. No one can pull themselves up by their own bootstraps. Instead, we should realize that prejudice is, rather, a resource for the historian; it provides “steam in the engine of historical thought.”Footnote ¹⁷⁶ However, the attentive reader will notice a subtle distinction in Collingwood’s quote. He is not saying that historical thought is prejudiced, but that historical thought begins from prejudice. The distinction is crucial.

Thinking historically, according to Collingwood’s arguments in this lecture, means caring for the integrity of the past. Such care implies that historians are committed to the idea that past realities existed in their own right. As a result, there is always a question of whether our interpretations do justice to the past realities they deal with. Our interpretations and past realities are two different things – if we think about the past as historians and not as novelists. This fundamental distinction meant that Collingwood was not at all concerned about ideological interpretive frameworks in historical research. As long as we are still engaged in historical thinking, our prejudices will be broken and reshaped in the encounter with the source material. Collingwood enthusiastically claimed that “the power of historical thinking” makes him convinced that investigations carried out for political purposes also lead to historical results beyond mere ideological confirmation.Footnote ¹⁷⁷

One may not share Collingwood’s optimism about the power of historical thinking to break the chains of prejudice.Footnote ¹⁷⁸ The principal conceptual distinction at issue, however, is not affected by psychological factors. Although all historians are prejudiced, it will still be possible to conceptually distinguish between interpreting historical actions and events from the horizon of the past versus the horizon of the historian’s present. It was that very conceptual distinction that Oakeshott was articulating in his argument for the autonomy of historical thought.

Contrary to practical understanding, Oakeshott argued, the historical mode of understanding attends to material from the past as survivals to be treated “as evidence of past happenings.”Footnote ¹⁷⁹ The aim of this mode of understanding is to establish the “authentic meaning” of such remains in relation to the historian’s questioning-activity.Footnote ¹⁸⁰ On this view, history does not, contrary to what Mudrovcic and others suggest, rely on the absolute separation of historical time, but only on the reconstruction of a hypothetical past that explains the evidence.Footnote ¹⁸¹ The historical past should not be mistaken for a tool of temporal othering. Clearly, difference from the present is not a necessary condition for the historian’s explanation of evidence.Footnote ¹⁸² The historical past is premised neither on the break between past and present, nor on gazing the material for “what really happened,” but on constructing a picture of the past based on what present evidence obliges one to believe. In this process, the historical mode of understanding uses analytical concepts and narrative sentences to construct a past that never existed except in the writings of historians.Footnote ¹⁸³ Consequently, it is also incorrect to equate the historical past, as Tamm suggested, with the idea of a forever determined and fixed past – unless one is fighting the ghost of Ranke. Nevertheless, the historical past is not an arbitrary construction but is based on the rational commitment to believe only what the evidence obliges. As Bernard Williams argued, belief aims at truth, and this means that historians can never legitimately decide to believe whatever they want about the past.Footnote ¹⁸⁴

The ideal-type interpretation of historical understanding is crucial for appreciating the relation between history and testimony. This interpretation allows for the carving out of an independent conceptual space for historical relations to the past. The conceptual space of history is one in which critique and evidential inference are, categorically, not employed for practical concerns in the present.Footnote ¹⁸⁵ Oakeshott’s original aim with his distinction was to show that the practical mode is not the only possible relation to the past.Footnote ¹⁸⁶ In contrast, advocates for the rapprochement of testimony and history deny this very possibility. This is the essence of Tamm’s culturalist argument as well as the arguments for history as (merely) a product of non-historical relations to the past. The decisive fallacy here is to assume that if something is a construction in the present, then such constructions are all of the same kind.Footnote ¹⁸⁷ This idea was aptly articulated by the oral historian Roseanne Kennedy:

Oakeshott’s distinction proves the contrary. He shows that it is conceptually possible – even within a thoroughly constructivist view of historical knowledge – to distinguish categorically between historical and practical modes of understanding past material. Consequently, arguing for the rapprochement of testimony and history based on the truism that “all evidence is constructed” is a non-starter. Not all constructions of meaning are of the same kind, and interpreting past material considering practical concerns in the present is not the only alternative available. Recognizing practical understanding as a specific perspective for understanding past material entails, ipso facto, that it is not the only perspective we can have. For acknowledging something as a perspective means that one can “make room for other modes of understanding.”Footnote ¹⁸⁹ In other words, embracing constructivism does not bring historical thinking closer to the understanding of the past found in testimonies.

4.6 The Claim of Testimony: Experience and Understanding

As we have seen, there are no compelling arguments for the rapprochement between historical understanding and testimony. Arguments for testimony in historiography need not, however, rely on the failed project of conceptual merger. Instead, one may argue for the inclusion of testimony on ethical grounds, justified theoretically by the consequences of the linguistic turn. Gabrielle M. Spiegel has summarized the result of this reasoning in an illuminating way:

Here, Spiegel is not claiming that historians and witnesses are on par epistemologically speaking. Rather, she is arguing that at stake is a question about “an ethical response to the catastrophes of the last century and … a turn from epistemological to ethical commitments in the study of the past.”Footnote ¹⁹¹ Consequently, the rapprochement between testimony and history is not about the epistemic dimension, or, in Spiegel’s terms, the “responsibility to seek to ‘get it right’ in our investigations of the past.”Footnote ¹⁹² Instead, it is in relation to conditions of understanding – the dimension of meaning – where history and memory come together.

Spiegel’s account is nuanced. The argument is not that historiography and testimony are identical as forms of understanding either, but rather that their shared condition justifies the admission of the “ethical” understanding offered by memory accounts in “normal historiographical discourse.” Considering that arguments for conceptual merger have already been dealt with, the following focuses on two other premises for including the “ethical” dimension of testimony in historiography. The first is the premise that the form of understanding that testimony offers is indeed relevant for, and compatible with, historiography, when the latter has been freed from objectivist pretensions. For investigating this premise, one must scrutinize the kind of understanding that testimony offers and how it relates to historical understanding. After that, I will consider the second premise; the idea that testimony offers a pertinent framework for articulating the ethics of historical research.

What exactly is the ethical and existential contribution to understanding that testimony offers historiography? The salient concepts in this discussion are authenticity, voice, experience, immediacy, trauma and embodied/existential truth – all of which are related to the often invoked but conceptually imprecise goal of understanding what it was like to live through certain historical events.Footnote ¹⁹³ This very interest is, of course, related to the fact that almost all discussions are framed by victim testimony about so-called limit events that, by definition, pose the problem of understanding and representing extreme atrocities and genocide, among which the Holocaust is still the paradigm example.

The contribution to understanding that testimony supposedly offers to history comes in two different forms. First, there is the idea that testimony offers unmediated access to the witnesses’ experience. This is sometimes framed as the embodied or existential truth of a testimony.Footnote ¹⁹⁴ From this perspective, the witness is to be interpreted not as someone transmitting a ready-made understanding in their representation of the events. Instead, much like a footprint in sand, one should relate to testimony not as a conscious representation but as an imprint of the event itself. This means that the words of the witness, and their silence (the void), testify directly to the events by way of the immediacy familiar from traumatic experiences.Footnote ¹⁹⁵ Second, and this is the form of witnessing most discussed in historical theory, there is the idea that testimony is a vehicle for transmitting a ready-made, authentic understanding or experience. Giving voice to this experience, expressed in the witness’s representation of the events, is allegedly a central part of the historian’s ethical responsibility.Footnote ¹⁹⁶ Furthermore, by offering understanding of the events, witnesses become participants in the historian’s game of creating historical representations, although testimony has the unique possibility of conveying the authentic experience of what it was like for the witness.

Several influential historical theorists have recently endorsed the idea of testimony as a vehicle of authentic understanding, beyond the objectivist confines of historiography. For example, in a discussion about Primo Levi, Hayden White wrote, “Levi’s Se questo è un uomo … derives its power as testimony, less from the scientific and positivistic registration of the ‘facts’ of Auschwitz, than from its enactment in poetic utterance of what it felt like to have had to endure such ‘facts.’”Footnote ¹⁹⁷ Annette Wieviorka appeals to the “extraordinary riches” that testimony brings to historical research as “the encounter with the human voice that traversed history and, in oblique fashion, not factual truth but the more subtle and also indispensable truth of an epoch and of an experience.”Footnote ¹⁹⁸ Similarly, Verónica Tozzi argues that paradigmatic survivor testimony, such as Levi’s, is akin to historical research since it addresses specific social questions, such as how collaborationism and the breakdown of the human condition in the death camps were possible. Tozzi contends that both professional historical research and witness accounts offer ways of understanding historical and social processes, and both depend on information chains that rely on relations of trust that are subject to communal checks appropriate for all testimonial reliance.Footnote ¹⁹⁹ As Spiegel has pointed out, Tozzi’s account is at risk of obliterating conceptual distinctions between the understanding and knowledge of the past provided by historical research and witness accounts respectively.Footnote ²⁰⁰

That survivor testimony does offer ways of understanding life in the death camps is not in doubt. Accounts such as Levi’s are indispensable for every attempt to comprehend the dehumanization that was an essential part of the Holocaust.Footnote ²⁰¹ Consequently, Levi’s account is an important resource also for historians when they try, like everyone else, to grasp somehow how the events were possible. The crucial concern for historical theory is not, however, the general claim of understanding; rather, it has to do with determining how the understanding provided by witness accounts relates to the presuppositions, a priori concepts and interests of historical research. On this score, however, advocates of testimony have relied on images in which historical research is tantamount to the ideals of the classical Rankean paradigm. Consequently, history is described as antiquarian, objectivist, and in the pursuit of mechanistic contextualization to reveal merely “what happened.” Very little is said about history as a specific mode of understanding, which was the core of the evidential paradigm.Footnote ²⁰² White was, again, illustrative when he wrote that the historical past is useless since it can only tell us “what people in other times, places, and circumstances did in their situation at that time and place.”Footnote ²⁰³

Arguments for the rapprochement between history and testimony rely on the very same opposition that they want to dissolve. Historiography is reduced to ideals of objectivity, universality and reliability, while memory accounts represent the contrary.Footnote ²⁰⁴ Arguably, the linguistic turn showed that these classical ideals of historiography were illusory, and, therefore, one may dissolve the opposition by subsuming historiography under the same culturally mediated category as memory. This entire line of reasoning, however, begs one central question: Are the ideals of objectivity, reliability and universality – the founding myths of the Rankean paradigm – the best characterization of the distinctiveness of historical thought? As we have seen, Collingwood and many proponents of the evidential paradigm did not think so. In the following, I show how attending to the specificity of the historical past, produced by the historian’s questioning-activity, allows for a deeper understanding of the distinctiveness of historical thought in relation to testimony.

4.7 The Historical Past and Testimonial Understanding

One may exemplify the specificity of the historical past by attending to the ways in which the historian’s questioning-activity produces its own object of inquiry. Think of research questions about large-scale social and political processes, such as the causes of the Russian Revolution, the political consequences of the Unification of Germany, or the motivational background of the Holocaust. While facts about a singular event (say, whether there were three days of street-fighting in Brussels in September 1830) may be available via witness accounts, this is not the case for the processes to which the singular event belongs. Large-scale social and political processes, such as the Unification of Germany or the Holocaust, are notional totalities that no one could have possibly witnessed directly. This impossibility is logical, not psychological. Descriptions of historical events by historians employ concepts denoting events that are irreducible to one single happening but stretch out in time and space beyond the reach of the perception of any single human being. A person may witness the looting of a palace or the execution of prisoners at a death camp, but this is not to witness “the Russian Revolution” or “the Holocaust” in the historian’s sense of these terms. Furthermore, as narrativist philosophers emphasize, historians may understand historical events in terms of retrospectively constructed analytical concepts and narrative sentences that were not logically available to contemporary witnesses.Footnote ²⁰⁵

One must be careful not to overestimate the force of these considerations in relation to testimony. Advocates of testimony could argue, and correctly so, that historiography is not solely concerned with narrating large-scale processes and retrospectivity. Obviously, historians may be specifically interested in individual accounts of what it was like to live through large-scale historical processes – and an understanding of this subjective side of experience is precisely what testimony may transmit. This is true, but problems arise with the assumption that “understanding” is an item that can be transmitted between historical agents and historians. As we saw in previous sections, there are good arguments to be made for the idea that knowledge about facts is something that can be transmitted. If you tell me that the train leaves at five o’clock, then this knowledge, or at least the justification for the claim in question, is transmitted to me when I take your word for it.Footnote ²⁰⁶ It is, however, a completely different thing to say that testimony may transmit understanding or experience. Typically, by the latter terms, one means something that is irreducibly first personal – that is, a process that everyone must (logically) do for themselves. Witnesses cannot simply give their own understanding to the audience in the same sense as one may exchange goods. This is not to deny the fact that testimony may, of course, lead to changes in someone’s understanding.Footnote ²⁰⁷

The most coherent alternative is to frame the contribution of testimony as an invitation to shared understanding. In this respect, the idea is that witnesses may express an understanding or experience that induces the audience to enter the witnesses’ subjective perspective, and thereby get a glimpse of what it was like from their point of view. Now, the key question here is not whether that is possible but whether entering the subjective perspective of the witness is indeed a point of interest for history. Collingwood’s answer is affirmatively negative: “the real essence of historical thought,” he wrote, “is that it aims at discovering a past which no one remembers or ever did remember, for the reason that no one ever knew it.”Footnote ²⁰⁸ As previously discussed, the point here is not simply that historians may often retrospectively constitute a past. At issue is also the fact that historical thinking involves viewing the past not by submerging oneself in one individual perspective but rather by viewing a historical agent’s perspective as part of a larger whole.Footnote ²⁰⁹

Historical understanding is about viewing the past not from the perspective of victims or of perpetrators but from a view unavailable to both.Footnote ²¹⁰ This is the view of how these perspectives relate to the situation and, equally, how the perspectives themselves serve as the producers and subjects of the very dimension of meaning inherent in the situation they faced. Importantly, this does not entail that historians may never legitimately use testimony as a complement for appreciating the lived experience of historical events. This kind of use of testimony is, of course, ubiquitous in historical research and has been developed to perfection in Holocaust studies by scholars such as Christopher Browning and Saul Friedländer. Using testimony as a complement, however, should not be confused with substitution, which would be to claim that the text cited from a witness account is historical understanding.

There is a conceptual watershed between history and testimony in relation to experience. In testimony, the experience of the witness is both the origin and endpoint of an account. Experience/understanding is that which is to be transmitted, and the process is completed when the message is received as it was sent. For history, the process is reversed: the experience and understanding of the witness is that which is to be explained, the very starting point of analysis. The historian’s task is to view an experience critically as the product of a particular understanding of the world. Accordingly, entering the space of the much-discussed “subjectivity” and “immediacy” of witness accounts, which testimony may offer, is explicitly not what historical thinking is about. Critical historical research does not, as Joan W. Scott wrote in her critique of foundationalist claims for experience, deny the existence of subjects, but critical history is different from the endorsement of specific subject-positions from the fact that “it … interrogates the processes of their creation.”Footnote ²¹¹ Historical thinking involves, in essence, viewing the past as mediated by the conceptual framework of the agents – a framework that also shapes the experience and understanding that agents may try to convey via testimony. In other words, historical thought rejects the immediacy of experience in favour of focusing on the conceptual mediation and renegotiation of experience.Footnote ²¹²

This philosophical elucidation is crucial for understanding the ways in which historical thinking relates to an individual historical agent’s testimony. It is not uncommon to believe that there is a direct conflict between the understanding offered in witness accounts and historical research. Wieviorka provides a good example of how this conflict has been framed:

Contrary to what I have argued, Wieviorka’s dilemma assumes that there is harmony between the historian’s and the witness’s interest in experience. The witness remembers something in a particular way, expressing an understanding tied to her very identity, which, in turn, gives the audience a view as to “what it was like.” This account is, allegedly, in conflict with the historian’s “quest for the truth.” But what, exactly, is the conflict about? Is it that the witness is not expressing genuinely “what it was like” for her? That seems absurd – that is, unless the historian is also the witness’s therapist. Thinking historically about the testimony, the historian’s main interest would, perhaps, be to investigate and explain the ways in which the character of the account relates to questions about the cultural and social conditions of the witness. However, this interest is not in conflict with appreciating the testimony as an expression of what the events felt like for the witness. The witness’s interests to tell what her experience was like and the historian’s interest to explain the character of experience inhabit completely different registers, making it difficult to understand what the “moral dilemma” is supposed to be. In fact, Wieviorka’s alleged dilemma clearly illustrates problems caused by the rapprochement found in recent historical theory. The erroneous supposition that understanding offered by witnesses is in the same game as history produces (false) dilemmas about who should yield and which understanding to accept.

As we have seen, distinguishing the historical past as a product of the historian’s questioning-activity offers a way of articulating the distinctiveness of historical thought beyond objectivism. The historical past is the product of a specific form of understanding premised on conceptually distinct points of interest. Why does this matter? First, it shows that abandoning objectivism – which is perceived as a key consequence of the linguistic turn – makes no difference for the conceptual distinction between history and testimony as different forms of understanding. Second, it provides a bulwark against the conceptual rapprochement between history and memory, as detailed in 4.2–4.3. Even if we agree that both historians and witnesses are subject to the linguistic, political and cultural frames of present preconceptions, then from this it does not follow that the very idea of historical understanding, somehow, approaches testimony and cultural memory. Arguably, no human activity can escape preconceptions of the present, but this predicament will in no way diminish the importance of distinguishing between different modes of relating to the past. On the contrary, abandoning objectivism only highlights the urgency of finding more pertinent conceptual tools for distinguishing testimony and history. In the following, I show how the distinction between historical and practical past matters for discussions about the historian’s ethical responsibility.

4.8 The Ethics of Historical and Practical Pasts

For many influential historical theorists, testimony is ethically important as political and moral messages about action for us in the present and the future. Contrary to professional history, representations of the past in testimony and fictive works, White argued, provide accounts in which past and present are fused with the aim of answering the question, “What should I do?”Footnote ²¹⁴ As Tozzi (developing White’s account) wrote, testimonies are valuable “not as a journey to the past but as an action in the present.”Footnote ²¹⁵ As such, the systematic collection of testimony in the postwar era was undoubtedly shaped by specific social and political purposes.Footnote ²¹⁶ However, the implicit or explicit claim in the ethical turn of historical theory is not empirical but conceptual. To be “ethical” in historical representation is to incorporate testimonial accounts that provide practical pasts for aiding action in the present and the future. On this score, Keith Jenkins has explicitly encouraged historians to “take up the cause of human emancipation” by abandoning “truth” in favour of accounts of the past that are politically useful, which means that “to be ‘ethical’ … perhaps signals … the possible end of a history ‘of a certain kind.’”Footnote ²¹⁷ One finds a similar juxtaposition of history as objective versus testimony as a source of the ethical in Assmann’s work:

The so-called ethical turn in historical theory is a movement away from historical pasts and toward practical pasts, and witness accounts are a popular vehicle for the latter.Footnote ²¹⁹ As a result, ethics in historical research seem to depend on the inclusion of accounts that are explicitly not the products of academic historiography and a relativization of the idea that historical understanding is qualitatively distinct from the understanding offered by witnesses. Becoming ethical apparently requires that historians forfeit their profession. This reduction of ethics to practical uses of the past is crucial for explicating White’s often quoted and rather puzzling denial of the ethical responsibility of historians qua historians: “I … deny that historians, in their current ‘professional’ capacity, possess the resources necessary for rendering ‘ethically responsible’ judgments on whatever it is we mean by ‘history.’”Footnote ²²⁰ This denial makes perfect sense on the premise that “ethics” is about the production of representations of the past that are practically useful, either as moral or political instruction for the present and the future. As White rightly emphasized, such practical representations are precisely what professional historians are not supposed to offer. In other words, historians can become “ethical” only by cutting and pasting accounts from practical pasts in testimony and fiction.

This line of reasoning assumes that the historian’s ethical commitments belong to the framework of the ethics of witnessing. The consequences are appositely expressed in Jenkins’ description of how historians may “ethically” relate to the catastrophes of the twentieth century. According to Jenkins, “the historian’s ethical responsibility” is to refrain from representing the horrors, which can only “drown out the screams of its victims,” and instead let the past speak “for itself” – in both the speech and the silences of the victims – in order “to ensure that the Holocaust/Auschwitz forever haunts us.”Footnote ²²¹ Perhaps unwittingly, Jenkins uses the exact same words that are typically used for describing the ethical responsibility of victim witnesses. Survivors of the Holocaust were often assigned the “duty to remember” as a moral obligation.Footnote ²²² Furthermore, considering the fact that “Nie Wieder – Never Again!” is perhaps the most prominent leitmotif in all Western postwar memory politics, it seems that the historian’s ethical responsibility amounts to nothing more than the propagation of cosmopolitan memory culture.Footnote ²²³

The untenability of this approach becomes apparent when one moves from abstract historical theory towards empirical cases of memory accounts. Doing actual research, the historian will face the question of whose testimony to cut and paste for becoming “ethical.” By what standards should historians judge whether a memory account does indeed support “emancipation”? The answer cannot simply be that historians should favour testimony from “the victims” – at least not since victimization became a popular method for doing politics. As a result, historical research reduces to politics, and ethics tantamount to the historian’s judgment concerning just and unjust testimonies. It would be wrong to assume that this position is peculiar to radical historical theorists such as White or Jenkins. Quite the contrary, the very same logic is present in the framework of narratives and counternarratives, where “history” is often described as the official narrative of hegemony and the nation-state, while memory accounts are styled as progressive counternarratives speaking truth to power. Especially in extremely contentious cases, such as the Israel–Palestine conflict or the history of indigenous peoples, it is commonplace to juxtapose history as the story of people in power against memory accounts as the story of the subjugated.Footnote ²²⁴ To consider history as always partisan is not problematic. As Collingwood’s lecture on impartiality articulated, political partisanship is indeed compatible with high scholarly standards in research. But if history is not to be reduced to ideology confirmation, then those scholarly standards cannot themselves be relative to the same partisanship.Footnote ²²⁵

Attending to testimony in relation to ethical concerns about “what to do” in the present and future is important.Footnote ²²⁶ Many labels are already in use for this kind of relation to the past. It is called memory, remembrance, commemoration, testimony – the practical past – all of which are possibly important concerns for the general health and stability of democratic societies.Footnote ²²⁷ But adding “history” to that list invites conceptual confusion. Furthermore, if one disagrees with reducing ethics in history to choosing the most progressive political partisanship, one must ask: What is the ethics of history qua history? Any sufficient answer to that question must both, (i) avoid outsourcing ethics to practical pasts, and (ii) acknowledge the distinguishing features of historical thinking that this Element has identified. While the key feature of practical relations to the past via testimony is their (perceived) quality of immediacy – as authenticity, experience and expression – the distinct feature of historical relations is to understand past experience as mediated by the conceptual framework of the agents, which is further investigated by way of the analytical concepts of the historian. The ethics of history qua history is about the ethics of understanding experience as conceptually mediated by both past and present historical agents.

The concept of historical past, as articulated in this Element, entails that there are two distinct but integrally connected ethical aspects of historical relations to the past.Footnote ²²⁸ First, in our attempts historically to understand experience as conceptually mediated by the perspective of the historical agents, there is a challenge to do justice to what that past perspective was. It belongs to the historian’s ethical responsibility to preserve the integrity of the historical past – without equating “the historical past” with the historical agent’s own understanding of their situation. Furthermore, the task of explaining others’ experiences demands scrutiny of the historian’s self-understanding. As the philosopher Peter Winch has emphasized, the task of historical understanding is neither to be seen as a process in which actions and utterances of others are subsumed under our own conditions of intelligibility, nor as a process in which the historian simply adopts the historical agent’s understanding. Rather, according to Winch, “seriously to study another way [of looking at things] is necessarily to seek to extend our own – not simply to bring the other way within the already existing boundaries of our own.”Footnote ²²⁹

The past-related ethical responsibility of historians is the demand for self-scrutiny. This is an ethical form of responsibility to the extent that historical agents may be wronged if we force the meaning of their words, actions and lives to fit the frames of our own preconceptions. Consequently, historians have the ethical responsibility to track how their own self-understanding enters their explanation of the experience and understanding expressed in testimony. This responsibility is distinctly towards the past, and not the present, since the ethical demand is to do justice to the lived world of historical agents. Dead people are as wronged by lies and projection as living people are. This ethical demand does not depend on the psychology of individual historians – whether they happen to “feel” any such responsibility towards the past or not. Rather, the (logical) possibility of being subject to responsibility towards the past follows from the normative implication of historical description.Footnote ²³⁰

There is also a second, present-related aspect of the ethics of history qua history. Considering that the contemporary world is dominated by practical relations to the past, the ethical responsibility of the historian is to critically examine the conditions of production of “voices of experience” – especially those that are rallying support for social and political action. History is endowed with the task of resisting the immediacy of “historical” experience in the present. The reason is not, of course, that listening to expressions of experience via testimony is unimportant. Rather, the reason is that the overwhelming power of complementary authenticities (via audiovisual testimony and artefacts) may, in fact, as Jones has pointed out, make us less disposed to thinking critically and deeply about the political, social and cultural structures that brought about the situation of violence in the first place – which is exactly what historically understanding the conditions of a certain experience may enable us to appreciate.Footnote ²³¹ Furthermore, testimony is constantly at the mercy of manipulative forms of remembrance in which victims end up being heard only as instruments for promoting concerns about one’s own distress in the present.Footnote ²³² Contrary to practical uses of the past, the task of history is not to call on ready-made images of the past in order to promote action in the present. On the contrary, the task is to critically examine the conditions in which the ready-made images of the past that populate the public sphere were produced.Footnote ²³³

The present-related task of history is ethical in the sense that critical examination of “experience” is a prerequisite for every just and reflective relation to conflicting collective pasts. The ethics of history is a form of critical cultural self-knowledge in which our understanding of the present is explicated through contrasts with the historical past. By explaining the conceptual mediation of the experience of historical agents, one is inevitably also articulating differences in relation to one’s own conceptual horizon.Footnote ²³⁴ As many historical theorists have argued, cultural memory is always-already part of our present understanding.Footnote ²³⁵ For this very reason, history is needed as the critical method for bringing to light the contours, conflicts and raptures of contemporary memory cultures. If one is not to blindly accept the established hierarchy of victim witnesses that memory cultures offer (that is, which witnesses should be heard first and last), then any ethically responsible decision about “what to do” presupposes the critical and reflective relation to the past that belongs to historical research. Historical thinking provides a form of understanding that witnesses qua witnesses can never give us.