3. The Romantic Paradigm: Historical Narrative

The historian organizes historical essences. The data of the past are the mass that is given a form by the historian, by empathy. Because of this, history is determined by the principles of empathy and organization, and as long as these principles are not yet there, historical artifacts, in the proper sense of the word, are not possible, and we discover only the traces of incidental acts of empathy where the unguided mind has exerted its influence.^[10] (my translation)

This statement was made by Novalis in 1798, exactly fifty years after Hume had made the statement quoted at the beginning of the previous section. Here we have entered a new and completely different world. The historian is no longer required to discover and to express (general) knowledge, but to organize it. The language used by the historian is no longer seen as a passive and immutable medium but as a Proteus adapting itself to the circumstances each time an individual historian depicts or pictures part of the past.

The general statement can be seen as shorthand for an (infinite) number of singular statements which are comparable in such a way that generalization becomes possible. Whether we are concerned with theoretical, inductive, or empirical statements is of no importance in this connection. As is the case with the general statement, a great number of singular statements "go into" historical narrative. The difference is, however, that in the

case of a historical narrative, the number of the statements is always finite; this number can be ascertained with absolute precision and, moreover, as far as their content is concerned, the statements of a historical narrative have no systematic similarities. If such similarities happen to exist, this is purely coincidental. These considerations already suggest that the singular statement is a kind of intermediary between the general statement and the narrative. If the singular statement describes or refers to a recurrent state of affairs (this may be due to the way the statement is formulated), it will have more affinity with the general statement; if not, it is a natural part of a narrative. We may draw the conclusion that the really interesting contrast is not, as is ordinarily believed, the contrast between the general and the singular statement, but between the general statement and the historical narrative. Here language is used for two quite different purposes—as will be shown below. The singular statement may serve two masters and is therefore, in a certain sense, essentially incomplete or unsaturated.

So let us concentrate upon the narrative instead of upon the singular statement. We can immediately discover an interesting asymmetry between the general statement and the narrative. The general statement is a generalization of a singular statement and can be obtained from the singular statement by means of a simple formal operation. The relation between the general statement and the singular statement is a formal and deductive one. The singular statement individuates the general statement. But historical narrative, consisting of a large number of different singular statements, can only be individuated by taking each of them into account. The number of singular statements tacitly referred to by the general statement is infinite, and yet only one is sufficient for defining the general statement and vice versa. However, the number of singular statements contained within a historical narrative is finite, and yet all of them have to be considered for individuating the specific narrative told by the historian. Or: there is a reversal in the relation between the singular statement and the general statement on the one hand and historical narrative on the other; due to the similarity of the general and the singular statement, we may say that the general statement defines the singular statement, whereas the singular statement defines the identity of the historian's narrative. Apart from the same element of formalization present in all general statements—thus not characteristic of each of them—no novel element is introduced when we go from the singular statement to the general statement; however, each time language is used narratively something new and unique will be created.

However, we can also claim a parallelism in the differences between the general statement and narrative. The general statement suggests the generalized, interchangeable knowing subject, and narrative the individual historian. Being heir to Romanticism, the individual historian has been

ejected from a sociohistorical reality shared by us all: each individual historian inhabits a sociohistorical "house" different from those of his fellow historians. There will be a systematic disparity between what one historian says or thinks about sociohistorical reality and the opinions of other historians regarding it. Each attempt to define (part of) historical reality may satisfy some historians but never all of them. In other words, the link between language (i.e., narrative) and reality can never be fixed in a way acceptable to all historians, thus becoming the knowledge of a generalized knowing subject. The fact that debate and discussion have a much more prominent place in historiography than in other disciplines and that historiographical debate rarely, if ever, results in conceptions shared once and for all by all historians should not be seen as a sad deficiency of historiography that has to be remedied, but as a necessary consequence of the linguistic instruments used by the historian.

All this may be illustrated by means of typical historical concepts like the Renaissance or the Cold War. As I have pointed out elsewhere, such concepts do not refer to historical reality itself but to narrative interpretations of the past.^[11] The term the Renaissance refers to a narrative interpretation and does not refer to historical reality, although the statements contained within the historian's narrative do so. It is therefore not surprising that the connotations of terms like the Renaissance are subject to continuous change. To require that a specific definition of the Renaissance should be accepted from now on by all historians would mean the immediate end of an important and interesting historiographical discussion. Therefore, words like the Renaissance or the Cold War show us that in historiography there is a systematic looseness or indeterminacy in the relation between language and reality. And this indeterminacy does not reflect some sorry state of affairs in historiography that has to be overcome at all costs, but is the condition for the very possibility of modern historiography.

To sum up once more, in historiography language is no longer a passive medium like the paperweight or a mirror, but it makes its presence felt in a way that cannot possibly be ignored. In history, language acquires a substantiality of its own; indeed—as we shall see below—historical narrative is a thing in the proper sense of the word. Narrative language does not have the transparency of the (social) scientist's language, but irresistibly draws the attention of the reader to itself. Due to this opacity, narrative language resembles the well-chosen word: in both cases, we can admire the linguistic instruments that have been used in speaking about reality, and in both cases the use of language has no other goal than to achieve this effect. This may also serve to justify the prominent role played by

stylistic considerations in historiography. In historiography, style is not just mere adornment but touches the essence of what the historian wants to convey. Gay was correct in saying that style concerns not only the manner but also the matter of historical discourse.^[12]

However, if narrative has a substantiality of its own, if it is itself a thing like the phenomena of the past described in it, no epistemological ties can be conceived of to fasten narrative to historical reality. There are no epistemological ties between things, only between things and language. We may now wonder what makes historiography still a profitable occupation—if it is one. What could a discipline without an epistemology possibly look like?