(p.421) Appendix One Tracing Sources: Toward Clarifying the Criteria for Detecting Sources
(p.421) Appendix One Tracing Sources: Toward Clarifying the Criteria for Detecting Sources
Abstract and Keywords
As a general literary principle, it is not possible to take a finished text and reconstruct sources that otherwise have never been seen – sources that are hypothetical. What is possible, though often difficult, is to identify a known text as one of the sources of another known text. Ancient writing consisted largely of rewriting, so the use of written sources was normal, and quite often both a text and some of its sources still exist. Modes of rewriting varied hugely, for instance: Midrash, Synthesis and transformation; Rewritten Bible; Rewritten Prophecy (the Pesharim); Rewritten Epic; and Imitation (mimesis). Criteria of dependence are external plausibility; significant similarities; the intelligibility of the differences.
As a general literary principle, it is not possible to take a finished text and reconstruct diverse sources that otherwise have never been seen—sources that are hypothetical. Those who discuss source criticism in literature at large (Morize, 1922, 82–131; Saunders, 1952, 162–191; Wellek and Warren, 1962, 257–258; Altick and Fenstermaker, 1993, 106–119; cf. Bateson, 1972, 192) never envisage the feasibility of reinventing lost sources.
What is possible, though it is often difficult (see the above‐mentioned authors), is to identify a known text as one of the sources of another known text. It is possible, for instance, to demonstrate that Virgil used Homer, that Chronicles used Genesis‐Kings, and, for most scholars, that Matthew used Mark.
Consequently the invoking of unknown documents—such as JEDP for the Pentateuch—is, at best, a last resort, to be undertaken only if there is no connection with known documents and if there are special indications that the hypothesis is working coherently and successfully.
The brittleness of hypotheses invoking unseen documents is illustrated in the nineteenth‐century effort to distinguish sources in the Epic of Gilgamesh: the only success depended not on an internal analysis of the Epic but on comparison with other, known, texts (Berlin, 1983, 132–134; cf. Tigay, 1982; Moran, 1995).
The primary purpose of this appendix is to clarify the criteria for claiming that one known text depends on another.
In the case of Genesis the situation may seem discouraging. For a long time the book looked unique—unrelated to other literature—and so there was no question of trying to trace its dependence on known writings. But now, research has connected Genesis with several kinds of literature, especially with historiography, epic, and prophecy, and so there is the prospect of finding specific texts used by the author of Genesis.
However, before embarking on this task, it is useful to stand back and look at the larger picture of using sources. The discussion is threefold: (p.422)
Writing as rewriting: ancient writing as a culture of preserving and rewriting.
The present gradual recognition of diverse categories of rewriting.
The clarifying of criteria for recognizing sources.
Writing as Rewriting: Ancient Writing as a Culture of Preservation—Of Imitating and Emulating
Almost the entire history of writing has been accompanied by a huge instinct for preservation. Whatever was written well was respected and, in some way, guarded. The Jewish care for used manuscripts—preserving them in a geniza—was but one symptom of a larger protective attitude.
The full reasons for this attitude are not clear. Apparently the whole apparatus of writing was regarded as rare and fragile. The craft was specialized, the manuscripts precious.
Later, however, the attitude changed. Soon after the printing press arrived, and especially after the mass production of encyclopedias, the emphasis on preservation began to wane (Ong, 1971, 277–279). The essential fact is this: until the eighteenth century—when the encyclopedias were multiplied—it was of the essence of writing that one engaged the work of previous writers. The Akkadians and Babylonians reshaped the work of the Sumerians, the early Greeks borrowed from Mesopotamia, the later Greeks rewrote the earlier Greeks, the Romans rewrote the Greeks, and the Renaissance reshaped the whole classical heritage (Burkert, 1992; Brodie, 1984, 17–19; Highet, 1949). In the Greco‐Roman tradition the process of rewriting and reshaping was refined by the doctrine and practice of imitation (mimēsis, imitatio; Abrams, 1953, 8–14, 30–47; Boyd, 1968; Brodie, 1984, 19–32).
In the eighteenth century, however, as the dawning of the romantic era glorified personal experience, the emphasis on imitation gave way to the promotion of originality. The change was articulated in 1759, in Edward Young's “Conjectures on Original Composition” (Kaplan, 1975, 220–250; Abrams, 1953, 198–203). In 1762, Rousseau attacked imitation in painting; artists should copy only nature (Webb, 1981,157). The centrality of the old writings began to fade (Boyd, 1968, 98–117). Then, toward the end of the twentieth century there was a further development: the old writings were not only allowed to fade; they were often explicitly condemned. They began to be seen, at least by some, not as classic but as oppressive.
What counts here is the basic fact of changing attitudes. The present freedom concerning former writings is not typical of ancient times. On the contrary, whatever was written well was preserved, in some form. The trouble is, the form varied, almost endlessly.
This phenomenon—limitless variation—is the heart of the problem. There was no one set form for preserving what was already written. Nor was there a set (p.423) of forms—a set of, say, three or four basic ways of reshaping a text. Morize (1922, 96–127), discussing French literature, lists seven kinds of sources but emphasizes (p. 127) “the impossibility of predicting every type of case that may arise.” On the contrary, there were no limits. At one extreme, one could rewrite the subtext (the original) almost word for word; at the other extreme, one could distill it to its essence and stand it on its head, mingling it with other sources. Thus at times it was easy to recognize the source(s); at other times, extremely difficult.
The Lack of Well‐Developed Criteria
One would imagine, given the central difficulty in recognizing sources, and given the modern deluge of dissertations and publications, that there would be numerous systematic studies for dealing with the problem, numerous guides setting out comprehensive criteria for establishing literary dependence.
Apparently there is none. At least, the present writer, despite prolonged searching in good libraries, has failed to locate one. Part of the search involved a telephone conversation with George Steiner (December 21st, 1979). Steiner said that when writing After Babel he looked for some such comprehensive survey but could not find one. Morize (1922, 128) envisaged a book of basic criteria and even gave the hypothetical book a title—Practical Handbook for the Investigation of Sources—but he was ambiguous toward the project: he dismissed the very idea, yet proceeded to give an inadequate substitute for it (1922, 128–130). Homeric scholar Dennis R. McDonald, speaking at the SBL (New Orleans, Nov. 25, 1996), said he had worked out five criteria for judging literary dependence, but he indicated, cautiously, that these five needed refinement.
To some degree the lack of a basic guide is understandable. Tracking literary dependence is like detective work; it needs instinct, and instinct cannot be learned, at least not in an organized way.
However, like detective work, literary tracking needs much more than instinct; it needs a huge proportion of methodical plodding. (Jonathan Z. Smith, in discussing the larger question of comparing religions, speaks of drudgery: Drudgery Divine: On the Comparison of Early Christianties and the Religions of Late Antiquity, 1990). Unlike instinct, drudgery can be organized. Detectives can learn the rules for methodical plodding. Yet when one looks for general guides to methodical plodding in tracing literary sources, there are none. There are indeed numerous specialized studies, usually about the most obvious cases—Babylon and Sumer; Chronicles and Kings; Matthew and Mark; Virgil and Homer; Seneca and Euripides; Racine and Seneca. There are books about writing in general, and about vast literary traditions, and about the whole phenomenon of literature, but there is no book that tries in a systematic way to identify both the dynamics between the texts and the criteria for identifying them.
The reason for this omission seems to lie in the modern antipathy to imitation—not only to practicing it but even to studying it. Until 1996 the Oxford Classical Dictionary had no entry for it, though it was foundational for the (p.424) classics, and the Princeton Encyclopedia of Poetry and Poetics (1975, ed. Preminger, 378–380) begins its discussion of imitation by noting that it is only now that the term is coming back after its banishment in the nineteenth century.
Also coming into use now is the term “intertextuality.” This can be misleading: the term refers primarily not to a text's dependence on another specific text but to its larger dependence on a whole cultural context (on the history and meaning of the term “intertextuality,” see Culler, 1981, 100–118; on Julia Kristeva, originator of the term, see Roudiez, 1980, 1–20; Lechte, 1994, 141–144). Intertextuality “has nothing to do with matters of influence by one writer on another, or with the sources of a literary work; it does, on the other hand, involve the components of a textual system such as the novel, for instance. It is defined . . . as the transposition of one or more systems of signs upon another, accompanied by a new articulation of the enunciative and denotative position” (Roudiez, 1980, 15). In discussing literary dependence therefore it is generally better not to use this broad term—to do so is like hitting a tack with a sledge‐hammer—but the term has been useful insofar as it has helped to alert researchers to a specific practice which, in ancient composition, was central.
In the case of Genesis, therefore, the situation is difficult. There is an emerging relationship to world literature and to the prophetic books, but in the tracing of possible dependence there are no agreed procedures or criteria.
The Tyranny of the Verbatim Criterion
The difficulty caused by the absence of wide‐ranging criteria is compounded by the obvious presence of one simple criterion: verbatim similarity. When modern studies neglected rewriting as a whole, verbatim similarity was sometimes so glaring that it could not be ignored, and so it led to the recognition of literary dependence. Chronicles was recognized as using Kings, Matthew as using Mark. Then, imperceptibly, the criterion of verbatim similarity became exclusive; in effect, it established a kind of tyranny. Other criteria were often not considered. “Literary dependence” became equated with just one model. The Matthew‐Mark model, standing at the entrance to the New Testament, sometimes sets the tone, even for scholars of the Old Testament.
The difficulty is illustrated in gospel studies: two of the inter‐gospel connections (Matthew‐Mark and Luke‐Mark) are so obvious that they make the third (John‐Mark) look strange. And so some scholars reject the third; implicitly using the obvious Matthew‐Mark model as a criterion of what is meant by literary dependence, they reject the idea of literary dependence between John and Mark. In a sense they are right; there is no literary dependence of the Matthew‐Mark kind. But there is another more complex form of literary dependence—a form for which there is abundant evidence (Brodie, 1993a).
The John‐Mark difficulty is a salutary warning. If biblical scholarship cannot bridge the short gap between John and Mark there is little chance that it will ever bridge the gap between Genesis and its extant sources.
(p.425) The Present Gradual Recognition of Diverse Categories of Rewriting: Transition from the Tyranny of the Verbatim Model
Since about 1950 there has been a gradual discovery of diverse categories of rewriting. To some degree these categories were already known, especially to Jewish and classical scholars, but now they have come closer to the forefront of research. Such discoveries have the implicit effect of moving the discussion of rewriting beyond the level of verbatim similarity. The process of discovery is still incomplete, but its existence and its gradual expansion establish a context that is conducive to discovering the categories possibly involved in composing Genesis.
The following categories of rewriting are particularly noteworthy.
“Midrash” means literally the “searching” of a text—searching its meaning and applicability to a new situation. In general, the searching or application of a text can be either explicit or implicit (overt or covert, quoted or merely implied), and there is a debate about whether the technical term “midrash” should be used of both forms (Le Déaut, 1971; Porton, 1979). But whatever the terminology, the reality remains essentially the same: in traditional Judaism, and within the biblical books themselves, there was a massive phenomenon of adapting existing texts. Despite its centrality this widespread practice remained largely unexplored (“domaine encore à peu près complètement inexploré”—Bloch, 1957). The exploration is still continuing (e.g., Porton, 1985; Hartman and Budick, 1986).
Synthesis and Transformation
Among recent studies of the inner‐biblical reworking of texts, Fishbane's work (1985) deserves special mention. Under the general heading of “inner‐biblical hermeneutics” he has emphasized the ideas of synthesis and transformation, and has linked prophetic texts with texts from the Pentateuch. Thus, in principle, his work opens the road to bridging the gap between Genesis and the prophetical books. But Fishbane's work, though it constitutes a breakthrough, leaves the breakthrough undeveloped. The emphasis is on a limited range of texts, and in particular there is no general discussion concerning the direction of the flow of influence.
The term “rewritten Bible” has been used to describe the way early noncanonical books—such as Jubilees, Assumption of Moses, Pseudo‐Philo's Biblical Antiquities—rewrite (p.426) the biblical texts, especially narrative texts. The process involves “a free rewriting of part . . . of Israel's sacred history” (Harrington, 1986, 242). This process is late (around the turn of the era) and emphasizes narrative (rather than prophecy), but its underlying concept—rewriting the sacred, freely—provides a further sense of the possibilities in reworking the sacred text.
Rewritten Prophecy: The Pesharim
Fifteen Qumran texts consist of pesharim, “interpretations”—systematic adaptations of prophetic texts. They quote the prophets explicitly—quite unlike the Pentateuch—but the freedom with which they adapt the text suggests possibilities for how the Pentateuch may have done so. The prophet's words about the Babylonians, for instance, are interpreted as referring to the Romans—a leap across time, space, culture, and diverse circumstances. The implied hermeneutical principles are diverse; Brownlee (1951, 60–62) counts thirteen, but Horgan (1979, 244–245) reduces them to four.
The Homeric epics, from the time of their completion in writing (ca. 700 BCE), were copied and variously adapted. Such adaptations of Homer, begun apparently in his younger contemporary Hesiod (O'Brien and Major, 1982, 51), flourished in fifth‐century drama, and remained a central feature of much writing in Greece and beyond. The Odyssey, in particular, was adapted repeatedly in multiple ways and places (Stanford, 1963)—including the early church (MacDonald, 1994, 17–34) and Joyce's Ulysses. Significantly, Homeric epic underlay the development of historiography (Lesky, 1966, 218–221).
Imitation (mimēsis), including Literary Imitation/emulation
In contrast to the modern search for originality, the ethos of much of the ancient world emphasized imitation. Nature itself was an imitation of a higher world (Plato, Timaeus 29); and all art—including literature—imitated nature (Aristotle, e.g., Physics, II, 2.194a22; II, 8.199a15–17). Literature, therefore, may be described as mimēsis, as mirroring the external world (cf. Auerbach, 1953). In Trible's words : “The most ancient and persistent [literary] approach holds that literature mirrors a world external to itself. The Greek word mimesis (imitation) denotes the concept” (1994, 10).
Literary mimēsis, however, was not limited to imitating nature; it also imitated other literature. Other literature provided a key to nature, an initial insight. There was no need, as with the romantics of the eighteenth and nineteenth centuries, to reinvent nature or the world every time one took up a pen. (p.427) The ancient writer worked with what other writers had already captured and expressed.
Imitation, however, was not slavish; verbatim imitation was unusual; the challenge was to combine imitation with emulation (zēlos)—in other words, simultaneously to absorb and surpass previous writers (Brodie, 1984, 19–26). This process of combining imitation and emulation applied also to historiography (ibid., 26–32).
The overall impression, from these emerging categories, is of a vast field of possibilities. As Steiner said in a broader context: “The [writer] need not cite [the] source‐text . . . [but] can treat it in a limitless variety of perspectives. . . from interlinear translation . . . to the faintest most arcane of allusions. . . . It is up to us to recognize and reconstruct the particular force of relation” (1975, 424–425).
That is the problem: it is up to us. If that is so, if we are to undertake this task, then it is first necessary to establish criteria.
Amid the Emerging Categories of Rewriting: Toward Establishing Criteria for Judging Literary Dependence
Most researchers, no matter how open‐minded, are not in the habit of dealing with complicated claims to literary dependence; and without long practice, it is difficult to judge such claims. It is particularly difficult to do so at speed—for instance, when reviewing books. It is tempting, very often, to reject such dependence or simply ignore it.
The burden for establishing criteria lies with the one making the claim. If that person does not indicate the criteria—often the case—there is little room for complaint when readers or reviewers do not engage the evidence.
To a limited degree other researchers have already suggested brief summaries of procedures and/or criteria (Morize, 1922, 87–96, 128–130; Saunders, 1952, 167–172; Altick and Fenstermaker, 1993, 108–118; Wellek and Warren, 1962, 258; Feldman, 1996, 14). The purpose here is to develop these suggestions.
The task is not easy—one is dealing with art not science, and at times the criteria or the ways of applying them are inadequate. But a beginning can be made.
Criteria are of two kinds—positive (positive indications) and negative (explanations that mislead).
There are three main kinds of indications that one text depends on another—external plausibility, internal similarities, and the intelligibility of the differences.
(p.428) External Plausibility (Context)
There is no point in trying to show that Genesis depends on Ovid's epic poem Metamorphoses (completed in exile, at Tomis, on the Black Sea). It is true that the Black Sea borders Asia Minor, and so the text would probably have been physically accessible. But without even glancing at Ovid's work it can be ruled out on the basis of an external factor—date (composed ca. 10 CE). Dependence can be invoked only if external factors make such dependence plausible. The Epic of Gilgamesh, in contrast, would not only have been accessible—versions have been found both in Asia Minor (in the Hittite capital, Hattusha) and in Babylonia—but would also have been old enough; those versions date from the second millenium (ANET 72–73).
Another external factor—already mentioned—is the pervading attitude toward previous writings. If the general attitude to preceding writings was one of independence or even hostility, then dependence would be implausible. If, for instance, the Gilgamesh epic, originally Akkadian, had been banned or ignored by others—by the Hittites, Hurrians, Assyrians, and Babylonians—then the idea of its use in Genesis would be correspondingly implausible. But the opposite was the case; the Epic of Gilgamesh was popular and accessible (Feldmman, 1996, 13).
Similarities are of diverse kinds, ranging from vague, isolated, or insignificant to crucial and decisive.
Similarity of theme. Similarity of theme can be an initial clue to dependence between writings. For instance, some of Shakespeare's tragic themes—such as the power of evil forces over life—can trigger awareness of his possible use of Seneca's tragic dramas. In the case of Gilgamesh and Genesis there is a shared emphasis on the theme of journeying. Such sharing of themes usually proves nothing—themes tend to be too general for drawing conclusions—but it can set the stage for a more probing investigation.
Pivotal leads or clues. Even if the relationship between two texts is complex or obscure the author may give some key indication of a link between the two. When Luke, for instance, uses the Elijah‐Elisha narrative, he provides initial clues to what he is doing: he invokes the example of Elijah and/or Elisha in two leading texts—the opening scene (Luke 1:5–25; see 1: 17) and the inaugural speech (Luke 4:16–27). To some degree, the same thing happens in Genesis: whether intended or not, some episodes, especially the flood story, are so strikingly similar to well‐known epics that they act as a leading indication that perhaps Genesis as a whole is using further epic material.
Action/plot. Similarity of action can be a strong clue to literary dependence. The clearest example here is the Atrahasis story, which “displays the same basic plot as Gen. 2–9” (Clifford, 2:2; for more detail, see Tsumura, 1994a, 46–47). In the case of the Gilgamesh epic and the Enuma Elish the evidence is not as easy, but it is significant.
Completeness. If similarities are confined to an isolated passage, or if only some passages from a possible source appear to be reflected in the finished writing, then a problem arises about the nature of the relationship between the texts. Why should some be missing? And does this absence cast doubt on the relationship to the text as a whole? (A classic problem is the impression, at first sight, that Luke omits almost two chapters of Mark.) But if all the passages of the possible source are reflected in some form in the final text then the case for direct dependence is greatly strengthened.
Order. When random elements occur in two documents in the same order the similarity requires explanation. Similarity of order does not occur easily. If two people, independently of each other, arrange the numbers 1 to 5 at random, the chance that they will arrange them in the same order is less than one in a hundred. If the numbers are 1 to 10 the chance is less than one in a million.
Telltale details (linguistic or otherwise). The link between two texts is sometimes given away by a small detail. If the detail remains isolated, it generally proves nothing; but if a series of details emerge then they become more significant—especially if clustered together or in the same order.
Complex coherence. The relationship between some texts is so complex and coherent that the only plausible explanation is literary dependence. There is a degree of complexity which oral tradition cannot handle. As Ong says in a related context, “closer plotting requires writing” (1977, 254).
The Intelligibility of the Differences
In judging possible questions of dependence, the issue is not whether there are differences but whether the differences are intelligible. If two people, one black and one white, have a child, the child will be different from both. The differences may be great, but they do not exclude dependence; in the circumstance, they are intelligible.
Likewise with two texts. For instance, Jesus’ refusal to call down destructive fire from heaven (Luke 9:54–55) is in direct contrast to Elijah's killing of more than one hundred soldiers (2 Kings 1), but the difference fits with Luke's wider portrayal of Jesus. Likewise with the other war‐like incidents in the Old Testament: they have been transformed to accord with a different strategy—a focus not on a specific land but on the kingdom of God. What counts, then, is not (p.430) difference, but intelligibility. Genesis is massively different from Herodotus, and even more so from the prophets and the Odyssey. But differences, no matter how great, do not clinch the issue. Some writers want to contradict an earlier text at every level. What counts is whether the similarities are significant and whether the differences are intelligible.
Negative Criteria: Principles That Can Mislead
In assessing evidence about literary dependence there are a number of factors that sometimes mislead or cause confusion.
Some Connections Are Weak
Total analysis is virtually impossible, and so in almost every comparative analysis of two texts there are points where the comparison is uncertain, where the connections are weak—obscure, questionable, apparently nonexistent. If the discussion focuses on these weak elements, then confusion follows—like judging the strength of a country's defenses by its babies. Insistence on what is weak, whether by the one presenting the comparison or by someone questioning it, obscures the decisive issue: are there arguments that are strong? One is dealing not with a single chain where the overall strength depends on the weakest link, but with a whole series of chains. That some are weak does not matter as long as there are enough that are strong.
Weaknesses occur in an analysis not only because total analysis is almost impossible, but also because the analyst makes mistakes or is inexperienced. Again, however, the issue is not whether there are mistakes but whether there is enough that is true to make the overall connection credible, or at least to make it worthy of further study. As in detective work, what the investigator needs is the ability to sift through evidence that is weak or misleading and to see whether there are a few good leads, or at least one.
The Differences Are Too Great
This question has been mentioned already but it bears repeating for it is a matter of great confusion. The differences between texts may be misleading; they may give the false impression that one text cannot possibly depend on the other. But differences do not decide the issue. The purpose of writing—as distinct from copying—is to say something that is in some way different. Difference, therefore, is of the essence of writing.
There is no limit on how different a writer may be from predecessors. Even the most sacred text can be rewritten or reversed: “It was said to you of old. . . . But I say. . . . ”As already mentioned, the books called “the rewritten Bible” contain “a free rewriting of [a] part . . . of Israel's sacred history” (Harrington, 1986, 242).
(p.431) The issue, therefore, when comparing texts is not whether there are differences, but whether, as already indicated, the differences are intelligible. In comparing Ezekiel with John 10, for instance, Bultmann (1971, 367) points to the dissimilarities, and then decides against dependence. But for Brown these dissimilarities are not decisive; the issue rather is “whether there is sufficient similarity to suggest that the OT supplied the raw material for . . . [a] creative reinterpretation” (1966, 397).
Creative reinterpretation—that is central to the discussion. It is creative reinterpretation that first causes the differences and it is the concept of creative reinterpretation that subsequently makes them intelligible.
The Differences Preclude Dependence and Point Rather to a Shared Tradition
There are times when, even if two writings show clear similarity and difference, direct dependence may not in fact be the explanation. Perhaps, instead of borrowing directly, both writings used a shared source—and interpreted it differently.
This is possible, and cannot be directly disproved. But very often it is a gratuitous claim and cannot be proved. And since it bears the burden of proof—it claims documents that no one has ever seen or sources that are unverifiable—it is in the weaker position.
One of the results of this claim to a third entity is that it avoids the phenomenon of direct dependence and thereby avoids dealing directly with the problem—the difference between the two documents. At a superficial level it makes the problem easier. The difference, with all its difficulty and richness, disappears under the cover of a third unproven element. The moment in which the difference occurred is pushed into an inaccessible background.
This removal of the difficulty can create the illusion that the difficulty has been solved. But it has not. To avoid saying A changed B, this theory says A and B both used C (A changed C, and possibly B changed C as well). But why should these hypothetical changes be more plausible than that of A changing B? The problem will not go away: at some stage someone changed a source significantly. If that change is accepted in principle—in fact it is unavoidable—then why not accept it immediately? The simplest hypothesis that accounts for the data is to say A changed B. The old adage applies: “Things should not be multiplied without necessity.”
The Similarity May Be Due to General Familiarity Rather Than Direct Literary Dependence
If someone absorbs the works of, say, Pascal or Dickens, it is likely that at a later stage the thoughts and phrases of these authors will reappear in the person's speech and writing. Much of the process is unconscious and it is not a (p.432) question of direct literary dependence. It may seem therefore to follow that identifying direct literary dependence is almost impossible.
The conclusion that follows, however, is not that identifying literary dependence is impossible but that the researcher needs to be doubly careful. The positive criteria, given above, need to be applied rigorously, checking not only for broad similarities, for style and for occasional phrases and details, but for the whole range of factors that might indicate direct dependence—from external plausibility to a sequence of details.
A Source Is Already Present, Therefore a Second One Cannot Be Admitted
Having established the use of a major source there may be a tendency to deny the presence of other sources. But a text may be a complex synthesis; “a poem or a page of prose is often a sort of mosaic” (Morize, 1922, 107). “Text” is derived from texere, “to weave” and so is at home with interweaving. The presence of one source does not exclude another.
The ancient method of composing was unlike the modern. Today's writer, in using sources, often has a basic choice—acknowledge or ignore. In this modern view, knowledge and writing are like personal property. But for the ancient writer, what was written was not personal property; it was so to speak, out there, and was not to be ignored. One built on previous writings and did so without acknowledgement. The issue therefore concerning the use of previous writings was not whether but how—how to take what was best in existing writings and how to reshape them to one's own vision.
There were no limits on the how—on the ways of reshaping an older document. Consequently, the ways in which some sources were used will probably forever escape detection, even when those source documents are extant. This is especially so in the case of minor sources.
But other sources have played a greater role, and, even if radically rethought and reshaped, can be detected. The criteria outlined above are not complete but, in this search for sources, they provide an initial orientation. The task is difficult, but not out of reach. At least it is possible, building on the work of others, to give a partial framework.