SOURCE: “Techniques of Quotation in Clement of Alexandria: A View of Ancient Literary Working Methods,” in Vigiliae Christianae, Vol. 50, No. 3, 1996, pp. 223-43.

[In the following excerpt, van den Hoek examines Clement's use of borrowed and quoted passages, including his accuracy, his method of giving credit, and his characteristic way of incorporating the material.]

INTRODUCTION

Borrowed material embedded in the flow of a writer's text is a common phenomenon in Antiquity. Since Clement's writings have so many borrowings, his case is of almost emblematic significance for this aspect of ancient literary technique. The problem has many facets. The way he accumulated his borrowed material deserves attention, and there the testimony of other ancient writers can be of great value. The sources that Clement quotes should be investigated, as well as how accurately he uses them, and how often and in what way he credits them. Turning the issue of accuracy on its head, the way he subtly or unsubtly transforms his borrowed material should also be explored.

NATURE OF CLEMENT'S WRITINGS

Clement of Alexandria is known to the modern reader as a difficult author. The stigma of being difficult is earned in part because of the “obscure” ways,¹ in which he expresses himself and also because of his numerous digressive references to other writers, which often tend to obstruct rather than to clarify his thoughts. This unclear style may be intentional. Clement warns the reader that knowledge of the ultimate truth is not to be obtained easily.² His obscurity may even have a pedagogical implication; the faithful need to grow spiritually, and the road toward knowledge can be travelled only by dint of hard labour.³ Obscurity may also be connected with the literary genre that Clement prefered: “Stromateis,” literary weavings.⁴ Loose and digressive structure is especially evident in his massive work, the Stromateis, but it is not totally absent from his other works either.

Much has been written about Clement's borrowings in general, especially around the turn of the century, and it is unnecessary here to repeat all the different theories that have been advanced; a few, however, should be mentioned.⁵ Some scholars have maintained that Clement was essentially copying his works from anthologies, epitomes and handbooks.⁶ Others have suggested that he in part transcribed oral traditions derived from his teachers.⁷ Unfortunately, neither of these speculations, interesting as they are, can be proven: the written words of anthologies and handbooks have largely been lost, and the oral traditions of Clement's teachers are equally impossible to retrace. Research during this century has generally taken a more productive course and has analyzed and has come to value Clement's literary creativity in its own right. Scholars have grown more interested in the organization of his material, including the borrowings and their placement in the flow of his verbiage.⁸

As is often the case with theories (even if they are ultimately discarded), something useful can be learned from them. The theory that handbooks and anthologies were important for Clement, for example, remains intriguing and cannot be dismissed lightly. It cannot be coincidental, for example, that some of Clement's selections of poetry can be paralleled in the work of later anthologists such as Johannes Stobaeus.⁹ Even though Clement does not seem to have been a slavish copyist as once was suggested, a part of his material must have come from other compilations. It remains to be seen whether these compilations were anthologies that provided abridged selections collected around specific topics or whether they were writings of single authors whose works already contained much borrowed material, such as Dio Chrysostomus,¹⁰ or Plutarch¹¹ a century earlier. Using compilations and creating new ones was certainly in the air, as can be seen from contemporaries of Clement, such as Sextus,¹² Athenaeus,¹³ or Diogenes Laertius.¹⁴

Some borrowings certainly came in a more direct way, namely through first-hand acquaintance with individual authors. Some may initially have come through memory; Clement was highly literate and belonged to a time when memorization was much valued. On the other hand, not all of these “first-hand borrowings” could have been memorized; it would seem impossible for a human mind to have retained such masses of material, and it is clear, as will be shown later on, that Clement consulted sources directly or else took notes while reading them. The technique of note-taking is itself of interest in this context and deserves some attention in its own right.

TECHNIQUES OF NOTE-TAKING IN ANTIQUITY

… Literature provides us with examples of how notes were made and how they were used. Pliny the Younger in his well-known letter to Baebius Macer explains how his uncle managed to write so many books in spite of his busy life in public affairs.¹⁵ He reports that books were read to his uncle at various moments of the day, during meals, during spare time, during travel, in short, on all possible occasions. During such readings, the elder Pliny used to make notes and excerpts.¹⁶ He kept a secretary at his side with book and notebook, and the two might exchange the roles of reader and note-taker.¹⁷ At his death, Pliny left his nephew one hundred sixty notebooks of selected passages, densely written in a minute hand on both sides of scrolls. An interesting detail is that reading was done from scrolls and final notes were written on scrolls, but the secretary's initial notes were compiled on tablets, which he calls “pugillares.”¹⁸

An example of notes in the form of excerpts, as described by Pliny has come to light in modern times. The papyrus of Toura¹⁹ discovered in 1946 contains a collection of excerpts from several works of Origen, some of which stem from his Contra Celsum. It is not known what the reason for this “Readers Digest” version was, but it dates from around the time that Origen's works were being condemned officially.²⁰ Since the full text of Contra Celsum is extant, the technique of excerpting can be closely observed. Jean Scherer notes many interesting details. The length of the fragments is variable; it can cover a word, some lines or several pages. The excerpts become shorter and more “hurried” towards the end of a book. The excerptor seems to have had a particular interest in biblical quotations, these, of course, being readily available throughout Origen's works. The process is one of constant abbreviation, and sometimes the text does not seem to contain more than simple reading notes.²¹

For both Pliny and the papyrus of Toura the concept of “notes” is equivalent to “excerpts made from other works.” Other reports do not spell out what kind of notes are meant. Lucian, for example, writes that every historian should take notes before composing and writing a book.²² Notes seem also to have been important for the composition of the Gospels. The Ecclesiastical History of Eusebius informs us that Papias refered to notes as a first phase in the composition of the gospel of Mark.²³ If one accepts the authenticity of the fragmentary letter of Clement to Theodorus, yet another report can be added; Clement says that Mark carried with himself both his own notes and those of Peter, which he used for his first Gospel and for another more spiritual Gospel.²⁴

Notes could also circulate, as is mentioned in connection with a Greek Hellenistic author, Apollonius of Perge.²⁵ Apollonius, who was a mathematician, reportedly sent out notes that formed the first draft of one of his books to be reviewed and commented on by friends.²⁶ Passing around such provisional texts without the author's name attached, as Devreesse points out, was not without its risks in those days (as perhaps in our times as well); sometimes manuscripts managed to get lost and could end up in the wrong hands or even reappear under the wrong name.²⁷

Yet another variety of note-taking can be identified from early Byzantine sources, as Marcel Richard has shown.²⁸ He discusses in great detail the Greek words… which have survived in the titles (or also sometimes at the end) of manuscripts, and which mean “according to the teaching of” or “taken during the course of,” preceding the name of the lecturer.²⁹ This kind of note-taking reflects a custom that existed much earlier. In his Church History Eusebius tells how late in his life Origen permitted shorthand writers to take down his public discourses, a practice that he had not allowed in his earlier days.³⁰

It would be interesting to know more about such lecture notes, which were apparently intended for circulation. Was the text produced from the notes also edited by the note-taker, and was it sent back for approval to the lecturer?³¹ Unfortunately, these details are almost never explicitly described. Apparently much depended on the scrupulousness or carelessness of a particular lecturer. Some authors have been quite conscientious about their written production, but others simply handed over the organization and the editing to a trusted pupil. Arrian took notes on and published the lectures of his teacher, Epictetus.³² Amelius did the same for his teacher, Plotinus.³³ Hermotimos, according to Lucian, was an eternal student who spent his time editing the lectures of his teachers.³⁴ There apparently were other celebrated but even less responsible intellectuals who just left behind them a messy bunch of raw notes.³⁵

To summarize these observations on note-taking, an author or his/her entourage could take notes in the form of excerpts from other manuscripts as material to be incorporated in future learned books, as did Pliny the Elder. Notes could also be made systematically to produce an abridged version of a manuscript, as in the papyrus of Toura. Lucian, Papias and Clement mention notes that were taken before the composition of a book; they do not specify whether these notes were excerpts from other works or different kinds of notes. Notes could form the text of a book in statu nascendi, as in the case of Apollonius. Finally, notes could be taken during lectures and then written down not by the lecturer himself but by someone in the audience. Most relevant for understanding the background to Clement's borrowed quotations is the first and second example of notetaking, the cases of Pliny and the papyrus of Toura, where notes were equivalent to excerpts.

DEFINING A QUOTATION

Clement borrows passages from numerous sources that reflect not only biblical and early Christian writings but also the whole span of Greek literature from Homer to his own time.³⁶ Material from Philo, analyzed previously by this author, will form the backbone of the discussion of techniques of quotation, but a few words should be said about some of Clement's other sources to avoid giving the misleading impression that Philo was Clement's main focus of attention. Philo is, indeed, prominently represented in the columns of Stählin's index,³⁷ but Plato is even more so.³⁸ Among the poets Homer is the most frequently cited,³⁹ followed by Euripides.⁴⁰ The champion in terms of popularity, however, is not to be found among the likes of Plato, the Stoics,⁴¹ Homer, Hesiod,⁴² Euripides, or Herodotus,⁴³ but is, by quite a wide margin, good old Saint Paul.⁴⁴

In his selection of non-Christian and non-Jewish sources, Clement is a typical representative of the Hellenistic-Roman tradition. The same sources are cited in about the same proportions by other authors, whether pagan or Christian.⁴⁵ Clement compares closely with other “bookworms” such as Plutarch and Eusebius, both separated from Clement (in opposite directions) by a century; they too were writers who reportedly loved books and libraries and planted many borrowings throughout their works.

Up to this point the neutral term “borrowing” has been used, because not all material taken from other writers is a clear-cut quotation. This problem is well-known to anyone who works with quotations in authors of almost any period of the past. At some point terminologies must be formulated for these varying kinds of “recycled materials,” and a definition of the word “quotation” must be included. Whether dealing with Antiquity, the Middle Ages, or the Enlightenment, it must be determined how close the correspondance is with our twentieth century definition of a quotation. General handbooks of Greek literature and language do not give much help, but general linguistic studies have contributed to the theoretical framework for the study of quotations.⁴⁶

Individual scholars in the classical area have tried to create their own classifications. Some need to be mentioned. Particularly useful is a compact monograph on Plutarch's quotations by William C. Helmbold and Edward N. O'Neil.⁴⁷ The booklet consists primarily of an alphabetical list of quotations beginning with Acesander and ending with Zopyrus (both previously unknown to this writer). There are another 495 names in between. Without counting the exact number of quotations, it was estimated that there was a total of 6840, not including the places where Plutarch quotes himself.⁴⁸ The book has a small introduction of barely four pages, in which the authors address the problems of identification, classification and terminology. They do not offer a definite solution, but only pose the problem. Nonetheless, their separation of borrowed material into quotations, reminiscences, references and paraphrases is very helpful. They do not, however, define the borders between these terms.

Other studies create their own classifications.⁴⁹ Some of them were so detailed that they became too complicated for the purposes aimed at here. In analyzing Clement's borrowings from Philo, a scheme was required that would do justice to Clement's various techniques but that would not result in too many distinctions. The forest of quotations should not be hidden by the trees of classification. The simple distinctions offered by Helmbold and O'Neil therefore proved to be the most serviceable.

As has been shown elsewhere,⁵⁰ the distinctions in Clement's Philonic borrowings can be considered threefold on one level and fourfold on another. First, the various borrowings can be usefully separated into quotations, paraphrases and reminiscences. Quotations should be defined as having a considerable degree of literality. They need not be verbatim in a modern sense, but they should follow the source to a considerable extent. A paraphrase distinguishes itself from a quotation in that only a few words of the original source (sometimes only one or two) are present. Reminiscences, in turn, are different from allusions by having no literal correspondences but merely resemblences in theme or thought.

There are, however, some vaguely Philonic bits in Clement that do not fit into any of these three categories. A large number of passages are philosophical or literary commonplaces, and a new approach seems necessary to assess them and put them in relation to the other borrowings. To deal with the loose ends, all putative borrowings were reclassified from a slightly different angle: namely, according to their degree of dependence on Philo. This second method of classification consisted of four categories, rating from A to D.⁵¹ Again, not too many options were permitted since that would obstruct the clarity of the system. A meant a certain dependence (which in the other terminology was a quotation or paraphrase); B was probable dependence (paraphrase or reminiscence); C unprovable dependence (reminiscence); D no dependence. The two classifications were then brought together, partly overlapping and partly supplementing one another. The advantage of the second system is that it also reflects the uncertainties in the relationship between two authors, about which there had been so much speculation over the centuries.

Much could be said about why quotations were used, what function they had for an author, whether they added something to the argument or if they were primarily embellishments, and what effect they had on the reader.⁵² These are important considerations in the study of borrowings, but they can only be mentioned here, since they do not directly relate to the technique of borrowing itself.

CLEMENT'S TECHNIQUES OF QUOTING

Quotations are, of course, most identifiable when the ancient authors themselves inform us that they are quoting. It would be a good subject for a dissertation to investigate the practice of indicating quotations by name and/or book. It has often been said that in ancient rhetorical traditions, citing by name was not customary or even polite, because the educated audience was supposed to know their classics. In the same way that educated audiences are—or rather were—supposed to know their Shakespeare, Molière, or Vondel, depending on the country involved.

Let us, therefore, have a closer look again at some of the Shakespeares of Antiquity to whom Clement refers and try to analyze how regularly he does or does not name them when he quotes them. As already mentioned, Paul, Plato, Homer and Euripides are the most frequently used, and therefore will be our primary subjects. For this kind of analysis a quatitative method can be extremely helpful. Via Stählin's index and the assemblage of Greek literature on computer, the Thesaurus Linguae Graecae, it is possible to gain some numerical insight into the relationship between borrowings and citations of the name. A preliminary caveat: the computer index of the TLG is a wonderful and accurate means to gain a quick insight into words in their context; the index of Stählin, however, is less accurate since it is a vast vessel of very diverse materials that were collected over the centuries. Experience suggests that it contains too many parallels, not all of which are valid. Weeding out Stählin's index systematically, however, would be a thesis-like job for every author involved, and has not been attempted here. Since all the numbers were used in the same way, namely the index versus the TLG, it seemed a legitimate method, since the relative deviation factor is (presumably) the same for most authors, although the absolute numbers may not be totally reliable.⁵³ The results of this “census” are tabulated at the end of the article.

According to Stählin's index, there are 1273 Pauline borrowings in Clement.… 296 are connected with borrowings, and in the other 13 cases, Clement just mentions Paul's name without borrowing anything: in other words, it was just namedropping. Thus 24٪ of the references to Paul in Stählin's index are clearly identified by a mention of his name, and the majority of these 296 references so distinguished represent distinct quotations,⁵⁴ namely 269. On only 28 occasions the letter from which a quotation was taken was also named, which represents 9٪ of the explicit references to the apostle.⁵⁵ Seven of Paul's letters were cited.

Stählin's index lists 618 borrowings from Plato. In 139 cases Plato's name is mentioned…⁵⁶ that is, in almost 22٪ of the passages, a figure quite close to Paul's 24٪. Plato's name, however, was merely dropped more frequently than Paul's was: on 41 occasions. On 98 of the occasions when Plato was cited by name, a particular thought or text was brought into Clement's discussion, and 70 of these passages were direct quotations. Occasionally Plato is quoted verbatim without his name being mentioned—how often, however, still remains to be seen. Also remarkable was the high frequency with which the source of the Platonic borrowing was specified. A particular book, dialogue or letter was cited 39 times: that is, almost 29٪ of the times that Plato's name comes up,⁵⁷ in contrast to only 9٪ in Paul's case. As it turns out, 17 different writings of Plato are involved, and on six occasions Clement even refers to an individual volume, particularly when he quotes from the Politeia.

Among the poets cited by Clement, Homer is the most popular.⁵⁸ According to Stählin's index, Homer is the source for 243 passages, 143 drawn from the Iliad and 100 from the Odyssey. Homer's name or a reference to “the poet”… turns up on 82 occasions, which is 34٪ of the total number. Of these 82 mentions of the poet's name, 69 are connected with borrowings, and 54 of them are word-for-word. The proportion of literal quotations is relatively high, as can be expected in borrowings from poetry. As with Plato, some literal quotations occur without Homer's name being mentioned. Homer's name was mentioned unaccompanied by a borrowing 13 times; this occurs, for example, in the context of chronologies or linked to Hesiod. Very rarely does Clement refer to the specific source of his borrowings from Homer.⁵⁹

The second most-popular poet for Clement is Euripides.⁶⁰ Stählin attributes 117 passages to him, and the TLG provides 59 occurrences of his name.… Sometimes Euripides’ name is placed between two quotations from his works, and therefore there are fewer “unacknowledged” borrowings than appears at first glance. In this quick survey, two such cases turned up, and there must be some more. Thus, there are 61 acknowledged borrowings for Euripides (so far), which is 52٪ of Stählin's references: a figure higher than for Homer and more than double the percentages for Paul or Plato. Almost all the borrowings from Euripides were literal quotations, namely 58 out of 61.⁶¹ Clement cites the source within Euripides 16 times, and mentions a strikingly large number of writings: namely, 13 different plays.⁶² If one counts the total number of Euripidean works, including the ones that Clement does not identify, he quotes from an even higher number: namely 31 different plays, 12 of which have survived in complete form. That means that a large part of Clement's quotations come from lost plays, some of which could be identified either because Clement mentioned them by name or because they are identified by other authors who preserve parallels; 14 fragments remain unidentified, and some of them are of dubious origin.

Much Philonic material appears in Clement. Stählin lists 279 references: more than Euripides or Homer, though less than half of the number for Plato. Strangely, Philo himself is mentioned only four times (only 1.5٪ of the total). Philo certainly does not fit into the general pattern, and scholars have rightly been puzzled by the rare occurrence of his name.⁶³ There may be a specific reason for this silence, but for the moment we only can guess what it was. On the other hand, if one uses a different method of calculation, the silence may be somewhat more apparent than real. As with Euripides, one citation of the name may be coupled with numerous borrowings. In the case of Philo, there are long sequences of borrowings from one book, and one citation of the name may have been deemed sufficient by Clement. Thus his name is mentioned in three of the four long sequences and left out in one of the long sequences and the four short sequences of borrowings. Thus, it can be argued that Philo is credited as much as 38٪ of the time, if one looks only at the coherent passages. None of the disconnected scraps that Stählin connects with Philo are, however, associated with a name. The figure, thus, can drop back substantially.

Various explanations have been advanced for Clement's unwillingness to acknowledge his debt to Philo. Eric Osborn sees it in terms of a duel with the Marcionites; Clement found it prudent to downplay his Jewish sources to avoid loosing adherents to a group well known for its hostile attitude towards the teachings of Moses.⁶⁴ David Runia approaches the problem from a different angle. Philo may not have been mentioned explicitly because Clement saw himself in the same theological tradition.⁶⁵

There are other authors that Clement does, indeed, credit rarely. Tatian is mentioned only three times, and not always where one would expect it;⁶⁶ his name is, for example, left out in one of the extensive borrowings in book one of the Stromateis. It has been suggested that Tatian was one of Clement's teachers before he came to Alexandria. Something similar may have happened to another teacher, Pantaenus, of whom we know very little.⁶⁷ Pantaenus' name is mentioned only once by Clement: namely, in his Eclogae Propheticae.⁶⁸ In his Stromateis Clement professes to be greatly indebted to him, but he does not refer to Pantaenus by name but in metaphoric terms only;⁶⁹ on other occasions he brings up the authority of an “elder,”…which may refer to Pantaenus as well.⁷⁰ Eusebius informs us that Clement mentions his teacher by name in his Hypotyposes and that he refers to Pantaenus' interpretations of Scripture.⁷¹ Unfortunately, the Hypotyposes are mostly lost so that Eusebius' remarks cannot be verified. Even if we include the Hypotyposes, the name of Pantaenus comes up very rarely.⁷²

In her recent dissertation, Denise Buell notes the special rhetorical function that omitting his teachers' names could have had for Clement.⁷³ Not the individual identities of the teachers, but their roles as mediators between the apostles and Clement's own time would have been important. To put the names of his teachers in the foreground would have overemphasized “their importance as individuals, a charge that Clement makes against the followers of Marcion, Basilides, and Valentinus (see Strom. VII 108,1).” Logically then, omitting Philo's name can be seen as placing him in the ranks of Clement's direct mentors.

To return to the subject of techniques, there are many ways to introduce a quotation; the most obvious examples are: “so and so said” or “according to so and so,” and all the possible variants of this model. Writers can address themselves also to their source in a more direct way, for example, through the vocative: “O Paul,” or “O Homer,” which is a good rhetorical device found in many ancient authors. Sometimes literal quotations can be identified by introductory expressions.…

Only a small fraction of Clement's very numerous quotations, however, are highlighted by katà léxin (only 24), but it is interesting that in every instance Clement also accurately identifies the author and the work from which he quotes.⁷⁴ He seems to use the phrase katà léxin for quotations from any kind of work except for the O.T. Most often it occurs in quotations from gnostic writers (12 times), various Greek authors in general (8 times, 3 from Plato), a few times from the Lucan Acts or Pseudo-Apostolic writings (3 times), and once from the Jewish author Aristobulos. It is particularly striking that Clement acknowledges the works of his gnostic adversaries in such an accurate way, naming author and book. This scrupulousness stands in sharp contrast to his practice in borrowing from authors to whom he apparently felt a kinship, such as Philo, Tatian, and Pantaenus.

Another introduction to a quotation is formed by the words ōdé pōs.⁷⁵ The phrase could be translated as “thus,” or “in this way,” and sometimes it need not be translated at all and can simply be rendered by a colon before the quotation. The adverb pōs apparently loses its indefinite flavor when it functions as an introduction to a quotation and becomes a kind of technical device.⁷⁶ The proof of this usage can be found in Str. III 9, 2, where Clement surprisingly introduces his quotation with… “(Epiphanes) thus goes on literally”…Although the expression ōdé pōs may also occur without being connected with a quotation, this happens only rarely (6 times). In the other 69 occasions it is linked to a quotation. This time the O.T. is well represented, particularly the Psalms and Proverbs,⁷⁷ but only a few gnostic authors are so addressed.⁷⁸ Some logia and apostolic writings are introduced with the phrase,⁷⁹ but hte majority of quotations distinguished in this way comes from Greek literature.⁸⁰

A third word that often occurs in connection with a quotation is the adverb αντικρυs, “straight on” or “openly.” It can be found either before or after a quotation. It serves to emphasize a particular interpretation or admonition. It occurs 60 times, 7 of which are unrelated to a borrowing. Of the remaining 53 occurrences, 43 are related to direct quotations. Greek literature, again, is prominent (23 times), followed by the O.T. (14); in this biblical group, books of wisdom stand out (Psalms 5; Prov. 2; Sap. 2). Of the N.T. (16 references in total), a remarkably high number (8) are from the Sermon on the Mount, and 7 are from the letters of Paul.

The adverb… “expressly” or “explicitly” is also frequent in this context. It occurs 21 times, 18 of which in combination with a quotation and 3 referring to a general opinion of a philosopher or prophet. The divisions are not clearly defined; of the 18 occurrences, 5 are related to Greek literature, 5 to the O.T. (2 of which to the deuterocanonical Sirach) and 8 to the N.T.

Finally, there are some characteristic ways in which Clement incorporates his borrowings, whether they are literal quotations or freer transpositions. These peculiarities emerged during the course of close comparisons between Philo and Clement.⁸¹ It was a great advantage, of course, that most of the works of Philo have survived, making them available for comparison; so many of the quotations in Clement and other authors can not be fully analyzed, since nothing remains to compare them with. Also the large volume of material represented by Clement's Philonic borrowings offers an advantage, since their sheer quantity makes it easier to discover recurring techniques. Results suggested by authors represented in only a few borrowings must be viewed with greater caution. As in all statistical surveys the accuracy of the findings correlate directly to the numbers involved.

Clement often quotes in sequential order; that means that the borrowings stay in the sequence in which they occurred in the original source.⁸² This pattern appears not only in the borrowings from Philo, but in biblical material, as Claude Mondésert has pointed out⁸³ and in other quotations as well, such as those from the First letter of Clement to the Corinthians incorporated into book IV of the Stromateis.⁸⁴

In some sequences, a peculiar phenomenon could be observed. Clement did not always start from the earliest point within the source he was using; he would begin with a reminiscence and then leap back to the beginning of his source and restart with quotations in a sequence, selecting a few lines from each column until he had run through the whole scroll.⁸⁵ The practice could be explained in a visual way; the author first cited from memory and then looked for the specific text; leafing through the manuscript, or rather, unrolling the scroll, he became more and more interested in it and read through the whole work. This method could be observed several times. On one occasion the process was reversed; instead of rolling the scroll forward, from beginning to end, Clement rolled it backwards.⁸⁶ It is quite possible that while reading through the manuscripts the author would have taken notes or would have had them taken for use in his final text.

The sequences tend to have a certain inner rationale; they often started out with relatively literal quotations and would subsequently decline to a less literal and more abbreviated rendition of his source.⁸⁷ After some transitional and introductory phrasing of his own, Clement would then turn to his source again, repeating the process of moving from literal to loose. This phenomenon suggests the use of notes as a basis for composition in a way that is similar to what we heard earlier from the letter of Pliny the Younger.

Without exception, all borrowings are heavily abbreviated and condensed, which again indicates the use of notes. Clement tends to draw on some authors, such as Philo, Barnabas and Hermas, for their quotations from the O.T. His technique is to focus on a biblical passage but to include phrases of the author he is consulting as a kind of wrapping material, so that some of their words remain attached to the biblical quotation.⁸⁸ This is also how the source could be recognized; the shreds of alien wrapping supplied by Philo, Hermas or Barnabas make it clear that the O.T. was not consulted directly.⁸⁹ Sometimes, however, when a quotation within a quotation occurs, Clement extends the biblical component from his own memory; the biblical passage becomes longer than it had been in his intermediary source.⁹⁰

When one compares a truncated “chunk” of borrowing to the text of Clement's source, as can be done with the treatises of Philo or the Letter of Clement to the Corinthians, it is striking how abruptly the material is sometimes presented.⁹¹ Abbreviations, discontinuities and modifications can give sentences a strange twist or an illogical turn. Repeatedly, confusion and disorder are created; words are altered in strange ways, and sentences are chopped into cryptic fragments. The development of Clement's thought at times would be incomprehensible if the original text was not at hand.⁹²

In spite of his rather brutal cut-paste-and-twist technique, his illogical insertions and his abrupt transitions, Clement is also capable of clever and ingenious inventions.⁹³ At times, he subtly turns the words of his source to serve his own purposes. Thus material is transformed by conversion and rearrangement. An impressive ability to vary and juggle is persistently manifest.⁹⁴

As was already noticed, Clement makes use of another technique: that of addressing himself to his source. He sets up an implied dialogue that makes it clear that he questions his model and uses it critically.⁹⁵ The implied dialogue may be created by the addition of a conjunction or an adverb that turns the borrowed material into a hypothesis rather than an assertion; Clement may also frame quotations in interrogative sentences. All these additions, have, of course, a considerable effect on the meaning and intention of the borrowed words.

Another distinctive technique is the process of accumulation.⁹⁶ This process, which is characteristic of Clement's working method in general, occurs with particular evidence in quotations. Accumulation is often a product of his associative way of thinking; one word, as it were, evokes another without the support of a logical connection; various layers of imaginative thought are piled on each other. The technique is particularly common in the construction of allegories. Schematically described, Clement departs from a biblical starting point; he introduces a first layer of allegories, derived from one source, and then follows with more interpretations taken from other sources or that he invents himself. He usually closes with a distinctly Christian allegory. At other times the development of a theme and not an allegory is involved. He seems to be working toward a biblical text, which he cites at the end, and he chooses his quotations from Greek authors in the light of the biblical text.

CONCLUDING REMARKS

It can be seen how a consideration of general problems involving borrowing in Antiquity can be a helpful background for viewing techniques of individual authors. It is, in fact, difficult to understand how an author like Clement of Alexandria worked without being aware of the interaction of memory, intermediate sources, direct consultation of texts, and notetaking. Other authors provide helpful models to reconstruct how Clement balanced these elements. In turn, classifying Clement's borrowings can throw light on other authors. Key issues for his method of work are his variable way of acknowledging his sources, which leaves strange gaps in the credits to authors that seem to have been dear to his heart. His way of introducing quotations also varies, with contrasting techniques for introducing apostles and philosophers, poets and prophets, friends and foes. His tendency to work in terms of sequences from a single work while still producing a discontinuous effect is also characteristic. The very abundance of borrowings in Clement makes him an especially useful subject for this kind of analysis.⁹⁷

Notes

1. Jaap Mansfeld has pointed out that speaking or writing in unclear and hidden ways has a long tradition in Antiquity; he cites Galen as his main example but refers to many others, including early Christian authors. Mansfeld argues that “obscurity” in philosophical or poetic texts justifies, as it were, methods of exegesis used by the interpreters of these texts. Mansfeld calls these interpretive methods “creative” and includes allegorical interpretations among them, see Jaap Mansfeld, Prolegomena. Questions to be settled before the study of an author, or a text, (Leiden, 1994), 155-161.

2. See, for example, Str. I 2,2: 20,4; 56,2.

3. See, for example, Str. VI 2,3-4; 96,4.

4. On the meaning of the word see, André Méhat, Étude sur les “Stromates” de Clément d'Alexandrie (Patristica Sorbonensia 7, Paris 1966), 96-98.

5. For a historical survey, see Otto Stählin, Clemens von Alexandreia, Bibliothek der Kirchenväter Bd VII/1 (München, 1938), Introduction 47ff.

6. See A. Elter, De gnomologiorum Graecorum historia atque origine (Bonn, 1893-1895); H. Diels, Doxographi Graeci (Berlin, 1897). Others argued that Clement copied one particular source, see, P. Wendland, Quaestiones Musonianae (Berlin, 1886); J. Gabrielsson, Über die Quellen des Clemens Alexandrinus, 2 vols. (Upsala, 1907 and 1909).

7. See W. Bousset, Jüdisch-Christlicher Schulbetrieb in Alexandria und Rom (Göttingen, 1915).

8. For a bibliography, see my Clement of Alexandria and his Use of Philo in the Stromateis (Leiden, 1988), 3-4.

9. For some examples, see Henry Chadwick, art. “Florilegium,” RAC 7 (1969), 1131-1160, esp. 1144.

10. 40-(after) 112 C.E.

11. 50-(after) 120 C.E.

12. Floruit circa 200 C.E.

13. Floruit circa 200 C.E.

14. First half of the third century C.E. For his working technique, see Jørgen Mejer, Diogenes Laertius and His Hellenistic Background (Hermes Einzelschriften 40, Wiesbaden, 1978), 16-29.…

15. Pliny the Younger, Epistulae, III 5,7 ff., see also Colin H. Roberts, “The Codex,” Proceedings of the British Academy vol. XL (1954), 169-204; Colin H. Roberts and T.C. Skeat, The Birth of the Codex (London, 1983).

16. Dorandi, “Den Autoren,” 14-15, points out that Pliny's sequence of “legere,” “adnotare” and “excerpere” reflects a variety of working techniques. Thus “adnotare” would refer to making marks in the text of parts that had to be excerpted later on.

17. As an illustration of how serious he was about study, he made sure that his secretary was protected from the cold of winter by long sleeves so that no working time would be lost (Ep. III 5, 15). On another occasion, one of the guests asked the reader to go back and repeat a word that he had mispronounced. When Pliny the Elder asked, “could you not understand him?” the friend admitted that he could. Pliny then said, “then why make him go back; your interruption has cost us at least ten lines (Ep. III 5, 12).”

18. Small enough to hold in the hand, see the word “pugillus”: “what can be held in the fist.” For a discussion of the use, see Joseph van Haelst, “Les origines du codex,” in Les débuts du codex (Alain Blanchard, ed., Turnhout 1989), 13-35; Horst Blanck, Das Buch, 48-52. Tiziano Dorandi, “Die Authoren,” 14. 32.

19. Museum of Cairo, papyrus no. 88747; the excerpts from Contra Celsum have been published by Jean Scherer, Extraits des livres I et II du Contre Celse d'Origéne d'aprés le papyrus no 88747 du Musée du Caire (Institut français d'Archéologie Orientale, Bibliothéque d'étude, XXVIII, Cairo, 1956), 26-29.

20. Early seventh century, see Jean Scherer, Extraits, 1-2.

21. Jean Scherer, Extraits, 26-27.

22. Lucian, Quomodo historia conscribenda sit, 47-48.

23. Eusebius, HE III 39, 15-16, cited by Wayne Meeks, “Hypomnemata from an Untamed Sceptic: A Response to George Kennedy,” in The Relationships Among the Gospels. An Interdisciplinary Dialogue, William O. Walker ed. (San Antonio, 1978), 157-172.

24. Clement, Letter to Theodorus, GCS Clement IV/1, XVII-XVIII.

25. Second half of the third century B.C.E.

26. Cf. Apollonius of Perge, Conica, book I, 2-4 (ed. J.L. Heiberg, Teubner, 1891). For the preview of books by friends, see also Horst Blanck, Das Buch, 118 ff.

27. Robert Devreesse, Introduction á l'étude des manuscrits grecs (Paris, 1954), 77. There are various reports about authors whose writings circulated without their consent; see Cicero, Letter to Atticus (3,12,2). Origen received a request from alarmed friends to publish the “authentic” version of some debates he had had with opponents, since inaccurate reports apparently were in circulation; see Jean Scherer, Entretiens d'Origéne avec Héraclide et les évêques, ses collégues sur le Pére, le Fils et l'âme (Cairo, 1949), 50-51. H. Marrou calls attention to a case concerning Augustin (Retractationes II 13 [39]) in which his personal notes had been surreptitiously brought out against his wishes, see H.I. Marrou, “La technique de l'édition á l'époque patristique,” VigChr 3 (1949), 208-224, 209.

28. Marcel Richard, “απò pωνηs,” Byzantion 20 (1950), 191-222.…

29. This meaning of the phrase απò pωνηs is connected with a particular period, namely from the end of the fifth until the seventh century, thereafter the words continue to occur in manuscripts but without the addition of a name so that the meaning of the words changes.

30. See Eusebius, HE VI 36. Earlier in Origen's career we also hear about shorthand writers, but, that was slightly different, since he intentionally dictated to them, see Eusebius, HE VI 23.

31. At times this was not possible since the author had died, see Epictetus, Dissertationes 5 (ed. Schenkl, Teubner, 1894).

32. Epictetus, Dissertationes 5.

33. Porphyry, Vita Plotini 3,46-47; 4,5; see also, Marie-Odille Goulet-Cazé, “Lécole de Plotin,” in Porphyre, la vie de Plotin I (Luc Brisson, Marie-Odille Goulet-Cazé, Richard Goulet and Denis O'Brien eds., Paris, 1982), 231-257.

34. Lucian, Hermotimus 2.

35. See Robert Devreesse, Introduction, 78, who refers to Pamphila (Photius, cod. 175).

36. In the indices of Stählin (BKV 7 and 20) 462 sources are listed: O.T. (42); N.T. (25); early Christian (32); non-Christian 363. Tollinton counts 348 sources, see R.B. Tollinton, Clement of Alexandria. A study in Christian Liberalism, 2 vols. (London, 1914), 157; he refers to a study of P.A. Scheck, De Fontibus Clementis Alexandrini (Augsburg, 1889), 15. Tollinton was unable to consult the index by Stählin, which appeared only in 1936 (GCS); 1933-36 (BKV).

37. 279 putative references; 9 columns in the BKV editions, which are more complete than Stählin's original Index in the GCS; one column in the BKV contains an average of 40 quotations.

38. 618 possible references and 15.5 columns.

39. 243 possible references and 6 columns.

40. 117 possible references and 3,5 columns.

41. Chrysippus is represented by 4 columns.

42. Hesiod is represented by 1 column.

43. Herodotus is represented by 2 columns.

44. 1273 possible references and 27 columns, pseudo-Pauline letters included; only followed at a large distance by Matthew 11 columns; Luke 7.5 columns; John 5 columns; Mark 3 columns; Acts 1.5 columns.

45. See, for example, Plutarch. For poetry quotations in Early Christian authors, see Nicole Zeegers-vander Vorst, Les citations des poètes grecs chez les apologètes chrétiens du II^e siècle (Louvain, 1972), 31-44.

46. For further bibliography, see Antoine Compagnon, La seconde main ou le travail de la citation (Paris, 1979); also Wilhelm Krause, Die Stellung der frühchristlichen Autoren zur heidnischen Literatur (Wien, 1958), 51-58, esp. note 1, in which (primarily German) literature is cited.

47. William C. Helmbold and Edward N. O'Neil, Plutarch's Quotations (Phiological Monographs XIX, APA, 1959).

48. There are an average of 45 quotations per column and two columns per page, which makes a total of 76 x 90 = 6840. Stählin's index of Clement contains approximately 7300 references; O.T. (1600); N.T. (2100); Early Christian authors (380); Greek literature (3180). From the numbers for Philo it turned out that of the 279 references 125 (45٪) were true borrowings, of which 93 (33٪) were clearcut quotations. Even if we apply the percentages for Philo (45٪ and 33٪) to the total numbers of Stählin's index, the outcome is impressive: 3285 borrowings, of which 2409 are quotations. Compared to Plutarch's number Clement is overshadowed, but Plutarch's oeuvre is after all much more extensive than that of Clement.

49. Cf. P.J.G. Gussen, Het leven in Alexandrïe volgens cultuurhistorische gegevens in de Paedagogus (boke I en III) van Clemens Alexandrinus (Assen, 1955); Wilhelm Krause, Die Stellung (see note 53 above). André Méhat, Kephalaia; Recherches sur les matériaux des “Stromates” de Clément d'Alexandrie et leur utilisation (thèse dact., 1966). Nicole Zeegers-vander Vorst, Les citations (see note 52 above); F. Castincaud, Les citations d'auteurs grecs profanes dans le Pédagogue de Clément d'Alexandrie (mémoire pour la maîtrise, Poitiers, 1976).

50. See my Clement of Alexandria, 20-21.

51. See my Clement of Alexandria, 22.

52. For these questions, see also the studies of Compagnon and Krause mentioned above, (note 53).

53. Stählin's index may be more reliable for poetry than for prose.

54. How many quotations are hidden among the remaining 76٪ of Stählin's references is uncertain; this survey was primarily concerned with the occurrence of Paul's name and writings.

55. A negligible 2٪ of the total number.

56. The claim that Plato's philosophy originates from the teachings of the Hebrews and in particular of Moses fits into a general apologetic argument; for an extensive documentation of this subject, see Heinrich Dörrie, Der hellenistische Rahmen des kaiserzeitlichen Platonismus (Der Platonismus in der Antike, Bd. II H. Dörrie and M. Baltes eds., Stuttgart, 1990), 190-219, nos. 69-71 (texts), 480-505 (commentary).

57. 6٪ of the total occurrences.

58. 38٪ of all poetic references are to Homer.

59. On only one occasion was a reference to the Odyssey found, see Protr. II 35,2.

60. 18٪ of all poetic references are to Euripides. Hesiod represents 5٪, and all others are less than 5٪.

61. On only one occasion the name was dropped without quotation.

62. In 26٪ of the passages with the name a source also is mentioned, which represents 14٪ of the total.

63. For a discussion of the issue, see Eric Osborn, “Philo and Clement,” Prudentia 19 (1987), 35-49; most recently David T. Runia, Philo in Early Christian Literature. A Survey, Assen, 1993.

64. Osborn, “Philo,” 35 ff.

65. Personal communication.

66. Out of 13 occasions (which still results in 23٪); for a portrait of Tatian in Clement, see Alain Le Boulluec, La notion d'hérésie dans la littérature grecque II^e-III^e siècles, 2 vols. (Paris, 1985), II 346-348.

67. See Pierre Nautin, “Panténe,” Tome commémoratif du Millénaire de la Bibliothéque d'Alexandrie (Alexandria, 1953), 145-152.

68. In connection with a quotation, see Ecl. 56,2.

69. In Str. I 11,2 (the true Sicilian bee).

70. Protr. 113,1; Ecl. 50,1; Fragm. 22 (III 201,26); Fr. in I Joh. 1,1 (III 210,1).

71. Eusebius, HE V 11,2; VI 13,2. Pierre Nautin, “Fin,” 296. 281, suggests that the Eclogae Propheticae (and also the Excerpta) are a part of the Hypotyposes, and that Eusebius is actually referring to Ecl. 56,2. Nautin maintains that a scribe excerpted and abbreviated the writings that occur after book VII of the Stromateis as preserved in the Codex Laurentianus V 3; these writings are: book VIII of the Stromateis, the Excerpta, and the Eclogae. A majority of scholars, such as J. Munck, R.P. Casey, G. Lazzati, Cl. Mondésert, F. Sagnard, O. Stählin and A. Méhat consider these writings, however, as Clement's reading notes for later works. Since the character of these writings seems to correspond with his technique of note-taking, there is ample reason to attribute them to Clement himself and not to a later scribe.

72. Monique Alexandre (personal communication) calls attention to a similar situation in the Pedagogue, where Clement is very close to the first century author Musonius Rufus, whose name, however, is never spelled out.

73. Denise Kimber Buell, Procreative Language in Clement of Alexandria (Ph.D. Dissertation, Harvard University, 1995), 108-109.…

74. On the accuracy of these quotations, see also William C. Helmbold and Edward N. O'Neill, Plutarch's Quotations, p. IX.…

75. Variants… occur similarly.

76. This observation is also made by Liddell and Scott s. v. πωs: “sometimes merely to qualify their force, when it cannot be always rendered by any one English equivalent.”

77. A total of 13 times, of which are 4 from the Psalms and 4 from Proverbs.

78. Only 3 times.

79. A total of 12 times.

80. A total of 37 times, including Plato 6, Euripides 4 and Homer 1.

81. See my Clement of Alexandria, 214-217.

82. The most clear example of this technique can be observed in Str. II 78-100 where Clement quotes extensively from Philo's De Virtutibus. Similar patterns have been observed in the working techniques of Diogenes Laertius, see Jørgen Mejer, Diogenes, 18.

83. See Claude Mondésert, Clément d'Alexandrie; introduction á l'étude de sa pensée religieuse á partir de l'Écriture (Paris 1944); also André Méhat, Kephalaia; Recherches sur les matériaux des “Stromates” de Clément d'Alexandrie et leur utilisation (thése dactyl.), 1966.

84. Str. IV 105-119.

85. See, for example, Str. II 78,2-3 and 81,1-2; V 32,2 and 32,3; I 29,10 and 30,3; II 51,3 and 51,4.

86. See Str. II 5,3-6,4.

87. This characteristic was also observed in the papyrus of Toura, see above note 25.

88. Similarly in the papyrus of Toura, see above note 25.

89. See, for example, for Philo, Str. II 5-6; for Hermas, Paed. II 83,4-5 for Barnabas, Str. II 67,2-3; also Annewies van den Hoek, “Clement and Origen as Sources on ‘Noncanonical’ Scriptural Traditions,” Origeniana Sexta (Gilles Dorival and Alain Le Boulluec eds., Leuven, 1995), 100-101.

90. See Str. II 5-6; 94,1; 96,3.

91. Especially the quotations from Philo's De Virtutibus fall into this category (Str. II 81,3; 83,3; 85,3; 94,3; 95,1). Philo had already offered disconnected precepts from the Pentateuch, which become even more chunky through Clement's treatment; for a detailed description, see my Clement of Alexandria, 69-115.

92. For example, when Clement mistakenly or absent-mindedly takes one word from a following sentence of his source and connects it with the previous sentence, see, for example Str. II 90,2 (in a quotation from Philo); Str. VI 131,2 (in a quotation from Hermas).

93. See, for example, in Str. II 81,1-3 where two disconnected sentences are used to project a different viewpoint. Philo had stated that a “true” man has to behave in a manly fashion. Clement alters: the man who devotes himself to the “truth” behaves in a manly way; also Str. II 97,2 where two fragmentary sentences are combined ingeniously.

94. See, for example, Str. II 78,2-3; 80-81; 94,3-5.

95. See, for example, in borrowings from Philo Str. II 81, 3 92,2 94, 4. On occasion he also rejects the interpretation of his source, as in Paed. II 83,4-5 (in a quotation from Barnabas).

96. This technique can be observed throughout Clement's works.

97. Many thanks go to Alain Le Boulluec and Jean-Daniel Dubois, at whose seminar at the ÉPHÉ (1994) in Paris an earlier version of this paper was presented. Similarly the Boston Area Patristic Group and the seminar of Richard Thomas and Greg Nagy at the Harvard Classics Department helped to define some of the questions; at Harvard the response by Sarolta Takács was very helpful. David Runia gave constructive advice and provided important bibliography. Finally, thanks go to François Bovon, Brian Daley and John Herrmann for their support and their critical eye(s).