The previous chapters have examined the macrostructures of imperial period and late antique works of scholarly erudition – their genre – before moving to microstructures and looking at their organization, and their formation, and how these structures may relate to and explain the Babylonian Talmud. The present chapter will examine the structure of one distinct part of erudite compilations: the story or narrative.

The tale (mythos/fabula) and its close relative, the narration (diēgēsis/narratio), were crucial components of imperial period and late antique education.Footnote ¹ In order to assess, from scratch, how stories, and in particular the erudite story, were constructed, this chapter will start with a brief discussion of the Greco-Roman curriculum, its singular nature, and its impact on other language cultures in the Mediterranean area. A brief survey of the time’s aesthetics will show in what ways they are reflected in the conceptualization and makeup of stories, including talmudic stories. It will be shown that the same methods employed to produce complex erudite books such as the Talmud are at work, in miniature, in stories. This significantly facilitated their integration into a composite text.

The Late Antique Story and Progymnasmatic Training

Recent scholarship has called new attention to the fact that the Babylonian Talmud, as well as rabbinic literature in general, show clear signs of the use of rhetorical patterns in the composition of arguments (sugyot) and homilies.Footnote ² This interest goes hand in hand with a renewed focus on rhetoric in classics as well as in patristics, which increasingly engage each other in the concept of the “Third Sophistic.”Footnote ³ The consensus that rhetoric, despite having originated in the law courts of the Greek polis, later became “the bedrock upon which the composition of orations, speeches and sermons was built at a time when opportunities for public speaking were numerous” is growing.Footnote ⁴ Indeed, rhetorical speaking was an essential part of late antique entertainment, the court system, education, and politics.Footnote ⁵ Different reasons have been identified for this increasing popularity of rhetoric and its spread beyond the courts, including the “massive administrative organization of the Roman Empire and its cultural system,” which necessitated and produced a standardized way of communication, or the public competition between “sophists, bishops, philosophers and other public figures.”Footnote ⁶ These deeply connected factors ultimately led to a comparatively uniform curriculum shaped by “publicly funded and managed schools … attested all across the Empire.”Footnote ⁷

Administrative needs alone can be met with alphabetization, grammatical training, and the introduction of basic protocols, that is, the type of writing necessary for everyday bookkeeping, formal letters, and documents.Footnote ⁸ There is no practical need to go beyond such basic abilities, particularly because there existed no obvious link between wealth and education prior to the imperial period: trades and specialized crafts, which did not necessarily require full literacy, were passed on within the family. The wealth of those who held offices was also inherited or acquired through booty, the discovery of mining deposits, or tax revenues.Footnote ⁹ The increased competition among the elite in the wake of the Roman Empire’s expansion may have been one reason for the investment in literacy and, especially, literary production as a means for building reputation.Footnote ¹⁰ Another reason was the competition between Alexander’s heirs about the true successors of the Greek heritage, to which the Ptolemies responded with an unprecedented investment in intellectual sponsorship.Footnote ¹¹ Patronage would continue to allow people to have a profitable occupation as an orator, author, or even “literary manager of others,” as imperial education could include everyone, “slaves and freedmen as well as the elite.”Footnote ¹² With literacy becoming a prestigious social pursuit, it is not surprising, then, that the imperial-period curriculum went far beyond basic alphabetization.

Rhetoric, which penetrated speech as well as writing, promised defense and persuasion, both crucial abilities in a world where quarrels were likely to end up before a judge. Rhetoric made people believe that the right words, written or spoken, had the power to subdue enemies, physical or metaphysical, and to effect substantial change.Footnote ¹³ Written amulets, so-called voces magicae, and theurgy testify to the importance of persuasive language in what might be termed “cosmic courtrooms.”Footnote ¹⁴ These were skills of interest to everyone.

Yet rhetoric served not only the purpose of persuasion but also entertainment. Rhetoric changed the way in which diverting texts were written and presented. One influential example is the sophist Lucian, an author famous for his pastiches of well-known scenes, full of allusions and comic exaggerations. Lucian saw an excellent mastery of language as a form of delightful acumen. This is perhaps best illustrated in the hommage he wrote for his teacher Demonax, whom Lucian characterizes primarily as a master of quick-witted responses.Footnote ¹⁵

The appealing promises of the late antique, Greek-based rhetorical curriculum were manifold, and there does not seem to have been an alternative curriculum in the Mediterranean and adjacent areas. Thus, while some bishops may have considered adapting the curriculum, that is, replacing Greek and Roman myths and tales that served as exercises with biblical ones, they did not and could not think of replacing its rhetorical goals.Footnote ¹⁶

The critical steps in shaping students’ ability to craft their own texts was taken in the formative process between the beginner’s curriculum – that is, alphabetization and acquaintance with grammar – and the more advanced study of rhetoric for juridical or deliberative purposes with an accomplished orator. This formative process, in which students learned how to go about their own written compositions, has been described in various progymnasmata, “preliminary rhetorical exercises.” Still extant progymnasmata in Greek are those ascribed to Hermogenes (second century), Aphthonius (fourth century), Libanius (fourth century), Nicolaus (fourth to fifth centuries), and Aelius Theon (fifth century).Footnote ¹⁷ Except for the exercises ascribed to Libanius, however, no treatise comes with an actual set of exercises. Rather, they describe the literary forms to be studied and how to teach them only in technical terms, usually on the basis of a single example. This implies that teachers were forced to choose their own examples based on which they would teach rhetorical principles. Considering the above-mentioned discussion among bishops, it may therefore be assumed that teachers who taught Hebrew or Aramaic composition chose examples from the Hebrew Bible or the Mishnah to teach rhetoric.Footnote ¹⁸

The progymnasmata did not train students to freely write their own stories. Quite the opposite: they were taught how to transform other stories or to enhance a maxim (gnōmē) with an action and a speaker, thereby creating a chreia. The bulk of the plot was thereby already given, forcing students to practice not originality but exegetical flexibility. In a juridical context, this flexibility served to transform the argument of an adversary into its contrary by artfully highlighting and enhancing certain points, or to reveal contradictions. Although this method could produce quite creative outcomes, the art of bending meaning was taught very mechanically at an early stage.

People wrote (and still write) according to their training. The somewhat mechanical methods applied to transform motifs were certainly responsible for the enormous number of books and treatises that late antique authors were able to produce. Lucian, for example, wrote variant after variant of motifs and twists found in earlier stories, and then he wrote again variants of these variants. Graham Anderson illustrates Lucian’s stereotyped reworking by showing how Aphthonius suggested reworking a maxim.Footnote ¹⁹ According to Aphthonius, a maxim could be altered by adding praise for its author, paraphrasing it, explaining the reason for the maxim, proving it right by virtue of the truth of its contrary, comparing it to a similar case, giving an example of a situation/action in which the maxim fits, adding a similar statement by another person as testimony to its truth, or by appending an epilogue to it.Footnote ²⁰ A maxim (gnōmē/sententia) could be true, plausible, or hyperbolic in its content, and simple or composite in its style.Footnote ²¹

Similar exercises were also suggested by different progymnasmata as exercises for the chreia. The chreia, a saying attributed to a person and sometimes enhanced with an action, can be considered the most prominent form of a miniature story in late antiquity. In modern scholarship, it has often been translated incorrectly as “anecdote,” which is likely to provoke incorrect assumptions regarding the truthfulness or the amusing character of its content.Footnote ²² Most of all, a translation of the chreia as “anecdote” hides the highly technical makeup of these short stories. The set of methods for slight change (exergasia) of the chreia was the same for maxims or sayings. Hermogenes provides a nice example of how these methods affect a saying (logikon):

The tale (mythos/fabula) and the narration (diēgēsis/narratio) were similarly subject to this systematic change by students. The narration had as its distinctive features the identification of the protagonists and indications as to where and when a certain event took place. The event itself did not need to be true but had to be plausible. The tale, on the other hand, is described as plainly fictitious in the progymnasmata. But the quality of a tale was likewise assessed by its plausibility. Such plausibility could be achieved by associating certain traits with suitable characters, like beauty with the peacock and cleverness with the fox, or by adapting a given plot to accommodate new protagonists.Footnote ²⁴ For some writers of progymnasmata, a tale, by definition, featured animals, a definition refuted by Theon and Aphthonius.Footnote ²⁵ Aphthonius further distinguished between the rational, the ethical, and the mixed tale. In the rational tale, humans do something, while the ethical tale transfers human characteristics to animals, and both features appear together in the mixed one.Footnote ²⁶

According to the progymnasmata, tales have a didactic purpose, and the lesson should be highlighted either in the beginning or in the end, in the form of a saying or maxim. These sayings or maxims can be replaced at will to give a tale a different direction, depending on the point somebody wishes to make. An epimython, a tale followed by a gnomic statement, might then read as follows:

Tales could further be enhanced with descriptions and additional dialogue, or they could be shortened by the omission of these very features.Footnote ²⁸ Apparently, then, late antique stories were defined and structured by a distinct and qualified set of methods for variegation.

The language of a tale should be clear and simple, as Nicolaus emphasized, and deviate “little from that used in ordinary conversations.”Footnote ²⁹ The moral value of the tale or its consistency could be acclaimed or refuted. This was an important exercise, as Theon explained, since it prepared students for the refutation or confirmation of a juridical argument.Footnote ³⁰ The purpose of these exercises was to help students realize that statements and actions could be used independently from their original context to create a different meaning. At this stage, sayings and actions from a “textual witness” were used as stand-ins for the juridical argument or case. The examples were wisely chosen so that their moral and instructive content offered an additional pedagogical benefit. This instructional habit of using quotations as proof obviously left its mark on late antique writing culture. Chreia and maxims have been found to underscore arguments in texts as diverse as private letters, amulets, or incantations, where they were used as claims to tie someone down (as in defixiones) or to set someone free (as in amulets).Footnote ³¹ The same sort of intercessional authority was transferred to whole books, which were worn as pendants for apotropaic purposes.Footnote ³²

It appears that the late antique story was built with or around sayings and maxims: they constituted a small unit (the chreia), introduced or appended the plot, or appeared in dialogues in the form of quotes or direct speech. Yet the progymnasmatic curriculum had more to offer regarding crafting a plausible story. In later stages, students were trained in “vivid description” (ekphrasis) and in speech in character (ethopoeia). Accomplished students of progymnasmatic training thus had a useful and certified set of tools at their disposal that allowed them to make sense of a text, to use it as proof for or against an argument, or simply to compose a plausible new version of an old story. They were also able to mix and match the learned writing methods for different purposes. In fact, students had no choice but to write along these lines even if they did not choose a career in the courtroom; this was how they were trained, what they knew, and what the audience expected.

One result of progymnasmatic training in a nonjuridical context is what I will call the “exegetical story.” This type of story combines the inquiry (thesis, discussed in detail in Chapter 1) with a tale or narration. It is found specifically in rabbinic and monastic literature. Like the “exegetical inquiry,” the exegetical story takes as its starting point one or two conflicting sentences from works considered to be “textual witnesses,” such as the Hebrew Bible or the New Testament. These conflicting “testimonies” are explained through a story, that is, a rather elaborate simile, before concluding with a maxim, saying, or quote from a “witness.” As I will discuss an exegetical story from the Talmud further below, I will illustrate this point here with an example from the Sayings of the Desert Fathers (Apophthegmata Patrum), a work written around the end of the fifth century:

The story starts with two verses from the New Testament (John 6:27; Luke 10:42). These verses do not conflict with each other but, rather, with the physical labor that monks must perform. The story needs to bring proof against the two textual witnesses. This proof is found in the physical needs of man and in a contextual interpretation of the story of Mary and Martha (Luke 10:38–42). Exegetical stories in the Talmud proceed in the exact same way, always aiming at proving or disproving a conflict, either between textual witnesses or between such a witness and daily experience.Footnote ³⁴ These conceptual parallels between rabbinic and monastic “exegetical stories” are suggestive of a literary training that adhered to the same stylistic concepts and objectives.

The Larger Impact of the Progymnasmata

There is no evidence of a conceptually different curriculum that would compare to the progymnasmata. It was the only curriculum proposing a continuation of literary training after basic alphabetization. In addition to the preserved progymnasmata, many others, now lost, seem to have circulated. The uniformity among the proposed exercises in the extant treatises, however, suggests that the lost curricula must not have differed much in content either.Footnote ³⁵ The exercises covered rhetorical subjects such as the attributed saying or action (chreia); the maxim (gnōmē); the reminiscence (apomnēmoneuma); the fable (mythos); the narration (diēgēsis); refutation (anaskeuē); confirmation (kataskeuē); the amplification of a brave or faulty deed (topos); vivid description (ekphrasis); the introduction of a speaker (prosōpopoeia); praise of living people (encomion); of the dead (epitaphios); of the gods (hymn); invective (psogos); comparison (syncrisis); imitation and speech in character (ethopoeia); inquiry (thesis); the introduction of a law (with focus on refutation or confirmation); as well as paraphrase and elaboration (exergasia).Footnote ³⁶

The curricula were translated into other languages in the Mediterranean. There is an Armenian translation of Aelius Theon’s progymnasmata, carried out by the so-called “Hellenizing School” (ca. 570–730 CE).Footnote ³⁷ A Latin translation was produced by Priscian (ca. fifth century CE) of (pseudo-)Hermogenes’s treatise. Worth noting is the fact that Priscian substituted the Greek examples with examples from Latin authors, such as Terence, Sallust, Virgil, and Cicero.Footnote ³⁸ It was indeed not difficult to substitute the exercises’ few suggested literary examples with those from a different body of literature, since the treatises, with the exception of Libanius’s progymnasmata, were mainly theoretical. As mentioned above, even instructors who wanted to teach based on Greek examples had to come up with additional examples themselves. Based on the popularity of the progymnasmata, it seems feasible that the curriculum was translated into Aramaic, the Sasanid lingua franca, as well. Jewish teachers would have substituted the original examples from Greek (or Persian) poets with examples from the Hebrew Bible, Aramaic texts written by Jews (“Judaized texts”), and maybe early rabbinic texts (pending the teacher’s involvement in this tradition).

Admittedly, there is no evidence of an Aramaic, Syriac, or Coptic translation of these curricula. Then again, there is ample evidence in Jewish Aramaic, Syriac, and even Coptic texts for the application of the writing standards taught through progymnasmata.Footnote ³⁹ In the case of Syriac, even without proof of an extant translation of progymnasmata, by the sixth century, “Syriophone education” spread “even amongst the lower end of the literacy spectrum … being increasingly assigned a prestige equal to that of Greek.”Footnote ⁴⁰ Teachers might also have translated the Greek standards directly into the local language without writing a formal translation of the curriculum. In any event, bilingual learning had become the standard rather than the exception since the translation of the Greek curriculum into Latin.Footnote ⁴¹ Many texts testify to bilingualism and immediate translation of Greek script, grammar, and style into another language.Footnote ⁴² Transfers of Greek idioms to late Hebrew and Jewish Aramaic have been observed in abundance, and Mesopotamian incantation bowls bear witness of people with the ability to write in several languages and scripts.Footnote ⁴³

Learning in late antiquity involved traveling, which meant that many students acquired their knowledge from different teachers and in various settings.Footnote ⁴⁴ In the case of Sasanid Babylonia, cultural fluidity was also promoted by geography: the Mesopotamian plain was a flat border area between the Roman and the Sasanid Empires. There was no exact demarcation or closable frontier, and the region allowed for and benefited from considerable exchange.Footnote ⁴⁵ Indeed, cultural boundaries seem to have been defined less by territorial frontiers than by foundational myths that were, again, the basis of grammatical learning. For Priscian, it was important that students studying the Latin language became simultaneously acquainted with the Latin poets. Similarly, there are no translations of Greek myths found in rabbinic literature, only isolated and reappropriated motifs barely recognizable as such.Footnote ⁴⁶ Clearly, as Blossom Stefaniw observed, “texts which were the object of grammatical study bound their readers into a historical and cultural lineage: the reader was connected to the past, in that she was brought into a relationship with the moral and literary patrimony passed down through ancient texts.”Footnote ⁴⁷

A fully historical understanding of late antique texts is only possible if the formative training of their authors is considered. Based on Theon’s remark that “training in exercises is absolutely useful not only to those who are going to practice rhetoric but also if one wishes to undertake the function of poets or historians or any other writers,” Robert Penella posited an “abiding influence of these rhetorical exercises on the ancient mind.”Footnote ⁴⁸ Indeed, the formative impact of the progymnasmata is most evident in the uniformity with which late antique literary culture presents itself. It was this uniformity, in fact, the overall notion of borrowing, fragmentation, and heterogeneity, which generated the long-held assumption of an intellectual decline in late antiquity, in contrast to the perceived originality and creativity of ancient authors. More recently, however, this mannerism has been acknowledged for its own beauty, which simultaneously coined and expressed the taste of the time.Footnote ⁴⁹

In his seminal article “The Treatment of Narrative in Late Antique Literature” (1988), Michael Roberts pointed to some shifts responsible for the distinct style of late antiquity as compared to classical antiquity. These shifts, he argued, are not only visible in the way the structure of narratives changed but, tellingly, also in works of art. Late antique art seems to contrast the harmony and internal order of antiquity with discontinuities, fractures, and a “preference for juxtaposition over continuity.”Footnote ⁵⁰ These changes, which seem to have occurred in tandem in art and speech, did not escape the notice of attentive contemporaries. Quintilian (first century), for example, criticized an overuse of sententiae (gnōmai in Greek), which led, in his assessment, to an uneven style.Footnote ⁵¹ The trend nevertheless continued. Sententiae saturate late antique literature not only in the form and content of classical maxims with their general and moralizing character but also in the form of statements of a more technical nature (i.e., medical, architectural, agricultural, and so on).Footnote ⁵² In general, there was an increasing trend toward the concise, short text, a phenomenon referred to by some scholars as “miniaturization.”Footnote ⁵³ Preference was given to the condensed work and the short treatise. The fact that many brief stories or small excerpts eventually added up to multivolume breviaria sometimes masks this trend.Footnote ⁵⁴

The diversity of the short sententiae that made late antique literary culture look like a patchwork quilt also found its reflection in fashion trends. Thus, Quintilian further lamented the new trend of exchanging the classical purple stripe of the toga for “multi-colored patches, panni or segmenta, applied to or embroidered on clothing.”Footnote ⁵⁵ The taste for colorful and variegated “patches” apparently penetrated several areas of life. In literary compositions, these patches were reflected on a macrolevel by excerpts and on a microlevel by proverbs, recipes, or brief technical instructions.

Like taste in clothing, taste in mosaics changed. Irving Lavin describes late antique mosaics as appearing amorphous and depthless at first, since they aim at assuring not one but multiple possible viewpoints.Footnote ⁵⁶ This tendency is also reflected in the time’s philosophy and its approach to truth and truth claims: whereas “Plato taught that truth is singular, objective and unchanging … the sophists viewed reality as being multifaceted, relative and in constant flux.”Footnote ⁵⁷ Similarly, mosaics are characterized by their division of space into single blocks, while they also maintain a superior thematic unity. It was a design that “offered yet another possibility of which the classic system was incapable .… The composition could be extended infinitely in any direction without prejudice to the unity of the surface as a whole.”Footnote ⁵⁸

These mosaics, then, mirror the already familiar literary pattern of miniature units, such as excerpts, short stories, or chreia, which can – but do not have to – be strung together endlessly. Indeed, stories often seem to be constructed from individual scenes that make independent points. With only minor changes, one or more of these scenes can easily be used in another catena-like story. Roberts linked the possibility for dissection and the focus on the description of single parts to the progymnasmatic exercise called ekphrasis (description), and especially to the process leptologia, or descriptio per partes. Leptologia is the division of a scene “into its constituent parts which will then be enumerated in elaborate detail.”Footnote ⁵⁹ Ekphrasis marked the interface between art and narrative, vision and text, since it aimed at describing something so vividly that listeners and readers turned into spectators.Footnote ⁶⁰ Training in ekphrasis seems also to have been the reason for the increasingly sensual and graphic stories in late antiquity.Footnote ⁶¹

The late antique story preferred thematic unity even over chronological order.Footnote ⁶² This observation finds support in Theon’s progymnasmata, where he writes regarding the narrative:

Similarly, the talmudic story deviates “from the strict temporal order, most often through flashbacks, such that events that occur later in the story are recounted earlier in the text.”Footnote ⁶⁴

These few but significant observations show how much of late antique writing culture can be understood through the lens of the progymnasmata. They provided students with the intellectual tools for purpose-driven writing, classification, and problem-solving, and, in many ways, culture tout court.

The Talmudic Story

In the last sixty years or so, considerable scholarly effort has been devoted to explaining the nature and purpose of rabbinic stories. Comparative historical approaches spurred by Henry Fischel and literary analysis promoted by Yonah Fraenkel emerged simultaneously in the late 1960s.Footnote ⁶⁵ These two schools, the former of which has since been criticized for its positivism and the latter for its decontextualizing approach, appear to have merged in recent years. They produced an approach that is critical regarding the historical reliability of the stories while also being sensitive to their cultural context.Footnote ⁶⁶ The purpose of the remainder of this chapter is to enhance these prior studies with an explanation of why the stories appear in their particular shape, how they were composed, and why we usually only find near or quasi parallels of certain motifs or stories in other works.

Most of the above-mentioned stylistic features of late antique stories have also been observed in talmudic stories but, to date, have not been linked to these. Based on the conclusions drawn in the previous chapters, an analysis of “the talmudic story” must consider the fact that the stories do not necessarily appear in the Talmud in the form in which they were originally composed. Instead, the stories may appear as excerpts in the shape that best suited the composers when arranging an adequate commentary on a certain lemma. In fact, the chain-like structure of late antique stories, with their easily detachable segments, makes them suitable for exactly such breakups and rearrangements. This obviously complicates the assessment of the actual story, especially since, as we shall see, the composers will apply – by default and similar training – the very same compositional methods to the story as the original author did. They will substitute dialogues or characters, if necessary, interrupt the story with associatively fitting excerpts, or add sayings to the concluding moral of the story. Then again, it is precisely the fact that the composers use the very same methods as the story’s author that enables an analysis of the story behind this tampering: the possibilities are limited and repetitive.

I will illustrate this point with a lengthy story found in the commentary to the lemma qordiaqos (b. Git. 67b–70b), with which the reader is already familiar through the discussion in Chapter 3 of that commentary’s structure and its assigned keywords (“cure,” “meat,” and “wine”). Familiarity with the story’s co-texts will facilitate its analysis from a compositional point of view. The story begins with two enigmatic biblical verses, which are used to explain each other. One of the verses is 1 Kgs. 6:7, where it is written that Solomon built the temple in Jerusalem out of stones so perfectly hewn that no sound of a chisel was heard while it was being built. In Eccl. 2:8, a book ascribed to Solomon’s authorship, we read that he busied himself with shidah and shidot.Footnote ⁶⁷ Both words, shidah and shidot, are unintelligible, biblical hapax legomena. The story’s author(s) will interpret the terms as referring to male and female demons, based on the Aramaic word for demon, shed. This interpretation will first be contrasted but finally harmonized with a somewhat older rabbinic tradition, which claims that Solomon achieved this temple miracle by using the shamir, a mysterious “something” that carves even the hardest of stones.Footnote ⁶⁸ Thus, like the above example from Sayings of the Desert Fathers, this story follows the pattern of the “exegetical story” that takes as its point of departure biblical verses but also conflicting statements by long-gone teachers.

For the sake of clarity in the discussion following the story, distinct parts are labeled with letters. The very fact that it is possible to label the story’s components in this way points to a similarity with the above-discussed feature of mosaics. Both, story and mosaic, are composed of distinct units. These units are, as shall be shown in due course, extractable and rearrangeable without disrupting the story – with the exception of the proem. Just as Theon suggested in his progymnasmata regarding proems in general, the proem in this story is a unique composition (Progym. §76).

The Case Example: Solomon, Ashmedai, and the Building of the Temple

The composers placed the following story about Solomon and the demon Ashmedai in the narrative section of the composition. It follows upon a story about Rav Sheshet’s dining at the exilarch’s house, where the latter’s servants tried to harm or even kill him, mostly by means of meat. The lengthy story, or rather story cycle, seems to have been chosen based on the keyword “wine,” or “wine” and “cure.” Within the commentary on qordiaqos, the story cycle will be followed by the “proofs,” a list of recipes. The story comprises eight units (A–H), two of which belong together (D^G and G^D).Footnote ⁶⁹

I have already suggested in the introduction to this section that this story cycle made it into the commentary based on the keyword “wine.” Based on the fact that this story is followed by medical recipes, several scholars have argued that the story was added to the Gittin commentary on qordiaqos because of Solomon’s reputation as a healer and subduer of demons.Footnote ⁸⁹ This argument may be strengthened by the fact that the proem to the commentary (not the proem to the story!) reads qordiaqos as Qordiaqos, the name of a spirit (ruha).Footnote ⁹⁰ This interpretation was, however, already refuted in the proem itself (see Chapter 3), and Ashmedai, with whom Solomon has to deal, is called a demon, shed, not a spirit, ruha. Based on the clear distinctions between types of demons made elsewhere in the Talmud, this link seems weak.Footnote ⁹¹ Moreover, the story makes no effort to depict Solomon as a healer and only makes a feeble attempt to portray him as a master of demons (B). And it is Ashmedai who cures the king’s hangovers with his remedies (E). The story itself may indeed be satirically reversing the positions, but by way of the keyword method a purposeful satirizing of a whole talmudic commentary would not have been possible; if it happened, it was by coincidence. Based on the keywords, it may have been the recipes provided by Ashmedai, situated at the apogee of the story (E), that were responsible for the inclusion of the story cycle into the commentary on qordiaqos. Indeed, the first of these recipes refers not just to one but to two of the three assigned keywords: “cure” and “wine.” Although the keyword method may appear mechanical and somewhat uninspired, it is the method’s reliance on association and chance that, nevertheless, creates numerous exegetically inspiring links between excerpts.

Like a typical late antique story, the narrative has a thematically compelling plot, while at the same time being divided into independent scenes or miniature stories. Each miniature (labeled with letters above) contains a distinct plot and contains an exegetical, instructive, or moral statement of its own. Even the parts D^G and G^D, which clearly belong together, have been fashioned as two pieces that can be used individually, just as they actually are. It seems as if the author of the story was already generating excerpts for the next composer. Indeed, not infrequently, authors would reuse such excerpts themselves in future compositions.Footnote ⁹²

The story’s patterning further seems reflective of the writing surface on which it was composed: wooden tablets. Tablets were not only suitable for the composition of such concertina-like stories but dictated this very style. Indeed, if we think of the story’s original surface as a wooden codex or a concertina-like notebook (polyptychon) that consist of tablets that are strung together, we can easily imagine each tablet to contain one scene. If so, excerpting and rearranging was not only a natural but also a noninvasive act. The tablets, each carrying a miniature story, could easily be detached, rearranged, tagged with a keyword, stored, and repurposed.

These material factors were simultaneously the cause and effect of late antique compositions, suggestive of the constant reuse of – literally – bits and pieces of stories in other compositions. For the proem of the qordiaqos commentary in b. Git. 67b, for instance, the composer of the Talmud excerpted one of the miniature stories of the Solomon-Ashmedai cycle, changed the name of the main protagonists, and substituted a medical recipe for the dialogue. This claim obviously needs substantiation since the story about Rav Amram being chased through the snow by the exilarch’s household (b. Git. 67b) seems at first unrelated to the Ashmedai-Solomon cycle. The connection appears only after a thorough analysis of the characters and their verisimilitude. In late antique stories, “verisimilitude” refers to conformity with what is known about a certain character from earlier works. In the case of Solomon and Benaiah, the sources are the Hebrew Bible and earlier rabbinic traditions; in the case of Ashmedai, it might be the book of Tobit together with earlier rabbinic traditions.

As pointed out previously, verisimilitude was highly valued and encouraged by the progymnasmata because it was the decisive factor for the success of a story, the focal point of critics. In what follows, I will show how carefully, and in how much alignment with known biographical “facts,” the characters in the present story were created. The connections between the above piece and the Solomon-Ashmedai story cycle can then be shown to have been based on careful research into its protagonists’ prior literary lives.

Creating a Plausible Character

Solomon is a well-known biblical figure, and he appears in the above narrative in accordance with the biblical description of him as the king who built the temple in Jerusalem. He is also regarded here and elsewhere in the Babylonian Talmud as the author (and “I”) of the book of Ecclesiastes/Qoheleth (Shabb. 30a–b). Other than that, Solomon is rarely mentioned in rabbinic works, as Gerhard Langer’s study has shown.Footnote ⁹³ The portrayal of Solomon is, in fact, ambivalent throughout rabbinic literature, which is consistent with the biblical account of his persona. The Bible says that Solomon built the temple (1 Kgs. 9), and God blessed him with incomparable wisdom (1 Kgs. 10), yet he loved women more than the God of Israel and committed idolatry in his later days (1 Kgs. 11). The story in Gittin presents Solomon exactly along these lines; he has a wanton lifestyle, including many women (E/H), but he is also wise enough to interpret Ashmedai’s riddles (E) and to equip Benaiah with the tools necessary to catch the demon (C). Then again, he is not able to expound Scripture and foolishly asks Ashmedai for explanation (H). In addition, Solomon repeatedly depends on the help of the rabbis, first to tell him about the shamir (A) and then to restore the kingdom to him (H).

The dominant motif in segment (H), Solomon’s replacement on the throne by a nonhuman being, is borrowed from a story recorded in the Palestinian Talmud.Footnote ⁹⁴ Again, the Babylonian story takes a different turn than the Palestinian one, where the rabbinic sages beat up Solomon (y. Sanh. 2:6) rather than helping to restore his kingship. Instead, part (H) is clearly designed to support the quote with which it ends, a saying attributed to the Babylonian sage Samuel: Solomon was first “a king and [then] a commoner and [again] a king” (// b. Sanh. 20b). This move is suggestive of the author’s obligation to incorporate everything that was known about a situation or a person. Indeed, the way in which Solomon’s character is constructed points to the author having thoroughly examined biblical and extrabiblical sources concerning Solomon before he (or she) began to compose the story – just like the composers of the Talmud who went through their archive. Interpretations of 1 Chr. 29:23 and 1 Kgs. 5:4, now rendered in b. Sanh. 20b and b. Meg. 11b, for example, state that Solomon reigned first over upper and later over lower beings. This “fact” has been incorporated into Solomon’s constant struggle over power with Ashmedai. Both Talmuds also state elsewhere that Solomon used the shamir to build the temple from hewn stones, while Moses used it to make the efod (b. Sotah 48b; y. Sotah 9:13–14). There was even more information available about the shamir: In order to be stored safely and without harming anyone, the shamir needed to be wrapped in tufts of wool, placed in a box made of lead that was filled with bran from barley (b. Sotah 48b). Even that knowledge shaped the course of the present story. Although the story does not specify how the shamir was hauled to Solomon after it was obtained from the bird, the Aramaic equivalent of the Hebrew term for “tufts of wool” in the baraita in b. Sotah 48b appears in the list of equipment given by Solomon to Benaiah (C).Footnote ⁹⁵ The connection of a bird with the shamir is made in a baraita in b. Hul. 63a, where a dukifat-bird is said to have brought the shamir to the temple.

Athough the story alters and recontextualizes some elements, it stays completely within the plausible, taking into consideration what has been said earlier about a certain character or topic. In fact, everything that was previously said about a pertaining subject is taken into consideration at some point or another in the story. It looks like the author of this particular story had an archive at his disposal that was at least very similar to the one used by the composers of the Talmud. It is also conceivable that there existed tables and lists, created by users of the archive as they were studying, and that indicated which documents contained information on a given subject or even on a mishnaic term or lemma.Footnote ⁹⁶ Whatever the auxiliary tool may have been, it seems that both the author of this story and the composers of the Talmud had access to the same sources: the story neither misses a reference to Solomon present in the Talmud, nor does it add a completely novel feature to his character. The same is true for the other two main characters, as we shall see.

Compared to King Solomon, Benaiah son of Yehoiada is a rather marginal figure in the Bible, with “only” eighteen mentions. Correspondingly, he is not often mentioned in the Babylonian Talmud, but when he is, the references are very precise. In fact, every single one of Benaiah’s outstanding actions is remembered in the Talmud: that he smote the two “lion-like men of Moab,” that he killed a lion in a pit on a snowy day, and that he slew an armed Egyptian with the latter’s own spear (2 Sam. 23:20–21, expounded in b. Ber. 18a–b).Footnote ⁹⁷ He is also compared to a robber (b. B. Qam. 79b) and to the Jewish court, the Sanhedrin (b. Ber. 3b–4a, in an interpretation of 1 Chr. 27:34). Both of these instances refer to the fact that Benaiah served as Solomon’s assassin (e.g., 1 Kgs. 2:25, 29). Only the biblical mention of Benaiah as famous among the thirty brave men at David’s court is not reiterated in the Talmud (2 Sam. 23:22–23).

In complete agreement with these biblical characterizations, Benaiah serves in the story under discussion as Solomon’s hero, whom the latter sent to catch a lion-like figure, Ashmedai. Indeed, the king of demons seems to be of gigantic size when he is described brushing against trees and bending over a house (D^G), and animal-like when he hides his nonhuman feet in shoes (H). In addition, just as the biblical Benaiah kills a lion in a pit (Hebr. bor, 2 Sam. 23:20), Benaiah catches Ashmedai by digging one pit (Aram. bira) above and one beneath the cistern.

Ashmedai, by contrast, is absent from the biblical plot. His name is the Aramaic translation of the Greek name “Asmodeus,” a demon who appears in the Hellenistic novel Tobit. The Asmodeus in Tobit kills the female protagonist’s newlywed husbands on their wedding night, seven in a row. By contrast, the Ashmedai of the Gittin story cries over the bride who will soon be a widow and will have to wait thirteen years to marry the yavam, the groom’s younger brother (D^G/G^D). He is portrayed as studying his daily portion of Torah in the heavenly academy, as well as in the academy on earth (B). He lives in a sober manner (C) and reasons based on biblical verses (C). Rather than destroying others, he hurts himself (D^G) and gives medical advice (E), while Tobit had to be given advice by an angel on how to get rid of Asmodeus. In his secluded lifestyle and seemingly deep and supernatural knowledge, Ashmedai comes much closer to the description of an anchorite monk in stories of the time than to contemporary portrayals of demons.Footnote ⁹⁸ He predicts Solomon’s end in the same way it is described in the Bible (F). When he tricks Solomon and usurps his throne, Ashmedai exposes the king’s foolishness as much as his own deviousness, and when he solicits Solomon’s wives, it is Solomon’s accumulation of women rather than Ashmedai’s behavior that is criticized. The only thing that is, from the perspective of rabbinic teaching, really worthy of contempt is that he does not care whether or not the women he is soliciting are menstruating.

If Ashmedai is not or at least not only Asmodeus by character, then who or what is he? The only other instance in which the Talmud mentions Ashmedai is quite telling. In b. Pesah. 110a, Ashmedai is said to be the king of the demons.Footnote ⁹⁹ A subsequent commentary discusses the nature of a king; some hold that “king” does not refer to someone causing harm, while others hold that a king is quick-tempered and does what he wants. The passage seems suggestive of the ambivalence that surrounded Ashmedai’s kingship, as well as kingship in general. In line with the excerpt now found in tractate Pesahim, the author of our story describes Ashmedai as unstable and untrustworthy like Solomon. Like Solomon, Ashmedai gets drunk, lusts after women, is at the same time pious and friendly, and knows remedies and the future. He is wise and foolish, kind and evil, rises to power and loses it again.

Additionally, Ashmedai is portrayed as being Persian throughout the story.Footnote ¹⁰⁰ He seals the cistern with his gushpanqa, while Solomon gives his yzqta to Benaiah. Both terms refer to a signet ring, but one word is of Persian origin (gushpanqa), while the other is Aramaic (yzqta). Ashmedai also wears Persian mōg-shoes to hide his feet. He is obviously Persian, even a Persian demon, as other details reveal: He has wings like Persian demons and dragons do, and he swallows Solomon, just as, in Middle Persian literature, the demon A̅z swallows Xēšma, or Ahriman swallows Tahmuras.Footnote ¹⁰¹ Like Persian demons, Ashmedai dwells on a mountain.Footnote ¹⁰² In fact, even the name “Asmodeus” is apparently a Greek translation of “Aēšma,” the name of the Persian demon of wrath. Based on the above-outlined late antique habit of writing a story based on another, it should be asked if maybe a story of Persian/Sasanid origin was decisive for Ashmedai’s character and the plot.

Finding the Template

The progymnasmata did not encourage students to write freely. Instead, students transformed stories by substituting characters or dialogues, adding a moral, or merging one plot with another. If the authors of talmudic stories had a background in rhetorical training, then they would have used templates for their stories as well. Indeed, many stories in the Babylonian Talmud appear to have such templates in the Palestinian Talmud. Then again, many do not. Except for the already discussed motif of Solomon’s replacement by a nonhuman being in (H), for example, the scenes in the story cycle under discussion have no parallels in the Palestinian Talmud. We must therefore look elsewhere to find the model story or model stories for the scenes. Judging from Ashmedai’s Persian attributes, at least some templates might come from Sasanid lore.

Sasanid lore is replete with stories of human heroes fighting demons, and one of these stories is indeed a near-complete match to one of the miniature stories in the Gittin commentary on qordiaqos. Yet, surprisingly, the episode is not found in the Solomon-Ashmedai story cycle (b. Git. 68a–b) but, rather, in the proem of exactly this commentary (b. Git. 67b). Reference is made to the above-cited story about Rav Sheshet, who flees from the exilarch’s servants. Although the Solomon-Ashmedai story cycle seems to be sound and complete in its present state, one piece had been excerpted, and is now part of the proem together with several other excerpts that foreshadow the major “arguments” that will follow (see the discussion in Chapter 3). As discussed above, the distinct shape of late antique stories allows excerption of scenes without damaging the composition as a whole.

The template for this excerpt is found in Ferdowsi’s Šāh-nāma, a tenth-century collection of stories concerned with the deeds of Sasanid heroes and demons. Although the source is considerably later than the Babylonian Talmud, it is obvious that Ferdowsi did not invent the stories from scratch, since the Avesta already alludes to some of them. Rather, Ferdowsi collected the stories, wrote them down or rewrote them in verse form, and “composed the innumerable speeches he put into the mouths of his heroes, as well as the many long letters written at the dictation of the kings and other principal characters.”Footnote ¹⁰³ What may, therefore, be used for comparison with talmudic stories is the very basic storyline of Ferdowsi’s narratives, not his embellishments.

The story in question, however, does not feature a figure mentioned in the Avesta but a popular Iranian hero named Rustam. Rustam, whose origins are still unclear, became the hero of many fights with demons. The onomastic and literary evidence points to the fact that “the Rustam legend was fully formed and well known in Western Iran by the seventh century.”Footnote ¹⁰⁴ The parallel in the Talmud would represent an early testimony to these Rustam legends.Footnote ¹⁰⁵

In this particular story, King Tahamtan sends Rustam to capture the demon (dı-v) Akvān, who had appeared among the king’s herds in the shape of an onager. Rustam is given a royal lasso in order to capture the demon alive. When Rustam sleeps near the cistern, Akvān digs a hole around him until he can seize Rustam together with the soil on which he sleeps and carries him high up into the air.Footnote ¹⁰⁶ Up in the air, the demon asks Rustam where he wants to be dropped: on a mountain or into the sea? Rustam reasons to himself that the demon will most likely do the exact opposite of whatever he tells him.Footnote ¹⁰⁷ Thus, while secretly choosing the sea, he tells the demon to cast him onto a mountain. Immediately the demon drops him into the sea. Rustam swims back, fighting off the sea monsters with sword in hand. When he comes back to the cistern, he catches Akvān with his lasso and brings him to King Tahamtan, who finally slays him.Footnote ¹⁰⁸

Based on this narrative, it is possible to suggest that when the Gittin story cycle originally reached the hands of the composers of the Talmud, it depicted Ashmedai chasing Benaiah through the snow, in an episode between what are now segments (B) and (C). Accordingly, Benaiah initially failed to capture Ashmedai but then outwitted him by filling up his cistern with wine and only then succeeded (C). The excerpt under question that was detached from the story cycle and used in the proem reads:

If Rav Amram the Pious is replaced with Benaiah, we also find an explanation for the somewhat unusual appearance of snow in the story. As mentioned above, the Bible states that Benaiah killed a lion in a pit on a “day of snow” (2 Sam. 23:20). Considering the fact that the Aramaic word for “snow,” talga, appears only five times in the entire Talmud, this cannot be a coincidence. We may therefore conclude that it was not Rav Amram who was chased through the snow by the exilarch’s servants; it was Benaiah who was pursued by Ashmedai. Benaiah, the biblical slaughterer of lion-like men and a lion, seems, indeed, to have been a fitting and carefully chosen cultural translation of Rustam, who, in turn, is famous for seven heroic deeds, one of which is the killing of a cruel lion.Footnote ¹⁰⁹

Based on the above analysis, it appears that the composers excerpted one of the miniature stories of the story cycle, changed the names of the protagonists and the dialogue, and placed a recipe against freezing into the new main protagonist’s mouth. It remains to be asked why the composer specifically chose to substitute Benaiah for Rav Amram the Pious. After all, the two figures do not seem to have much in common. But the choice was most likely not motivated by Rav Amram’s character but by the excerpt the composers wanted or had to use after the one about the chase through the snow. Such a technical and practical motivation would also explain the friction between the two excerpts, which remains despite the alignment of names. As the two excerpts are arranged now, we learn that after Rav Amram successfully tricked the exilarch’s servants into giving him food that sustained his body temperature (fatty meat on coals and undiluted wine), “Yalta heard. She brought him into a bathhouse. She made him stand in the water of the bathhouse until the water of the bathhouse changed and became blood, and it fell off from him ‘coin by coin’” (b. Git. 67b).Footnote ¹¹⁰ There is an obvious inconsistency between Rav Amram successfully tricking the exilarch’s servants into giving him what he needed to stay warm (fatty meat on coals and undiluted wine) and the woman Yalta hearing that Amram was in pain. In an obvious rupture with the prior plot, the sequel negates the prequel’s punchline. It is in instances like these that we encounter the sort of paper-cut method implemented by the composers: the Yalta sequel must be an excerpt or literal piece from another story. This notion is further substantiated by the lack of the usual semantic puns that often connect independent scenes.Footnote ¹¹¹

If the thesis outlined in this book is more or less correct, and the composers worked with excerpts, it should be possible to find the rest of this Yalta and Rav Amram story. As it turns out, there is indeed a story that is a much better fit for the scene in which Yalta takes Rav Amram to the bathhouse:

Although one might argue that Rav Amram was already free from his evil inclination by the end of the story in Qiddushin as well, several things tie in better here. First, it is a woman, Yalta, who wants to make sure that Rav Amram has really eradicated his uncontrolled lust for women. Second, the therapeutic measure, water, stands in direct relationship to the cry “Fire in the house of Amram,” which, Reuven Kiperwasser has shown, is a metaphorical reference to Rav Amram’s body.Footnote ¹¹³ Third, a therapy against extended sunstroke, rendered in the proem of the commentary on qordiaqos and placed right before the story of Rav Amram in the snow, states that the patient should “go down and stand in water until his world becomes weak” (b. Git. 67b). This is exactly what Yalta does here, with the telling difference that Rav Amram is cold in the Gittin story, and only in the Qiddushin story is he hot (“on fire”) and in need of the indicated treatment. Fourth, there is a pun on the word “flesh” (bissra): while Rav Amram claims that his flesh is preferable to the fiery yetser, Yalta exposes the very weakness of this very flesh. It seems, therefore, more in accord with the literary ambition and aesthetics of talmudic stories to view the excerpt as an original scene of the Rav Amram story in Qiddushin rather than of the one in Gittin. This observation raises questions regarding the procedures and aspirations of the composers: how did this mix of story lines in Gittin happen, and how does it tie in with the composers’ modus operandi?

As discussed in Chapter 3, the composers, based on clues in the Palestinian Talmud, reached the conclusion that qordiaqos was a disease equal to a sunstroke that lasted three days. Accordingly, qordiaqos could be cured with the same therapy, “red meat on coals and diluted wine.” The composers then had to look for case stories with which to substantiate their claim, just as the Palestinian Talmud did with the story about the Tarsian weaver seized by qordiaqos. Although the text is corrupt, it appears that they gave him “red in something” and then “something in red” to drink. Going through their excerpt collection based on the selected keywords, they came across the Solomon–Ashmedai story with the scene in which Ashmedai asks Benaiah where he wants to be dropped, or something similar. The scene lent itself perfectly to such a recipe reversal. The fact that there was snow in the story allowed the composers to reverse not only the therapy but also the indication, as is reiterated by way of summary right before the excerpt: “Against the ‘sun’: red meat on coals and diluted wine; against the ‘snow’: fatty meat on coals and undiluted wine” (b. Git. 67b). The composers excerpted the passage with Yalta and Rav Amram in the bathhouse from the Qiddushin story, which was apparently classified under “cures.” They did this at first, perhaps, because they wanted to use the excerpt as proof of the efficacy of the recipe for extended sunstroke (“stand in water, etc.”) and then because it allowed them to show that the water therapy could also be used in reverse, against cold. The rest of the Rav Amram story, the part in which he is seduced by captive women, must then have been reclassified under “women” without the therapeutic part.

The Solomon–Ashmedai cycle is thus a good example of how the concise and independent nature of catena-like late antique stories allowed for migration into other contexts. In a few steps, including a plausible change of characters and a different dialogue, a whole new story could be created. Certainly, the composition of such a lengthy narrative required much research, since everything had to remain plausible and in harmony with earlier traditions, including the Hebrew Bible. Recognition of authors’ careful investigations into the prehistory of characters in talmudic stories, in turn, may help explain features that have left scholars puzzled. It may explain, for example, why Rav Kahana is called “Rav Kahana” in b. Bava Qamma 103a and not simply “Kahana,” as in the Palestinian parallel (y. B. Metz. 5:6, 10c): the Babylonian author aligned his take on the Palestinian with other traditions in the Babylonian, in which Kahana is called “Rav” throughout.Footnote ¹¹⁴ Inquiry into a protagonist’s previous literary life to stay in character further explains why the Babylonian exilarch seems to be modeled on the Palestinian one.Footnote ¹¹⁵

Summary and Repetition: Potential and Limits of Talmudic Stories

Jacques Fontaine has pointed out that late antique literature reveals an “extreme refinement” in the methods applied to generate allusion.Footnote ¹¹⁶ In the above example, we saw how allusions to other stories and motifs came into being, even though allusion was not what the authors of the stories had in mind. Rather, a careful search into the previous literary life of a character and the joining of these characters in any one story, together with the author’s method of writing based on templates, inevitably produced this dense and complex web of allusions.

The procedures applied by the authors of talmudic stories were complex but consistently the same. One was expansion by addition of a dialogue, namely, speech in character as taught in the progymnasmata.Footnote ¹¹⁷ This is a frequent feature of stories in the Babylonian Talmud that becomes obvious when (quasi-) parallel stories from both Talmuds are compared. But examples of such expansions by means of dialogue can also be found among stories within the Babylonian Talmud. The following example is a rare case in which the statement in Hebrew (baraita) that served as the basis for the expansion, as well as two expanded versions, ended up right next to each other due to their identical keyword (“demon”). The statement in Hebrew is as follows: “A single person should not go out at night, not on the night of the fourth day or the night of the Sabbath. Because Agrat bat Mahlat and eighteen myriads of angels of destruction go out [on these nights], and each and every one has permission to destroy in his own right” (b. Pesah. 112b). The account is brief and without dialogue. The only named character is the demoness Agrat bat Mahlat. Right after this baraita follows an expanded version:

The purpose of this short inquiry is to explain why these demons would swarm only two nights per week. The reason is given in the demoness’s encounter with Hanina ben Dosa, a figure known from the Mishnah. Hanina ben Dosa is a “man of the deed,” a man through whom and for whom God performs miracles.Footnote ¹¹⁹ This was apparently reason enough to choose him to stand up against the demoness.

Before analyzing the dialogue, a quick look at the progymnasmata may clarify the author’s task at hand, namely, ethopoeia, the “imitation of the character of a proposed speaker” (Aphthonius, Progym. §34R).Footnote ¹²⁰ The authors of the progymnasmata agree that the speech must be written from the perspective of the speaker in accordance with – and this point is most elaborated by Theon – age, gender, social status, and occasion, that is, with everything “aiming at what fits the speaker and his manner of speech and the time and his lot in life” (Progym. §116).Footnote ¹²¹ The exercise basically asks authors to step into the shoes, as it were, of the conversation partners in the plot and, in our case, must figure out how a demoness would speak, and what she would say to Hanina ben Dosa, and what he would reply. The result is a combination of what is known about the demoness from the baraita, namely, that she is extremely dangerous and harmful, and what is known about Hanina ben Dosa, a wonderworker in special proximity to God. Thus, while the demoness expresses her desire to kill Hanina, heaven refers to the rabbi as “son.” The motif of intermediary beings’ hearing announcements in heaven about humans is a recurring one (see passage G^D in the Solomon–Ashmedai story above, b. Git. 68b). The heavenly decree gives Hanina the authority to negotiate with the demoness, leaving her certain nights to roam. In this way, the expanded story remains in agreement with the baraita stating that the demoness and her army roamed the earth two nights per week. Thus, here as well, depending on how much is known about the characters, speech in character has many restrictions but also innovative potential through the choice of character.

A version of this story underlines the restrictive nature of speech in character when the Palestinian “man of the deed,” Hanina ben Dosa, is replaced by the Babylonian teacher Abaye:

The announcement in heaven emphasizes Abaye’s learning by identifying him with his teacher, Nahmani, thereby relating Abaye’s authority to his teacher. The learned Abaye, then, does not leave “room” for the demon. In that sense, the Babylonian looks more heroic than the Palestinian Hanina ben Dosa, but the story as such fails, since it does not prove the baraita right. This problem is fixed by the composer, who adds a commentary to the story. In a question-and-answer format, this commentary clarifies that the demoness would nevertheless roam about on the fourth night and the night of the Sabbath, yet only on narrow streets.

I would like to return now to the issue of reusing lines (exergasia), which is ultimately a miniature version of the use of templates and excerpting. For example, a structurally identical sentence to the one attributed to Agrat bat Mahlat, “If it were not publicly announced in heaven …,” is placed into the mouth of Satan in b. Qidd. 81a. Satan says: “If it were not publicly announced in heaven: ‘Beware of Rabbi Meir and his Torah!’ I would value your blood [only] as much as two coins!” In spite of their identical formulation, however, the three statements each make recourse to another threat. The example exposes the creativity inherent in the convention of reusing well-made and successful scenes or even sentences in a different way. Most of all, the method is, again, timesaving and economical. At the same time, the method may easily result in tedious repetitions.

Yet there seems to be a certain restriction in play with regard to repetition: parallelisms are usually executed in sets of three, even if they are dispersed all over the Talmud. To give some examples: In the story discussed above, Rav Amram’s skin falls off “coin by coin [peshitti peshitti].” In another story, a man swallows a snake that comes out after the treatment “piece by piece [guva guva],” while jaundice leaves an affected Arab “little by little [purta purta]” after the appropriate therapy.Footnote ¹²³ Two men, a demon, and a cedar “burst.”Footnote ¹²⁴ Illness, discharge, and evil inclination each issue forth like branches, but like different types of branches in each case.Footnote ¹²⁵ In three different stories, a wooden slip (pitqa) falls from the sky with a heavenly note on it.Footnote ¹²⁶ There are so many such stock phrases and familiar motifs that they evoke the impression that everything is connected to everything in the Talmud.Footnote ¹²⁷ According to the thesis for the formation of the Talmud put forward in this book, the sense of a web spanning the whole Talmud was generated by the detachment of stories and other pieces of information from their original compositions, in which the same idea or phrasing was reused several – mostly three – times, each time making a somewhat different point.

Late antique authors often reused catchphrases, and their stories repeatedly took similar turns and describe encounters between comparable protagonists or places. With regard to Lucian’s works, Graham Anderson referred to this feature as “self-pastiche.” Lucian repeatedly took his own literary creations as templates for new ones. He subjected his stories to the same methods of alteration and adaptation he had used when first fashioning individual scenes on the basis of scenes written by others.Footnote ¹²⁸ This constant alienation of the same plot prevents the establishment of a chronology between the stories and the reconstruction of their actual source or sources: One might be tempted to propose an external source when, in fact, Lucian simply reused his own work.

Lucian’s external sources seem to have been quite a manageable supply of works, including the Greek myths of Homer, some comedies, and the Platonic myths.Footnote ¹²⁹ Others, too, departed from and relied on such a comparatively modest set of works. The Neoplatonic philosopher Porphyry, for example, was able to list the set of works with which his teacher Plotinus had been engaged when writing his commentaries (Plot. 14.3).Footnote ¹³⁰ The authors of the stories included in the Talmud may have had a similarly limited supply of works at their disposal, including the Hebrew Bible, baraitot, and the Palestinian Talmud, along with some story and maxim collections.

So much literature has been lost, however, that it is impossible to even approximate the stories and collections available to these authors. In the Institutio oratoria, for example, Quintilian mentions the value of excerpting passages from the works of a comedian named Philemon (10.1.72). Said Philemon apparently wrote ninety-seven plays, of which not a single work survived in its entirety; only fragments or merely the titles of fifty-three of his works are extant. He also had a son who, under the same name, wrote a total of fifty-four plays, of which only two fragments and no titles survived.Footnote ¹³¹ The paradoxographical work Rivers and Mountains and What Is Found in Them, written around 300 CE, cites a wealth of authorities and works that are mostly unknown.Footnote ¹³²

One may wonder how such works would relate to the concise stories that we find in the Talmud, which often have the shape of plot summaries rather than fully fleshed-out stories. Indeed, even comparatively long stories in the Talmud as the one discussed above only cover a few folia. Yet it seems inconceivable that authors of talmudic stories would, after conducting exhaustive searches for plot, summarize whole books only to obtain a concise template for a story they then rewrote. The possibility of collections of ready-made plot summaries seems more reasonable. Such plot summaries began to emerge in the second century as prefixes to comedies.Footnote ¹³³

Because of the previously described loss of many relevant sources, it is rather rare that we come across an obvious case of plot-summary use. Yet a story in one of the talmudic commentaries discussed in Chapter 3 appears to build upon what may originally have been such a plot summary, in this case a summary of (pseudo-)Lucian’s version of Lucius’s metamorphosis into an ass:

Short as it is, the story contains the most prominent scenes of (pseudo-)Lucian’s novel Lucius or the Ass. These include the transformation of a human being into a donkey, a person of the opposite sex’s riding on it, and the public humiliation following the revelation of the donkey’s actual human nature.Footnote ¹³⁴ Like Lucius, Yannai is not at home when the transformation happens and, in both stories, women are involved in the metamorphosis in some way. Significantly, the actual process leading to transformation is different, most likely because (pseudo-)Lucian describes a method unmentioned by rabbinic literature. In this scene, Lucius rubs himself with oil, a practice that could too easily be confused with the biblical anointing of a king. Rather, substantial change is brought about in the Mishnah and elsewhere in the Talmud by murmuring.Footnote ¹³⁵ In keeping with these literary standards, here, too, murmuring charges the porridge with change-effecting potency.

All in all, it seems that the more rhetorical the role of the story is, the shorter it becomes, since the story is used as an argument, and not primarily to entertain people. The story about Lucius, how he turned into an ass, and his long period of suffering until he finally regained his human form, makes the same point as the short story about Yannai in the Talmud, namely, that witchcraft is not to be engaged in lightly. However, while (pseudo-)Lucian’s version of the story and the even longer Latin one by Apuleius elaborate on their morals in a verbose style that requires several hours of serious reading, the talmudic version makes an instant point, allowing for even more proof to be added to the same argument in a fraction of that time.

In conclusion, it can be said that the distinct style of the late antique story – a concise but apt scene that can be attached to others to create lengthy story cycles – seems to be the result of at least three major factors: (1) the territorial expansions of the Roman Empire that led to an increase in knowledge, and which then had to be condensed again in order to remain useful (Chapter 1); (2) the limitations imposed by accessible, convenient, and cheap writing material such as wooden tablets and other flexible, portable writing surfaces (Chapter 2); and (3) the impact of the rhetorical, and still court-influenced, curriculum, which focused on the argumentative potential of the story.

The content of the stories was shaped against two factors that could be seen as restricting the story’s potential: a set of authoritative texts that dictated and framed plausibility, and the author’s use of templates. Yet an author’s in-depth inquiry into characters and plot could unearth unexpected connections to other topics and lead to a substantial and informed transformation of the template. Authors, talmudic and else, seem to have been supported in their search for plot and moral by collections of stories, gnomologies, sayings, and glossaries.Footnote ¹³⁶ Both the limits and the potential of the late antique story resulted from the fixed set of methods outlined in the progymnasmata, all the way supported and pressed by the materiality of writing.

Conclusion

The purpose of this chapter has been to show that the Talmud is not only an erudite construct in its entirety but that this erudition is likewise mirrored in its parts, that is, the excerpts used by the composers. The compositional processes evident in talmudic stories, for example, are not much different from the ones applied by the composers of the Talmud to compile the work. Thus, the author of a story similarly started off with pieces of other people’s writings, such as a template story and one or more maxims or sayings, arranging the two in the most plausible way and enhancing them with details of an inquiry into the chosen topic’s or protagonist’s prior literary life. Flexible and size-limited writing surfaces, such as tablets, ostraca, papyrus scraps, and the like affected the morphology of the story, its concise style and individual scenes, but also facilitated the arrangement of several such scenes into a whole, and their possible subsequent rearrangement or exchange.

Like Chapter 3, this chapter has highlighted the deep connection between rhetoric and writing, as well as the relationship of late antique rhetoric to its original purpose, namely, advocacy and defense in court. The stories of the Talmud mostly – if not always – argue for something. The dialogues they feature are elaborate, sharp, and filled with clever repartee: exemplary rhetoric, in sum. We might, therefore, ask whether some, if not all, of the texts used in the Talmud are the leisurely product of rhetorically trained men, or men in rhetorical training.

Indeed, Catherine Hezser has described the most typical talmudic stories as “case-stories consisting of a case-description … and a decision part.”Footnote ¹³⁷ The following passage, which has two quasi parallels in the Palestinian and one in the Babylonian Talmud, is a good example. The parallels illustrate not only the adjustments made by the respective composers – a corrective discursive note in the Palestinian Talmud (y. Pesah.) and a discursive introduction in the Babylonian – but also how scribal methods and content were taught.Footnote ¹³⁸

As is the case with most stories in talmudic literature, the stories of the following example are not integrated without slight friction into the commentaries in tractates y. Bava Metzi’a, y. Pesahim, or b. Pesahim. The texts were obviously written for their own sake and not to fill their present spots. This justifies thinking of them as individual texts, that is, school exercises. If so, they were not excerpts taken from longer texts but, rather, were stored directly with the tablet or ostracon on which they were written.

I would suggest the following scenario to explain the shape of this and many similar short exercises. First, the teacher discussed the case (now found in y. Pesahim) with the students. Then they dictated the case to the class and asked the students to devise a resolution for it: Somebody deposits bread with a sage and does not collect it before Passover Eve. Who would the sage ask for advice and what would be the ruling?Footnote ¹³⁹ Such an example mostly tested the ability to build a conclusive argument. All three examples succeeded, which was most likely the criterion for their inclusion in the Talmuds. In terms of elaboration and style, the Babylonian example clearly surpasses the other two with its miniature ekphrasis, a vivid description, regarding the bag damaged by mice, as well as with the suspense created in the last part, when the answer is withheld until the right time has come to act. Rhetoric is about the ability to produce a sound argument. The topic with which rhetoric is most profitably matched is law, its original source. Rabbinic concern for law and rhetoric were a perfect match.

In conclusion, I would like to suggest that scholars of rabbinic texts increasingly treat rhetorical training as the cause and effect of rabbinic literature, instead of thinking of rhetoric as something that merely left traces in that literature. Rhetoric defined literate and argumentative thought; it was not merely a method for composing and performing orations. In 1949, David Daube made a similar claim: “Hellenistic rhetoric is at the bottom both of the fundamental ideas, presuppositions, from which the Rabbis proceeded and of the major details of application, the manner in which these ideas were translated into practice.”Footnote ¹⁴⁰ There may be a wealth of different cultural influences in the Talmud, but the way in which they were analyzed, scrutinized, and matched with older traditions is clearly based on educational principles outlined in the progymnasmata. These, as we have seen, were adopted and translated freely by other language cultures, who detached them from their basis in Greek grammar and myth.