Composing Arugat ha-Bosem: How Piyyut Commentary Became Associated with Ḥasidei Ashkenaz

Based on Ivan Marcus’s concept of “open book” and considerations on medieval Ashkenazic concepts of authorship, the present article inquires into the circumstances surrounding the production of SeferArugat ha-Bosem, a collection of piyyut commentaries written or compiled by the thirteenth-century scholar Abraham b. Azriel. Unlike all other piyyut commentators, Abraham ben Azriel inscribed his name into his commentary and claims to supersede previous commentaries, asserting authorship and authority. Based on the two different versions preserved in MS Vatican 301 and MS Merzbacher 95 (Frankfurt fol. 16), already in 1939 Ephraim E. Urbach suggested that Abraham b. Azriel might have written more than one edition of his piyyut commentaries. The present reevaluation considers recent scholarship on concepts of authorship and “open genre” as well as new research into piyyut commentary. To facilitate a comparison with Marcus’s definition of “open book,” this article also explores the arrangement and rearrangement of small blocks of texts within a work.

composed by known (or anonymous) authors, on the other. Scholars have invested great effort in search of the respective Urtexts, while consideration has also been given to the possibility that transmission might account for major variants. 4 Unlike medieval Christian vernacular literature, where the quest for the author follows from the search for the stable text, 5 in the study of Jewish culture in the Middle Ages, the Hebrew medieval texts are oftentimes the less prominent entity; they serve as vehicles for gaining knowledge about the author. 6 To account for major differences between versions while preserving the authority of the author intact, Israel Ta-Shma suggested the term "open book" in describing multiple editions of the same work created by an author who continued to develop, edit, and emend his work even after publishing a version that circulated in manuscript. Owing to certain characteristics of Ashkenazic textual transmission and aggressive intervention by editors and scribes, Ta-Shma had to allow for other actors involved in the variant versions. 7 In pursuit of the process of production of an Ashkenazic  UK, 2006), 77-102, 79: "Textual philology requires an author. Without an author, there can be no philology." 6 Cf., e.g., the approach in Avraham Grossman, The Early Sages of Ashkenaz: Their Lives, Leadership and Works (900-1096) [in Hebrew] (Jerusalem, 1981); Avraham Grossman, The Early Sages of France: Their Lives, Leadership and Works [in Hebrew] (Jerusalem, 1995); Ivan G. Marcus, Sefer Hasidim and the Ashkenazic Book in Medieval Europe (Philadelphia, 2018). For a critique of this approach, see the contribution of David Shyovitz to this collection. 7 Israel M. Ta-Shma, "The 'Open' Book in Medieval Hebrew Literature: The Problem of Authorized Editions," Bulletin of the John Rylands Library 75, no. 3 (1993): 17-24. This understanding of "open book" is not related to Umberto Eco's concept of "open book" in which the reader contributes to the meaning of the text; see Umberto Eco, Opera aperta: Forma e indeterminazione nelle poetiche contemporanee (Milano, 1962). The additional actors in the work, Ivan Marcus recently redefined the term "open book" as follows: The meaning I am giving to "open book" in the case of Sefer Hasidim and of many other Ashkenazic books refers to an author: • composing a work in short text units that he sometimes rewrites; • combining them disjunctively (without linear coherence); and • producing more than one parallel edition that the author or someone else revises one or more times. The term "open book" here refers to writing parallel editions of a book so that there never was only one original edition from which the others are derived. 8 In his definition, Marcus emphasizes the role of the "author," yet also identifies two other important features regarding the composition of texts in Ashkenaz: small units that, following a suggestion by Simcha Emanuel, 9 he compares to Lego blocks, and parallel editions that do not appear to be derived one from another. Marcus envisions a process of assembling parallel but independent editions that becomes possible, on the level of textual composition, when there is no strong, predefined order to guide the author's hand, thus giving him carte blanche to select those units he wishes to combine. The author who sorts his own collected material-even before Luhmann's Zettelkästen-has the power to shape it into different works or different editions of the same overarching idea. The author, it seems, survived Barthes' proclamation of his death, 10 but philology has to be reinvented since the quest for the work (opera), be it an Urtext or the variants created by transmission, is futile for the works that Marcus associates with his definition. Since he searched for works consisting of small units, easily shifted and relocated within the text, most of his Ashkenazic formation of texts would function as "model author." For this concept, see Umberto Eco, Six Walks in the Fictional Woods (Cambridge, MA, 1995). 8 Marcus,Sefer Hasidim,5. 9 Ibid., 7, 127 n. 25. The comparison is apt because it refers not only to the stability and instability of the product, but also includes the fact that the multifunctionality of the blocks is limited only by their shape. Hanna Liss, "Copyright im Mittelalter? Die esoterischen Schriften R. El'azar von Worms zwischen Traditions-und Autorenliteratur," Frankfurter Judaistische Beiträge 21 (1994): 81-108, 87, suggests the simile of a kaleidoscope in which the fragments do not have a defined place in the text and can be reshaped in ever new formations. 10 Virginie Green, "What Happened to Medievalists After the Death of the Author?" in idem, The Medieval Author in Medieval French Literature (Basingstoke, UK, 2006), 205-44. Hans Ulrich Gumbrecht, The Powers of Philology: Dynamics of Textual Scholarship (Urbana, IL, 2003, 6-7, stated this as principle: "Text editing conjures up the desire of embodying the text in question, which can transform itself into the desire of also embodying the author of the text." See also the contribution of David Shyovitz to this volume. examples are normative texts, and only a few examples are exegetical, a type of text that, in any case, is often bound to the text it seeks to explain. 11 One of the few exegetical works that Marcus mentions is Sefer Arugat ha-Bosem, a collection of piyyut commentaries written by Abraham b. Azriel. 12 The following inquiry into the possible conditions of its production examines the complicated relationship between a group of textual witnesses usually referred to by the title of a work (Arugat ha-Bosem) and the person addressed as its author (Abraham b. Azriel).
Abraham ), and whose oral teachings he quotes several times in his work. He did not, however, quote from Sefer H . asidim, to which it appears he did not have access. He did have access to Sefer ha-Kavod, ascribed to Judah he-H . asid, which he cites numerous times, apparently from a copy that also contained interpretations of Eleazar of Worms. In fact, the many references to Eleazar b. Judah in Arugat ha-Bosem indicate that Abraham b. Azriel studied under him and considered him to be one of his primary teachers, but it is not clear when and where he studied with the Western Ashkenazic sage. Unlike all other piyyut commentators, Abraham ben Azriel inscribed his name into his commentary and claims to supersede previous commentaries, asserting authorship and authority. His relationship to both Judah he-H . asid and Eleazar b. Judah and his attitude towards authorship make Arugat ha-Bosem an excellent example with which to study the reception of Judah's understanding of authorship and authority. 13 In the 1939 introduction to the first volume of his monumental edition of Arugat ha-Bosem, Ephraim E. Urbach suggested that Abraham b. Azriel might have written more than one edition of his piyyut commentaries, basing 11 Marcus, Sefer Hasidim, 78-86. Clearly, commentaries compiled from different sources are also composed from rearranged short blocks, but compilers were often anonymous and did not claim authorship. The notable example is H . izkiya b. Mano'ah . , who proudly lists the many sources he compiled in his Torah commentary; see Haym Dov Chavel, ed., H . izkuni: Perushe ha-Torah le-Rabbenu H . izkiah bar Mano'ah . (Jerusalem, 1981) and Sara Japhet, "Perush ha-H . izkuni la-Torah: Demuto shel H . ibbur u-Matarato," in Sefer ha-Yovel le-Rav Mordekhai Breuer: Collected Essays in Jewish Studies, ed. Moshe Bar-Asher (Jerusalem, 1992), 91-111. 12 Marcus, Sefer Hasidim, 82. In note 40 (p. 156) Marcus offers brief descriptions of the presentation of texts in the two manuscripts and points to Urbach's more complete description of the manuscripts. 13 See Ephraim E. Urbach, ed., Sefer Arugat ha-Bosem Auctore R. Abraham b. R. 'Azriel [in Hebrew], 4 vols., (Jerusalem, 1939(Jerusalem, -1963 his theory on evidence of the two different versions preserved in MS Vatican 301 and MS Merzbacher 95 (Frankfurt fol. 16). 14 He developed this idea further in the fourth volume of the edition that served as his introduction to the corpus, which was published in 1963. 15 Urbach noted the differences in the length and accuracy of quotes from rabbinic texts in the two versions, as well as the number of commentaries on verses and midrashim introduced as proof texts. While acknowledging that these could be due to additions and deletions by scribes, Urbach claimed that the differences in the system of cross-references are the result of conscious editing and rewriting by Abraham b. Azriel himself. MS Vatican has many cross-references to commentaries on other piyyutim, both before and after the current commentary. MS Merzbacher, by contrast, has hardly any cross-references to explanations that occur after the current commentary and only a few to piyyut commentaries that occur before the insertion point of the cross-reference. According to Urbach, this systematic difference could not have been created by later scribes. Moreover, while in MS Vatican there are many cross-references to commentaries on piyyutim that follow in the manuscript, and the comments exist in the later place referenced, these comments occur in MS Merzbacher only in the first occasion, i.e., where the reference is found in MS Vatican, but are missing in the place to which MS Vatican references. Therefore, Urbach concluded, MS Merzbacher represents an earlier edition, and MS Vatican a later, enlarged edition, of the "authored" text. 16 In other words, after having "published" the first edition, Abraham b. Azriel added new sources, commentaries, and additional references to his commentary, indicating their character as secondary explanations. In instances where he found to have repeated himself, he deleted the commentary in one place and placed a cross-reference to 14 See ibid, 1:13* n. 1. The Vatican manuscript was known and mentioned in several bibliographical works, but the first publication that treats this piyyut commentary from the manuscript source was Abraham Berliner, "Aus den Bibliotheken Italiens I," Magazin für jüdische Geschichte und Literatur 1 (1874) the location where he considered the commentary to fit best or where his version of the commentary was more complete. 17 The process as analyzed and described by Urbach is similar to the "open book" as defined by Ta-Shma. However, in the case of Arugat ha-Bosem two factors complicate the situation: first, the open genre of piyyut commentary that was an important source for Abraham b. Azriel, and second, piyyut commentaries related to Arugat ha-Bosem, such as the commentary in the Nuremberg Mahzor. While Urbach did take note of both these factors in his 1963 introduction, a reevaluation is necessary that will take account of recent scholarship concerning concepts of authorship and "open genre" and new research into piyyut commentary. Additionally, to facilitate a comparison with Marcus' definition of "open book," the issue of the arrangement and rearrangement of small blocks of texts within a work, a feature of Marcus' concept, will also be explored.

Piyyut Commentary as Open Genre
Piyyut commentary was a very popular genre in medieval Ashkenaz during the eleventh to fourteenth centuries. 18 A few explanations of piyyutic phrases and lines have been transmitted that can be attributed to early scholars such as the Provençal Rabbi Menahem b. H . elbo (uncle of the more famous eleventh-century Rabbi Joseph Kara) and Meir b. Isaac, but the beginning of written piyyut commentary as a productive genre actually took place in Rashi's academy in Troyes and is associated mainly with Joseph Kara and Shemaya. 19 In addition to these two, it is clear that other scholars and students present at the academy took notes when Rashi explained piyyutim, resulting in several complete commentaries. 20 Transmission in this 17 Ibid., 4:132. Since Urbach's analysis, several fragments have been identified that contain parts of Arugat ha-Bosem and that confirm that different editions existed, even for passages that have been transmitted only in MS Vatican. One example was published by Simcha Emanuel, "Supplement to Arugat ha-Bosem, the Piyyut Commentary of Abraham ben Azriel" [in Hebrew], Kobez al Yad 25 (2017): 385-91. Emanuel was able to identify eight leaves from the same copy of Arugat ha-Bosem and suggests that two other fragments could have belonged to another copy, but both copies were used in binding fragments in Vienna. This is additional evidence for the Eastern Ashkenazic character of the collection. 18 Elisabeth Hollender, Clavis Commentariorum of Hebrew Liturgical Poetry in Manuscript (Leiden, 2005), lists more than 12,000 individual commentaries from Ashkenaz and S . arfat, transmitted in almost two hundred different manuscripts. On the genre and its compilatory nature, see eadem, Piyyut Commentary in Medieval Ashkenaz (Berlin, 2008). 19 Grossman, Early Sages of France, 256 and n 7. 20 See, for example, the case described in Elisabeth Hollender, "Commentary on a 'Lost' Piyyut: Considering the Transmission of Teachings and Texts in Rashi's Bet Midrash," in phase was mainly anonymous, even though commentators often attributed individual explanations to their teachers or other well-known scholars. 21 This anonymity likely contributed to the openness of the texts that were copied and edited freely. 22 The small textual units, which in Joseph Kara's piyyut commentaries were usually paraphrases, lexical explanations, and reworked quotations from midrashim, 23 result in a structure that allowed for insertions and deletions by later copyists and editors at many points in the text. 24 Neither the first commentators nor the later compilers and editors claimed that the texts they produced were "closed" works, nor did they claim authorship, largely preferring to remain anonymous. Instead, different versions of very similar commentaries existed, in which mostly anonymous editors and compilers not only chose what to include and what to exclude, but also reacted to, criticized, and rectified explanations in the commentaries they transmitted. In the thirteenth century, some compilers mentioned their names when adding their own voice and exegesis to the commentaries; others remained anonymous but identified themselves as the students of particular, named scholars. 25 However, this did not change the general character of piyyut commentary as an anonymous, open genre. Even when copied into independent manuscripts instead of onto the margins of a mahzor manuscript, 26 these commentaries were not "works" (opera): they had no title, named no author, showed no attempt at textual integrity, and were not quoted as a h . ibbur or sefer. 27 This changed, however, with the "publication" of Abraham b. Azriel's Sefer Arugat ha-Bosem. Considered to be the most famous work of piyyut commentary, it was the only such text to have a title and an author, the only text quoted as a sefer, and the only one that was subsequently published in a single critical edition based on two manuscripts.

Arugat ha-Bosem and the Open Genre of Piyyut Commentary
It is evident that Abraham b. Azriel considered his enterprise as a work (opera), giving it a title and wanting to be remembered as its author, as can be seen in the short introduction transmitted in MS Vatican: "I will bless him, who teaches intelligence to man, [to the] son of Rabbi Azrieli. I was helped by the master of heavens in my editing of this delight ( ‫ח‬ ‫פ‬ ‫ץ‬ ) which is called in Israel by the name 'Flowerbed of Fragrance'." 28 That the name and the conception of the text as "a work" were accepted by the audience is evinced by the fact that it was quoted as Sefer Arugat ha-Bosem by Hayyim b. Isaac Or Zarua' (Responsa, § 185). 29 Although this attitude was an innovation in the genre of piyyut commentary, Abraham b. Azriel was part of an extensive tradition of piyyut commentators and based large parts of his work on previously existing piyyut commentaries, selecting from among explanations that he liked and adding both linguistic explanations and exegesis of the texts that were considered to be the sources of the piyyutim (hypotexts). He acknowledged this in his introduction, where he wrote: 26 Except for individual glosses written on the margins of liturgical manuscripts and interlinear glosses, neither of which necessarily reflect the commentary traditions, the different physical forms of transmissions, i.e., full marginal commentaries or independent texts filling the whole page, did not influence the style of commentary. From this, we may also conclude that most marginal commentaries were not copied by the h . azanim who used the liturgical manuscripts. 27 The same holds true for earlier liturgical works known today as "works," such as Mah . it cannot be determined whether it did include an introduction or not. The fact that Abraham b. Azriel did preface his commentary with an introduction is a strong indication of his selfawareness as author. The few introductions to commentaries before him addressed questions of interpretation, such as Rashi's introductory paragraph to his commentary on Song of Songs. Previous piyyut commentators had identified themselves in auctorial notes in their commentaries but did not compose introductions. 29 See Berliner, "Aus den Bibliotheken Italiens I," 2.
And this garden which is sealed but not locked, this is its quality and its type: because with this the Elevated One broke [the wall of the garden] and disclosed the wisdom that was hidden from many for those who understand, listed and closed, old words, sweeter than honeydew, beautiful and precious, some of them cold [water] for tired ones, bound to the language of every man. And from among them is [taken] what I wrote in my own language, sometimes I extended, sometimes I shortened, so that the reader can hurry through them. 30 This dependence on earlier commentaries is less visible in his commentaries on yos . erot, since he includes many commentaries on piyyutim from the Eastern Ashkenazic rites that were not commented on in the earlier collections of commentaries. On the other hand, the commentaries on the qinot and selih . ot are closely related to the Northern French tradition of piyyut commentary attributed to Joseph Kara. These commentaries are short and mostly lack the long linguistic insertions and exegesis of the piyyut sources that characterize Abraham b. Azriel's commentaries on yos . erot, and they also contain many French glosses. Based on the difference, Avraham Grossman has argued that this section is not part of Abraham b. Azriel's composition and was added by a later scribe. 31 According to Urbach, the style of the section nevertheless shows the "voice" of Abraham b. Azriel, including some glosses in lashon Kenaan (Old Czech) and references to his teacher Eleazar and to Eleazar from Bohemia. 32 Whether by Abraham b. Azriel or by a later, anonymous compiler, it is apparent that these commentaries were produced in a manner similar to that of other editors and compilers of piyyut commentariesnamely the deletion of unwanted commentary elements and the insertion of some elements taken from the compiler's own studies. 33 As Urbach noted, the relation between MS Vatican and MS Merzbacher in the commentary on the selih . ot is different from that in the other parts of the work. Here, the differences are not those of a longer versus a shorter edition, as is the case in the commentaries on the yos . erot. Instead, according to Urbach, the versions could be considered drafts of parallel commentaries, as if Abraham b. Azriel had prepared an individualized copy from a traditional Vorlage with only minor changes. 34 While the incorporation of earlier commentaries into his own compilation is no different from what other editors and compilers did, it takes on a new meaning when Abraham b. Azriel turned his compilation into a book that bears his name, breaking with the traditional anonymity of the genre. It should be noted that the resulting commentaries clearly-and perhaps proudly-show their Eastern Ashkenazic origin, using Czech vernacular and references to Eastern Ashkenazic scholars, including Judah he-H . asid, to transplant the genre from its Western Ashkenazic origins to the new Eastern Ashkenazic centers in Regensburg and Prague.

Piyyut Commentaries Related to Arugat ha-Bosem
The collection of piyyut commentaries most similar to Arugat ha-Bosem is the marginal commentary in the Nuremberg Mahzor. This illuminated manuscript, probably copied in Regensburg in 1331, contains an extremely rich Eastern Ashkenazic rite and commentaries on more than five hundred piyyutim. 35 Even before Urbach, Bernhard Ziemlich, in his comprehensive description of the manuscript, compared these commentaries to the Merzbacher manuscript of Arugat ha-Bosem. 36 Following Perles's description of MS Merzbacher, 37 Ziemlich differentiated between its two distinct sections. He listed thirty-five commentaries that appear both in the first part of Arugat ha-Bosem and in the Nuremberg Mahzor, two of which he then excluded for various reasons. 38 Regarding the other thirty-three commentaries contained in MS Merzbacher, Ziemlich stated that the Nuremberg Mah-34 Urbach, Arugat ha-Bosem, 4:135. Following Grossman's argument for a later addition, one can only explain the differences between the two manuscripts by assuming that the addition was made to the manuscripts' Vorlage and that the two copyists exercised different editorial freedom in this section. 35 Dr. David Jeselsohn, the owner of the magnificent Nuremberg Mahzor, has initiated a new collection of studies that will contain a full description of the rite and the piyyutim by Yonah and Avraham Fraenkel. For a Hebrew version see their Prayer and Piyyut in Mahzor Nuremberg, https://web.nli.org.il/sites/nlis/he/manuscript/Documents/fraenkel_j_a.pdf). The following discussion summarizes parts of my contribution to this yet unpublished volume. 36 Bernhard Ziemlich, Das Machsor Nürnberg: Ein Beitrag zur Erforschung des Ritus und der Commentarliteratur des Deutschen Machsor (Berlin, 1886), 47-76. 37 Perles, "Das Buch Arûgath habosem des Abraham b. Azriel," 362. 38 He stated that they are very short or only the beginning of the commentary is identical; in the latter case the version in the Nuremberg Mahzor is longer than that in Arugat ha-Bosem. zor often skipped lengthy passages from Arugat ha-Bosem, namely the halakhic and grammatical excursa, quotations, references to names, and many of the Bohemian glosses. 39 Furthermore, in many cases, the Nuremberg Mahzor's commentary added supplementary material to passages from Arugat ha-Bosem. Ziemlich inferred that these parallels should be characterized as abbreviations of Arugat ha-Bosem. Another proof he cited are the two different commentaries on the ofan: yeh . av lashon by Judah ha-Levi transmitted in MS Merzbacher, the shorter of which is parallel to the commentary in the Nuremberg Mahzor. 40 He also identified eleven parallel selih . ot commentaries and rated these as bearing a closer resemblance to Arugat ha-Bosem than the other shared commentaries on piyyutim, which he explained by returning to the idea of an abridged version of the selih . ot commentary in MS Merzbacher. 41 It is possible, however, to read the evidence from the Nuremberg Mahzor differently. In the case of the selih . ot commentary, we may assumewith Urbach-that the texts in the Arugat ha-Bosem manuscripts are slightly edited versions of some tradition of piyyut commentary. This tradition could have (independently) served as the Vorlage for the editor of the Nuremberg Mahzor. Its use, therefore, does not in itself indicate the direct influence of Abraham b. Azriel's work on the Nuremberg Mahzor. 42 The situation becomes more complicated in the case of the commentaries on the yos . erot. Regarding several commentaries, the best explanation is that the Vorlage for the Nuremberg Mahzor was an early, shorter edition of Arugat ha-Bosem, lacking many of the long grammatical and exegetical passages so typical of the printed edition of the work. In most of those cases, more similarities exist with the edition in MS Merzbacher than with that in MS Vatican.
There is, however, also a group of commentaries in the Nuremberg Mahzor not closely related to Arugat ha-Bosem, whose features identify them as belonging to a H . asidei Ashkenaz tradition. They contain many references to Eleazar of Worms, 43 Ibid.,[52][53] Ibid., 53-55. This was obviously owed to the search for an Urtext and need not be discussed further. 42 Ziemlich wants to identify Abraham b. Azriel as author based on two phrases in one commentary and one Bohemian gloss in another commentary, where the parallel in MS Merzbacher has only the Old French gloss; see ibid., 54. This is not enough to assume a direct influence. 43 The references to him mention both oral teachings and written works, including piyyut commentaries and other compositions, such as a Sefer ha-Kavod attributed to him. the name of Shoh . er Tov, and refer to other, less common rabbinic works as well as medieval works such as Tosafot, Sefer Terumah, and Sefer Zerubavel. These commentaries show greater interest in gematriyot and other numerological explanations than earlier commentaries and refer more often to Sefer Yes . irah and Sefer Hekhalot. These features, unknown in earlier piyyut commentary, suggest an attribution to the second or third generation of H . asidei Ashkenaz, anonymous students of Eleazar of Worms, who were probably Eastern Ashkenazic contemporaries of Abraham b. Azriel. 44 Similar commentaries can be found in a number of other manuscripts. While they quote the tenth century Iberian Menahem b. Saruq's Mah . beret (a text already known to and used by Rashi in his piyyutim and biblical commentaries) and occasionally discuss the correct vocalization of a word, these commentaries do not exhibit the same interest in linguistic questions that characterized Abraham b. Azriel's work. Like most other editors and compilers of piyyut commentary, their compilers chose to remain anonymous, and they did not arrange their commentaries into a book. This makes Abraham b. Azriel's decision to compose a titled book published under his own name even more noteworthy. It also raises the question of the sources he may have used and edited to include almost contemporary commentaries in his composition. In addition to the many piyyut commentaries from Western Ashkenaz in existence by the end of the twelfth century, were there also commentaries by Eastern Ashkenazic authors or even by the first generation of H . asidei Ashkenaz that he could have used? If so, in what form would they have existed? 45

Small Blocks and Textual Order
The topic of materials for his composition from preexisting sources takes us back to the small units (the Lego metaphor) defined by Marcus as central to 44 That the Nuremberg Mahzor transmits commentaries from this tradition on piyyutim to which the principal other commentary, Arugat ha-Bosem, shows neither parallels nor even similarity, demonstrates that these commentaries are not based on an earlier and less developed edition of Arugat ha-Bosem. For examples, see the abovementioned forthcoming chapter on the commentaries in the Nuremberg Mahzor. 45 Most of the evidence that Urbach, Arugat ha-Bosem, 4:83-100, quotes for Samuel he-H . asid and Judah he-H . asid relates to prayer commentary rather than piyyut commentary. The prayer commentary attributed to Eleazar b. Judah was published by Moshe Hershler and Judah Alter Hershler, eds., Perushei Siddur ha-Tefila la-Roqeah . : Perush ha-Tefila ve-Sodoteiah le-khol Yamot ha-Shana le-Rabbenu Eleazar bar Yehuda me-Germaiza, 2 vols. (Jerusalem, 1992). A survey of commentaries explicitly attributed to Eleazar of Worms in manuscripts shows no traces of the style that characterizes the later H . asidei Ashkenaz commentaries and Arugat ha-Bosem. the composition of the Ashkenazic book. In the case of piyyut commentary, it is possible to define the commentaries on individual piyyutim as units in and of themselves, but it is obvious that they were in fact compiled from yet smaller elements. 46 Commentary elements are not as independent as other textual units in Ashkenazic books. Textual units such as normative statements and narratives, for example, were sorted by soft criteria of content, shared keywords, or ideas. Commentary elements, on the other hand, had to be arranged according to the order of the texts they explain. 47 For anybody but the initial commentator, the first step in compiling a commentary therefore would be to decide whether to include an element from the Vorlage at the proper place. This would be followed by inserting additional material, depending on the preferences of the compiler and his intended audience. 48 While it is possible, on a general level, to describe the preferences of a compiler based on his choices in the first step, it is the second step that turns him into an identifiable commentator shaping his own text.
Abraham b. Azriel and the other, anonymous students of Eleazar of Worms who composed piyyut commentaries invested considerable effort in this second step, bringing additional material that was not essential for the understanding of the words of the piyyutim (such as normative digressions, exegetical inserts on the quoted sources, gematriyot, etc.). Explaining piyyutim was apparently no longer their main function in these texts. Instead, commentaries became vehicles for discussing issues only alluded to in the piyyutim but significant for the educational agenda of H . asidei Ashkenaz. Topics vary from the importance of donning tefillin to discussions of Torah portions for special Sabbaths and Rosh H . odesh as well as moralistic matters such as the evil inclination. 49 Not all the commentaries that can be attributed to the 46 See the definition of possible commentary elements in Hollender, Piyyut Commentary, 59-171. On the order of commentaries in Arugat ha-Bosem, see below. 47 Joseph Kara introduced a "normal" order for different commentary elements on the same lemma, namely a paraphrase followed by lexical explanations and reworked quotations from midrashim; see Hollender, Piyyut Commentary, 37. Urbach pointed out a similar preference in Arugat ha-Bosem; see Urbach, Arugat ha- Bosem, In the case of Arugat ha-Bosem this would be the exegetical units on proof texts quoted in the commentary, grammatical explanations, Old Czech vernacular explanations, etc. 49 Public worship and its correct implementation was a major topic for H . asidei Ashkenaz, as seen in the sections on prayer in Sefer H . asidim. See also Talya Fishman, "Rhineland Pietist Approaches to Prayer and the Textualization of Rabbinic Culture in Medieval Northern Europe," Jewish Studies Quarterly 11, no. 4 (2004): 313-31. Ivan G. Marcus, Piety andSociety: The Jewish Pietists of Medieval Germany (Leiden, 1981), 98-101. On the importance of the concept of evil inclination, see ibid., 30-32. Usually, the connection between the piyyut and the insertions is discernible, such as a discussion of the norms and reasons of tefillin in the commentary on yedid 'eliyon, an ahava (i.e., a piyyut) for Ahava rabba in the blessings of the Shema, by Judah ha-Levi, used in Eastern Ashkenaz for parashat 'eqev (Urbach, Arugat traditions of H . asidei Ashkenaz generate this exegetical surplus, but the proportion is sufficiently high to consider this a defining feature that emerged in their circles. These commentaries open up the textual order established by the text commented on, introducing additional topics, new lines of exegesis or thought, or any other material that draws attention away from the original explanatory purpose of the commentary. A similar disregard of previously established order occurs in Arugat ha-Bosem on the level of the larger units-the commentaries on individual piyyutim. Piyyut commentary usually follows the order of the liturgy and liturgical calendar, even in manuscripts that contain only commentaries. Since fewer commentaries exist on piyyutim for life cycle events such as a Shabbat H . atan and Shabbat Brit Mila, 50 their position in the manuscripts is less fixed, but most piyyutim are commented on in the liturgically appropriate position. Minor variations of order exist in the case of hoshaanot, qinot, and selih . ot, since these were subject to local traditions that varied. Arugat ha-Bosem follows an entirely different principle, and instead consists of a few large blocks (sections). The first block may be described as commentary on the piyyutim of the siddur, in that it contains commentaries on piyyutim for minor festivals and special Sabbaths. Here, Arugat ha-Bosem follows a calendar that starts after Passover, in itself an unusual choice for liturgical manuscripts, which usually start with the special Sabbaths before Passover or with Rosh Hashanah. 51 This section skips most of the established piyyutim for the festivals, commenting mainly on piyyutim not contained in standard Western Ashkenazic mahzor manuscripts. It also groups the commentaries for any given day according to their poetic genres, e.g., ofanim or zulatot. 52 ha-Bosem, 2:36-89). They also connect to the biblical reading of the day when a piyyut is recited, which is especially obvious in the commentary on the meora beginning Shney zetim, ibid., 2:89-94, which does not connect to Hanukkah as the more prominent liturgical setting but refers to the second use of this piyyut in Eastern Ashkenaz, on Shabbat Behaalotekha. 50 Only sixteen piyyutim for Shabbat H . atan have been commented on, with a total of less than eighty known commentaries. Given the number of piyyutim for this occasion and the number of known piyyut commentaries from Ashkenaz (more than 12,000), this is very few. Only seven Ashkenazic piyyutim for Shabbat Brit Mila have been commented, for a total of twenty-three commentaries. 51 Manuscripts of piyyut commentary often start with Hanukkah, followed by the special Sabbaths before Passover, including Purim. 52 This type of arrangement is attested in a few collections of piyyutim, such as those attached to Mah . zor Vitry manuscripts (e.g., Haim Brody, Kuntras ha-Pijutim nach der Machsor-Vitry-Handschrift (Brit. Mus. Add. 27200 u. 27201) [Berlin, 1894]). If the section followed a Vorlage, this would have been an Eastern Ashkenazic manuscript, but only a few such collections have survived, among them MS JTS 8972. Arugat ha-Bosem starts with one full yoż er composition for the first Sabbath after Passover, compiled from piyyutim by different poets, and then a collection of zulatot, followed by some geulot. This is followed by two compiled composi-In MS Vatican the commentaries on piyyutim for the Ninth of Av come first; the commentaries on selih . ot follow. In the manuscript a few commentaries on other piyyutim, such as maaravot for Passover and a zulat for Hanukkah, are situated in different places between the selih . ot, the last one being a group of piyyutim from a qedushta for Yom Kippur. None of these commentaries is introduced by the scribe as being misplaced because it was forgotten in the proper place in the manuscript. By comparison, MS Merzbacher is defective in the beginning; it lacks the section on the Ninth of Av, and its section on the selih . ot is less often interrupted by "displaced" commentaries. 53 These moments of disorder certainly confirm Urbach's claim that the section of selih . ot commentaries in the two manuscripts are less related to each other, and possibly are even two different drafts. It also shows that at the time these texts were being copied and cited under the name of Sefer Arugat ha-Bosem, the work did not have a defined gestalt and could be recreated by combining elements of a certain type, much like unnamed, open piyyut commentaries.

"Open Genre" versus Opera
Two final differences between Arugat ha-Bosem and other collections of piyyut commentary still must be considered: the presentation of the compilation as an authored work with a title and the two different editions that reflect at least one reworking or recompilation of the collection by Abraham b. Azriel himself. If the version present in MS Vatican was consciously produced as a "second edition," Abraham b. Azriel would likely have considered tions for the Sabbath after Pentecost, two yos . erot for Shabbat Nah . amu, and several parts of compositions for this day. After that there follow piyyutim for parashat 'ekev, Shabbat Shuva, for Sukkot, Shabbat Bereshit, a piyyut for parashat va-yera ', Hanukkah, "Shabbat Hafsaqa," Shabbat Rosh H . odesh, Shabbat H . atan, and Shabbat Brit Mila,  , in the binding fragments in Vienna, Erzbischöfliches Diözesanarchiv, Kirnberger Bibliothek der Wiener Domprobstei C-8. This demonstrates that the order of piyyutim was not completely fixed in copies of Arugat ha-Bosem. 53 It is tempting to compare the larger sections in Arugat ha-Bosem to the thematic blocks in Sefer H . asidim and the commentaries on individual piyyutim to the individual small units, be they normative or narrative. However, the sections in Arugat ha-Bosem are defined by an outside order, namely liturgy, whereas those in Sefer H . asidim seem to be defined mainly by topic. The similarity lies in the approach to constructing a compiled text from larger units built from smaller blocks, with omission and insertion possible on both levels. it a work (opera) that could be improved upon. However, Marcus's previously quoted definition of "open book" makes it possible to reconstruct a different order of events in the case of Arugat ha-Bosem. Given the similar but anonymous commentaries in the Nuremberg Mahzor and the missing beginning of MS Merzbacher-which allows for the possibility that an earlier version of Abraham b. Azriel's piyyut commentaries did not have an introduction that claims authorship and provides the title-it is possible that Abraham b. Azriel did not set out to compose a work (opera) when he first compiled and edited piyyut commentaries. It might have been the process of reworking his own compilation of commentaries that made him realize that his piyyut commentaries, enriched with the teachings of Judah he-H . asid and Eleazar of Worms, as well as the teachings of others, could stake a claim for greater authority than previous compiled commentaries.
In his introduction, he claims that he decided to produce Arugat ha-Bosem because many wrong and confused piyyut commentaries existed, referring probably not only to written commentaries but also to what he considered faulty understandings of liturgical texts. Discussing a rabbinic exegesis of Songs 2:4 that claims God loves even the ramblings of those who study Torah, he posits that only those who do not know better can earn God's love by rambling in their study of Torah, while those who do know how to study need to be diligent. Therefore, those who know how to interpret should do so in an orderly fashion, while those who do not know should refer to the interpretations of those who do. 54 A systematic approach to piyyut commentary, evidenced in the many carefully inserted cross-references that present piyyut commentary as more than the explanation of difficult words and phrases, positioned Arugat ha-Bosem as edifying, educational literature rather than transmitted exegetical and normative knowledge. Such a work could claim a new type of authority, one that was associated with an author and merited a title.
In his claim to authorship, Abraham b. Azriel does not follow Eleazar of Worms, who demanded that "every man needs to note his name in his book" and suggested that the title of the book should contain a gematriya of the author's name. 55 Instead he chose a title for his work that is semantically close to the name of Eleazar of Worms's most famous normative work, ha-Roqe'ah . (the perfumer). 56 Since the short introduction in MS Vatican of Arugat ha-Bosem is the only passage that allows us a glimpse into Abraham b. 54 Urbach, Arugat ha-Bosem, 1:1. 55 Cf. beginning of Sefer ha-Roqe'ah . . Instead, Abraham b. Azriel alluded to his name in this introduction (Avarekhekha ha-melammed), although Sefer H . asidim praised those who did not claim authorship in their books (SHP § 1052); see Urbach, Arugat ha-Bosem, 1:1; Berliner, "Aus den Bibliotheken Italiens I," 2. 56 Eleazar of Worms is often quoted in Arugat ha-Bosem and might have provided the inspiration to produce a titled work. In the introduction Abraham b. Azriel explains why he chose Azriel's arguments for a semi-closed work of piyyut commentary, a detailed comparison with the conscious claim of authority that Daniel Abrams claims for Eleazar of Worms in committing esoteric knowledge to writing and the reasons for doing so is not possible. 57 We can, however, place Abraham b. Azriel, together with Eleazar of Worms, Isaac b. Moses Or Zarua', and other scholars of the thirteenth century, as scholars who claimed authority through authorship.