Al-Āmidī’s al-Muwāzana and the Size of Knowledge

This paper first surveys the structure, content and purpose of al-Āmidī’s al-Muwāzana , the largest work of poetic criticism of the fourth/tenth century, then its medieval reception, which grants it quasi-encyclopaedic authority. Modern reception has often dehistoricized it by viewing it only as criticism, but book studies suggest broader contexts of interpretation. Looser than “encyclopaedia,” the notion of the “big book,” like those by al-Āmidī’s contemporaries and neighbours Abū l-Faraj al-Iṣfahānī and al-Tanūkhī, fits well with al-Muwāzana . The chronology of al-Āmidī’s life, and the idea of his “big book” as a highly personal work of lifelong scholarship, endorse M.Z. Sallām’s reading of al-Muwāzana as an intellectual ego-document. Finally, the citation practices of fourth/tenth-century “big books” anticipate aspects of the “archival” outlook of Mamluk scholarship and suggest how the “big books” discussed here sought to validate their own versions of cultural memory.


Introduction
Al-Ḥasan b. Bishr al-Āmidī, d. 370 or 371/980 or 9811 (his birth date will be discussed later), was a well-educated, middle-ranking Iraqi functionary (kātib), of no recorded family or geographical background despite his nisba, who moved 6 Volumes 1 and 2 were edited by al-Sayyid Aḥmad Ṣaqr 1961-5, volumes 3 (i) and 3 (ii) by ʿAbdallāh Ḥamad Muḥārib in 1990. Ṣaqr's volumes have no index. Muḥarib indexed them in his volume 3 (ii) on the basis of the first edition and second printing (1972). The pagination differs in subsequent printings. References will be to the first edition. 7 For editions of al-Muwāzana, see abridged lists in Heinrichs, al-Āmidī, and Ṣaqr, preface to Āmidī, Muwāzana, I, 5. Editions up to and including that of Ṣaqr, and their manuscript sources, are discussed in more detail in Ashtiany [Bray], The Muwāzana, 96-

Bray
Journal of Abbasid Studies 7 (2020) 121-144 ʿilm or knowledge.8 (He does not attach words that correspond to my "data" or "information," or indeed any specific terms, to the matter of the preliminary stages in the teaching/learning process.) In al-Āmidī's view, however, the potential of data to yield either information (for example, that a line of verse is metrically incorrect)9 or knowledge (such as the understanding that this is an example of what makes a given poet's output uneven in quality)10 is not realized until he himself has activated the data's content for readers through his comments. His own knowledge and judgments are constantly to the fore, even in the second part of al-Muwāzana where the reader is supposed to be left to his own devices to apply the knowledge that has been demonstrated to him by al-Āmidī in the first part. How does the foregrounding of authorial opinion interact with the organization of content? How important a factor is it in deciding how to classify al-Muwāzana? This leads to a last question, that of the relationship between the arrangement of content and access to information, or, in this case, chiefly data. Al-Muwāzana gives its readers tools for locating materials: programmes of topics the author intends to treat, and cross-references;11 but the same processes of producing knowledge, or highlighting information, through authorial commentary, are used throughout al-Muwāzana, and the same data are sometimes repeated in different chapters.12 Wherever they are placed, topics and data predominantly yield the same kind of information: an analysis or demonstration of expertise, and/or an evaluation by al-Āmidī, which is to be accepted by the reader as knowledge. Organization and location are therefore non-functional insofar as they neither differentiate types of information nor configure it so as to locate it in a structure of knowledge. The ʿilm that al-Āmidī displays or invokes consists of various kinds of knowledge of Arabic lexicon, grammar and poetry, but has no apparent hierarchy.
Against a checklist that has so far yielded only queries or negatives, it might seem wrongheaded to persist in trying to discuss al-Muwāzana as an encyclopaedic enterprise. Nevertheless, it is worth looking more closely at the issues raised by the attempt. As regards size and scope, al-Muwāzana consists of two 8 Which is attributed to "those knowledgeable about poetry" (ahl al-ʿilm bi l-shiʿr); see, e.g., Āmidī, Muwāzana, I, 424-429, For a discussion of the reception of al-Āmidī's conception of ʿilm and its authority, see Ashtiany [Bray], The Muwāzana, 38-39. 9 Āmidī, Muwāzana, I, 287-290 and 386-388: metrical flaws in the poetry of Abū Tammām and al-Buḥturī respectively. 10 On the notion of istiwāʾ (consistency) see Āmidī, Muwāzana, I, 52, ll.3-9 and Ashtiany [Bray], The Muwāzana, 29, notes 8-10. 11 See below,notes 18,20,[22][23][24][25][26]35. 12 See Muḥārib's footnotes. main parts. "Part" is a word I use for convenience, in a loose, non-technical sense, to avoid confusion with the shorter sections that al-Āmidī terms juzʾ. The two parts are reflected, approximately, in the publication history of the text. The early printed editions (from 1287 A.H. onwards, see note 7) contained only the first part and the very beginning of the second, and it is the first part that has continued to attract the most attention, even though Mandūr had early argued, on the basis of what little he knew of it, that the second part was the essential part,13 an opinion vindicated by the publication of the greatly augmented critical edition by Ṣaqr and Muḥārib, which contains all that is now known to exist of the second part. Muḥārib's continuation of Ṣaqr's edition identifies, among others, a manuscript source referred to, but not disclosed by, Ṣaqr and ʿAbbās, and not known to other scholars and critics, the problematic Cambridge University Library MS Qq286 (see note 7). The first part of al-Muwāzana is polemical, outlining first the adversarial positions of the amateur supporters of Abū Tammām and al-Buḥturī, then al-Āmidī's own positions on the faults ascribed to one or the other poet by previous critics. Its structure betrays successive, unfinalized revisions. It is divided into an uncertain number of subdivisions (four?) which al-Āmidī calls ajzāʾ. (Yāqūt describes al-Muwāzana as being in ten ajzāʾ,14 but there are fewer than ten in the text that we have, and al-Āmidī's own references to a given juzʾ are not always distinct.) Here is an overview of the contents of the first part. In a brief but pregnant pseudo-dedicatory introduction,15 al-Āmidī outlines expert consensus on the characteristics of the two poets Abū Tammām and al-Buḥturī, enumerates the classes of people that each poet appeals to, and asserts their incommensurability, introducing the key critical term ʿamūd al-shiʿr al-maʿrūf ("what is well known to be central to poetry").16 Al-Buḥturī always adheres to the criteria that al-Āmidī goes on to unpack from this term, whereas Abū Tammām's poetry fluctuates wildly.17 Since, however, tastes in poetry are always irreconcilable, al-Āmidī says that he will not pass overall judgement -this task falls to the dedicatee or reader. Instead he will analyse and compare, giving only interim evaluations of how each poet performs on a given theme (maʿnā) or in passages and poems sharing a topic, 13 Mandūr,107. 14 Yāqūt,II,852. 15 The dedicatee is unnamed and probably non-existent, a common convention. 16 Āmidī, Muwāzana, I, 5-7. The term ʿamūd al-shiʿr al-maʿrūf occurs 6, l.8. On this term and its short-lived importance in Abbasid criticism, see Heinrichs,and Ṣubḥ, The introduction to this section has been partly translated by Ajami," 31,and by Naaman,Literature and the Islamic Court,[201][202]

Bray
Journal of Abbasid Studies 7 (2020) 121-144 metre and rhyme scheme. This is the general plan of the book. First, though, he will "report what I have heard of the contentions that each set of partisans directs at the other when they argue the respective merits of their poets." The aim of this exercise is didactic: "It will enable you to ponder the matter, and pass judgement -or else, if you wish, to hold to your convictions -more forcefully and with greater insight." The editor, Ṣaqr, numbers the exchanges, which are of variable length, from 1 to 22; they cover some forty pages.18 The first introduction and the imagined debate is followed by a second introduction, which slightly revises the plan of the book. Al-Āmidī will survey the faults characteristic of each poet and end with their merits (it is not clear whether he means that he will end the section or end the book with this topic). He will begin with the faults of Abū Tammām, before going on to compare him with al-Buḥturī "poem by poem (qaṣīdatan wa-qaṣīdatan), when they match in rhyme and metre, then theme by theme (maʿnan wa-maʿnan)."19 Then there will be chapters on themes unique to each poet, and on simile (tashbīh) and apothegms (amthāl), "which will conclude the treatise (risāla)." The whole will be rounded off with an anthology of their poetry (al-ikhtiyār almujarrad min shiʿrayhimā) organized alphabetically "for convenience and ease of memorization."20 Al-Āmidī adheres to the first part of the plan, starting with sections on the plagiarisms of Abū Tammām (120 items in Ṣaqr's numbering) and those he was rightly or wrongly accused of by Ibn Abī Ṭāhir (he presumably means Ibn Abī Ṭāhir Ṭayfūr and his lost "Plagiarisms of the Poets"),21 a further forty-six items. But then comes "juzʾ Two," and a further preface which summarizes what has gone before, describing the preceding section on plagiarism as an appendix to "juzʾ One,"22 and outlines a further interim plan: al-Āmidī will investigate and, if he can, dismiss the accusations of errors of conception and expression brought against Abū Tammām (mā ghaliṭa fīhi… min al-maʿānī wa-l-alfāẓ) which he has "heard from the mouths of men and experts on poetry (ahl al-ʿilm bi-l-shiʿr) in conversation and discussion, as well as what I have myself… deduced."23 The forty-five items in this section of around hundred pages are the subject of lengthy analyses, and are followed by a passage that 18 Al-Āmidī (Muwāzana, I, 7) introduces the debate passage with the words: wa-anā abtadiʾu bi-mā samiʿtuhu min iḥtijāji kulli firqatin min aṣḥābi hādhayni l-shāʿirayni ʿalā l-firqati l-ukhrā. serves as a postface to this juzʾ and a preface to the following one,24 which consists of a brief section on faults in Abū Tammām's poetry that cannot be explained away,25 followed by sections on the faults in al-Buḥturī's poetry that correspond to those in Abū Tammām's, beginning with plagiarism.26 A scribal note between the end of the last section on Abū Tammām's faults and the beginning of the one on al-Buḥturī's plagiarisms states: "This is the end of the second book (sifr) of al-Muwāzana according to the way the author divided it up."27 The lead-up to what I have called the second part of al-Muwāzana is complicated. First al-Āmidī announces a modification to the general plan: "In this juzʾ I will discuss the themes common to both poets"28 -the idea of combining common themes with shared rhyme and metre has silently been dropped. There follows a discussion of epistemology. If the reader wants to know exactly why al-Āmidī expresses a preference for one or the other poet in the case of a given maʿnā, he should work it out from the examples he has already been given. Al-Āmidī will explain whatever can be explained, but "there remains that which cannot be put into words or reasoning and can only be known from experience"29 This is illustrated by an excursus on the nature of those kinds of knowledge that are based on experience and natural ability. The latter is not given to everyone. In the case of poetry, such knowledge is best achieved by studying the consensus of past masters -a central idea in al-Āmidī's thought.30 Confusingly, this epistemology is followed by two short passages on the respective merits of Abū Tammām and al-Buḥturī, each attributed to the partisans of the other,31 merging into another excursus on epistemology, also central to al-Āmidī's thinking. The philosophy of the Greeks, the wisdom of the Indians and the adab of the Persians are not poetry, he says, for poetry must follow "the way of the Arabs (ṭarīqat al-ʿarab)."32 (Al-Āmidī never explains who he means by "the Arabs," or when or where they had, or have, their existence.)33 Nevertheless, there are four conditions of creation, which al-Āmidī enumerates and discusses, that are common to everything, including the verbal arts, 24 Ibid., 243-244. 25 Ibid., 345-390. 26 Ibid., 291-388. 27 Ibid., 290. 28 Ibid., 388. 29 Ibid., 388-389. 30 Ibid., 389-396. 31 Ibid., 397-399, 400-401. 32 Ibid., 401. 33 His reasons for being unspecific are discussed in Ashtiany [Bray], The Muwāzana, 87-93. and for which al-Āmidī cites the Persian sage Buzurjmihr.34 "And now," says al-Āmidī, "I come to the comparison (muwāzana), and although it would have been best to compare lines or passages in the same metre and rhyme, this rarely coincides with shared meaning/theme (maʿnā), which is our goal."35 In fact, only common maʿānī will be compared. Overall, maʿnā is used by al-Āmidī in the loosest possible way, of any theme, motif or image that is part of the established poetic repertory as he conceives it, or of any poetic idea.
In the second part of al-Muwāzana, he tries to use it in such a way as to work through the structure of the qaṣīda from beginning to end. The whole of the second part of al-Muwāzana is given over to the thematic comparison. Large as it is, occupying more than three volumes of the Ṣaqr/ Muḥārib edition's four, this second part usually attracted little attention during the first half of the twentieth century, for earlier editions contain only thirteen of al-Āmidī's heads of thematic comparison, as against eighty-two in Ṣaqr's critical edition and a further eighty-five (some with subdivisions) in its continuation by Muḥārib. The severely truncated earlier editions give the impression that the comparison was tacked on in half-hearted support of al-Āmidī's exposition of his critical principles. Our perspective on al-Muwāzana alters if the second, much longer (even though incomplete) part is given equal prominence with the first, and is used interactively with it as al-Āmidī intended. Our impression would be different again if we still had, or if al-Āmidī had ever completed, the final sections of the "epistle" on unique themes and on tashbīh and amthāl, and the anthology of select verses, "alphabetically organized for convenience and ease of memorization."36

Al-Muwāzana in the Context of Third to Fourth/Ninth to Tenth-Century Criticism
The topics and structure of the first part of al-Muwāzana respond to critical and, less overtly, to theoretical issues most of which emerged in the late third/ ninth century and reached a high point of development in the earlier part of al-Āmidī's lifetime, that is, the first two decades of the fourth/tenth century (see "Al-Āmidī's life and times" below). The Abbasid prince, poet, critic and literary historian Ibn al-Muʿtazz (247-296/861-908) had claimed, in his Kitāb al-Badīʿ ("The New Poetic Style") that, when he wrote it in 274/887-8, he was 34 Āmidī, Muwāzana, I, 402-405. 35 Ibid., 405. 36 See note 20 above.
the first person to define the "five sorts of badīʿ."37 By doing so, he historicized and theorized, within a continuous poetic tradition, those features whose frequent occurrence distinguishes "modern" poetry, including that of Abū Tammām, from that of the "ancient" Arabs.  By way of contrast with the topicality of the first part, the second part of al-Muwāzana gives the impression of taking its arrangement neutrally and naturally from the primary data under consideration, that is, the entire qaṣīda output of Abū Tammām and al-Buḥturī and the repertory of poetic themes and motifs, maʿānī, as established by "the Arabs," from the vaguely antique past of Arabic poetry until some equally unspecified later point in time.53 In this part of al-Muwāzana, al-Āmidī simply works his way through the themes and motifs treated in Arabic poetry through the ages in the sequences in which they occur in Abū Tammām and al-Buḥturī's grand ceremonial pieces (qaṣīdas) -opening motifs, transitional motifs and so on -citing other poets of their own and earlier periods as exemplars against which the two poets are to be measured. 46 Āmidī, Muwāzana, III (i), 28. 47 Ibid., I, 9, l.7. 48 Ibid., I, 12, ll.9-10, 13; 13, ll.3, 6; 133, l.8; II, l.7 (reporting from Ibn Qutayba); 176, ll.

The Organization of al-Muwāzana: A Tool for Navigating Data, or for Producing Knowledge?
Insofar as it is dictated by the data that it presents, the organization of the second part of al-Muwāzana could be seen as inert, and it should be noted that the formal organizational criterion of shared maʿānī is distinct from the critical criteria, based on close reading, which al-Āmidī demonstrated in the first part of al-Muwāzana and returns to intermittently in the second part. The tool given to readers to locate material in this second part is a prefatory list of chapter headings summarizing maʿānī more or less broadly.54 It is a minimalist tool. Conceptually, however, it is powerful, because it implies that al-Āmidī's rubrics capture the whole Arab or Arabic poetic world, even though his scheme excludes the numerous poetic genres other than the qaṣīda that were now well established, and sometimes practised by Abū Tammām and al-Buḥturī.55 Al-Āmidī's poetic world has no precise location in period or geography even though, as the author of al-Muʾtalif wa l-mukhtalif, a work on early poets with similar names and how to distinguish them, it is to be presumed that he was not vague about such matters unless he wished to be. As he seeks to demonstrate in his detailed analyses, the poetic world, or rather the world of true poetry, exists in literary convention, established by linguistically and culturally exclusive Arab or Arabic precedent. When a poet perceives and expresses a maʿnā through that convention and the audience's horizon of expectation is shaped by it, there is a perfect and fulfilling fit between what the poet sees and says and what the audience hears and visualizes, which is disrupted when a poet introduces notions or procedures that do not belong to the canon and are therefore not fully poetic.56 Ostensibly serving as a tool for navigating data, maʿnā headings therefore serve to give an air of established consensus to al-Āmidī's theory of what constitutes the Arabic poetic universe. In the first as in the second part of al-Muwāzana, al-Āmidī implies that the data he sets before readers cover all the essential matter to which readers need access; in this case, all the critical positions relevant to the evaluation of Abū Tammām and al-Buḥturī and indeed poetry generally, and all the examples and analyses that they need in order to acquire ʿilm. In its subsequent reception, it seems that it is either its coverage of its specific topic or its grasp of general principles that gives al-Muwāzana a claim to be considered from the angle of encyclopaedism. cross-references work? (Yes, often.)60 More important to an overall understanding of his composition practice, in our present very imperfect state of knowledge of the manuscript sources of the standard edition, are the internal references by al-Āmidī to other works that he had written on the topics treated in al-Muwāzana, the titles of such works cited by later biobibliographers,61 and the external evidence of the existence of a body of material that Ibn al-Mustawfī used interchangeably with al-Muwāzana.62 Al-Muwāzana as we have it was evidently never finally revised; it seems likely that it was never completed, and that later authors, and perhaps al-Āmidī himself, viewed it as part of a larger corpus that could be treated as a single text, rather than as an autonomous book.63

De-historicizing and Re-historicizing al-Muwāzana
An interim summing-up will be useful at this point. Applying the question "is it/is it not an encyclopaedia/encyclopaedic" to al-Muwāzana clarifies that al-Muwāzana's size matters: how large it is, is part of what it tries to do, whether or not that is "encyclopaedic." Ignoring its size and the relationship between its two main parts has led to its being placed in too narrow a category, as simply "criticism," and this has often led to its being de-historicized. Al-Āmidī claims that he will not declare his own preference, but for many modern scholars his bias against Abū Tammām is patent. Hence al-Muwāzana tends to be measured, as criticism, against anachronistic notions of objectivity (which al-Āmidī equates with his own knowledge of poetry and with ʿamūd al-shiʿr) and the premium some modern readers place on poetic originality such as Abū Tammām Mansoor Mohammed Alharthey, in the most recent study of al-Muwāzana that I am aware of, bases his negative critique on twentieth-century theories of readerly horizons of expectation.65 But although al-Muwāzana is rooted in close reading, its fundamental concern is epistemological:66 its purpose is to make its readers experience and understand the nature of genuine poetry. Al-Āmidī's concern with the bases of knowledge and his exhaustive exposure of his readers to data in order to train them empirically could be seen as encyclopaedic, if by this we mean an extensive didactic endeavour founded on a conception of the nature of knowledge and the means of attaining it. Setting al-Muwāzana in an encyclopaedic perspective by paying attention to its size, scope and structure in turn introduces new questions shaped by book studies: questions about how al-Āmidī wrote his book and how we conceptualize it as a book. These may help us to rethink older questions more concretely and less contentiously. Al-Muwāzana responds to scholarly written works, draws on al-Āmidī's own writings and probably, for poetry, on his memory, but it also uses manuscript sources to check readings, and imaginatively recreates, in essay form, the kind of viva voce disputes that took place in literary gatherings. In other words, methodologically, it has a varied and complex relationship to both written and oral scholarship and to literary connoisseurship as a social practice. Historically, it adds substantially to what we can derive from Gregor Schoeler's examinations of oral-written continuum and exchange and Beatrice Gruendler's explorations of what books could do that performance could not.67 It also adds to our thinking about what we mean by a "book." Even though it may not have been completed or closed as a text, al-Muwāzana is largely finalized as a project. Yet it is also part of a cluster of works, now lost, in which al-Āmidī seems to have used the same techniques to attack aspects of the same topic. Ibn al-Mustawfī makes no distinction between these works and al-Muwāzana when he cites al-Āmidī. For his purposes, they are all part of al-Āmidī's distinctive critical discourse. How did al-Āmidī himself see the relationship between them? Is his oeuvre intentionally fluid, a concept left open to expansion? If so, should we extend the idea of encyclopaedism to the cluster, privileging the idea of a book as the working out of an idea, regardless of what forms it takes, over that of the book as a self-contained entity and a physically finite thing, which is how al-Muwāzana is usually understood? Or do we need 65 He does, however, concede that al-Āmidī may have reflected the readerly expectations of his own culture (Alharthey,Literary Reception,(209)(210) to find a new frame of reference for al-Muwāzana? In my conclusion, I shall suggest that the best working framework might be not that of a category of book, but simply of a size of book -"big books" with, ostensibly, nothing else in common but their size. I use this term as a place-holder. Unlike "encyclopaedia," it is naïve and untheorized and brings no baggage with it.

Al-Āmidī's Life and Times: Cultural Rupture or Cultural Continuity?
Meanwhile, it is time to look at the author as well as his book. So far we have problematized al-Muwāzana in relation to a loose notion of the "encyclopaedic" or, in relation to its literary historical context, as criticism or epistemology; but Muḥammad Zaghlūl Sallām also problematized it in relation to historical context when he suggested that al-Āmidī's Arab or Arabic purism and objection to cultural intrusion, which we touched on earlier, is what explains its genesis and is its raison d'être.68 Sallām saw al-Muwāzana as a response to and rejection of the fact that the Persian Buyids had taken over Iraq and brought the Arab caliphate and heritage under their heel. Such a reading would make al-Muwāzana both a psychological and a historical ego-document, in addition to whatever else it may be. Sallām's analysis rested, ideologically, on nationalist premises, but could equally be restated today in terms of cultural studies. It is true that in the very first pages of al-Muwāzana al-Āmidī says that al-Buḥturī's supporters include "the Bedouin/pure Arabs" (al-aʿrāb), whoever they might have been at this time in urban Iraq, alongside secretaries (kuttāb; al-Āmidī was himself a secretary), while Abū Tammām's poetry, which "is not like that of the ancient Arabs (al-awāʾil)," is characterized, among other things, he says, by "miscegenated themes" or "motifs that are only part-Arab" (al-maʿānī al-muwallada),69 and its partisans include people fond of (by implication) un-Arab "subtlety and philosophical discourse" (man yamīlu ilā l-tadqīqi wafalsafiyyi l-kalām).70 In matters of poetry, al-Āmidī is indeed an Arab purist, borrowing the idea of the innate truthfulness of the old Arab poets from Ibn Ṭabāṭabā, but applying the idea of truthfulness as intrinsic to the true Arab(ic) tradition more widely than Ibn Ṭabāṭabā did, to a range of modern poets71 68 Sallām,I,206. 69 "Newfangled" would be an anodyne translation of the literary application of this term, which of course is also used of human beings, most notably women slaves of mixed parentage.

Bray
Journal of Abbasid Studies 7 (2020) 121-144 (though none is later than the third/ninth century; al-Āmidī does not cite his own contemporaries as poets, only as lovers of poetry). It will be remembered that Qudāma b. Jaʿfar, whose theory of the principles of poetic adequacy could well be called "philosophical discourse,"72 quoted supporting examples almost exclusively from ancient Arab poets.73 Arab or Arabic purism does not therefore seem to have been peculiar to al-Āmidī, or indeed situational in the straightforward way suggested by Sallām, either in al-Āmidī or in the critical tradition on which he drew. It is not tied to ethnicity as a living issue. Indeed, al-Āmidī brought the notion of authentic poetic Arabness into interaction with the complexities of modern times by linking it to modern social classes and cultural camps: the secretaries like himself, who favour al-Buḥturī, and the philosophizers who support Abū Tammām. Having done this, he does not explain why a modern culture which is complex and diverse ought to hold to a simple vision of an exclusively Arab past when it comes to poetry. To pursue Sallām's diagnosis of cultural humiliation from another angle, by the second half of the fourth/tenth century, the Arabic linguistic and critical tradition was at one and the same time hegemonic and colonial on the one hand, and post-colonial and potentially subaltern on the other: hegemonic, since the caliphs had ruled over a swathe of conquered peoples and had made Arabic the imperial language; post-colonial or subaltern, since from the end of the third/ninth century the caliphs continually lost power to local rulers, and by the second half of al-Āmidī's life Iraq, the heartland of Arabic linguistic and literary scholarship, had become a vassal state ruled by Persians. Yet Arabic remained the language of the post-imperial administrations and did not lose cultural prestige, but gained it, and its users continually affirmed their filiation with the linguistic standards and scholarly traditions laid down in imperial times.
To begin to understand how al-Āmidī may have experienced those traditions, we must look at the chronology of his life. The sources do not say how old he was when he died in his home town of Basra in 370 or 371/980 or 981, but the formative period of his younger years seems to have predated the Buyid seizure of Baghdad in 334/945 by some decades. He had studied in Baghdad with the best grammarians and literary scholars,74 one of whom was the leading pupil of Thaʿlab (200-291/815-904), Abū Mūsā Sulaymān b. Muḥammad b. Aḥmad, known as al-Ḥāmiḍ on account of his sour temper, who died in 305/917. This would place al-Āmidī's birth in around 288/900 or earlier, and shows that his education stood in the living tradition of the third century. We 72 See, e.g., Heinrichs, Qudāma ibn Jaʿfar. 73 See Heinrichs, Qudāma ibn Jaʿfar. 74 Yāqūt Udabāʾ, II, 851; Heinrichs, Al-Āmidī.
have already seen that one of al-Āmidī's sources for al-Muwāzana had been a friend of al-Buḥturī, while another had known Ibn Qutayba.75 He tells us that in 317/929, he began working on an anthology of the best lines by Abū Tammām and al-Buḥturī, a project to which he kept returning (no doubt this anthology was what he intended to re-use as the final section of al-Muwāzana).76 After studying and working in Baghdad as a kātib in the bureaucracy of al-Muqtadir (r. 295-320/908-932), al-Āmidī returned to Basra, by Heinrichs' calculation "at some point before the year 334/945"77 when the Buyids took over Iraq. While there is evidence such as the above to show how al-Āmidī experienced continuity, there is not much to show how he experienced change. In Basra he worked as kātib to a succession of judges, and came to know the judge, poet and man of letters al-Muḥassin b. ʿAlī al-Tanūkhī (327-384/939-994). The two men were quite close, it seems, since al-Tanūkhī is the main source for facts about his life78 and for autobiographical anecdotes told by al-Āmidī about his relations with his Basran employers,79 including a faux-naif comic poem in which al-Āmidī anthropomorphizes a judge's hat of office (qulansiyya), which is moved to plaintive speech by finding itself placed on an unworthy head80 -not an "Arab" maʿnā as set forth in al-Muwāzana, any more than are those of the other occasional poems by al-Āmidī that survive.81 Was the life of the salon, for al-Āmidī, distinct from that of the intellect? Or did "Arabness" as the fundamental criterion of modern as well as "ancient" poetry apply only to the qaṣīda?

Bray
Journal of Abbasid Studies 7 (2020) 121-144 (284-ca.363/897-ca.972) -a kātib in the caliphal bureaucracy, like al-Āmidīwhose unfinished Kitāb al-Aghānī ("Book of Songs") was one of the fourth/ tenth century's biggest books of all. Big books were not a new phenomenon, and went on to become increasingly characteristic of Arabic culture; but of all the big books of the fourth/tenth century, the three just cited, along with al-Muwāzana, are the ones that it seems most appropriate to look at together, by virtue of their authors' propinquity.
One feature common to them all is that even if they had their origin in a commission, like Kitāb al-Aghānī or al-Muwāzana (the latter probably fictitiously), they took little or no notice of it.82 They were free compositions. Consequently, even though as written books meant for publication they had a plan and design, they stayed open to remodelling and expansion. They could and did grow with their authors, not just as compositions but as personal projects. Al-Tanūkhī explains to his readers that his two compilations held deep personal meaning for him and came out of his own relationships and experiences.83 As his successive prefaces show, he improvised Nishwār after laying down a first plan,84 and his introduction to al-Faraj describes how he struggled, over years of tinkering, to reduce it to publishable proportions.85 Abū l-Faraj al-Iṣfahānī seems to have been so attached to writing Kitāb al-Aghānī that he could not make himself bring it to a close,86 and al-Āmidī's often vehement wording leaves no doubt that he too was deeply invested in his critical project, to which he returned again and again and of which al-Muwāzana appears as the, albeit unconcluded, summation.
All of these big books were therefore lifelong projects, and each of them strives to convey not "the world in a book" (in the words Elias Muhanna applies 82 If al-Muhallabī (d. 352/963), as Kilpatrick thinks probable, had commissioned Kitāb al-Aghānī, his fall would have deprived Abū l-Faraj al-Iṣfahānī of a patron: "Well into his sixties by then, he must have abandoned any hope of finding a new patron." Moreover, the third part of Kitāb al-Aghānī no longer refers to previous song lists (though with "fundamental differences"), but "is constructed around a core of songs chosen by Abū l-Faraj himself" (Kilpatrick,Making the Great Book of Songs,19,29,28). 83 Tanūkhī, to a Mamluk encyclopaedia),87 but a personal, highly selective vision of a world: a world of their authors' fashioning, in the shape of a book that could be spun out, lived with, and perhaps lived through, indefinitely, by their authors. ʿAbbās seems to argue that in al-Muwāzana al-Āmidī had constructed a fantasy world inhabited by no-one but himself,88 and al-Āmidī's absorption in his own world of ideas is exactly what makes al-Muwāzana interesting and important as a "big book" of the fourth/tenth century. The literalism with which al-Āmidī parses qaṣīda imagery and interprets whatever he feels is truthful in it as both true and real,89 and on that basis develops a theory of realism which is almost entirely aesthetic, literary and asocial,90 expresses an intensely personal understanding of meaning and value, as do the vision of the poetic and musical arts and of artistic dignity expressed in Kitāb al-Aghānī, and the vision of human experience expressed in al-Faraj and Nishwār.
We are still a long way away from knowing how to understand historically the passions that drove these writers to their grand enterprises, but the field of enquiry is widening. Whether malaise with contemporary reality contributed to al-Āmidī's views on the purity of Arab poetry, as Sallām suggested, is a question that need no longer be seen in isolation, and rather than taking simple racial self-identification as its underlying argument, it can now be framed by a growing body of enquiry into Abbasid historical constructions and cultural uses of Arabness.91 Sallām's reading of al-Muwāzana as a psychological egodocument, which he framed in the rather narrow context of the history of criticism, has also become more widely pertinent. If the fourth/tenth century produced encyclopaedias, it also produced big books which were repositories of learning, experience and knowledge that had certainly been compiled in order to share knowledge, but whose authors' purpose was not to hand over their knowledge to readers merely as information, but rather to bring it alive to them as their own vision of a particular world within the greater world and 87 Muhanna , priority belongs to generic typology or to authorship and its circumstances.

"Big Books" and the "Archival Turn"
Finally, while reading al-Muwāzana against three other "big books" of the fourth/tenth century emphasizes that it and they are highly personal, interpretative assemblages of knowledge which should be read within the frame of their own time and circumstances, it also reminds us that a prominent common feature that goes with their size is that, by design, they recirculate previous materials and shape them to new purposes. Thus al-Muwāzana acts as a repository that both preserves and repositions poetic texts and critical texts and approaches. Many of these texts would have been in wide circulation, but a few would probably have been less commonly available or generally familiar to readers of the type al-Āmidī claims to be addressing, lovers of poetry who were not specialists in criticism. From this point of view, al-Muwāzana's position among the "big books" of the period could be set in a technical perspective, for it could be viewed as an archive: that is, a store of information and documentation. But what kind of archive? Archives can be institutional or personal, deliberate and planned, or passive, even accidental repositories. Al-Muwāzana is not passive, but a constructed, specialist archive, limited to a specific cultural topic. The materials that it puts on record are not there simply because they happen to be available; on the contrary, al-Āmidī links his citations to topics within a critical debate in which he himself is a participant, which he builds into the argument and design of al-Muwāzana. Nor is al-Muwāzana a closed or "dead" archive, for al-Āmidī's own ongoing writings are part of it. We could perhaps therefore define it as a personal archive, in which he stores both his sources and the output that he derives from them. At the same time, it is meant for public use by readers. Does thinking about al-Muwāzana not only as a "big book" but also as an archive add to an understanding of it? "Archive," like "encyclopaedia," is a notion that has undergone radical revision in the light of the past decade of Arabic book studies, particularly in respect of Mamluk historiography. Most recently, Fozia Bora has argued that, even though increasing numbers of physical documents are coming to light, we should concern ourselves equally with the immaterial, individually-curated archives latent in Mamluk historians' choice and use of sources. Individual writers selected sources that they considered of continuing relevance and used them to shape their own versions of what they wished to become common memory.92 Al-Āmidī similarly tries to create cultural memory and cultural history with his theories of the Arabness of Arabic poetry and of ʿamūd al-shiʿr al-maʿrūf, bolstered by the alleged critical consensus on poetic values laid down by the past masters he quotes, notably Ibn al-Muʿtazz.93 For lack of independent material evidence, twentieth-century scholarship was prone to view the immaterial archival practices of Arabic authors as traditional and institutional, that is, as points of consolidation and general agreement, thereby reading the sources of Abbasid intellectual history backwards as consensus, instead of as a forward-moving process of enquiry. This reading emphasized the archive as a passive deposit at the expense of considering how and why it was formed and intended to be used. Thus al-Muwāzana was viewed by ʿAbbās as a "leap forward" in the field of criticism, but only insofar as it surveys and preserves earlier key texts and uses the gamut of available critical tools with unprecedented (and in ʿAbbās's view misapplied) rigour.94 This account positioned al-Muwāzana conveniently on an evolutionary scale, at a time when medieval Arabic poetics and poetic criticism were being rediscovered and needed a defining historical narrative. It does not map so well on to the less narrowly conceived, more diverse critical landscape that recent scholarship has started to explore.95 Al-Āmidī's use of the archive, on the other hand, aligns him with a scholarly practice that is not discipline-specific and is found in all the "big books" under discussion here: a concern with the detailed and accurate citation of sources and informants, with establishing a persuasive version of the past by 92 Bora, Writing History. 93 Some factors are common to both physical and immaterial archives. Anyone who has worked as an archivist and been witness to despoiled files will know that "weeding" material regarded as otiose or undesirable is a major factor to take into account in assessing the integrity or representativeness of an archive. Curating an archive can involve destruction and attempted obliteration as well as preservation. 94 ʿAbbās, Tārīkh al-naqd, 157-158. 95 See, e.g., James White, Anthologists and the Literary Market.