Modarres-e Reḍavi’s Edition of Anvari’s divān : A Critical Assessment

The aim of this paper is two-folded: 1) to discuss Modarres-e Reḍavi’s edition of Anvari’s divān in order to show that this edition, although still very valuable, should be used cautiously: even for non-philological, literary-oriented studies manuscripts should be checked. These should include not only the newly-discovered codices, not used by the editor, but also the manuscripts he used, which must be double-checked; 2) to give a solid starting point to any scholar attempting to investigate Anvari’s divān from a philological perspective, by showing in which areas Modarres-e Reḍavi’s edition is lacking and to what extent.

5 As stated by MR himself in Anvari 1959-61, 2: 162.It is not unusual for editions of classical poetry printed in Iran to be bowdlerised.How satire/obscene poetry (hajv/hazl) has been treated by editors is described in Zipoli 2016, XXXI-XXXL.
7 No complete list of the known manuscripts of Anvari's divān exists.Partial ones, sometimes with very brief descriptions, can be found in de Blois 2004, 221-5; Monzavi  1971, 2235-42 and Derāyati 1389, 5: 58-68, as well as in the editors introductions to the two printed editions.Dublin, Chester Beatty, Per.103 is extensively described in Arberry, Minovi, Blochet 1959, 4-11.London, British Library, Or. 3713 is also extensively described, in Rieu 1895, 141-3.Brief descriptions of the two Istanbul manuscripts, Istanbul, Fāteḥ, 3784 and Istanbul, Fāteḥ, 3786 can be found in Ritter, Reinert 1986, 119-20.8 This book is de facto a traditional commentary written in the twentieth century.
Šahidi relies heavily on the classical commentaries of Farāhāni and Šādiābādi, and nearly always concerns himself with the explanation of individual lines, with a particular focus on lexicon, while never analysing poems as a whole.9 He relied on: 1.An incomplete manuscript of which M. Minovi gave him a reproduction, to which he refers to as nosxe-ye minow.Thirteenth-fourteenth century CE; it corresponds to Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 4113.
Šafiʿi Kadkani, as part of a series of annotated anthologies meant to enable modern day students to read classical poetry, published a selection of Anvari's poems, with an introduction and many annotations (Šafiʿi Kadkani 1993).He also did not follow MR's text, but many times preferred the readings given by old, unused manuscripts. 10n recent years a number of articles where written by Iranian scholars (Karami, Amini, Kowṯari 2014; Ḏabiḥi 2016; Nurāyi, Aḥmadpur 2016), 11 all of them stating the need of a new edition of Anvari's divān.This is mainly justified by enumerating the multiple old manuscripts that were not used by MR, 12 the five main and oldest ones being: 13 • Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 2615, not dated.• Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 4113, not dated.• Tehran, Ketābxāne-ye majles-e šurā-ye eslāmi, 86666, dated 680 Q./1281-82 CE.It is the oldest known dated copy of Anvari's divān. 14 Another incomplete manuscript owned by A. Afšār Širāzi, to which he refers to as nosxe-ye afš.Thirteenth-fourteenth century CE; it corresponds to Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 2615.3. Tehran, Ketābxāne-ye majles-e šurā-ye eslāmi, 86666; dated 1281-82 CE. 4. MR's printed edition.
11 Each of these articles gives a different list of manuscripts that should be used for the new edition.No one tries to outline a criterion on the basis of which manuscripts should be selected.For the lists of manuscripts given in those articles, see appendix 2.
12 Karami, Amini, Kowṯari 1392 clearly states the need of a new edition with similar arguments to the ones made in this paper; the authors also quote MR's claim to have registered only the "important and necessary" variant readings and take note of the lack of a proper method.However, their approach is not consistent.They state the impossibility of working only on the newly found codices (138), but then they proceed to mention among the nosax-e mowred-e estefāde dar taṣḥiḥ-e tāze-ye qaṣāyed-e Anvari (witnesses that ought to be used in the new edition of Anvari's panegyrics) only manuscripts that MR has not used.Moreover, in spite of what was written earlier about MR's work, in the lines edited as samples of their method, the editing is made bā tavajjoh be nosxebadal-e divān o nosax-e xaṭṭi-ye yādšode (referencing the variant readings of the divān [edited by MR] and the aforementioned manuscripts [not used by MR]).Ḏabiḥi 2016 states the need of a new edition, lists both manuscripts used and unused by MR, but does not discuss neither MR's edition (except by alluding in the abstract to naqṣ dar qerāʾat-e nosxehā-ye estefāde šode 'mistakes in reading the utilised witnesses') nor any methodological issues.
13 Brief descriptions of these codices can be found in "Appendix 2".
14 The last page is damaged and the day and month of completion are not readable, only the year is.
Among western scholars similar statements were made by A.L. Beelaert, who, after discussing the manuscripts not used by MR, wrote: "Anvarī's Dīvān is badly in need of a new edition" (2017).

2
Modarres-e Reḍavi's Method To begin a discussion on why MR's edition, and especially its apparatus, should be used by scholars with caution, it is first necessary to summarise what MR says about the method he followed (as stated in Anvari 1959-61, 1: 15-22).1.He made a list of poems "on the basis of some witnesses".2.He ordered the poems alphabetically, by rhyme. 16.He wrote down the whole divān on the basis of this list, then he added the poems present in "a number of witnesses" to have a complete version of the divān.4.He compared the divān thus obtained with everyone of the selected witnesses, 17 then choosing Istanbul, Fāṭeḥ, 3784 18 as his base manuscript (siglum 'lām' in the apparatus), describing it as: "older and more solid than the other witnesses". 5.He states that: extelāfāt-e nosax-rā ke dar natije-ye moqābele peydā šod ānče-rā ke mohemm o lāzem did dar ḏeyl-e ṣafḥāt yāddāšt kard (Among the differences between manuscripts [i.e.variant readings] that became evident during the process of comparison [i.e.collation] the ones that seemed important and necessary were recorded in the footnotes [i.e. the apparatus]). 196.He adapted the text to the modern orthographical standard.7.He compared the text with "thirty old witnesses and anthologies", 20 and included the poems found in them that were not already present in his text.
16 Tackling the highly complex issue of what part of a line precisely is to be considered its qāfiye (rhyme) is far beyond the scope of this paper.MR simply ordered the poems alphabetically, considering the letters of the first line in reverse order, beginning from the last.The expression 'by rhyme' will be used for simplicity.
17 By this he means the manuscripts selected to be used for the edition.
18 Dated 1309 CE.For a brief description see "Appendix 1".Furthermore, the editor states that he checked the form in which Anvari's lines were quoted in classical dictionaries, such as the Jahāngiri.Another important information is provided in the methodological additions in the introduction to the second volume (Anvari 1959-61, 2: 158-63): 22 even when all manuscripts were clearly wrong (e.g. a line did not fit the metre), MR never included personal conjectures in the text, but always chose between what he found in the manuscripts.This is probably also true for the headings.Many poems have a heading mentioning a clearly wrong addressee (as the name of another patron is mentioned in the poem itself).MR simply chose one of the titles mentioned in the manuscripts, even if he knew they were wrong. 23n regard to the outlined method, a series of remarks must be made.a.The editor's starting point was a text he himself compiled (from unidentified sources), not a specific copy to which he subsequently collated other witnesses.This makes the starting point inevitably biased.No criterion that guided the selection of witnesses is ever given, nor the editor explains how he chose between variant readings.MR also does not give any information on how manuscripts could be related to each other.b.In many manuscripts used by the editor 24 poems are not arranged 'by rhyme' (according to the last letters of each beyt) as it happens in other manuscripts and most modern editions.They are instead ordered in a much different way, loosely grouped by addressee or theme (de Blois 1995). 25In listing and describing the manuscripts he used, MR

Giacomo Brotto
Modarres-e Reḍavi's Edition of Anvari's divān: A Critical Assessment any indication to where a poem is situated in manuscripts not arranged 'by rhyme' is provided anywhere.This could have been useful not only to understand the criterion by which poems were originally organised, but also to know which poems were traditionally thought to be linked together.This information could prove important in understanding the poems themselves.c.Istanbul, Fāteḥ, 3784, siglum 'lām', was chosen as the editor's nosxe-ye aṣli (base witness) partly on the false assumption it was the oldest among the collected manuscripts.London, British Library, 3713, siglum 'te', is older, 27 but it was at first described imprecisely by MR, who states: "This is a photographic reproduction (nosxe-ye ʿaksi) of a collection of divāns of different poets kept in the Paris national library" (Anvari 1959-61, 1: 18); he also probably did not know the date. 28This mistake was fixed in the introduction to volume II, 29 but the choice of lām as nosxe-ye aṣli had already been made.This is not the only instance of MR giving misleading information on a manuscript: he mentions the date 753 Q./1352 CE for Tehran, Ketābxāne-ye bayāni, 54/2, siglum 'dāl' (Anvari 1959-61, 1: 17), although the scribe's subscriptio clearly mentions 768 Q./1367 CE. 30 d.The editor did not record every variant reading in the apparatus, but only the ones he deemed "important and necessary".Even trusting the editor not to have excluded any potentially correct variant reading, this makes it impossible to forward any hypothesis on the relation between manuscripts using only MR's apparatus.e.The "thirty old witnesses and anthologies" to which the editor compared his almost finished work are never described, nor any siglum is assigned to any of them in the introductions.The use of these unidentified witnesses could explain but it seems that another order is being observed.Panegyrics dedicated to a patron follow each other, and in the same way qeṭʿes addressed to that patron follow the panegyrics".In the descriptions of some of the other manuscripts the editor quotes the maṭlaʿ 'opening couplet' of the first qaṣide, from which the order of the poems can be guessed.the presence in the apparatus of sigla such as 'sin', to which no witness corresponds; it seems that some of those witnesses were assigned a siglum by the editor during his work, but he did not state them anywhere. 31

The Apparatus
As widely known, the most important part of a critical edition is the apparatus.Even if some of the variant readings chosen by the editor are proven to be wrong, a well built apparatus allows any scholar to study the textual tradition of the edited work independently of the editor's opinion.In a case such as Anvari's divān, where the relationship between manuscripts has not yet been studied, it is even more important to be able to systematically compare the variant readings given by different witnesses.
A brief description of these four manuscripts is given in "Appendix 1".According to their descriptions as given in both introductions, MR seems to have heavily relied on them; this is especially true of the two Istanbul manuscripts lām and ʿeyn.
31 'sin' is present, for example, in the apparatus of qaṣides 71 and 77.
32 Since the editor, at times, records in the apparatus every kind of non merely graphical variant reading (e.g. the difference between spellings ‫کی‬ and ‫که‬ for ke), including the presence or absence of the conjunction o and the confusion between the demonstratives ān and in, every non merely graphical variant reading has been taken into consideration.
33 Istanbul, Fāteḥ, 3784 has pencil written numbers on the upper left corner every 10 ff.The count appears, from 20 onwards, to be off by 1 (20 is 19, 30 is 29 and so on).This seems to be a simple miscounting issue, I could not find a place where a page might have fallen among the first 20.

97
The two qaṣides are addressed to the same patron, Ṣadr al-din Moḥammad, likely a great-grandson of the famous vizir Neẓām al-Molk (Anvari 1959-61, 2: 67-8).They share the same rhyme, metre and general structure; moreover, in both prologues the poet describes his love sickness (in 80 his beloved, maʿšuq, is far away, in 84 the poet chooses to depart from him) from which he finds comfort at his patron's court.The transitional lines, gorizgāh, are also very similar.
This comparison has shown that many variant readings were not reported in the apparatus by the editor: e.g.qaṣide 80, l. 8, the variant reading nāle] nowḥe is recorded only as given by witnesses ṭā and mim; this variant reading is also present in lām and te; qaṣide 84, l. 42, šaxṣ-e ajal] aʿdā-t-rā in ʿeyn is not recorded at all.Recording the presence of variant readings only in some of the manuscripts in which they are actually present can cause the reader to underestimate their importance and diffusion.
Even things as potentially relevant as absent or misplaced lines are at times not recorded properly: e.g. in qaṣide 80 ll.6-10 appear in a completely different order in lām and te, 8>6>9>7>10; this is not recorded at all.Of course any formal analysis of a poem cannot ignore the order of the lines.In regard to qaṣide 80, l. 14, MR adds a note saying that in witness te only the first 14 lines are present: this is not true, in te the whole poem can be read.The presence of qaṣide 84 in manuscript te is ignored altogether; the manuscript is not listed in the apparatus among the ones in which this poem is present.
Tehran, Ketābxāne-ye malek, 5267, has many interlinear corrections (in which a word perceived to be wrong is corrected by writing the 'correct' one above it, between two lines of text) and additions: lines that were thought to be missing were added outside the main body of text, in the margins.This is clearly signalled in the manuscript itself since before both qaṣides 80 and 84 the Arabic letters ‫صح‬ are written near their headings, indicating that the text was 'corrected' (Arabic ṣaḥḥa) by comparing it with another copy (Déroche, Sagaria Rossi 2012, 216-22).MR's approach to these additions is not coherent: sometimes the editor acknowledges them, sometimes he records the variant readings present in these interlinear corrections or added lines without mentioning that they do not belong to the main body of text, some other times he just ignores them.
For qaṣide 84, l. 7, MR records: ʿazm] dar ḥāšiye-ye kāf (in kāf's additions): ʿoḏr.This shows that MR was aware of the issue.For qaṣide 80, ll.19-20, MR records eḥkām] farmān and farmān] taʾid in kāf.These two lines are actually absent from kāf's main text, and are added in the margins; it is in these additions that these variant readings are present.Examples of the editor not recording these corrections and additions are plenty, see "Appendix 3".
Even though proposing a new text is not the aim of this paper, the editor's choices do appear at times questionable: e.g. in qaṣide 80, l.
3, havā siyāh be kerdār-e qirgun xaftān // falak kabud nemudār-e nilgun meġfar (The air [was] as black as a tar coloured kaftan, the sky [was] as blue as a Nile coloured helmet), 34 MR chooses nemudār instead of another be kerdār in the second meṣrāʿ, even with be kerdār being present in all four witnesses above mentioned and maintaining a perfect parallelism between the two meṣrāʿs. 35In qaṣide 84, l. 7, bahāne-ye safar o ʿazm-e raftan āvardi // del-at ze ṣoḥbat-e yārān malul gašt magar (You have brought forward excuses for a voyage and the intent to go, has your heart grown tired of the company of friends?), MR chooses ʿazm (intent) over ʿoḏr (forgiveness/pardon), which is present in all four witnesses above mentioned. 36In qaṣide 84, l. 23, če dast-e u be saxā dar če abr dar neysān // če ṭabʿ-e u be soxan dar če baḥr-e bi-maʿbar (His hand in generosity [i.e.his generous hand] resembles a cloud in the month of Neysān, his aptitude for speech resembles a sea without crossings), 37 MR chooses abr dar neysān (the cloud in the month of Neysān) over abr-e bi-noqṣān (the cloud without fault): this second reading, present in te and lām (MR's nosxe-ye aṣli), 38 preserves a closer parallel between the two meṣrāʿs, with both the patron's hand and his aptitude for speech compared to a noun described by an adjective build with the prefix bi-, 'without'. 39 'Nile coloured helmet' indicates the deep and shining blue-gray of the night sky.
The translation of meġfar is at times problematic, for it can indicate both the helmet and the mail or network of steel worn under it.
35 In the apparatus he records be kerdār as a variant reading present in lām, his base witness, but not in the other three, see "Appendix 3".The correctness of the second be kerdār is reinforced by Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 2615 and Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 4113 which both have this reading.The meaning or translation of the line does not drastically change.
36 This might have happened because MR did not record ʿoḏr as a variant reading present in lām, ʿeyn and te.See "Appendix 3".The correctness of ʿoḏr is reinforced by Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 2615 and Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 4113, which both have this reading.The line thus reads "You brought forward excuses for a voyage and [asked for] forgiveness for going away...".
37 Meaning the patron's hand is as generous as a spring cloud gifting rain and his aptitude for speech is as wide as an immeasurable sea (the cloud being an established symbol of generosity, the sea, with its width and richness of pearls, an established metaphor associated with speech).The Persian construction with two čes per meṣrāʿ was translated quite freely.
38 This is not acknowledged in the apparatus, where only čo abr dar neysān] in ṭā and če: čo abr-e bi-noqṣān is recorded.
39 Moreover Neysān, the month traditionally associated with spring, of common use in spring descriptions, is identifiable as lectio facilior.The reading bi-noqṣān is also present in both Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 2615 and Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 4113.Furthermore, the juxtaposition between abr-e bi-noqṣān and baḥr-e bi-maʿbar is also present in other lines of Anvari's divān, e.g.qaṣide 81, l. 11 (Anvari 1959-61, 1: 199).The beyt would thus read: "His hand in generosity [i.e.his generous hand] resembles/it is more generous than the cloud without fault, his aptitude for speech resembles/it is more rich than the sea without crossings".

Giacomo Brotto
Modarres-e Reḍavi's Edition of Anvari's divān: A Critical Assessment 99 Crucial information about manuscripts, even key ones, seems to be lacking: e.g.Istanbul, Fāteḥ, 3786, one of MR's main witnesses, is never mentioned to be in disorder, as shown by facts such as qaṣide 59 beginning on f. 5v, being interrupted and then continuing on what is now f. 40r.In this manuscript, also, qaṣide 68 is written twice: ff.33v-34v have the first 24 lines while ff.126r-127v have the whole text, albeit with a completely different maṭlaʿ, duš čun čašmeye xʷoršid-e sepehrdavār // gašt az časm nehān dar pas-e parde šab-e tār (Last night, when the sun's disk became invisible to the eyes beneath that veil that is the dark night...).The maṭlaʿ printed in the edition is: di čo beškast šahanšāh-e falak nowbat-e bār // v-az sarāpardeye šab gerd-e jahān kard ḥeṣār (Yesterday, when the king of kings in the sky [i.e. the sun] interrupted court time [i.e.set] and the world became encompassed by night's veil...). 40Neither the double presence of this qaṣide nor the different maṭlaʿ are recorded in the apparatus (Anvari 1959-61, 1: 154).

Conclusions
MR's edition has been used for many years without taking into account or discussing its multiple problems, especially among western scholars. 41This is somewhat surprising since another of MR's editions of medieval Persian poetry, Sanāʾi's Ḥadīqatu l-Ḥaqīqah (in Arabic, The Garden of Truth), was the object of a very harsh review by H. Ritter, who wrote: As praiseworthy as this edition is, this preface already makes clear that the editor, much differently from the great master Muḥammad Qazvini, is one of those Persian scholars who do not give much value to precision and proper documentation.(Ritter 1952, 190)   Ritter proceeded to point out to multiple instances of poorly described manuscripts, the absence of any criterion in the evaluation of variant readings and, judging from the facsimile reproductions of the folia of some manuscripts printed in the edition, many inaccuracies in the recording of variant readings in the apparatus.I hope to have shown that anyone who wishes to study Anvari's divān has to take into account how MR's edition was made, and 40 Both maṭlaʿs are temporal clauses, linked by enjambement to the following line (ruy benmud mah-e ʿid... "ʿId's moon showed its face...").An eżāfe particle seems to have fallen for metrical reasons between parde and šab (or parde is to be considered an apposition of šab-e tār).
41 Up until now: still in de Bruijn 2019 Anvari's poems are quoted verbatim from MR's edition, without saying anything about it.proceed accordingly.Although MR's work is still the essential starting point for any research on Anvari and certain aspects of it still are very valuable, especially the learned and extensive introduction, the text and the apparatus must always be closely scrutinised, not only by reading old manuscripts not used by MR, but also by double-checking the ones used by the editor.This is especially important for future philological works on Anvari's poetry, which is highly necessary.

Appendix 1 42
Manuscripts that are not ordered alphabetically, by rhyme, are all organised in approximately the same way: they begin with qaṣide 60, addressed to Sanjar, and tend to group together poems (both qaṣides and qeṭʿes) addressed to the same patron.This tendency, however, is not always followed: more than one block of poems having the same addressee can be found in the same manuscript while poems addressed to a different patron are, at times, inserted in the 'wrong' block; moreover, some texts seem to be grouped by 'theme' (e.g.qeṭʿes of similar subject).
It is to be noted that the order of poems in these manuscripts is never exactly the same.
The four manuscripts used for the sample collation of qaṣides 80 and 84 are: 1 42 For all the following manuscripts, references to Monzavi's catalogue, Monzavi 1971, have been added when possible.To avoid repetitions manuscripts will be ordered consequentially across appendixes 1 and 2; when a manuscript appears in more than one list, the number of its first mention will be referenced, where it will have been briefly described.
43 Literal translation.The line means: If the state of things/the world does not change according to qaḍā (Fate, divine decree), why does not it change in the way that humans want?It is a rhetorical question, of course things change per divine decree, and so things do not go how mortals want them to go.This conventional wisdom theme is developed throughout the qaṣide's prologue.and a hand were a sea and a mine they would be the sultan's heart and hand).It seems to be an unfinished product, f. 110v is the last page for which the jadval has been traced; many headings are not written.Some folia are not in the original order and some seem to have fallen.Corresponds to item 21659 in Monzavi (1971, 2235) and a hand were a sea and a mine they would be the sultan's heart and hand).MR describes it as 'full of mistakes'.It was compared with another manuscript and there are many corrections, in which the correct word is written above the perceived mistake, between the lines; lines that were thought to be missing were added in the margins.Corresponds to items 17451 and 21658 in Monzavi (1971, 1847

Appendix 2
The following section lists the manuscripts mentioned by the three above quoted papers arguing the necessity of a new edition.
The original was part of the private collection of Afšār Širāzi.The final page was lost and with it the copyist's subscriptio.
Neither the name of the scribe nor the date are present, it could date back to the seventh Q./thirteenth century CE. 46 Layout of the page: two columns, 17 lines per page, one meṣrāʿ per column.Poems are not ordered by rhyme.Lines that were thought to be missing have been added at times in the margins.The first poem is qaṣide 60, maṭlaʿ gar del o dast baḥr o kān bāšad // del o dast-e xodāygān bāšad (If a heart and a hand were a sea and a mine they would be the sultan's heart and hand).It is Šahidi's 'nosxe-ye afš' and it seems to be now lost (Beelaert 2017).Corresponds to item 21710 in Monzavi (1971, 2238). 476.Tehran, Ketābxāne-ye majles-e šurā-ye eslāmi, 86666. 48

Giacomo Brotto
Modarres-e Reḍavi's Edition of Anvari's divān: A Critical Assessment its writing is more or less of the eighth century Q./fourteenth century CE" (Anvari 1959-61, 2: 146). 69Layout of the page: two columns, 15 lines per page.It begins with what in MR's edition is l. 17 of qaṣide 16 z-ānke emruz az ulū l-amr-i 70 o yazdān dar nobi // hamčonin gofta-st o ḥaqq in-ast o digar torrahāt (since today you are among the ulū al-amr, thus has God spoken in the Quran, this is the truth, all other things are insignificant). 71 3) = 2) Nurāyi, Aḥmadpur 2016 Tehran, Ketābxāne-ye majles-e šurā-ye eslāmi, 86834. 72It is an anthology containing a selection of poems from 17 different poets, including Anvari.It has no date, it is said to have been written at the end of sixth Q./end thirteenth-beginning fourteenth CE. 73 = 10) It is to be noted that although all three papers underline the necessity of a scientific method in compiling a new edition, no such method is described.This can already be seen by the different lists of manuscripts suggested by the authors: a criterion on the basis of which manuscripts should be selected is never stated, nor a philological argument to justify using a witness over another is ever made; witnesses are included or excluded seemingly at random.l. 5a rox-am] zax-am in te; 77 jān 2 ] in kāf: del.This variant reading is recorded in the apparatus, but it is not mentioned that jān, MR's chosen reading, is written above del.ll.6-7 omitted in ʿeyn and kāf.MR only records the absence of l. 7 from ʿeyn.In kāf both lines are absent from the main text but they are added outside of it, l. 6 on the left margin and l. 7 on the right one, without any variant reading.ll.6-10 The order of these lines is different in lām and te: Referencing this line, in footnote 8 of the apparatus, MR writes: "Witness te does not have more than the first 14 lines of this qaṣide, which, from here onwards, has fallen from this witness" (Anvari 1959-61, 1: 196).This is not true.In fact, in manuscript te, beyt 15 follows beyt 14 in the very same line of text, with the first beyt occupying columns 1-2 and the second columns 3-4.Together they make up the last line of f. 58v, the poem continuing without any interruption on the next page.

Giacomo Brotto
Modarres-e Reḍavi's Edition of Anvari's divān: A Critical Assessment 109 only present in the added lines, not in the main text.Also, in those added lines, l. 19a be nik] bar nik is not noted in the apparatus. 79l. 21a qaḍā] in ʿeyn: falak; l. 21b qadar] in kāf: qamar, with qadar, MR's chosen variant reading, written over qamar.The apparatus notes l. 21a towfiq] in kāf: towqiʿ.In kāf ‫قیع‬ is written above be towfiq, showing that, according to whoever made the correction, the correct reading was be towqiʿ.However, this footnote does not show that MR's chosen variant reading is actually part of kāf's main text.Also, the apparatus notes qaḍā] in kāf: falak, but not that qaḍā, MR's chosen variant reading, is written above falak.l. 22a qaḍā] in kāf: qadar, with qaḍā, MR's chosen variant reading, written above qadar; l. 22b qadar] in ʿeyn: falak (MR's apparatus erroneously records qadar] in ʿeyn: qaḍā).MR's apparatus records qadar] in kāf: falak, but not the full variant reading: qadar bepičad] in kāf: falak betābad, with qadar bepičad written above the writing line.l. 24 konand] in lām: ‫,کنبذ‬ in both meṣrāʿs; 80 l. 24b v-az ān] in kāf: v-az in, with v-az-ān, MR's chosen variant reading, written over v-az in.MR's apparatus records l. 24a k-az-in] in kāf: az ān, but not that k-azin, MR's chosen variant reading, is written above az ān; l. 25b navāl-aš] in ʿeyn and kāf: navāz-aš (in kāf ‫,لش‬ l-aš, is written over navāz-aš, showing that, according to who made the correction, the correct reading was navāl-aš.MR's apparatus records l. 25a ʿetāb-aš] in kāf: nehib-aš, but not that ʿetāb-aš, MR's chosen text, is written over nehib-aš.l. 26a ān] in lām and te: in; in] in lām, ʿeyn and te: ān; baxur-e ʿabir] in lām: baxur o ʿabir; l. 26b in] in lām and te: ān; ān] in ʿeyn and te: in, omitted in lām (where in is added in the margin); boxār-e šarar] in lām, kāf and te: boxār o šarar.ll.27 and 28 omitted in kāf, where they are added in the margins.MR's apparatus records l. 27b ke] in kāf: čo, without mentioning that this variant reading is only present in the added line, not in the main text.In l. 28a the variant reading va gar] agar, present in the added text, is not recorded.In the same added line the variant reading gah] gar could also be present.l. 27a baḥr-e saxā] in te: baḥr o saxā; hami] omitted in te; l. 27b ke] in te: čo.l. 29a ze sim o zar] in ʿeyn: ze sim-e zar; deram] in lām: gohar.MR's apparatus also does not record that in kāf gohar is written over deram.The variant reading deram] in te: gohar is recorded, which 79 In kāf the end of the added l. 20 could have the variant reading bebaste kamar] na baste magar, but the text is small and partially erased, making it very difficult to read.80 Maybe to be read gonbad (dome).The variant reading konand, chosen by the editor, gives a much better meaning.makes the editor's statement in footnote 8 of the preceding page, here quoted in regard to l. 14, even more puzzling.l. 30b be rafʿat o hemmat] in lām and te: be hemmat o rafʿat.The variant reading baxšeš] in kāf: rafʿat is recorded in the apparatus, but it is not shown that baxšeš, MR's chosen variant reading, is written over baxšeš.l. 31 omitted in ʿeyn and kāf; in the latter it is added in the margin.l. 32 ma-rā] in kāf: to-rā.The situation in kāf is peculiar: l. 31 was originally omitted, but afterwards someone tried to integrate l. 31 by writing some words above l.32, the first meṣrāʿ of which has been written between the two columns.In the manuscript the line reads to-rā sazad ke bovad gāh-e naẓm-e medḥat-e to // bayāḍ ruz o siyāhi šab o qalam meḥvar with tāʿat o farmān written above the first meṣrāʿ, marā sazad ke bovad gāh-e naẓm-e medḥat-e to written vertically between the two writing columns and falak ġolām qaḍā bande qadar čākar written above the second meṣrāʿ.l. 33a meh az jahān agar andar jahān] in te: hazār jān be jahān dar agar; 81 agar andar jahān] in lām: be jahān dar agar; andar jahān] in kāf: andar u (be jahān dar agar is written above the line); 82 l. 33b bed-u andar] in te: be-d-u-yi dar MR's apparatus records l. 33b be-d-u andar] in lām: be-d-u dar.This is incorrect.This manuscript, like te, has the variant reading be-du-yi dar, which, in opposition to the recorded be-d-u dar, fits the metre.MR records the variant reading be-d-u dar also for witness kāf.This is also incorrect: kāf's main text reads be-d-u andar, the variant reading chosen by MR, with be-d-u-yi dar written over it.ll.34-7.MR records the omission of these lines in kāf, but not that they are all added in the margins. 83ll.35-9.Line 38 directly follows l. 35 in lām, where the order is: 35>38>36>37>39.l. 35b diyār] in lām: zamāne.This variant reading could also be present in kāf's added lines.ll.36-7 omitted in te.
83 The text is small and hard to read, the conjunction o seems to have been omitted both after ḥekmat and hešmat, in the first and second meṣrāʿ of l. 34 respectively, as does the o after hešmat in l. 35b. to in l. 37a seems also to be omitted.

Qaṣide 84
MR does not mention at all that this poem is also present in manuscript te.All the variant readings from this manuscript are recorded here.l. 1a čo] in kāf: ke, with čo, MR's chosen variant reading, written above ke.l. 4a čo andar ātaš ʿud] in ʿeyn and te: čo ātaš andar ʿud.l. 5b.The variant reading nadāram] in kāf: natābam is recorded in the apparatus, but that nadāram, MR's chosen variant reading, is written above natābam, is not.l. 6b nāraside] in te: nāgoḏašte.l. 7a ʿazm] in ʿeyn, lām and te: ʿoḏr.In kāf ʿoḏr is written over ʿazm, MR's chosen variant reading.This seems to be one of the few instances in which the editor explicitly acknowledges the existence of kāf's additions in the apparatus, recording: ʿazm] in kāf's additions (ḥāšiye): ʿoḏr.l. 8a raftan] in kāf: ġeybat (raftan, MR's chosen variant reading, is written above ġeybat); kardan] in kāf: raftan (kardan, MR's chosen variant reading, is written above raftan).MR's footnote 4 (Anvari 1959-61, 1: 209), recording the variant reading vaqt-e raftan o 84 This variant reading could be also present in ʿeyn, but it is not easy to see if the copyist actually wrote the be's body ‫بی(‬ ‫سخا‬ vs. ‫.)سخای‬85 Line 38 has two '20's written above it, referencing footnotes.The second, after bi in the second meṣrāʿ, is probably a misprint for '21', since footnote 20 records the variant reading jud o saxā-ye kaf, corresponding to the first 20 of l. 38, written after kaf, and no '21' is present above any of page 197 lines despite the presence of a footnote with this number.This is wrong.The variant reading present in witness kāf is molkat o din that, contrary to mamlakat o din, fits the metre (also, mellat o molk, MR's chosen variant reading, is written above molkat o din); l. 20b ebtedā be ʿadl] in kāf: ebtedā-ye ʿadl, with be ʿadl written above it, indicating that, according to the person who made the correction, the section was to be read ebtedā be ʿadl.MR's apparatus notes ebtedā be ʿadl] in ʿeyn: ebtedā-ye ʿadl; this is also wrong, witness ʿeyn has ebtedā be ʿadl.l. 22 omitted in ʿeyn and kaf; l. 22b basāyaṭ] in lām and te: basāṭat.In kāf this line is added in the margin, with the variant reading basāyaṭ] basāṭat.l. 23a dar neysān] in lām and te: bi-noqṣān; l. 23b baḥr] in te: rud.l. 24b be taqviyat] in ʿeyn: ze taqviyat, in kāf: ze tarbiyat.l. 25b šur o fetne] in lām, ʿeyn, kāf and te: šur-e fetne.MR's apparatus records šur o fetne] in ʿeyn: suz-e fetne.This is probably wrong, as the dot of the supposed ze most likely refers to jāh's jim in the line above, and the relevant word is to be read šur, albeit with no dots over šin.MR's apparatus also records l. 25a vasan] in lām: rasan.Due to the way the scribe writes res and vāvs, it is possible that this variant reading does not exist.ll.26-7.The order of these two lines is reversed in lām and te.l. 26a če šir o če gorg] in lām and te: če kabk o če ‫.کرک‬ 87 86 In lām there seems to be a dot above mānd mim, but nānd does not exist.l. 42a ḥesām-e qahr-e to šaxṣ-e ajal zanad be do nim] in ʿeyn: ḥisām-e qahr-e to aʿdā-t-rā zanad be do nim. 91l. 43a qahr-at] in ʿeyn and kāf: ‫;محنت‬ in manuscript kāf qahr-at, MR's chosen variant reading, is written above ‫.محنت‬ 92l. 44a dāruy o taryāk] in lām, ʿeyn and te: dāruy-e taryāk, in kāf: dāru va taryāk; in Farroxi's poetry and in other lines of Anvari's divān), reading karg 'rhinoceros' might be correct: "When his [the patron's] falcon is hunting, what's a partridge and what's a rhinoceros?(He catches both preys, even if one is easy and the other should be impossible.)When his horse is advancing on his path, what is a sea and what is a plain?(He traverses both, even if one is easy and the other should be impossible)".Furthermore, the rhinoceros is mentioned as the mamduḥ's prey, being very difficult to capture, in Farroxi's poetry; see de Fouchécour 1969, 152.The line has a comprehensible meaning also by reading gorg, but the hyperbole and the symmetry between meṣrāʿs do not work as well (and, to the extent of my knowledge, the very commonly mentioned wolf is never a prey of the mamduḥ's hawk, see de Fouchécour 1969, 156).The unusual mention of a small prey, the partridge, alongside a very big one, the rhinoceros or the wolf, could explain why kabk 'partridge' present in many old manuscripts (lām, te, Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 2615 and Tehran, Ketābxāne-ye dānešgāh-e Tehrān, microfilm 4113), was changed to šir 'lion' by copyists who did not understand the line and mechanically replicated the hendiadys šir o gorg 'lion and wolf'.It is likely that the reading šir, chosen by MR, is not the correct one.
88 Small and difficult to read, the extra letter, that makes the line not fit the metre, is surely a mistake.89 In l. 37, in MS lām there seems to be a meaningless dot below ‫رد‬ ‫.ک‬ 90 ġalm is an Arabic maṣdar meaning 'being lustful', nearly never used in Persian.This dot is very likely a mistake.91 ḥisām does not exist and it is not metrically possible, but the yā's two dots are clearly written.
95 Could be read moḥanjer, a rare Arabic word for 'throat', mentioned for example in the Majmaʿ fī bihār al-anwār, a sixteenth-century dictionary by Muḥammad al-Fattinī (s.v.ḥanjar, consulted online through the website arabiclexicon.hawramani.com,available at http://arabiclexicon.hawramani.com/).96 The first letter could be a mim with a large head as well as a ḥe (e.g.jegar).In any case the correct reading is kamar, the one chosen by MR.

44
Both in the margins and between the two columns of text, vertically.In Mozavi 1971 the first time is mentioned under kolliyāt-e Anvari the second under divān-e Anvari.45 He was not the only one to work on this manuscript.According to Rieu (1895, 143), Hamgar's quatrains were copied by Eṣḥāq b.Qevām Moḥammd Hamgar, the poet's grandson.
maṭlaʿ gar del o dast baḥr o kān bāšad // del o dast-e xodāygān bāšad (If a heart and a hand were a sea and a mine they would be the sultan's heart and hand).
8.He compiled the glossary and the indexes.21

Giacomo Brotto Modarres-e Reḍavi's Edition of Anvari's divān: A Critical Assessment
does not systematically give information on how poems were arranged, 26 nor 21 He describes the process in detail; this part is summarised since it is not relevant to the present discussion.22Inparticular 160.23This is shown, e.g., by qaṣide 'panegyric' 67: the title mentions abu al-Ḥasan Layout of the page: two columns, 29 lines per page.Poems are ordered alphabetically, by rhyme (radif excluded).The first poem is qaṣide 18, maṭlaʿ agar moḥāvel-e ḥāl-e jahāniyān na qaḍā-st // čerā majāre-ye aḥvāl bar xelāf-e reḍā-st (if the constant turning [i.e.changing] of the mortals condition is not [dictated by] Fate, why the flowing of events is different from what is wanted [by humans]?). 43MR declares it to . lām: Istanbul, Fāteḥ, 3784.Copied by Moḥammad b. ʿAbdallāh b.Moḥammad al-Ḥāfeẓ and completed in avāxer-e māh-e Šavvāl 708 Q. (late Šavvāl 708 Q.)/April 1309 CE.

Reḍavi's Edition of Anvari's divān: A Critical Assessment 101 be
his nosxe-ye aṣli.Corresponds to item 21660 inMonzavi  (1971, 2235).2. ʿeyn: Istanbul, Fāteḥ, 3786.Neither the name of the scribe nor the date are present since it lacks the final pages and with them the copyist's subscriptio.MR dates it to the end of seventh Q.-beginning of eighth Q./end of thirteenth-beginning of fourteenth century CE.Layout of the page: two columns, 19 lines per page, one meṣrāʿ per column.Poems are not ordered by rhyme.The first poem is qaṣide 60, maṭlaʿ gar del o dast baḥr o kān bāšad // del o dast-e xodāygān bāšad (If a heart . 3. kāf: Tehran, Ketābxāne-ye malek, 5267.Neither the name of the scribe nor the date are present since it lacks the final pag- es and with it the copyist's subscriptio.MR considers it very old, written in the seventh Q./thirteenth century CE.This dating is also given in the catalogue (Afšar et al. 1975-76, 264).Layout of the page: two columns, 20 lines per page, one meṣrāʿ per column.Poems are not ordered by rhyme.The first poem is qaṣide 60, maṭlaʿ gar del o dast baḥr o kān bāšad // del o dast-e xodāygān bāšad (If a heart and 2235 respectively). 444. London, British Library, Or. 3713.Collection of seven divāns, Anvari's is the fifth, ff.36v-125r.Copied by Moḥammadšāh b. ʿĀli b.Maḥmud al-Eṣfahāni, 45 completed on the 6th of Rabiʿ II 693 Q./13th of March 1294 CE.Layout of the page: four columns, 31 lines per page, one meṣrāʿ per column, two consecutive beyts (couplets) per line.Poems are not ordered by rhyme.The first poem is qaṣide 60,

Reḍavi's Edition of Anvari's divān: A Critical Assessment 12
Layout of the page: two columns, 19 lines per page, one meṣrāʿ per column.Lines that were thought to be missing have been added at times in the margins.Poems are not ordered by rhyme.The first poem is qaṣide 60, maṭlaʿ gar del o dast baḥr o kān bāšad // del o dast-e xodāygān bāšad (If a 46 Described by Nafisi in the introduction to his edition of Anvari's divān.He considers it written in the thirteenth-fourteenth century.47 The date Monzavi refers to, 1086 h.q./1675-76 CE, is on the new last page, clearly added later when the manuscript was restored.48 This number is the šomāre-ye ṯabt.The šomāre-ye fehrest is 13503. .Širāz, Ketābxāne-ye ḥeḍrat ʿAlā al-din Ḥoseyn, 1214.Dated 1248 Q./1832-33 CE. 58 13.Tehran, Ketābxāne-ye majles-e šurā-ye eslāmi, 8431. 59Copied by Moḥammad ʿEbrat Nāyebi and finished in 1242 Q./1826-27 CE. 60 Ketābxāne-ye dānešgāh Tehrān, microfilm 4113.It lacks both the initial and the final part, some ff.seems to have fallen inside also.Neither the name of the scribe nor the date are present, estimated to have been copied at the end seventh-beginning eighth Q.
Dated 680 Q./1281-82 CE, it is the oldest known dated copy of Anvari's divān.The last page is badly damaged, and much of the copyist's subscriptio is not readable, including the copyist's name and the day and month in which the manuscript was completed./endthirteenth-beginningfourteenthcenturyCE.It is Šahidi's 'nosxe-ye minow' (seeBeelaert 2017) and it is also the witness used by Šafiʿi Kadkani (1993, 69, fn.69).Layout of the page: two columns, 17 lines per page, one meṣrāʿ per column.Poems are not ordered by rhyme.It begins with the last 5 couplets of qaṣide 160.A heading found in f. 128v, in bāb ašʿār o moqaṭṭaʿāt-e parākande dar ḥaqq-e xʷāje Ṭāher bā ḏamm-e šāʿeri o esteġfār az u (This part: scattered poems in honour of lord Ṭāher and 58 Non vidi.No further information given.and'askingfor forgiveness to Ṭāher'; being that 'ḏamm-e šāʿeri' is mentioned immediately before, the first option seems better.62Thus in Arberry, Minovi, Blochet 1959, 10.63 The same scribe who copied London, British Library, Or. 3713.
l. 12a The conjunction o is added after derāz in lām, ʿeyn and te; do čašm-am] in kāf: ze čašm-am (do, MR's chosen reading, is written above ze); ze nowk] in ʿeyn and te: be nowk; hami] omitted in lām.MR's apparatus records ze nowk] in kāf: do nowk, but it does not show that in this manuscript be nowk is written above do nowk.l.13b bar] in lām and te: dar.MR's apparatus shows aṯar] in kāf: xabar.It does not show that aṯar, MR's chosen variant reading, is written both above xabar and in the left margin.l.14a ʿešve] in lām and te: ʿešq; be dast-e ešve hame šab gerefte dāman-e del] in ʿeyn: be dast-e ʿešq gerefte omid dāman-e del (MR's apparatus shows the variant reading dast-e ešve hame šab gerefte dāman-e del] in kāf: be dast-e ʿešq gerefte omid dāman-e del, but not that the editor's chosen text is added above the line).78

Reḍavi's Edition of Anvari's divān: A Critical Assessment
87 I am not sure ‫کرک‬ is to be read gorg 'wolf', as in MR's edition.Even if it is not an animal frequently mentioned by Persian poets (although it appears in Ferdowsi's Šāhnāme,