SOURCE: Davis, Dick. “Narrative and Doctrine in the first story of Rumi's Mathnawi.” Journal of Semitic Studies, Supplement 12 (2000): 93-104.

[In the following essay, Davis examines the first story of the Mathnawi and suggests that the interruptions and interpolations within the story undermine the entire structure of explication, leaving the reader with a poem that simultaneously demands and resists interpretation, subverts allegorical meanings as soon as it establishes them, and remains stubbornly grounded in human experience even as it attempts transcendence.]

I first read medieval Persian in Tehran, in the early 1970s, under the tutelage of my friend, later my PhD adviser, Norman Calder. We began, hubristically enough, with the first book of Jalāl al-Dīn Rūmī's (d. 1273) Mathnawī, which Norman guided me through with exemplary kindness, patience, enthusiasm and erudition. Now, almost thirty years later, I look back with deep nostalgia and pleasure on those afternoons we spent together in Tehran, parsing Rūmī and talking about poetry both Persian and Western. As I certainly owe my subsequent professional involvement with Persian to Norman's initial guidance, I gratefully offer this essay on those same opening pages of the Mathnawī to his memory.

I

Rūmī is explicit in acknowledging his debts to his predecessors in the genre of the mystical mathnawī, Hakīm Sanā'ī (d. c. 1130) and Farīd al-Dīn ‘Aṭṭār (d. c. 1220), and his own mathnawī is clearly written within the tradition that they had established and developed. The couplet form is used to string together anecdotes interspersed with mystically oriented homily; the anecdotes refer to people in many walks of life (kings, merchants, servants, peasants, etc.), so that the audience/reader both has a sense that all human life contributes to the vision adumbrated, and that the vision is in turn applicable to individuals at all levels of human society. The anecdotes refer to the famous and the anonymous, to figures of religious significance and to secular princes, to historical personages and to mythological ones, so that the human panorama not only permeates all social levels but all imaginative levels as well. And within these all-inclusive concerns, one can detect in the work of all three authors a particular affection for the poor, the humble and the anonymous as they figure in the stories. These characteristics of the mystical mathnawī as a genre, laid down by Sanā'ī and elaborated by ‘Aṭṭār, are shared by Rūmī's work.

But, despite this acceptance of the overall features of the genre as laid down by his predecessors, Rūmī's Mathnawī-yi Ma‘nawī begins, when compared with the mathnawīs of Sanā'ī and ‘Aṭṭār, in a very peculiar way. The first significant mystical mathnawī in Persian is Sanā'ī's Hadīqat al-Haqīqat, and Rūmī's mathnawī is in many ways closer to Sanā'ī's pioneering work than it is to the intervening and apparently more sophisticated examples of the genre created by ‘Aṭṭār. His poem, like Sanā'ī's, lacks a frame narrative (a device ‘Aṭṭār particularly favoured); like Sanā'ī's it tends not to separate homily and anecdote into the aesthetic compartments habitually set up by ‘Aṭṭār, but to mingle them apparently promiscuously; and there is frequently a kind of impatient or blunt roughness to the diction of Rūmī's Mathnawī which looks like a deliberate harking back to the relative asperity, jaggedness even, of Sanā'ī's verse, rather than to the more mellifluous and poetically beguiling work of ‘Aṭṭār. But Rūmī does not follow Sanā'ī's example when it comes to the exordium to his work.

The Hadīqat al-Haqīqat begins with a direct address to God (ay darūn parwar birūn ārā-ī) and this leads into an introduction, of approximately one hundred and sixty lines, devoted to God's praise. This introduction is divided into five sections; the first is on God's unity (tawḥīd), the second on the knowledge (ma‘rifat) of God, the third is again on his unity (waḥdat) and glory (‘azmat), the fourth on his transcendence (tanzīh), and the last on the necessity for purity (ṣafā) on the part of those who would approach and know God. The rhetoric is overwhelmingly directed towards establishing the majesty and omnipotence of God, and the inviolable certainty of God's inexpressible but insistently invoked attributes. Only after this rhetoric, centred almost exclusively on the divine, does Sanā'ī turn to the first story of his poem, that of the blind men of Ghūr who attempted to describe a visiting king's elephant. In so far as the human has entered the poem before this story it is as a foil to the divine, as ignorant and weak, and unable to comprehend or approach the nature of God (an insight forcibly underlined by the first story).

Let us now look for a moment at how a number of the mathnawīs by his other predecessor, ‘Aṭṭār, begin. In the exordia to his Musībatnāmah, Ilāhīnāmah, Asrārnāmah and Manṭiq al-Tayr, ‘Aṭṭār starts with the invocation of God as the omnipotent creator (ḥamd-i pāk az jān-i pāk ān pāk-rā [Musībatnāmah]; bih nām-i kardagār-i haft aflāk [Ilāhīnāmah]; bih nām-i ānkih jān-rā nūr-i dīn dād [Asrārnāmah]; āfarīn jān āfarīn-i pāk-rā [Manṭiq al-Tayr]). In the first lines of ‘Aṭṭār's mathnawīs God is, as in Sanā'ī's poem, invoked by epithets that focus attention either on his capacity as creator, or on his purity and transcendence, or on both, and the subsequent lines elaborate the same topoi. ‘Aṭṭār's normal practice is to provide a link between this praise of the divine and the fallen world of quotidian humanity by passages in praise of religious leaders, especially though not exclusively the Prophet Muḥammad. Only after some hundreds of lines of such praise (over 550 in the case of the Manṭiq al-Tayr, smaller but still substantial numbers in the case of his other mathnawīs) does ‘Aṭṭār move to his poems’ narrative anecdotes and the world of human life unilluminated by divine revelation.

About the only thing that the opening of Rūmī's Mathnawī shares with these exordia is that, like Sanā'ī's Hadīqat al-Haqīqat, it opens in the second person, with a direct address (Bishnaw az nay chūn ḥikāyat mīkunad).¹ It is far shorter than its predecessors (thirty-five lines, as against over one hundred, and over five hundred in the case of the Manṭiq al-Tayr) and, most significantly of all, it ignores the subject matter of all previous mathnawī introductions. This is clear as soon as we look at the opening address, which is not to God, or to a prophet illumined by God, but to the poem's audience/reader; i.e. to fallen humanity.

It is difficult to overestimate the effect of these simple but transfiguring innovations. The first line is spoken by an unmediated authorial voice (i.e. by ‘Rūmī’), but with the second line we move to a poetic persona, that of the reed, who speaks of the pain of separation from its reed-bed, and which we immediately allegorize as the pain of the soul's separation from the divine. With line 9 Rūmī himself again speaks, and at line 27 he compares himself putatively to the reed. The two speakers, ‘Rūmī’ and the reed, are thus conflated as mouthpieces of the soul's longing for the divine. The divine is invoked obliquely, as an absence; this is wholly different from the exordia of Sanā'ī's and ‘Aṭṭār's mathnawīs, in which the divine is invoked explicitly, as an overwhelming, omnipotent presence. The focus of Rūmī's exordium is the turbulent emotional state of the human soul when divided from the divine: the focal points of the exordia to ‘Aṭṭār's and Sanā'ī's mathnawīs are not the human soul but God, not turbulence but the certainty of theological truth, and not the specifics of emotional longing but the order of God's creation. In ‘Aṭṭār's and Sanā'ī's poems there is a wide gap between the subject matter of the exordia and the poems' audiences, and this gap itself, in the emphasis placed on God's transcendence, functions partly as a confirmation of that subject matter. In Rūmī's poem, the subject of the exordium, the confused and unhappy human soul, is also the addressee of the exordium. The genre has been brought from the heavenly, the divine, and the eternal, to the earthly, the human, and the quotidian. Above all it has been brought from the unknowable to the known, and from transcendent truth to immediate experience. The trope by which Rūmī effects this transition is in itself significant. In ‘Aṭṭār's exordia the intermediary between the world of secular humanity and the divine is the existence of the prophets: in Rūmī's poem we do not move up the scale from unillumined humanity in order to find a mouthpiece for truth, but down from it, to the vegetable world, to the reed, and it is less a clear, divine, eternal truth that is invoked than a confused, human, experiential one. We descend into the mud of the reed-bed. (It is perhaps worth remarking here that our own romanticization of the reed, partly because of this very passage, and also because of the association of the reed-flute's tones with samā‘[musical audition], would have been largely irrelevant in Rūmī's time. The reed was a folk instrument, played by the uneducated, and gathered and fashioned in the wilderness, and largely retaining its association with rural life. It was an emblem of a hard and often subsistence pastoral existence, i.e. of human life at its most recalcitrant and unregenerate. The reed metaphor at the opening of the Mathnawī needs to be stripped of some of the mystical and transcendent patina that has subsequently accrued to it, in order for us to feel how forcefully Rūmī locates the opening of his poem in the dirt—in ‘the foul rag and bone shop of the heart’, to quote a poet from another tradition²—instead of locating it, as his predecessors had located theirs, in the divine.)

In so far as Sanā'ī and ‘Aṭṭār draw attention to themselves in their introductions, it is as generic poets or generic Moslems: Rūmī's insinuation of himself into his poem is less emblematic and more personal. His reference to his own love-longing is of course to be read as a longing for the divine, but also, specifically and notoriously, for the divine as mediated by Shams-i Tabrīzī; this is hinted at in the exordium and made explicit in the poem's first tale. We see him locating his poem in the human, the autobiographical and microcosmic, rather than in the divine and universal. It is this personal concern that propels the poem so quickly into its first narrative: if we may use an architectural metaphor, Sanā'ī's and ‘Aṭṭār's mathnawīs are preceded by elaborate facades that are, or at least certainly appear to be, radically different in kind from the buildings to which they provide entrance. Rūmī's entrance is relatively homely and abrupt, and of a piece with the building as a whole: one is no sooner on the threshold than in the body of the house as it were.

When ‘Aṭṭār and Sanā'ī embark on the anecdotal portions of their mathnawīs, they too turn from assurance to longing, and from omnipotence to weakness, but these all too human states literally follow on from, and are thus to be read in the light of, the massive unchanging certainties of the poems' openings. Rūmī begins not with the certainties of theology, but with the fallibility of the human, and not with the dignified invocations of eternal truth, but with an in medias res emotional turmoil associated generically with the ghazal rather than with the mathnawī, and with fiction rather than doctrine. He abandons the certainty of structure (and, without being too cute about it, we can also say he abandons the structure of certainty) evoked and elaborated by Sanā'ī and ‘Aṭṭār. The Mathnawī-yi Ma‘nawī is a mystical mathnawī that begins by flouting the conventions of mystical mathnawīs, and by drawing on apparently extraneous traditions to make its very striking rhetorical point (so striking that it's probably safe to say that its opening lines are among the three or four best-known pieces of Persian verse ever written).

II

When we turn to the first narrative of Rūmī's Mathnawī we see that, as with his treatment of the conventional prologue, he has quite radically recast the way in which his poem's opening narrative is presented. Again a comparison with the work of Sanā'ī and ‘Aṭṭār makes this plain.

Sanā'ī's story of the blind men of Ghūr (a tale Rūmī himself was later to use) is presented in a very simple, straightforward manner. The tale is told as a brief uninterrupted narrative, and it is concluded by three lines of general commentary (summed up by the line az khudā-i khalāyiq āgāh nīst / ‘aqlā-rā dar īn sukhan rāh nīst: ‘Men have no understanding of God / In this discourse reason can make no headway’). A separate section of eleven lines, with its own heading, then follows, explaining the meaning of the tale and its application to man's search for religious truth. The first narrative within the overall frame tale of ‘Aṭṭār's Manṭiq al-Tayr is again a relatively brief story: it is told by the hoopoe to the nightingale, and is an allegory of the nature of false love. The hoopoe, (who stands in for ‘Aṭṭār in the poem) presents the tale without any exegetical commentary. The stories are short (much shorter than the exordia they follow on from), and are presented to the audience's understanding as either unequivocal in meaning or unmediated by authorial commentary. Sanā'ī explicitly tells us what his story means; ‘Aṭṭār, by not including any exegesis, implies either that the story's meaning is obvious, or that he intends it to work in the recipient's mind unencumbered by commentary or other kinds of extraneous material.

The most striking way that the first narrative of the Mathnawī-yi Ma‘nawī differs from the opening narratives of these two poems is in its relative length: it is much longer than the exordium it follows on from, so reversing the relationship set up by Rūmī's predecessors. One reason for this length is that, unlike the opening tales by Sanā'ī and ‘Aṭṭār, it is filled with extraneous material. Neither Sanā'ī nor ‘Aṭṭār leave their tales while they are in progress;³ Rūmī does so frequently, and at one point (i. 143) he has to admonish himself to return to the story he is ostensibly telling but has apparently forgotten about. Because of the fluid nature of his narrative technique, it is sometimes difficult to say exactly where narrative ends and commentary begins,⁴ but my rough estimate is that of the 210 lines devoted to the story, approximately 90 (i.e. more than 40 per cent) can be considered as commentary.⁵

The word ‘commentary’, although the obvious one to use for non-narrative additions to an avowedly mystical and didactic text, hardly does justice to the nature of Rūmī's interpolations. There are two extended passages that can be seen as commentary of the expected exegetical kind: these are the brief excursus on adab (ii. 79-85), and the closing presentation of the story's moral import, in the manner of Sanā'ī. A third passage, which looks at first as if it will be of this kind, is the long interpolation on the nature of love, which begins at i. 109 and continues to i. 144. This passage is noticeable because of its highly personal character, in that it inserts Rūmī himself into the text, and it also introduces there two other historical individuals, Shams-i Tabrīzī and Rūmī's amanuensis, Husām al-Dīn Chalabī. Shams-i Tabrīzī is invoked as a beloved absence, and Chalabī, who was in a sense Shams's successor, is there as a necessary but provoking presence in that, in his role as amanuensis, he is the tale's transmitter, but one with whom the dictating poet has a fairly severe altercation. As well as literally being the tale's first audience, Chalabī here also seems to be a surrogate for the poem's putative future audiences, in that he appears to raise precisely the objections that ‘we’, the tale's subsequent readers, feel as the tale goes forward.

The frequency of the interruptions to the narrative flow, and their often personal nature, together produce a quite complex aesthetic effect. On the one hand, the tale's status as an edifying narrative is insisted upon, and the constant interruptions prevent our maintaining any Coleridgean ‘willing suspension of disbelief’ in the story as such. In fact we could say, staying with the same poet's explorations of the relationship between belief and fiction, that Rūmī functions as his own ‘person from Porlock’, in that as soon as the narrative flow has been established he interrupts it. This means that the reader is frequently diverted away from the story as a story, in order to be reminded of its existence as an illustrative moral anecdote: the story is lifted away from its status as narrative and placed in an interpretive, didactic context in which what finally matters is moral meaning, natījah.⁶ But, at the same time as this insistent abstraction of spiritual meaning is going on, the highly personal, unresolved emotionalism of the text, centred on three historical individuals (Rūmī, Shams, Chalabī—two of whom we see playing out a fraught, even angry, relationship before our eyes), regrounds the text in the specifics of particular human lives. The interpolations both cut through the fictional narrative in the interests of abstract interpretation, and at the same time subvert the abstract by insisting on an immediate and experiential ‘historical’ narrative (the circumstances of the tale's dictation). This is noticeably unlike the effects of both Sanā'ī and ‘Aṭṭār, which tend to be altogether simpler. Sanā'ī has little interest in the experiential as such, as virtually always foregrounds the natījah of his tales. ‘Aṭṭār is of course also concerned with natījah, but while he is narrating a story (especially a love story) he rarely interrupts it and appears concerned to maintain narrative continuity. and hence a ‘willing suspension of disbelief’. During the process of narration the aesthetics of the fictional narrative tend to be primary for ‘Aṭṭār, so that the reader is invited to enter into the fictional world virtually for its own sake, and the tale's didactic import is only insisted upon after the story is over.⁷ And in so far as Sanā'ī and ‘Aṭṭār introduce themselves into their mathnawīs, they present themselves as accomplished poets and knowledgeable mystics (i.e. as poets who ask for our attention precisely to the extent that they have risen above human imperfection). They do not present themselves—as Rūmī does—as individuals still caught within the experiential mire of quotidian human life. Both ‘Aṭṭār and Sanā'ī do this in their ghazals, where it is generically appropriate, of course, but it is one of the innovative characteristics of Rūmī's Mathnawī that he imports this topos (of a speaker caught within the toils of all-too-human need) and its attendant aesthetic, from the ghazal to a didactic work (in which the expectation is that the narrator's persona, in so far as it is present, is of one who has reached a stage beyond such spiritual and emotional bewilderment, and is thereby able to help others transcend them).

III

The ambiguous effect of the interpolations in the narrative is further complicated by another characteristic of Rūmī's tale which becomes very distinct when we compare it with those of his predecessors, and this is its detailed verisimilitude. The clearest instance of this is the scene when the visiting physician questions the girl as to her sickness. The way that his gentleness is insisted upon, the detail of his taking her pulse and noting her reaction to different names,⁸ the apparently gratuitous location of the love affair in specific places (Samarqand, Sar-i pul, Ghātafar) make the scene, in its plenitude of particulars and in its humane psychology, like something one would expect to find in a nineteenth-century European novel. The plethora and specificity of human detail here seem to resist any attempts to schematize what is presented as (mere) allegory. This location of the tale's details in the recalcitrantly quotidian is also clear in Rūmī's wonderfully exuberant and inventive imagery. Annemarie Schimmel,⁹ and more recently Fatemeh Keshavarz,¹⁰ have both drawn attention to the fecundity and inventiveness of Rūmī's original imagery (which is rivalled, and apparently valued, by few if any other medieval Persian poets), and such images play a large part in establishing the story's atmosphere of insistent realism.

As is usual with Rūmī, the imagery is frequently present as a brief moment of illustrative commentary, which both draws the reader away from the tale being told, but by its informality and immediacy resists any impulse toward interpretative abstraction. In this way it is similar to those interpolations in which Rūmī introduces himself, Shams-i Tabrīzī and Husām al-Dīn Chalabī, in that we leave the tale not in order to move toward abstract speculation but to insist on an experiential, physical, context or parallel. The images can have the pithy force of folk aphorisms (i. 107, bū-yi har hīzum padīd āyad zi dūd, ‘the scent of each wood is apparent from its smoke’; or i. 215, īn jahān kūhast u fi‘l-i mā nadā / sū-yi mā āyad nadāhā-rā ṣadā, ‘this world is a mountain and our actions a call / the echo of our calls comes back to us’); they can insist with some violence on a homely analogy for the inexpressible (i. 114, chūn qalam andar niwishtan mīshitāft / chūn bi-‘ishq āmad qalam bar khud shikāft, ‘as the writing pen was scurrying forward, it came to love and split open’); or they can refer to the ineffable with a kind of sly erotic humor (i. 138, mī nakhusbam bā ṣanam bā pīrhan, ‘I don't sleep with my lover in my shirt’). They can also be very tender, as when the physician says to the sick girl, at line 172, ān kunam bar tu kih bārān bā chaman (‘I will do to you as the rain does to the meadows’). Certainly the most extraordinary cluster of images in the story concerns the thom located first in a person's foot, when we are given what seems like a moment of ekphrasis that describes a Hellenistic statue with the sufferer propping his foot on his knee and probing for the thorn with a needle.¹¹ And then we have another of Rūmī's images involving violence, in that the thorn is thrust under a donkey's tail, and the animal's distracted pain is offered as a metaphor for the pains of love in general, and for the girl's secret agony in particular. What is especially noticeable here is the way that the metaphor almost brutally implies that the girl's love is a false physical (mujāzī) love. If the donkey here stands in for the girl, as it clearly does, zīr-i dum (under the tail) can be taken as meaning ‘in the vulva’, so categorizing the girl's love-longing as simply a gross and painful genital irritation. The cumulative effect of the imagery in the story, which is less conventional (and more frequent, vivid, and occasionally violent than the relatively traditional tropes favoured by Sanā'ī and ‘Attār), is to give the narrative a context that is folksy, homiletic and humanly recognizable (the man probing for the thorn with a needle), but also strongly and even aggressively reductionist (love is at basis a sexual wound) in its insistent physicality.

IV

How then is one to interpret the story, and the extraneous matter with which Rūmī loads it? Clearly to take the narrative at face value is to set oneself up for a severe disappointment; Rūmī himself acknowledges this in his admonitions to Chalabī during its course, and to the reader after it is over. He does not expect his audience to like his tale. We disapprove, but he says, we are like Moses reproaching Khiḍr in the eighteenth sūra of the Qur'ān, in that we have leapt to judgement because we don't understand what is going on.

The tale begins as a love story, of a traditional recognizable kind, but it soon betrays the expectations set up by such a story. As Margaret Mills has pointed out,¹² the tale ‘can be regarded as an antiromance or antifolktale’ in that it deploys motifs common to such tales and then subverts them. The type of tale in question involves ‘beautiful young people [who] conceive an idealized love for each other … [who are then] separated by another male, superior in strength and usually in age and status to the male lover’. In the normal development of such a tale ‘after some time and considerable struggle, the senior male adversary is neutralized or killed and the young couple are united’, and it is this dénouement which is of course reversed in Rūmī's version. One may also remark that the tale bears a close resemblance to another folk type too, the Cinderella story, in which a poor girl is chosen as his beloved/bride by a prince and after some tribulations the two are united. Here the difference in Rūmī's version lies not in the dénouement, but in the nature of the obstacles to love between the rich suitor and the poor girl, which are not external (the families of the lovers) but internal (the girl's previous commitment to another lover, a topos that subverts yet another folk motif—that first love is true love).

Familiarity with the didactic mathnawī tradition dictates that we read the story as allegory. But this by no means clears up a reader's difficulties. The usual expectation in allegory is that motifs and characters once assigned a symbolic meaning keep that meaning for the duration of the tale. This is as true of Western allegories (e.g Spenser's Faerie Queene) as it is of Persian mathnawīs. For example, in the tale of the blind villagers arguing about the nature of the elephant, in Sanā'ī's Hadīqat al-Haqīqat,¹³ both the villagers and the elephant retain their symbolic meanings (fallible ‘blind’ humanity, the nature of divine truth, respectively) throughout the tale. Similarly, in the story of the king who drives away an ungrateful dog but allows him to keep his jewelled collar as a reminder of his origin, in ‘Aṭṭār's Manṭiq al-Tayr,¹⁴ both the king and the dog retain their symbolic meanings (God, fallen humanity, respectively) while the tale continues. But if we consider the characters in Rūmī's tale, we see that matters are clearly much more complex and unpredictable. In overall terms it seems plain enough that the king in the story represents God, the girl the human soul, and the goldsmith the meretricious attractions of the physical world. As Abdolhossein Zarrinkub has put it, ‘until the soul is able to escape from its connections to the world of glittering appearances, which the goldsmith of Samarqand … represents … it will not find its way to the love that leads to truth/God (ḥaqq), represented by the king’, and ‘The story of the servant girl and the king … demonstrates that until the soul frees itself from its attachments to the physical world it cannot be united with the king, who represents God, and this represėnts the stage of detachment from the world (tahattul), i.e. freedom from the self’.¹⁵ The problem is not simply that the tale tends to run away with the allegory (so that we feel sympathy for the lovers, much as the reader of Dante's Inferno feels sympathy for the explicitly condemned Paolo and Francesca), a fact that Rūmī exasperatedly comments on, but that this overall allegorical meaning is not sustained in the story's details. For example, if we follow the most famous premodern commentary on the Mathnawī, that of Anqarawī,¹⁶ we find that during the meeting of the king and the visiting physician the king is to be taken as a murīd and the physician as his murshid.¹⁷ Similarly, when the king mounts his horse, Anqarawī reads this as an allegorical representation of setting out on the sufi path,¹⁸ so that the king is here, it would seem, to be taken as representative of the sufi aspirant. Common sense would suggest that a character cannot simultaneously represent an omnipotent God, a sufi aspirant, and a murshid's murīd, or if he does we are not dealing with a ‘stable’ allegory in the manner of Sanā'l or Spenser. And in the same way, in Rūmī's text, that one character can represent a multiplicity of concepts, so one concept can be represented by a plurality of characters: during the king's courtship of the girl, for example, it would appear that both in some sense represent the human soul, he in his need and she in her error. This shifting of allegorical referents, so that the reader must remain constantly alert for new meanings, is not entirely without precedent in some of the more complex tales told by ‘Aṭṭār,¹⁹ but neither ‘Aṭṭār nor Sanā'ī ever challenge the reader's attention with such a kaleidoscopically unstable set of referents within the one tale.

The multiplication of meanings and referents does not end here. At times motifs function with a simultaneous plurality of meanings (in contrast to the serial meanings located in, for example, the figure of the king), and this is most obviously the case in Rūmī's introduction of Shams in his discourse on love. The way that shams is used to indicate three different overlapping and mutually dependent referents (the historical individual Shams-i Tabrīzī, the sun, the ‘sun of the soul’ [khurshīd-i jān]) is well explicated in the contemporary commentary by Muḥammad Taqī Ja‘farī.²⁰ The ramifications of polysemy set out in Ja‘farī's commentary remind one at first sight of Dante's apparently similar explanation of the various layers of meaning to be found in the language of his Commedia,²¹ with the difference that Dante's meanings are considerably more stable, and as it were answerable to exegesis, than Rūmī's, which seem designed precisely to evade or at least complicate fixed interpretation.

Given this bewildering plethora of allegorical meanings present in Rūmī's text, it might be thought that the only way to proceed is to read the text with minute allegorical spectacles, explicating everything in sight. And indeed the monumental, multi-volume commentaries on the Mathnawī by Anqarawī, Ja‘farī, and others, indicate that there have been those more than ready to take up this challenge. But I referred above to the insistent specificity and physicality of Rūmī's narrative style, and one of the characteristics of such a style is that it recalcitrantly resists such treatment. When it comes to the location of the girl's beloved for example, Anqarawī is able to explicate parts of his address (Samarqand, sar-i pul) in ways that one might expect: Samarqand is explicated as ‘the natural world’ (shahr-i ṭabī‘iyyat), and sar-i pul is glossed entirely predictably (a bridge is not a place of rest, and in the same way the world is not a place of rest but of passing over), but he is stumped by other parts. Ghātafar is explicated as simply ‘the name of a place in Samarqand’.²² This inability to provide an allegorical meaning for Ghātafar might be thought a minor matter, but in reality it seriously undermines the whole edifice of explication. Because Ghātafar cannot be assigned an allegorical meaning it drags Samarqand back from being ‘the natural world’ to being simply Samarqand, a specific central Asian city. Another place at which the commentary falters is in describing the death of the goldsmith; the goldsmith's dying curse seems incapable of being given an edifying mystical explication (especially as, given the story's apparent allegorical scheme, it is God whom he is cursing). Like Ghātafar, it appears to be there simply as veridical detail. This is not at all of course to claim that the practice of allegorical explanation, of which Anqarawī's and Ja‘farī's commentaries are respectively among the most prized traditional and modern representatives, is in any sense mistaken. Clearly the poem calls out for allegorical explication, but equally clearly the tale's allegory is shifting, open and unstable; and at times it breaks down altogether into specific verismo detail that resists interpretation as anything but itself (Ghātafar is simply Ghātafar). The narrative refuses to be shifted wholesale into the realm of the spiritual, and this insistence on the quotidian, experiential and physical is further confirmed by the always homely and familiar, and often deliberately bestial and gross, imagery that Rūmī employs. In a sense Rūmī himself, within his tale, sets out these contradictory parameters, by including the little excursus recommending adab (ii. 79-85), which is set within a tale which is consistently, as it were, bī-adab, in that it adumbrates, and even appears to celebrate, conventions that it then cheerfully breaks. What one is left with is a text which simultaneously demands and resists interpretation, which points to meanings that are no sooner summoned than they are subverted, and which remains stubbornly rooted in the vulgar complexities of untranscended human experience.

Notes

1. Jalāl al-Dīn Rūmī, Mathnawī-yi Ma‘nawī (ed. R. A. Nicholson), Leiden 1925-33, reprinted Tehran 1350/1971-2, i, i. (All quotations from the Mathnawī are from this edition, and they are referred to by volume and verse number.)

2. W. B. Yeats, Collected Poems, London 1958, 392.

3. By ‘their tales’ I mean here the opening tales of the Hadīqat al-Haqīqat and the Manṭiq al-Tayr. It is true that both poets do interpolate commentary into some other stories in their mathnawīs, but they do so relatively sparingly and virtually never to the extent, or in the multiform ways, that Rūmī does in the Mathnawī's opening narrative. Sanā'ī's technique is often one of didactic homily illustrated by very brief anecdotes (i.e. the extended narratives are relatively few), rather than of narratives to which are added homilies as introductions and/or addenda (‘Aṭṭār's most frequent way of proceeding).

4. This is particularly the case when Rūmī resorts, as he frequently does, to passages of extended illustrative imagery, which tend to move from fictional verisimilitude to commentary/explication and back again within a few lines. For instance, the images of a thorn in someone's foot and another thorn placed under a donkey's tail (i. 150-56) are there at first sight as veridical detail, but it is clear that their main function is to provide illustrative commentary.

5. This is taking i. 36-246 as constituting the text of the tale. If we consider the concluding lines of commentary as a separate section (analogous to the commentary under a separate heading that follows on from Sanā'ī's first tale), we are left with c. 66 lines of commentary out of 186 lines of text. This is still over a third of the total.

6. The effect is not dissimilar from that which the twentieth-century German poet and playwright Brecht strove to achieve, in that he deliberately incorporated such distancing devices into his narratives, in order that the audience would not sink into the story as such but be forced to view it from a didactic/ideological perspective.

7. A good example of this occurs in ‘Aṭṭār's presentation of the tale of the Princess and the Slave (op. cit., 463-71, ll. 3794-870) in which the interests of telling a beguiling story almost obliterate the tale's ostensible didactic import.

8. Interestingly enough this precise vignette reappears, a hundred or so years later, in the Decameron of Boccaccio, an author who is often seen as the ‘father’ of the modern European novel (see the eighth Story of the second Day).

9. Annemarie Schimmel, The Triumphal Sun: A Study of the Works of Jalaloddin Rumi, London-The Hague 1978, 59-222.

10. Fatemeh Keshavarz, Reading Mystical Lyric: The Case of Jalal al-Din Rumi, Columbia, SC 1997, passim, but see especially 31-48 and 72-99. Dr. Keshavarz is mainly concerned with the Dīwān-i Shams-i Tabrīzī, but her illuminating remarks on imagery are in general equally applicable to the Mathnawī-yi Ma‘nawī.

11. The subject is part of the repertory of Hellenistic statuary, and a particularly fine example exists of a shepherd boy, in exactly the pose conjured up by Rūmī's description.

12. Margaret Mills, “Folk Tradition in the Mathnawī and the Mathnawī in Folk Tradition”, in Amin Banani, Richard Houannisian and Georges Sabagh (eds), Poetry and Mysticism in Islam: The Heritage of Rumi, Cambridge 1994, 145.

13. Hadīqat al-Haqīqat wa Sharī‘at al-Tarīqat (ed. Modarres Razavi). Tehran n. d. (3rd printing), 69-70.

14. Farīd al-Dīn ‘Aṭṭār, Manṭiq al-Tayr (ed. Ahmad Ranjbar), Tehran 1366/1987-8, 272-3, ll. 2183-94.

15. Abdolhossein Zarrinkub, Palah Palah tā Mulāqat-i Khudā, Tehran 1371/1992-3, 260 and 266.

16. Sharḥ-ī Kabīr-i Anqarawī bar Mathnawī-yi Mawlawī (tr. Akbar Behruz), Tehran 1348/1969-70.

17. Ibid., i. 127.

18. Ibid., i, 101. Anqarawī gives a similar interpretation to the physician's journey (123-4).

19. E.g in the last story of the Manṭiq al-Tayr the king and the boy at different points in the tale each appear to represent both God and the human soul. I have discussed this story and its allegorical technique in “The Journey as Paradigm in ‘Aṭṭār's Manteq al-Tayr”, Edebiyyar, 4 (New Series), 173-83.

20. Muḥammad Taqī Ja‘farī, Tafsīr we Naqd wa Tahlīl-i Mathnawī, Tehran 1366/1987-8 (11th printing), i, 111-5.

21. Dante, letter to Can Grande. The Latin next of the relevant portion of the letter (which interestingly enough includes the Greek borrowing ‘polysemos’ to describe the phenomenon), together with an English translation, may be found in William Anderson's Dante the Maker, New York 1982, 333-4. Dante distinguishes four possible meanings to be found simultaneously in his text; the literal, the allegorical, the moral, and the anagogical (although the last three may be subsumed under the general heading ‘allegorical’).

22. Op. cit., 153. Ja‘farī, op. cit., 117, repeats word for word Anqarawī's description (‘the name of a place in Samarqand’), and Karīm Zamānī in his commentary, Sharḥ-i Jāmi‘-i Mathnawī-yī Ma‘nawī Tehran 1374/1995-6, i, 103, has an almost identical phrase. None of these commentators assign an allegorical meaning to Ghātafar.