Ibn Gabirol's Religious Poetry and Sufi Poetry

It is well established that the startling innovations introduced into Hebrew poetry by poets in the circle of I:Iasdai Ibn Shaprut in tenthcentury al-Andalus were due to the influence of Arabic poetry and society. In the tenth century the influence of Arabic extended mainly to secular Hebrew poetry, the very existence of this poetry being one of the most interesting of the innovations. Liturgical poetry also changed, but not as radically; until Solomon lbn Gabirol (ca 1021-ca 1058). Early Andalusian Hebrew liturgical poetry can be understood as a simple evolutionary development of the liturgical poetry of ninth and tenth century Iraq. With lbn Gabirol, however, the Andalusian liturgy acquires new characteristic forms, expressing new themes, in many cases even serving new liturgical functions and occupying new liturgical sites • It has long been recognized that the Hebrew liturgical poets employed motifs from secular Arabic love poetry. Israel Levin has attempted to explain the introduction of such motifs and a number of other distinctive f eatures of Gol den Age religious poetry as the result of the influence of Arabic Sufi poetry • There is nothing


RA YM0ND P. SCHEINDLIN
religiously impossible about this hypothesis. Late twelfth and early thirteenth century Jews like Abraham Maimuni admired certain aspects of Sufism, adapted them to Judaism, and even incorporated quotations from classic Sufi texts into their own writings. Already in eleventh century Spain, the influence of mild Sufi pietism on Ba}Jya lbn Paquda is well established 3 . Levin's hypothesis would push back the influence of Sufi texts on Jewish writers by half a century, also not an impossibility. But how is such an hypothesis to be proved? Granted that Levin is able to point to sorne poems by Solomon lbn Gabirol that use phrases reminiscent of phrases in early Sufi poets, do such reminiscences justify us in making the leap from noting the parallels to positing influence? Implicitly Levin 's hypothesis raises the question of just what is meant by literary influence.
In order to take Levin 's hypothesis to the next step and to deal with the questions just raised, we would need to study the poetry of Sufi masters that was available for the Hebrew poets to read and be impressed by. But Sufi poetry predating the thirteenth century has received rather little attention, overshadowed as it is by the more flamboyant and abundant poetry of Ibn al-Fari9, lbn cArabi, al Sustari, and their successors in Arabic and Persian 4 • Yet even befare the age of these luminaries poetry was hardly lacking among 3 G. D. CoHEN, «The Soteriology of R. Abraham Maimuni», PAAJR 35 (1967) 75-98; 36 ( 1968) 33-56; G. V AJDA, La théo/ogie ascétique de Bat,ya Ibn Paquda, Paris the Sufis. In order to study the ways in which Sufi poetry could have influenced the Hebrew poets of eleventh and twelfth century al-Andalus, I have surveyed the major Sufi treatises that predate them and have assembled a corpus of the poetry preserved in these texts 5 • I have been studying these poems with a view toward deve loping criteria that will either validate or invalidate Levin 's hypothe sis, in the hope that this research may also cast sorne light on the methodological questions mentioned above.
Levin finds evidence of Sufi influence on Hebrew liturgical poetry both in Sufi poetry on the love of God and in Arabic zuhd poetry, dealing with asceticism and renunciation of the world. We take these up in sequence.

11.
It is commonplace to say that the history of Sufism begins with the introduction of the theme of love of God into Islamic asceticism. This assertion is completely borne out by the use of love poetry in Sufi treatises. Much of the poetry quoted in Sulami's sketches of famous Sufis in bis Tabaqat al-�ffya is simply love poetry; not religious poetry that draws on the conventions of secular love poetry, but love poetry plain and simple, sometimes by known secular poets. The early Sufi masters who used poetry did not necessarily compose it themselves but would recite lines of known secular poems as a kind of prooftext in confirmation of a point of their preaching. Sometimes they would use secular verses f or the startling effect of putting them in a religious context. The verse is not always treated allegorically, but is often used as exemplary; f or instan ce, if one who loves flesh and blood can be aff ected to the degree depicted in the poem, how much more is expected of one who would truly love God. Many are the anecdotes in which a master is thrown into religious ecstasy, goes mad, or even dies on hearing someone in the street reciting a secular love poem or recalling a secular verse. The master's behavior on such occasions often recalls the mad behavior related in the Majiiri c a/-'ussiiq (by Ja c far b. A}:unad al-Sarray, d.

500/1106) of persons especially sensitive to lave and lave poetry 6 •
The upshot is that we cannot always tell whether a line of poetry in the mouth of a master is of his own composition or not. When, as often, the poem is introduced with the word tama11ala (meaning «to make use of a line of poetry by someone else» ), it clearly is by someone else, whether a Sufi or a secular poet. But more often the introductory expression is ansha > a yaqül («he recited, saying» ), which gives no clue as to whether the master is or is not the author, and therefore whether the poem was composed with religious intent or not. By far the majority of the love poems in Sulami, like the Song of Songs, have no explicit religious theme. In this situation we have to treat every love poem said to have been recited by a Sufi master as a religious poem, whether or not the religious content is manifest.
We now turn to the poems likely to have been composed by the Sufis, as opposed to simply being quoted by them or mentioned in connection with them. They are mostly short, consisting of two to eight lines 7 • I found it convenient to sort them into broad categories determined by the overall intention of each poem, and to study themes like love in connection with the larger category in which they appear. I f ound it possible to sort nearly all the poems in my corpus into the following categories: 1. wisdom-type poems of ins truction; 2. theoretical statements about the nature of the mystic's religious experience; 3. personal statements purporting to describe a particular mystic's religious experience (sometimes shading off into a kind of religious fa"/s:r); 4. prayers. There is sorne mixing of categories, particularly inasmuch as theoretical poems and poems of personal experience sometimes end with a brief prayer. The famous poem attributed to Rabi < a, Uf,.ibbuka t,.ubbaini t,.ubba-1-hawa («1 lave 6 According to J. N. BELL, «Al-Sarráj's Ma;iiri < a/-'ushshiiq: A Hanbalite Work?►•, JAOS 99 (1979) 235-248, many of these secular stories were transmitted to Sarráy by a Sufi, demonstrating again the free flow of literary material on secular love into thc sphere of Sufism. The very purpose of the Ma;ári' -secular or religious-is the subject of scholarly debate. 7 A f ew are longer, especially in the árwan of l;lallay, and there is one of heroic length by �unayd in the 1/ilyat al-awffya > .
When I claimed to love, she said, « Y ou lie ! for if so, why do I see your limbs covered [with flesh]? It's only love when your skin sticks to your gut, when you are too deaf to answer when someone calls, when you wither so that love leaves you nothing but an eye with which to weep or to confide».

SEf LIV I (1994)
Sometimes love themes are used to demonstrate the extent of the Sufi's attachment to God. Themes from love poetry are used to show how the 'libid is different from other men, set apart from normal society, an object of criticism and scorn. Sometimes this theme shades off into fa/ir, when the poet boasts of his isolation and the hostility he endures from society. Here is a typical example, by Yal}.ya Ibn Mu' a cj 12 : Poems of the theoretical type tell how a lover of God acts and what he feels; poems of personal experience describe how the speaker acts and feels. But in almost all the poems it is the 'abid, the worshiper, and his spiritual suffering, not God, who is the center of attention. It is amazing how little of the hymnic strain is heard in this poetry. Most of it seems not to be intended as a vehicle for the adoration of God, but as an exploration of religious virtuosity, sometimes in the f orm of instruction on how to recognize it, at others as a description of the feelings of those who embody it.
Of the various dominant trends of Abbasid love poetry, it was the <u{jri that attracted the Sufis 16 , precisely because of the centrality in that body of poetry of the experience of frustration and suffering, and because of the preponderance of attention it devotes to the lover's virtuosity as a sufferer. The suffering of the lover who can never be united with his love was perfectly analogous to the suffering of the <abid who longs helplessly f or union with God, or f or return of the state of ecstacy he has experienced bef ore. Here the early Sufi masters found the human experience that most closely paralleled their religious experience. Not surprisingly we find them often quo ting Ma9'mln-Laila (or poems attributed to him) and Yamü-Bu,!ayna, and using verses from their poems f or their sama'-sessions 11• 14 An extremely rare exception: Qü-1-Nün, Maja/u quliibi '/. c arifiña bi-rau"atin (l�fahani, 1/ilyat, X, 369 and 391, which speaks of fulfillment; but other than the word «love», which occurs a few times in the poem, the imagery comes from outside the sphere of love poetry. J ust how closely are these sentiments paralleled in Hebrew poetry of the synagogue?
Since lbn Gabirol was the first of the Hebrew poets to compose liturgical poetry using Arabic prosody, we turn first to him. And in fact there are a number of striking liturgical poems by lbn Gabirol in which the erotic theme is highly developed. For example 18: How similar is this poem to the love poems f ound in early Sufi treatises?
The fact that this poem (and Ibn Gabirol's other resuyo1 on the Davidic theme) deals with the religious yearnings of a community rather than of an individual does not in itself argue against a genetic relationship. Synagogue poetry is by definition the prayer of the community. The separated state of lovers in the Song of Songs http://sefarad.revistas.csic.es regularly served the synagogue poets as a means of describing the exiled condition of the Jewish people and their longing for redemption; in adapting secular love poetry to the synagogue, the Jews, like the Sufis, would merely have been investing its conventions with the new meaning appropriate to their own set of religious problems. The question is, did the Jews adopt elements of secular love poetry in such a way that we have the right to conclude that they learned the technique from the Sufis?
To answer this question, I believe that it is not enough to look f or individual pairs of similar phrases. We need to try to understand the character of the religious experience expressed, or rather, embo died in the poem. Individual motifs can crop up in different cultures f or a variety of reasons, but evidence of meaningf ul influence must be sought in a similar mental set, a similar structure of experience that would create a climate favorable to the adoption of materials of one culture by another. The extent to which such similarities exist can often be deduced from the formal and emotional structure of the literary work. In the absence of such larger mental similarities, there would seem to be no way to explain how and why the individual literary motifs would cross the space between the two cultures.
lbn Gabirol 's poem is a f antas y in which Israel, represented as a woman, addresses a silent interlocutor, who may be God or the Messiah. Israel's speech is in eff ect a seduction. The speaker takes the aggressive role, promising food, wine, and beds, conjuring up an atmosphere of luxury and sensuousness. She alludes to similar ren dezvous that she has had with others in the past. There is not the slightest intimation that she f ears that this invitation will go unans wered 19 • The allusion to David at the poem's end intensifies the sensuous ness of the atmosphere, for David is one biblical figure whose physical beauty and sexuality are an important part of bis story. I think that the depiction of the glorious age of the Davidic monarchy as a lovers' rendezvous is unprecedented in Hebrew poetry. If the poem is addressed to the Messiah (who is to be a descendant of David) instead of God, that too would be without precedent. The depiction of the theme of Israel's marital union with God in the sanctuary of the Temple does have salid roots in rabbinic lore; but if the intended lover here is God, the poet's comparison of Him to a flesh-and-blood lover like David would also result in a daring heightening of the erotic aspect of this ancient tradition.
N ot one of the approximately two hundred Sufí poems in my corpus resembles this one and its fellows in any signifícant way. The reason is evident. The liturgical purpose of lbn Gabirol's poem is to express hope f or redemption. The speaker is hopeful, even confídent, and certainly anything but depressed, as she cites past pleasures as precedent f or a delicious future. There is no calling of attention to the speaker's own religious virtuosity; the poem is an ardent speech designed to draw the hesitating lover into the circle of the speaker's own warmth.
It is impossible to imagine that the Sufí poetry known to have existed bef ore Ibn Gabirol could have been the literary so urce of the imagery, style, or spirit of lbn Gabirol's Davidic liturgical poems. The only source that we can be sure of is one interna} to the Jewish community, that ancient mainstay of Hebrew synagogue poetry, the Song of Songs. Elsewhere I have discussed how lbn Gabirol developed the images of the Song of Songs by combining them with passages in the stories of the young David 20• The most that can be said in favor of Arabic influence is the novelty of lbn Gabirol's using Arabic prosody in a short religious poem on the theme of love, far short Arabic love poems in quantitative meter certainly are typical of the Abbasid period; it may be that that the sensuous associations of secular Arabic love poetry were carried into Hebrew with the form. Furthermore, similar seduction scenes can be found in secular Arabic poetry. But they are decidedly not f ound in the <u(jri poetry so dear to the hearts of the early Sufís. For overtly sexual themes in Sufí poetry we have to wait for later centuries. Given the emotional chasm between the sensuousness of lbn Gabirol's Davidic poems and the chasteness of the prior Sufí poetry, it is impossible to see any direct link between the two. We might even go further and say that it is unlikely that the kind of religious spirit that created Ibn Gabirol's poem was not even capable of being aff ected by Sufí love poetry, so diff erent in character are 20 In R. P. SCHEINDLIN, The Gazelle, pp. 90-107. http://sefarad.revistas.csic.es the two types of religious yearning that they express. But at this stage it would probably not be prudent to assume this much consistency on the part of any poet. At the very least we can be sure that these poems by lbn Gabirol are independent of Sufi poetry.
Ibn Gabirol can only have gotten the inspiration f or this series of poems from within the Jewish tradition. The liturgical site of the poem required that it deal with the love between God and Israel; the Song of Songs was already a mainstay of such poems. A poet who was so innovative in other ways could very well have had the genius to create these sensuous dialogues on bis own, drawing directly on secular Arabic poetic traditions. And although the in fluence of Arabic writing on lbn Gabirol's literary production in general is beyond dispute, in the particular case of the these poems Sufí influence is out of the question.

IV.
The themes and the tone of love poetry may be present even in poems in which the language is not so overtly sensuous as in those built around the Davidic theme described above. Yet, while many of these poems speak of the love of God, we do not find a single one that could be characterized as a love poem in the sense of being structured around the theme of love like the Sufi poems quoted earlier. Furthermore, rigorous comparative method would dictate that references to the love of God or even references to secular love motif s are not sufficient to establish a connection between lbn Gabirol and Sufi poetry; f or love of God is a theme of the Psalms, which even more than the Song of Songs was a mainstay of Jewish worship from its earliest beginnings, and secular love poetry was just as available to lbn Gabirol as it was to the Sufís. To decide for meaningful Sufi influence, we shall have to identify a similarity in the structure of the poems reflecting a similiarity in the structure of the religious mentality that the poems seek to express.
Levin off ers fíve Sufi poems and six Hebrew poems by lbn Gabirol to demonstrate the probability of lbn Gabirol having bo rrowed the theme of love from the Sufís. We will speak of the Arabic poems later. First we turn to the poems by Ibn Gabirol cited Hear, Lord of the world! Hear my praise; may my prayer be set like incense before You. My heart loves you very much and is so unable to conceal it that it reveals its love in my speech. At ali times I reflect that tomorrow I return to You, for the beginning of my start was with You. Lo, for Your sake, not mine, do I stand before You, and for the sake of Y our glory, not for reward for my deed. My soul has humbled itself before You when it carne to clay, so that ali my greatness has become no better than dust.
Like all resuyoJ, this poem is definitely a prayer, and like most resuyoJ its theme is the nature of prayer and the attitude of the worshiper as he stands bef ore God. Though love is not the theme of the poem, verse 2 does speak of love 22 and, as Levin correctly points out, ref ers to a convention of love poetry, the idea that the lo ver is una ble to conceal his leve because of its intensity.
But the way in which the lover unwillingly discloses his leve is not through weeping, emaciation, sighs, depression, or any of the other signs of leve so ably catalogued by lbn azm; it is through words, presumably through the poem itself, or at least through the prayer to which the poem serves as a preamble. lbn Gabirol often uses the resuJ to explore and to explain the impulse to pray, and in severa! poems he explains this impulse as a welling up of love for God within himself, overflowing in words 23  This poem and the others like it, though diff ering in significant respects from the reluyo1 of Ibn Gabirol, also resembles them in important ways. For one thing, it is a prayer: the address to God is not merely a f ormality, but is of the poem 's essence, providing the frame for an outpouring of simple, almost artless expressions of awe and devotion. lt resembles the rlluyoJ in form and tone, f or these too are short poems in plain diction that suggest a mood of intimacy

méfw"z fzef:,i wé-�ur ma{lsi wé-'5:osi mena1 fze/qi wé-górali we-fze!J.li
Although a note of anxiety can be detected in Rabi c a's poem, as noted by Badawi, there is in it also an undeniable warmth and security in the contemplation of the speaker's feelings that may be found in lbn Gabirol's resuyo1 as well.
But where was this mood when we were speaking of the love poems in the corpus of early Sufi poetry? Precisely what is lacking in that corpus is this warmth, this security, this sense that the speaker is sustained by love of God; rather, the prevailing tone is anxiety and pain. It seems almost as if my whole corpus stands on one side, and the eleven poems attributed to Rabi < a on the other across a great emotional divide, and that if lbn Gabirol was influen ced by any Sufi love poetry it could only have been by these eleven poems rather than by the more than two hundred in the corpus.
At this point it becomes necessary to examine the attribution of poems to Rabi < a. Levin is aware that these attributions are uncertain, for he begins his discussion of the examples taken from her work by saying that even if not all poems attributed to her are authentic, they «faithfully represent the style and spirit of her circle» 27• But the problem cannot be so cavalierly dismissed. If we merely assume that the poems attributed to Rabi < a are typical of her circle, we ourselves are arguing in a circle. And the sources on early Sufism do not provide any confirmation that the eleven poems do represent the style and spirit of her circle. If we could be sure that the poems attributed to Rabi < a were roughly contemporaneous with her, it would not matter for our purposes whether they were written by her or by someone else. But of the eleven poems generally attributed to Rabi < a, only one, the famous «two-loves poem», actually occurs in sources that definitely predate Ibn Gabirol. All the others occur only in later sources 28 • The earliest sources on Sufism present other problems in connec tion with Rabi c a. Sorne surprisingly do not mention her at all 29 • Others tell of another Rabi c a (or Rayiya), the wife of AQIOad b. al Harawi; her personality is suspiciously like that of our Rabi < a, yet different enough to pose the question of just how many Rabi c as there were and which stories belong to which? There are even diff erent traditions concerning her most famous poem. One of the earliest sources, Kalabacji, quotes it anonymously 30 • The Qüt al qulüb and the sources dependent on it 31 attribute it explicitly to her on the authority of Sufyan al-Tawri and others. But Abü Nu < aym al-I�fahani (who does not devote a chapter to Rabi < a) quotes it as  SEF LIV l (1994) having been recited by a mysterious woman who appeared to I)ü-1-Nün al-Mi�ri in the «wilderness of the lsraelites», while Zabidi's commentary on Gazzali's It,ya > (eighteenth century) has another version of this story, in which the mysterious woman appears to I)ü-1-Nün by the sea shore. I have not explored the Rabi < a tradition beyond this point; but what I have just reported is enough to suggest that her figure is at least partly built up of Sufi folklore, and that the poetry attributed to her must be treated with greatest suspicion, not only as to who composed it, but, more importantly, when and in what circle it was composed. The research does not yet seem to have been done that would demonstrate when the Rabi < a legend arase, and at what point which poems carne to be attributed to her.

RAYMOND P. SCHEINDLIN
Thus the only poem attributed to Rabi < a in a Sufi treatise that was composed early enough to have been available to lbn Gabirol is the «two-loves» poem, and even that is suspect. Furthermore, it has very little in common with lbn Gabirol's resuyoí, except for the idea of love of God in general. Even if lbn Gabirol knew this one poem, it seems unlikely that he could have had known the Rabi < a legend as a whole or that the one poem could have inspired him to create the new resúJ. This chronological argument is not watertight, f or man y treatises and Sufi traditions may have existed in lbn Gabirol's time that are completely unknown to us. But neither is it a weak argument, f or there are plenty of sources predating lbn Gabirol that could have cited poems by Rabi < a had they known them.
The poems attributed to Rabi < a have in common with Ibn Gabi rol's resuyoJ the theme of love of God and the tone of intimacy described above. But love of God is too general an idea to be seen in and of itself as the link between the two bodies of poetry. More impressive is the similarity in the poems' tone, the literary reflection of a religious mood of serenity that is quite uncharacteristic of early Sufi poetry. I gladly concede the similarity between Ibn Gabirol's poems and those attributed to Rabi c a, but I do not think that the religious experience expressed in these poems is characteristic of the Sufí poetry of the eighth to the tenth centuries. It theref ore seems to me unlikely that these poems or poems like them could have been lbn Gabirol's inspiration. ( The agony of separation, the martyrdom of waiting passively for union with the. bel o ved, the f eeling of rejection and worthlessness experienced by the one who can do nothing to bring about the return of the absent one without whom the waiter's life has no meaning: this central experience of all the great Sufís, indeed of many of the world's great mystics, found its natural secular analogue in <uqr1 poetry. In bis personal poetry, Ibn Gabirol too deals with the agony of spritual isolation and passive waiting for illumination, but neither the experience f or which he longs nor the literary f orm in which he expresses it is rooted in Sufísm; likewise when in bis religious poetry he turns to the predicament of the J ewish peo ple, bis imagery comes from other sources, primarily Jewish one as we have seen. Yet the cuqr1 poetry that was such an excellent model for Sufí poetry was also potentially a model for poetry on the Jewish condition; f or this was a people habituated to portraying itself as the abandoned beloved in the Song of Songs or the widowed wife in Lamentations, and to portraying God as an absent lo ver. It was only a matter of time before a Hebrew poet did come to make use of the Sufí model.
In a notoriously disturbing liturgical poem, J udah Halevi makes the leap 32 • Y ou have ever been the emcampment of love; my love encamped wherever You encamped. The reproaches of my enemies for Your name's sake are sweet, let them alone to torment one whom Y ou have tormented. My foes learned your wrath, so I love them, for they pursue a corpse whom You have slain. The day Y ou hated me I loathed myself, for I will honor none whom Y ou dispise. Until Your anger pass, and You send again redemption to this Y our inheritance which you once before redeemed.

SEF LIV I (1994)
embraced by the poet. It was the inestimable contribution of Israel Levin to discover that this poem is not an original composition in which Halevi merely exploited the conventions of the Arabic love poetry but that it is actually-except f or its last line-a translation of an Arabic poem 33 • The poet was Abü-l-Si$, a contemporary of Abü Nuwas in the court of Harün al-Rasid. Halevi's poem follows the Arabic quite closely, and in the same meter, quite a tour-de f orce; by adding the last line, Hale vi made explicit the connection between the attitude of the beloved and the proper attitude of Israel toward her suff erings 34 • Levin did not try to trace the channels through which Abü-l-Si$'s poem reached Halevi. The question might not even seem a fruitful one, f or the poem was quite well known and widely quoted in adab works. Abü-1-S� himself thought it his own best poem, and Abü Nuwas talked of plagiarizing it 35 • But alongside the many sources within the official literary tradi tion, I have identified another set of sources that is more intriguing for us. Sulami tells the following anecdote: Al-Murta'is (d. 328/940), a Sufi master of Baghdad, was asked, «How can a slave achieve lave [ of God]?». He answered, «By associating with the friends of God and making enemies of His enemies». Then al-Murta c is looked at one of his companions and said, «Recite the poem that you were reciting last night». In reply, the disciple recited two lines of the poem by Abü-l-Si$! 36 • Thus the Arabic source of Halevi's poem had been exploited f or its religious meaning long befo re Halevi. This is a typical case of a Sufi master using existing secular With the discovery of the passage in a contemporary Sufí work, it seems likely that Halevi knew Abu-1-Si f s poem in a Sufí context. Abu-l-Si$'s poem thus turns out to be a kind of missing link that establishes with fair certainty that at least Halevi was open to the influence of Sufísm in sorne of bis synagogue poetry. And precisely here, where there is an historical link between Arabic and Hebrew, there is also a thematic correspondence that goes beyond vague references to love. The religious impulse of certain Sufís and Halevi's view of Judaism were suffíciently clase in this case that the very same poem could serve both.
There are other moments in Halevi's synagogue poetry that recall aspects of Sufí religiosity.

Or,
Hurt me more and I will love Y ou more, for Your love is wonderful to me 39 •

VI.
We now turn to Levin's argument from zuhd poetry. From the start it is necessary to correct Levin 's misconception that zuhd poetry is a forerunner of Sufi poetry. To say that the theme of zuhd is an important to Sufism is not to say that the poetry of zuhd is prominent in Sufism. My corpus of early Sufi poetry, as mentioned, is almost completely devoid of zuhd poetry. On reflection, the absence of this type of poetry is not all that surprising. The early Sufi masters used poetry primarily to teach their own ideas. But while asceticism and rejection of the world are important to Sufism, they are not peculiar to it; rather they are part of the overall religious heritage of Islam. Furthermore, when early Sufis used poetry in their preaching, one of the main eff ects they aimed f or was paradox, as is especially evident in their use of secular love poetry. This rhetorical eff ect was · not possible with zuhd poetry, which is simply what it is-versified preaching. Finally, zuhd poetry could well have been regarded by true religious virtuosi as somewhat tainted, f or the most prominent zuhd poet was anything but a Sufi. Abü-1-'Atahiya was not a pietist but a court poet who made a specialty of zuhd poetry. According to a famous anecdote, he had a confrontation with Abü Nuwas in which he drew the line between bis territory and that of bis rival, taking zuhd f or himself and demanding that Abü Nuwas restrict himself to jamrfyat and the like. The anecdote 41 Ibn Gabirol's death date is contested, but the evidence favors 1058; see R. P. SCHEINDLIN, «El poema de lbn Gabirol y la fuente de los leones», Cuadernos de la Alhambra 29, [forthcoming]. Halevi is generally thought to have been bom before 1075; I have allowed him to reach the age of ten before writing poetry, but it may have taken him longer. In any case, lbn Gabirol and Halevi did not overlap. implies a certain cynicism toward the religious theme, as if Abü-1c Atahiya's commitment to it was strictly a matter of business, and that the distribution could very well have gone the other way, with Abü Nuwas becoming the zuhd poet and Abü-1-c Atahiya, the liberti ne. In fact, as we shall be discussing later, Abü Nuwas did write a few very interesting zuhdiyat 42 • Zuhd poetry is thus not a forerunner of Sufí poetry but a different stream altogether 43 • lbn Gabirol did borrow certain motifs from zuhd poetry, but these borrowings do not link him to Sufism.
But even more important than this technical argument is a more purely literary one. We can identify borrowings in lbn Gabirol's poetry from zuhd poetry; but these borrowings do not establish a common spirit that links his inner world with that of the zuhd poet.
One of lbn Gabirol's most fruitful innovations is the new tone of intimacy with which he invested the resuJ. This tone is the Golden Age's most distinctive purely poetic contribution to the Jewish liturgy. lt is a tone that is not attested in previous Hebrew liturgical poetry, which always expresses itself in collective style. The only precedent for lbn Gabirol's new tone is encountered through a leap backward over more than a millennium to the Psalms; only here do we find in early Hebrew literature the features that make the Golden Age resu1 so appealing: the constant play of first and second person, with the worshiper addressing God repea tedly and familiarly and speaking of his own personal experiences and needs (whether externa! and biographical or interna! and spiri tual) without reference to the community, as in all pre-Andalusian poetry of the synagogue. Now if there was ever any poetry that lacked the intimacy of lbn Gabirol's re!uyoJ, it is Arabic zuhd poetry.
On the assumption that one of the essential factors contributing to this tone of intimacy is direct address, and observing that lbn Gabirol's poems most celebrated for this tone are those in which the poet addresses either God or bis own soul, let us make sorne comparisons. We start with lbn Gabirol 44 : Lift up your eye, my soul, to your Rock, and think of your Creator in your youth. Cry out to Him night and day, and sing ever with your song to His name. Y our portion and your lot on earth, your stronghold when you leave your flesh. Has He not set for you a place of rest in His presence, has He not set your home beneath His throne? Therefore I will bless the Lord, just as all souls bless Him.
Toe poet addresses bis own soul, urging it to keep God constantly bef ore itself. The word f or «soul» (yet,idá) is grammatically feminine; sin ce each oí the first f our lines is addressed to her, all the verbs that apply to her have the second person feminine form. The rhyme syllable reinforces the address to the soul. It is rels:, in which -els: is a pronominal suffix meaning «yours». The poem follows the Arabic rule oí monorhyme with hemistich rhyme in the first verse, so that the first hemistich and the first f our verses all end in «your Rock, your youth, your song, your flesh, your chamber». But in case we might still overlook the pronouns, the poet goes out oí bis way to focus our attention on them in verse 3, with its interna! rhymes: t,e/qeli, weli,o"se/i ba-'adamá 11miºtafie/i be¡e'J.e/i mibbesareli,. That the verse is meant to be climactic is indicated not only by the internal rhyme but also by its rhetorical structure, since, alone among the verses oí the poem, it is built on antithetical balance. (The other verses are also parallel, but synonymous). It is also the middle verse oí the poem. The five-f old address to the soul is thus circled in red, as it were, f or our attention, and the soul is very much in the poem's foreground. Since the soul is advised to take action-to lift its eyes, to remember, to cry out, to sing-it almost seems to be personified.
Abu-1-< Atahiya has, as far as I know, no poem that is as com- In tbis example, Abü-I-c Atabiya's address to tbe soul is set in a narrative frame. Tbus, altbough the words ya nafs appear in verse 2, the pronoun embedded in the rhyme syllable that repeatedly brings us back to the soul is not second person but third. Tbe narrative in verse 1 calls attention from the very beginning to tbe discontinuity between the speaker and tbe soul, whereas lbn Gabirol calls attention to the self as speaker only in verse 5, after the soul has been well established as the center of attention. Abü-I-c Atabiya's /a < al/a in verse 1 suggests doubt as to whether tbe soul will even beed bis advice. But most telling is that after setting up the address to the soul as a near-personification that might have served to unify tbe poem, the poet drops her in the very next word following tbe address (fa-qultu taha yii nafsu mii kuntu iijir;!an ... ) and goes on to speak about himself, tbe nature of this material world, the brevity of bis own life. By tbe end, he drops even the pretense of addressing tbe soul, turning to a second person masculine addressee and warning   O soul, repent before you will be unable to repent; and seek forgiveness for your sins of God, the pardoner of sins.
In these verses we recognize the tone of the male confessor or other father figure who lectures a wayward woman, a rather common 48 /bid. p. 31. f eature of these poems. These two lines to approach the feel of our example from Ibn Gabirol.
As But even in the first three verses the tone is hardly intimate; the poem is a kind of harangue that could be addressed to anyone. In fact, such versified sermons make up a large part of Abu-1c Atahiya's áIWtin. The only difference between them and this example is that they are addressed outward, while this one is addressed inward.
The upshot of ali these examples is that in Abu-1c Atahiya the soul is a convenient addressee for the poet's message; but there is very little motion in the direction of the kind of near-personification that we saw in Ibn Gabirol. The soul is never developed into a character with an imaginary personality. We might have expected her to be portrayed far more negatively, perhaps as a temptress that leads man astray, an image that does occur in contemporary Sufi writing 50 • But in Abü-1c Atahiya that role is reserved for dunyii, which is portrayed as actively wicked. The soul is at worst merely heedless, lazy, and unreflecting, just like members of a congregation listening to a sermon. Like the individual person f or which it stands it must confront the thought of the last judgment as described in the Quran and in the Islamic tradition. lbn Gabirol can never be as neutral toward the soul as can Abü-1-< Atahiya because the soul occupies a much more prominent place in bis outlook on the nature of man. For him, the soul is part of a Neoplatonic scheme, a character in an elaborate myth of the nature of man and bis relationship to God. For Ibn Gabirol, nefes, the Hebrew cognate of nafs, never stands for the lower instincts in man, and only rarely is it used neutrally to mean «self». Like its synonyms yef,ít;Jd and nesamá, it stands for the higher or rational soul, emanated from God, divine in substance, and temporarily caged in a body. lts natural tendency is upward, toward union with its divine source. lt can be diverted from that upward tendency by temporal things, but it can also be set back on its course by admonition, precisely because that course is natural to it. Thus Ibn Gabirol can encourage the soul, as in the poem quoted at the beginning, by reminding her than God is her portion in the world as well as her stronghold after death, and that God has provided her a resting place under His throne. Being amenable to correction, Ibn Gabirol's soul is a more human-like, more sympathetic being, and this partly accounts f or the greater degree of personification we feel in bis poems.
lbn Gabirol does sometimes take a harsh tone in addressing the soul. Here are a few lines from a lengthy and stern t6/iet,á, a poem of rebuke to the soul 52 : Return to your resting place, my soul! For God has treated you kindly 53 • Did He not breathe you out of the breath of His mouth, and stretch you out from foot to head? And with a little of His tremendous might He established you, and taught you a little of His great wisdom. He taught you to explore the sciences, and gave you wisdom to know deep thoughts. He surrounded you with wisdom and understanding 51 This feature of moralizing poetry is distinguished from the poet's true address to himself in G. J. VAN  set the sickle against the corn of your sin.

SEF LIV 1 (1994)
The language is harsh, and much of it does recall the rebukes of Abü-1c Atahiya. But the speaker never fails to recall the higher source of the soul or to challenge it to live up to its potential. And when he says, «Therefore, listen, girl, observe, and give your ear», he sounds as if he is admonishing a willful daughter rather than cursing a harlot. Whatever details of imagery lbn Gabirol may have derived from Abu-1-< Atahiya's world, the poems themselves are utterly diff erent in spirit.
But the difference is not accounted for by the two poets' different religious affiliations. Most of the doctrinal content of Abü-1c Atahiya 's poetry would be perfectly acceptable to a Jew, and Hebrew poets did compose ordinary zuhdiyat in Hebrew. The diffe- (c) Consejo Superior de Investigaciones Científicas Licencia Creative Commons Reconocimiento 4.0 Internacional (CC BY 4.0) http://sefarad.revistas.csic.es rence lies rather in the two poets' different attitudes toward the soul; it is this diff erence that dictated a different literary style.
Our comparison of the address to the soul in lbn Gabirol to the address to the soul in Abu-1-< Atahiya may be taken as a model for the whole question of Arabic influences on lbn Gabirol. Levin is doubtless right that certain words, phrases, and images from Arabic and Islamic literature were adopted by Ibn Gabirol. But the bare philological fact that such borrowings exist is relatively trivial f or literary history, f or we airead y know that Hebrew poets were in fluenced by Arabic. We want to know the extent of the cultural dependence; we want to know how similar were the attitudes and sensibilities of Muslims and Jews. We cannot answer the deep literary and religious question by citing surface parallels. To gauge the extent to which acknowledged borrowings reflect a real intellec tual kinship we must consider the tone and attitudes of the poem as a whole, as I have tried to do with the examples cited. Here, tone can be relatively easily defined and correlated with known intellectual positions. By this criterion, lbn Gabirol seems quite far removed from zuhd, at least from the zuhd poetry of Abu-1-< Atahiya. And to try to connect him with Sufism is even more far-fetched.

VII.
So where did lbn Gabirol derive the peculiar intimacy of bis resuyóJ if not from zuhd poetry? Is it entirely bis own invention? Are there models in Arabic?
Though it is customary to say that Arabic does not have a poetry dealing with the inner life of religioi;i, and though there is no generally recognized poetic genre associated with religious themes except zuhd, it is nevertheless possible to compile a sizable collection of very touching Arabic prayers couched in poetic f orm. Such prayers turn up in severa! types of Arabic sources.
One source is my corpus of early Sufi texts, though this source is not as abundant in prayers as might be assumed. The bulk of the poems in this corpus are either expressions of the religious feelings of the ecstatic speaker couched in personal language (which I call religious f ajr), or theoretical statements about what it means or what it is like to be an ecstatic. There are sorne prayers, and also there are sorne verses couched in the f orm of prayers that are (c)  I am amazed at You and at me, O desire of the desirer! Y ou brought me so near to You that I think that Y ou are l.
But his conviction of the divine character of the soul leads him to compare it explicitly with God and to compase sorne rather striking paradoxes: For You I long with a thirsty heart; I am like a poor man begging at my door and my threshold 61• But we do not need to seek out extreme cases. Even outside Sufí literature we can find verse-prayers scattered in adab works. As mentioned above, Abü Nuwas, the arch-libertine, has severa!; ironi cally, these poems are the ones Levin cites as the link between lbn Gabirol and zuhd poetry. But they are not zuhd poems at all; they are simply prayers of an intimate type that are as far removed from zuhd in tone as are lbn Gabirol's Hebrew resuyoJ. Several such prayers are found in the áiwlin of Abü-1-< Atahiya; one is supposed to have been recited by him on his death bed 62: