Sound patterning and extra-prosodic sonority. The poetics of Hafez, perhaps more so than many Persian poets, depends on a sensuality of language and imagery. Smell, taste, texture, color and certainly sound imagery abound, often mixing synaesthetically. Enchanting music, bewitching beauty, intoxicating fragrance and delectable savor, even when not explicitly invoked, are often implied by the setting of the ghazals—the real or stylized wine symposium, with its locus classicus in Athens, its establishment in Persia during the Hellenistic period and its later development in the ḵamriyāt of the Islamic tradition.

Complementing this thematic sensuality in the ghazals of Hafez is an often tactile sonority created by a thick texture and complex patterning of sounds. We know Hafez to have been competitively conscious of the work and wording of other poets, which he often quotes, adapts and improves upon. While such judgments can be subjective, Moḥammad-Reżā Šafiʿi-Kadkani (1991, p. 439 ff.) offers a quantitative comparison of a ghazal by Salmān-e Sāvaji which Hafez emulated in the same rhyme and radif. He concludes that Sāvaji exhibits many more phonetic glitches and dissonances than does Hafez (p. 441) and proposes on the basis of existing manuscript variants that Hafez, who composed his poems carefully (at an average rate of no more than 10 per year), revised them later with an eye toward innovative or striking sonorities (Šafiʿi-Kadkani, 1991, pp. 425-27). The poet’s nom de plume indicates one who has memorized the Qur’ān; perhaps an intimate practical knowledge of the cantillation rules (tajwid and tartil)and the fourteen canonical recitations (čahārdah rewāyat, Ḵ. 93:10b; please note, the poems of Hafez are cited in parenthesis by ghazal and line number, as above, or by page and line number for the other genres, as per Divān-e Ḥāfeẓ, ed. Parviz Nātel Ḵānlari, 2 vols., Tehran, 1359 Š./1980, 2nd revised ed., 1362 Š./1984, abbreviated to Ḵ.) honed the poet’s sensibility to harmonizing melody and text, and also to phonetics, particularly the place of articulation of consonants and the rules (summarized in Nelson, pp. 18-31) for assimilation, pharyngealization, nasalization, etc.

Thus, beyond the traditional sound patterning (e.g., meter, rhyme, radif) and phono-semantic play (parallelism, paronomasia and other types of jenās) expected by the Persian manuals of rhetoric and prosody, Hafez often seems quite consciously to employ phonetic patterning, particularly consonance and assonance, which were not generally stipulated or even articulated as stylistic objectives in traditional Persian poetry. These include carefully crafted onomatopoeic passages, such as the gurgling of the juice of the grapevine as it pours from the jug in . . . ke ḵun-e ḵom / bā naḡma-hā-ye qolqola andar galu . . . (Ḵ. 32:5), with its repeated velar/uvular consonants, the plosives bracketing fricatives (keḵu ḵoaḡ qo qo ga). Note the phonoaesthesia of the condemnatory dentals (ṭ, d) and growling “r“s that are ultimately hushed by the semantic sweetness and relaxed phonemes at the end of the following hemistich:

samāṭ-e dahr-e dun-parvar na-dārad šahd-e āsāyeš (Ḵ. 273:3a)

Fate’s table fattens the base and affords no honey of repose


In the hemistich feḡān k-in luliān-e šuḵ-e širin-kār-e šahr-āšub (Ḵ. 3:3a), alliteration, consonance, assonance and alternating consonant-vowel parallelism (syllable initial . . . lu li . . . šu . . . ši . . . ša . . . šu and word-final ān, in, iān . . . in . . . as well as word-medial iri . . . ār-e . . . ahr-ā . . .) heighten the stirring sense of tumult visited upon the town by gypsy beauties with their cloying, clowning ways. Likewise, the textures of the bitter-sharp yet silky-smooth rose-red wine intoxicate the phonetic palate in the following hemistich: bāda-ye golrang-e talḵ-e tiz-e ḵᵛoš-ḵᵛār-e sabok (Ḵ. 303:6a). Finally, in šarāb-e talḵ-e ṣufi-suz bonyād-am be-ḵᵛāhad bord (Ḵ. 348:2a), the sibilants and fricatives (š, ḵ, ṣ, f, s, z, ḵ) blaze around the more solid plosives (b, t, b, d, b, d, b, d) to simulate with sound the bitter fiery vintage as it burns the Sufi’s frame to the ground. Similar examples are given by Hillmann (p. 91) and Eslāmi-Nodušan (p. 533).

Poems set to music vs. plain poems. In 1962, Homāʾi proposed in several footnotes to his edition of ʿOṯmān Moḵtāri Ḡaznavi’s Divān (p. 81, n. 5; 375, n. 1; 524-25, n.3, and the extended footnote on pp. 569-76) that poems which had been formerly described as being in baḥr-e nā-maṭbuʿ (literally, “unpleasant meters”), had originally been composed as lyrics for occasional entertainments, and performed to rhythmic musical accompaniment by singers (qawwālān, rāmešgarān). Homāʾi thought these metrically unwieldy poems, typically quatrains or short qeṭʿas composed to suit a specific occasion or to fit a particular musical piece, were similar to the early tarāna and do-bayti, or the modern taṣnif (cf. the songs of Robert Burns). Homāʾi linked these poems of “unpleasant meters” with the category identified by classical rhetoricians (including Šams-e Qays) as šeʿr-e malḥun, or poetry set to music, in contrast to šeʿr-e mojarrad, or plain poetry, which was not amenable to musical accompaniment. Though susceptible to prosodic scansion according to one of the various modifications of the classical quantitative meters (e.g., możāreʿ-e moṯamman-e aḵrab-e makfuf-e masluḵ, or rajaz-e mosaddas-e maškul, or wā-fer), a supplemental accent-based musical rhythm would be superimposed in performance on the šeʿr-e malḥun. Eventually, however, these poems gained status as literary qaṣidas, at which point they came to be preserved in divāns, and their musical associations were gradually forgotten. However, their metrical patterns did not readily correspond to meters commonly in use, and thus were felt to be “unpleasant” or uncustomary with respect to the elite literary forms, and could be subject to corruption by well-meaning scribes trying to harmonize the prosody to expectations. Heshmat Moayyad (p. 121), Julie Scott Meisami (pp. 142-43, n. 2), Šafiʿi-Kadkani (1992, p. 128, n. 112), and Sirus Šamisā (pp. 4-9, 15-21) have generally accepted the notion that some Persian poems were composed as song lyrics, though as yet this consensus has not been systematically applied to the study of Persian genre development and editorial theory.

Though Hafez does not use this term malḥun, he does use a variety of terms to indicate words set to music and/or the melodies that accompany lyrics. The Persian ghazal emerged in the Ghaznavid period as lyrics composed by poets for singers to perform (Lewis, 1995), and Hafez frequently associates the words ḡazal and qawl with music, musicians, singers and songbirds (see below). A generation after the death of Hafez, ʿAbd-al-Qāder Marāḡi (d. 838/1435) explains that the nawba, or musical performance suite, consists of four movements, structured as follows:

1) qawl, in which Arabic poems are sung

2) ḡazal, in which Persian poems are sung

3) tarāna, consisting of a robāʿi text sung in either language (Jāmeʿ, pp. 241-43)

4) forudāšt, in which Arabic poems are sung (Maqā-ṣed, p. 103)


Since Šams-e Qays Rāzi had used very similar terminology in his al-Moʿjam (630/1232), these meanings were indeed current throughout the lifetime of Hafez, who frequently pairs the words qawl o azal in a clearly musical context (Ḵ. 91:9, 141:2, 272:4, 370:8 and the moḡanni-naḡma, p. 1058:5b), with the word ḡazal often alluding to his own poem/song. This is surely an indication that Hafez’s own ghazals were composed with the idea that they would, or might, be sung (Lewis, 1995, pp. 91-95). Furthermore, the traditional introduction to the Divān of Hafez, composed after the poet’s death but not later than 824/1421, indicates that Hafez’s poems spread quickly (i.e., during his lifetime) to India, Central Asia, western Persia, Iraq and Azerbaijan, and were virtually obligatory in Sufi samāʿ sessions and wine symposiums. The speed with which the poems traveled, and the settings in which they were reportedly heard, strongly indicate musical performance as the primary mode of pop-ular diffusion (Lewis, 1995, pp. 229-33).

Despite this widespread singing of his ghazals, few if any of the poems of Hafez could be characterized as “unpleasing” or awkward with respect to the classical quantitative metrics, and one wonders if Homāʾi’s neat distinction between sung poems and plain poems pertains to this period. Nevertheless, it has been suggested (Hillmann, pp. 87-88) that some ghazals of Hafez exhibit certain regularities of stress, or patterns of accentual rhythm, overlaying and enhancing the quantitative literary meter. One might speculate that the strings of iambs in many hemistichs of the first ghazal in the Divān, e.g., ke ʿešq āsān namud awwal vali oftād moškel-hā (where vali, though normally accented in modern Persian on the first syllable, is sometimes sung in taṣnif with accent on the second syllable, as in the rendition of Šahrām Nāẓeri) reflect a stress patterning secondary to the quantitative hazaj meter, at least in musical performance. One type of taṣnif mentioned in the classical musical manuals is naqš, a term to which Hafez may allude on one occasion in a technical sense (Ḵ. 119:1b), perhaps signifying a one-to-one correspondence of tune and poem, such that each note and beat is matched to a single syllable (Mal-lāḥ, 1972, pp. 213-18; cf. Levin, p. 222 and track 20). Jiří Cejpek (p. 696) has noted the continued existence in the dialects of nomadic tribes, as well as the bāzāri colloquial in towns, of syllabic meter qaṣidas and ghazals (for recorded examples of contemporary Persian folk ghazals, see Levin, pp. 204-5 and track 18), varying between seven to fifteen syllables per line, in which assonance sometimes replaces rhyme.

With respect to metrics, Homāʾi himself has noted that Hafez composed most frequently in permissible variations of the ramal, mojtaṯṯ and możāreʿ meters, followed by variations of hazaj and monsareḥ, with motaqāreb, rajaz, moqtażab and ḵafif the least frequent (cited in Mallāḥ, 1972, p. 26), but the evidence to tell us whether any of these specific meters is more or less likely to be sung has not been marshaled. It would seem that most of Hafez’s poems circulated in a musical performance context (see below), so perhaps the majority of his ghazals were treated as though adaptable to musical settings. Whether and to what extent the aesthetics of musical composition may more subtly inform Hafez’s prosody and poetics remains to be studied in detail.

Musical terms in Hafez. Lexicons and general works on the links between music and poetry notwithstanding, the function and relative importance of music in various poets has not yet been investigated from a comparative point of view. Meneghini Correale has produced a full and scientific concordance of the ghazals of Hafez (see HAFEZ iv); once similar concordances and frequency lists of the complete oeuvre of other poets have been compiled, it will then be possible to make in-depth comparison of the relative importance specific words/themes hold for particular poets, and for particular formal/generic contexts (i.e., ghazals, robāʿis, qaṣidas, maṯnawis, etc.).

An inventory of the musical terminology of Hafez (as per the Ḵānlari ed.) follows, divided into categories and listing the total number of occurrences of each word in the Divān, as well as the genres/forms in which it appears. The frequencies are based upon Meneghini Correale’s concordance for the ghazals (indicated by “g”), and upon my own tabulations for the qaṣāʾed (qa),maṯ-nawiyāt (m), qeṭʿas(qe) and robāʿiyāt (r). Where a lexical item is used by Hafez in more than one meaning, a (subjective) effort has been made to isolate occurrences which pertain to music either explicitly or implicitly (i.e., as a double entente), from non-musical homonyms (e.g., 6/7 indicates seven occurrences of the word, six of which have musical meanings).

Musical sounds. Hafez uses a variety of words in the general meaning of melody, tune, air, song, or music, whether the sounds are produced by instrument or voice. These include: āhang 6 (5/7g, 1qe), āvāz 13 (6/9g, 5m, 1qe, 1r), bāng 10 (10/17g; exclusive of the non-musical sound bāng-e jaras), golbāng 10 (8g, 1qa, 1m), nāy (1/2g), navā 13 (9g, 1qa, 2m, 1qe), ṣawt (6g). Some of these have technical meanings, according to the pre-modern manuals of music, but the precise significance of most such words in the poetry of Hafez seems more context-determined than intrinsically fixed. Reverberation seems to be the primary meaning of ṣedā (2g), whereas words like zamzama (3g, of voices or instruments), bam o zir (1g), ziri (1r), or ṣafir (as it is associated with birds, 4/5g), seem to describe the timbre, pitch, or volume of the tones. Words such as faryād, nāla, and ḵoruš sometimes represent the emotional peaks of music, whether the sound comes from instruments, songbirds, or the human voice.

Musical modes, scales, melodic structures. Hafez does not use the technical terms given by Marāḡi for musical measure or beat (naqara) and rhythm (iqāʿ), but naḡma 12 (10g, 1m, 1r) would appear to convey at times the technical sense of “musical tone” or “note,” and at others the general sense of melody or tune. A particular arrangement of notes creates the framework of a mode, or maqām (Arabic for gāh, literally the “station” of the fingers on the instrument, but the term dastgāh is not yet attested and the repertoire of modes recognized in Persia since the 19th century were apparently unknown in the time of Hafez). Hafez uses this classical term for the modal system once, in describing a musician (moṭreb) as maqām-šenās (141:2a). Typically, however, Hafez uses the Persian noun parda, which occurs 14 times (13/58g, 1m) to indicate one of the established musical modes (parda can also mean the fret of a lute, a section of a reed flute, and the cords which fasten the strings to the harp frame, though it most often means veil or curtain). Although only a very small percentage of the occurrences of rāh/rah (“road, way”) in the Divān concern music, when this word occurs with the verb zadan, it frequently means to play a particular musical mode (“follow a melodic path”).

Hafez mentions several musical modes by their traditional names of ʿArāq(i) 5 (4/6, 1qa), Eṣfahān (1qa; qq.v. ʿERĀQ; BAYĀT-E EṢFAHĀN) and Ḥejāz (3/4g). He also appears to mention on occasion technical terms for melodic bridges or maneuvers, suites or movements within a composition, or fixed melodies, including ʿamal (1/14g), bāz-gašt (1g), Ḵosrovāni (2m), Pahlavi (1/2g), and Żarb-e oṣul (1m). This latter term is now applied to a reng in the mode Šur, but may simply indicate for Hafez musical rhythm or measure in general (Mallāḥ, 1972, pp. 156-58). A few of the 16 occurrences in the ghazals of the word ʿoššāq (“lovers”) may allude to the musical mode of this name (e.g., 119:2), as may the word navā (e.g., 26:4b). Mallāḥ (1972, pp. 80-82, 85-86) suggests that Jāma-darān (1g) and Taḵt-e firuzi (1g) may also constitute allusions to the names of particular modes or sub-sections thereof (he further mentions sāz-e Nowruzi in this capacity, pp. 141-42, but this term occurs in a line not considered authoritative by Ḵānlari).

Musical instruments: Musical instruments in general are called sāz, a word Hafez uses in this meaning per-haps 11 times (9/15g, 2qe), though sāz can also suggest harmonizing, or the tuning of, instruments(e.g., 150:1, 164:8). Several different verbs (in various conjugations) denote or connote performance on instruments in various contexts: zadan, sāz kardan/dādan/sāḵtan, less frequently navāḵtan 4 (3g, 1qe) and, metaphorically, ḵarāšidan (1g), etc.

String instruments or ālāt-e ḏawāt-al-awtār, of which Marāḡi (Jāmeʿ, p. 199) names 28 kinds: Hafez mentions rud 8 times (4/10g, 4m) in the meaning of lute, or string instruments in general. Marāḡi does describe two specific instruments with rud as an element of their name, viz., ṭarab-rud and šah-rud (Jāmeʿ, pp. 200, 206-7), the former apparently with sympathetic strings, the latter apparently revived by the author from desuetude. Hafez himself seems to use rud rather vaguely or generically. He does once (sāqi-nāma: 30b) explicitly link the rud with the celestial muse, Venus (Zohra), who elsewhere appears playing the ʿud or čang.

The number of lute strings varied by instrument, region and period, though Hafez apparently alludes to the proper names of four rud strings: the outer “low and high” strings (bam o zir, Ḵ. 251:3b) and between them the second (maṯnā, pl. maṯāni) and the third (maṯlaṯ, pl. maṯāleṯ) strings, mentioned in a qeṭʿa (Divān, p. 1083:4) and possibly also in one ghazal (Ḵ. 454:1b). Words used denotatively or connotatively for the strings of such instruments include abrišam, muy and rag. Possible allusions to the plectrum, zaḵma (zaḵm of Ḵ. 119:1b) or meżrāb (Ḵ. 313:6b), used in plucking the strings, have also been detected (Mallāḥ, 1972, pp. 130-31, 189-90). Specific lutes mentioned by Hafez include: Barbaṭ (q.v. BARBAṬ, where During suggests a Kushan/Gandharan origin, but note also the etymological similarity to Gk. barbiton, a lyre associated with Dionysus, Ionian lyric, and the sympotic poetry of Anacreon) 7 (6g, 1r), a short-necked instrument existing already in Sasanian times. In the time of Hafez this was probably identical with the ʿud, as Marāḡi does not mention the barbaṭ by name.

Čang (q.v.) 42 (38/42g, 1qa, 2m, 1qe), the most frequently mentioned instrument in the Divān. Though a word with multiple meanings (i.e., claw; clutches; grasp; harp), it usually occurs in the Divān in the sense of a harp, or of instrumental music generally. As depicted in the pre-Islamic period, the čang was an angular frame harp which, when strung, would form a triangle; in performance, the harp would be held against the chest or rest in the lap. In the Šāh-nāma, as apparently in Sasanian times, the harp was mostly played by women, which perhaps in part explains its association with Venus and suggests the metaphor of the harpist as Love personified, embracing the lover and running her fingers through his hair as she sings. The harpist, or muse (čangi 1g, 1m), is this archetypal Venus (nāhid-e čangi in the moḡanni-nāma, Divān, p. 1058:10b; zohra-ye čangi, 273:2b; čang-e zohra, 325:9a), playing the instrument by holding one edge of the frame to her chest (327:4a; see Mallāḥ, Ḥāfezá p. 90) and plucking the strings with her fingers (“claws” = čang). Marāḡi describes the čang as a well-known instrument (mašhur), typically with 24 strings (Jāmeʿ, p. 202). By the 16th century this instrument is said to have become obsolete, but it apparently retained its trigonal shape through the time of Hafez, who compares it to a person bent with age (čang-e ḵamida-qāmat, 122:6a), and a sage elder whose advice to be merry should be heeded (447:2, 470:5). The musical tones of the harp are variously described (bāng-e čang, naḡma-ye čang, nāla-ye čang, āhang-e čang, qawl-e čang, zam-zama-ye čang, ḡolḡol-e čang, čang-i ḥazin, samāʿ-e čang, navā-ye čang), and Hafez frequently mentions the instrument in tandem with another (nay, ʿud, rabāb, daf, čaḡāna), usually in contexts suggestive of dance and mirth, sometimes with ghazals being sung (e.g., Ḵ. 383:7).

Do-tāʾi (1g), evidently a two-stringed instrument (possibly an ancestor of the modern do-tār?). Marāḡi does not mention this, though he does have a section on the tuning of two-stringed lutes (Jāmeʿ, pp. 101-4), and in enumerating the kinds of instruments, he describes one named the yak-tāy and another called šaš-tāy, of which there are three types (Jāmeʿ, pp. 203 and 199-200).

Rabāb (7g), commonly robāb in Persian pronunciation, a precursor to the European rebec. A variety of zither or lute by this name is still used in Tajikistan and Afghanistan, but in classical Persian poetry it would normally appear to indicate a bowed spiked lute (similar to the modern Iranian kamānča). Writing within a generation after the death of Hafez, Marāḡi (who had himself traveled from Azerbaijan to Baghdad and at least as far as Herat) says the instrument enjoys its greatest popularity in Isfahan and Fārs. It is, he adds, strung with three, sometimes four [or even five] strings, tuned in pairs (mozawwaj) like the ʿud (Jāmeʿ, p. 202).

[Ṭanbur] occurs in one ms. of the sāqi-nāma, but in a line not canonically accepted by Ḵānlari (Divān, p. 1055); Marāḡi (Jāmeʿ,pp. 200-201) describes both a Turkish/Mongol and a “Šervaniān” [? šrwnyān] type of ṭanbur, as well as a nāy-e ṭanbur.

ʿUd (5/9g), a kind of wood from which incense and, apparently, the soundbox of a lute were made. Hafez uses the word variously, in the sense either of incense or as a musical instrument, of which Marāḡi mentions two varieties, the complete (ʿud-e kāmel) of five strings, and the traditional (ʿud-e qadim) of three or four strings, the respective tunings of which he explains in some detail (Jāmeʿ, pp. 101-9, 199).

Wind instruments, or ālāt-e ḏawāt-al-nafkò(on which, see Marāḡi, Jāmeʿ, pp. 207-9). Hafez does not mention any of the various bugles or horns (e.g., buq, karnā, šaypur, sornāy, ṣur, etc.), all of which usually pertain to a military context (razm) in the Šāh-nāma. In place of yarḡu (355:9b), some manuscripts record the word boṟḡu, a kind of royal ceremonial bugle, but if Hafez did in-deed use the latter, it was not in the context of a musical performance.

Arḡanun (2g), a pipe instrument (from Gk. organon), often used by Europeans (ahl-e farang), according to Marāḡi (Jāmeʿ, p. 209). It occurs both times in the gha-zals of Hafez in association with the harmonies of music and the celestial spheres. The word qānun, which, as an instrument belongs to the class of strings and denotes a zither, a kind of psaltery or dulcimer (Marāḡi, Jāmeʿ, p. 203), likewise comes from Greek, and may also connote a similar sense of “harmony” in one ghazal (161:4b).

Nay 16 (13/17g, 1m, 1r; once as nāy 1/2g), the reed flute, an instrument with particular significance in the semiotics of the classical Greek symposium (aulos, the double reed flute, frequently played by the female hetaera). Marāḡi describes many different kinds of nāy, but Hafez gives little indication of having a special one in mind. Hafez alternately gives the Persian nay a melodious sound (āvāz, naḡma), a plaintive one (faryād, nāla), or even that of an intelligible human voice (qawl, ḥadiṯ). After the čang, the nay is the most frequently mentioned musical instrument in the poetry of Hafez, and often appears paired with another instrument (daf, čang, barbaṭ, ʿud).

Percussive instruments. Drums and cymbals are not mentioned in the discussion of musical instruments in either of the books by Marāḡi, Hafez’s near contemporary. Similarly, Hafez does not depict any of the drums familiar to us from later or prior periods (e.g., tonbak/żarb, dohol, tabira, naqqāra, sanj) in a musical context, unless in one case (Ḵ 313:7a) we understand the sāqi to be drumming on an inverted drinking bowl (kāsa) as the poet recites (Mallāḥ, 1972, pp. 147-48). He does, however, mention a variety of drums and bells struck to inform and announce, including bells associated with travel and the caravanserai, jaras (4g) and darā (1g); the bell of a monastery, nāqus (1g, but Ḵānlari’s ed. gives nāmus here); and drums associated with towncriers, kus (2g, a kettle drum), or used in an idiomatic, non-musical sense, ṭabl (1g).

In a musical context, the only percussion instruments we find in the Hafez corpus are: Čaḡāna (q.v.; 2g), either a sort of castanets, or a dried gourd filled with seeds or pebbles to make a kind of rattle, closely associated in both occurrences in the Divān with the čang, probably in part because of its phonetic similarity (čang-e čaḡāna); and Daf 11 (10g, 1m), a tambourine or hand-held drum (see DAF(F) and DĀYERA). Both daf and čaḡāna occur only as accompanying instruments (usually with the čang or nay), apparently serving to create a dance rhythm.

Dance, motion and/or the emotive qualities of tone or the emotional effects of music. A gathering where music is heard is called samāʿ 13 (11g, 1qa, 1qe), especially if poetry/lyrics are involved and the audience responds enthusiastically by clapping, moving or dancing. Despite theological controversy over the permissibility of samāʿ, by the time of Hafez the practice seems to have been widely tolerated by the orthodox (see Māyel Haravi; and Lewis, 2000, pp. 309-13), though the word retained strong associations with Sufism, and can suggest a mystically transcendent state brought on by music and poetry, as do both wajd 2 (1g, 1m) and ḥālat (2/6g; cf. also ḥāl). Evoking a more earthly but no less enthusiastic pleasure in listening to music is ṭarab 16 (14/22g, 1m, 0/2qa, 1/3r), a word etymologically related to moṭreb (see below), the musician, and ṭarab-ḵāna 1 (1/2g) or ṭarab-sarāy (1g), the audience hall made joyful by the presence of music and/or the beloved. Words for dance and dancing include the noun raqṣ 11 (8g, 2m, 1qe), the verb raqṣidan (10g), the active participle pāy-kub (1g), and the adverbs raqṣ-konān (6g), pā-kubān (1g), and dast-afšān (4g).

Singers, musicians, muses: As generic nouns for musician and/or singer, we find moṭreb 43 (39g, 1qa, 3qe) and moḡanni 7 (1g, 6m), the latter also listed by Marāḡi as a kind of zither (Jāmeʿ, p. 202). Throughout an entire maṯnawi in the motaqāreb meter, Hafez addresses a moḡanni, who from the context is quite clearly a musician and singer. The moḡanni also appears in a ghazal (470:6b), apparently in the capacity of a singer, as the poet has earlier in the poem (470:4b) entreated a musician, moṭreb, to continue playing in the same mode, and asks the cupbearer to pour the wine in order to (or as soon as) s/he hears the Arabic words howa al-ḡani from the voice of the singer (ṣawt-e moḡanni). The moṭreb always connotes the presence of music, but is more broadly the musician of a symposion, frequently evoked in the same breath with wine (may, sāqi, bāda) or love (ʿešq). Sometimes the moṭreb appears in the same line with an instrument (e.g. Ḵ. 26:9, 343:2, 423:5), sometimes the context more explicitly suggests poetry and song (e.g., moṭreb-e širin-soḵan, Ḵ. 303:2a), in particular the singing of a ghazal of Hafez (moṭreb az gofta-ye Ḥafeẓ azal-i mast be-ḵᵛān,Ḵ. 169:8a). We also find words for the profession of singing, ḵonyāgari (1m), or the act of singing well or with a pleasant voice, ḵᵛoš-ḵᵛāni (1g).

Proper names associated with music include the famous Sasanian musician Bārbod (1m) or Bārbad (q.v.), and the muse, Venus,evoked respectively by her Per-sian and Arabic names, Nāhid 2 (1/2g, 1m) and Zohra 9 (7/8g, 1m, 1qe). An auspicious planet in the calculations of astrologers, Venus conducts the music of the celestial spheres (Ḵ. 4:8, 288:3) in her role as harbinger of music and play (laʿb-e zohra-ye čangi), as opposed to Mars, harbinger of war (merriḵ-e salaḥšur, Ḵ. 273:2). Venus probably also evokes the pagan practice, condemned in Islamic jurisprudence, of slave girls (kanizak, jāriya) as musicians (moṭreba) and singers (see, e.g., Māyel Haravi, p. 124).

Songbirds, often metaphorically representing the voice of human singers or lovers, include the famous bolbol 51 (49g, 1qa, 1qe; q.v.) or ʿandalib (7g), the hazār (e.g., Ḵ. 159:3b and 164:7b), the morḡ-e šab-kᵛān (2g), etc., and the generic “bird,” morḡ 59 (57g, 1qa, 1r), though not every occurrence is relevant to music. Adjectives describing singers or songbirds include ḵᵛoš-laḥja (3g), ḵᵛoš-ḵᵛān (2g), ḵᵛoš-alḥān (2g), šobḥ-ḵᵛān (1g), ḵᵛoš-āhang (1g), ḵᵛoš-āvāz (1g), ḵᵛoš-navā (1g), and naḡma-sarā (1g). “Songs of David,” naḡma-yeDāwud(i) (2g) describe two of the songs sung by birds, perhaps indicating a love song or a particular singing style. Verbs for singing include simple and compound instances of the following verbs, in conjugation or in the infinitive: ḵᵛān-dan, goftan, sorudan, sarāʾidan, sar dādan, etc.

Songs, lyrics, poems. The word ḡazal, which occurs 26 times in various forms in the Divān, strongly suggests poetic texts in Persian set to music, as noted above. Hafez uses the following forms of this lyrical word: ḡazal 17 (16g,1m), ḡazal-hā (4g), ḡazaliyāt (1g), ḡazal-ḵᵛān (3g), and ḡazal-sarāʾi (1g). As discussed above, ḡazal is often paired with the word qawl 11 (10/15g, 1m), also the words of a song, though in a more specific sense, an Arabic poem set to music. Other names for song, or styles of setting words to music, include pah-lavi (1/2g), which apparently describes a kind of ghazal (477:3b), sorud 13 (8g, 5m; meaning lyric, song, music), tarāna 6 (5g, 1qe, meaning melody or a Persian robāʿi that is sung), and zabur 1 (268:4, the psalms of David as love song/ghazal).

Modern musical versions of the poems. In the mode māhur, there is a guša known as Sāqi-nāma, to which poems in a certain meter addressed to the sāqi are frequently sung. This musical form would appear to be an innovation of the Qajar era, unrelated to the Safavid era performance of the poem going by that name in the Divān of Hafez (Mallāḥ, 1972, p. 276), though selected lines of the sāqi-nāma and moḡanni-nāma of Hafez have been recorded in this guša (e.g., Rowšanravān). It is, however, the ghazals of Hafez that have been especially important to the repertoire of traditional Persian music, and are probably more frequently sung, both in āvāz and taṣnif style, than the ghazals of any other poet. In the 19th century, Forṣat-al-Dawla (q.v.) specifies not only which modes are appropriate for particular poems, but also at what time of day and in what type of locations they should ideally be sung. An early folk/ethnomusicological specimen of Persian music was collected on a visit to Tehran by Nina Rosen, who apparently included transcriptions (in German translation and romanized Persian script) of two Hafez poems set to an Indian and an Afghan melody in her “Sieben orientalische Weisen: aus dem Munde des Volkes in Teheran” (Cologne, 1890).

Recordings of Hafez’s Persian poetry. The contemporary poet Aḥmad Šāmlu (d. 2001) made two notable recordings of selected ghazals of Hafez recited in his own voice and interspersed with music. A comprehensive list of recordings of Hafez poems in the original does not exist, but Niknām (pp. 131-37) gives a partial discography, identifying the specific poems sung. In the 1960s and 70s, the Iranian radio programs Golhā-ye rangārang and Golhā-ye tāza (see GOL-HĀ) were responsible for many recordings of taṣnif and āvāz performances of poems of Hafez. In the latter half of the 20th century, Hafez poems were recorded in a variety of traditional, folk and pop formats, and sung by performers including ʿAhdiya, Farāmarz Aṣlāni, Banān, Simā Binā, Delkaš, ʿAli-Reżā Efteḵāri, Golčin, Iraj, Maḥmud Ḵᵛānsāri, Mehr-dād Kāẓemi, Kāmrān Loṭfi, Marżiya, Šahrām Nāẓeri, Parisā, Ḥosayn Qawāmi, Ḥosayn Saʿādatmand, ʿAbd-al-Wahhāb Šahidi, and Moḥammad-Reżā Šajariān.

Hafez in European music. Translations, adaptations and inspirations from Hafez have repeatedly been set to music in songs/lieder of the western classical music tradition. Many of these survive only as sheet music for voice and piano, never having been heard outside private parlors, while others have entered the repertoire, some of them having been recorded more than once. Goethe’s (q.v.) West-östlicher Divan provided the inspiration for the Zurich-based Othmar Schoeck (1886-1957) in both his Op. 19b (1906-15) and his twelve “Hafis Lieder,” Op. 33 (1919-20, recorded in Lausanne, Gall, 2001). Goethe’s Hafez also inspired the “Symphonische Kantate nach Dichtungen aus Goethes West-östlicher Divan” of Waldemar von Bausnern (1866-1931), and the “Saki Nameh” (1970) of Hans Ludwig Schilling (b. 1927), who three years prior had published “Quintetto 67; Zeacis Hafis (on his beautiful daughter).”

Georg Friedrich Daumer’s translations, Hafis: eine Sammlung persischer Gedichte (Hamburg, 1846), inspired many to compose song-settings, including the singer-composer Bettina von Arnim (Brentano, 1785-1859), followed first by Adolf Jensen (1837-79) in 1863 (“Sieben Gesänge am Pianoforte aus dem Persischen,” op. 11), then by Johannes Brahms (1833-97) in 1864 (Liebeslieder, Op. 32, nos. 7-9) and again in 1868 (Fünf Lieder, Op 47, nos. 1-2). Daumer’s Hafis continued to provide lyric text for Emil Mattiesen in 1900 (“Zwölf Liebeslieder des Hafis,” op. 9), Erich Wolff (1874-1913), in several of his Lieder (nos. 18, 21, 25, 38, 54; recorded 1936), and Munich-based Rudolf Bode (1881-1970), for his “Hafis-Lieder” in 1939 (see also HAFEZ xi). The text of Hans Bethge’s Hafis (Leipzig: Insel, 1919) provided the lyrics for four songs by Richard Strauss (1864-1949) from the 1928 collection “Gesänge des Orients” (Op. 77), including Ihre Augen, Schwung, Die Allmächtige and Huldigung, and also for the Polish composer Karol Szymanowksi (1882-1937), with his “Des Hafis Liebeslieder,” Op. 24. Kaikhosru Shapurji Sorabji described this 1911 piece as “giving us in musical terms what we instinctively know and recognize as the essence of Persian art” (Sadie, 1995, s.v. “Szymanowski”). Szymanowksi also wrote a symphonic setting for a poem of Rumi, as well as his posthumously published “Das Grab des Hafis” (1937). Ludwig Hess (1877-1944) did “Songs of Hafez,” Op. 40, in German and English, featuring “Wenn einer mässig trinket,” which appeared in 1912, just as the temperance movement was picking up steam in America.

Hafez seems to have been such a staple of the Lieder genre that several composers began their career with him, inlcuding Frédéric Louis Ritter (1834-91), who published his Opus 1 for piano and voice as “Hafis: ein Liederkreis aus dessen Gedichten” (German, with English version by Fanny Raymond Ritter, Leipzig, 1866), and the Swiss composer Paul-Müller-Zürich (b. 1898), who published his “Acht Lieder nach Gedichten von Hafis,” Op. 1 (1920).

For some composers the notion of the exotic Orient seems more important than the poetry of Hafez: Sir Granville Bantock (1868-1946; Figure 1) composed “Five Ghazals of Hafiz” based on Edwin Arnold’s translations, as well as pseudo-eastern pieces such as “ Lalla Rookh” and “Omar Khayyam.” In 1926, the Viennese-born, Edinburgh-based Hans Gal, who also set pieces by Goethe, Rückert and Tagore, included Hafez in his “Herbstlieder,” Op. 25, for an a capella women’s choir. The Ameri-can Bertram Shapleigh (1871-1940) composed a “Hafiz Serenade: Fünf Lieder,” Op. 32 (1901), in addition to pieces such as “Dance of the Dervishes” (1905), “Ramayana” (1908), and “Vedic Hymn” (1910). Meanwhile, Mrs. Shapleigh provided English words for the German text of the “Hafis Lieder” of Theodor Streicher (1874-1940), sheet music published in Germany and New York in 1907.

Various European renderings of Hafez continued to inspire western composers well into the 20th century. American-born, Paris-based Blair Fairchild (1877-1933) contributed seven songs in 1914 based on the renderings of Paul de Stoecklin, “Les amours de Hafiz,” Op. 38, and ten years later the Danish composer Poul Schierbeck (1888-1949) composed a song for piano titled “Le Tombeau du poète Hafiz.” Richard Le Gallienne’s jaunty translations of Hafez appeared in both “Before the Dawn, a Persian Idyll,” a 1913 musical setting for a chorus of men’s voices by W. Franke Harling (1887-1958), and “A Caravan from China Comes,” a 1920 score by the American composer Elliott Griffis (1893-1967). Alan Hovhaness (1911-2000) used the John Hindley translations in 1938 for his “Seven Love Songs of Hafez,” Op. 33. In 1926 Anatolii Aleksandrov (1888-1977) set A. Fet’s Rus-sian adaptation of Hafez to music as “Tri stikhotvoreniia Gafiza,” Op. 2 (Rimsky-Korsakov may also have composed for a Fet version of Hafez). In the same year, Dirk Foch (1886-1973) did a cycle of three Hafez song settings as “Drei Stimmungen” (op. 20). Viktor Ullman, the Austro-Hungarian composer who died at Auschwitz in about 1944, likewise has a “Liederbuch des Hafis,” Op. 30 (recorded 1999), while the Austrian anti-Nazi, Gottfried von Einem set eight Hafez poems to music in his Opus 5 of 1945. Franz Alphons Wolpert (1917-78) set “Vier Lieder nach Gedichten des Hafis” to music in 1948. More recently, the Hafez tradition has been kept alive by Jón Leifs (1899-1968) and his “Drift Ice” in Icelandic, with text by Einar Benediktsson (Hafís: fyrir blanda an kór & hljómsveit op. 63; recording 1999); Ernst Piffner (b. 1922) with “Ein Hafis-Zyklus” (1983 re-cording); and Franco Donatoni (b. 1927) composed “Arie per voce femminilie e Orchestra,” a piece setting poems mostly of Khayyam, but also Hafez, as late as 1989.

For a music sample, see Agar ān tork-e Širāzi.

For a music sample, see Hejāz.

For a music sample, see Kāleqi, Mey-e nāb.



Ṭalʿat Baṣṣāri, “Eṣṭelāḥ-hā-ye musiqi dar Divān-e Ḥāfezá,” in Jašn-nāma-ye Modarres-e Rażawi, Tehran, 1356 Š./1977, pp. 79-109.

Jiří Cejpek, “Iranian Folk-Literature,” in Rypka, Hist. Iran. Lit., pp. 607-709.

Daniela Meneghini Correale, The Gha-zals of Ḥāfeẓ: Concordance and Vocabulary, Rome, 1988.

Idem, Lirica Persica Hypertext (HyperFolia 1, Lirica Persica 17), Venice, 2000.

Jean During, “Chang,” in New Grove Dictionary of Musical Instruments, ed. Stanley Sadie, New York, 1984, I, p. 332.

Moḥam-mad-ʿAli Eslāmi Nadušan, “Šiva-ye šāʿeri-e Ḥāfeẓ,” Irān-nāma 6/4, 1367 Š./1988, pp. 521-58.

Mirzā Naṣir Forṣat-al-Dawla Širāzi, Boḥur-al-alḥān dar ʿelm-e musiqi wa nesbat-e ān bā ʿaruż, ed. Moḥammad-Qāsem Ṣāleḥ Rāmsari, Tehran, 1367 Š./1989.

Michael Hillmann, Unity in the Ghazals of Ḥāfeẓ, Minneapolis and Chicago, 1976.

Jalāl-al-Din Homāʾi, ed. Divān-e ʿOṯmān-e Moḵtāri, Tehran, 1341 Š./1962, pp. 81, n5, 375, n1, 524-25, n3, 569-76 (extended footnote).

Theodore Levin, The Hundred Thousand Fools of God: Musical Travels in Central Asia (and Queens, New York), Bloomington, 1996 (with Compact Disc).

Franklin Lewis, “Reading, Writing and Recitation: Sanā’i and the Origins of the Persian Ghazal,” Ph.D. Diss., University of Chicago, 1995.

Idem, Rumi: Past and Present, East and West, Oxford, 2000.

Ḥosayn ʿAli Mallāḥ, Ḥāfeẓ o musiqi, Tehran, 1351 Š./1972, 2nd ed., 1363 Š./1984.

Idem, Manučehri Dāmḡāni wa musiqi, Tehran, 1363 Š./1984.

Idem, Payvand-e musiqi wa šeʿr, Tehran, 1367 Š./1988.

ʿAbd-al-Qāder b. Ḡaybi Ḥāfeẓ Marāḡi, Maqāṣed-al-alḥān, ed. Taqi Bineš, Tehran, 1344 Š./1965.

Idem, Jāmeʿ-al-alḥān, ed. Taqi Bineš, Tehran, 1366 Š./1987.

Najib Māyel Haravi, Andar azal-e ḵᵛiš nehān ḵᵛāham gaštan: samāʿ-nāmahā-ye Fārsi, Tehran, 1372 Š./1994.

Julie Scott Meisami, “Persona and Generic Conventions in Medieval Persian Lyric," Comparative Criticism 12, 1990, pp. 125-51.

Heshmat Moayyad, “Lyric Poetry,” in Ehsan Yarshater, ed., Persian Literature, Columbia Lectures on Iranian Studies 3, New York, 1988, pp. 120-46.

Šahrām-e Nāẓeri, singer on the audiocassette Gol-e ṣad barg in Bayāt-e Tork (Caltex Records C458), where the first line from Hafez’s first poem (alā yā ay-yohā . . .) is interspersed in Rumi’s ghazal delā nazd-e kasi benšin. Kristina Nelson, The Art of Reciting the Qur’an, Austin, 1985; repr. Cairo, 2001.

Mehrdād Niknām, Ketāb-šenāsi-e Ḥāfez, Tehran, 1367 Š./1989.

Kāmbiz Rowšanravān, orchestration, Sāqi-nāma/Ṣufi-nāma; voice: Šahrām Nāẓeri, 2 audiocassettes [Tehran], Sāzemān-e Farhangi o Honari-e Hārmoni, 1364 Š./1985 (?), side 4.

Stanley Sadie, ed., The New Grove Dictionary of Musical Instruments, 3 vols., London and New York, 1984.

Idem, ed., The New Grove Dictionary of Music and Musicians, 20 vols., London and New York, 1995.

Katāyun Ṣārami and Faridun Amāni, Ferdowsi wa musiqi dar Šāh-nāma, Tehran, 1373 Š./1994.

Moḥammad-Reżā Šafiʿi-Kadkani, Musiqi-e šeʿr, Tehran, 1357 Š./1978, 3rd revised ed., 1370 Š./1991.

Idem, Mofles-e kimiā-foruš: naqd o taḥlil-e šeʿr-e Anvari, Tehran, 1372 Š./1992.

Sirus Šamisā, Sayr-e azal dar šeʿr-e fārsi, Tehran, 1363 Š./1984.

(Franklin Lewis)

Originally Published: December 15, 2002

Last Updated: March 1, 2012

This article is available in print.
Vol. XI, Fasc. 5, pp. 491-498