PAGE 1

The Vocalization of Medieval Hebrew PoetryNumerous medieval manuscripts of Hebrew poetry are vocalized with Tiberian vocalization signs. Many of these display some amount of irregularity in the application of the graphemes, as well as evidence of non-standard vocalization practices. A large number of such manuscripts that deviate from standard Tiberian vocalization have been preserved in the Cairo Genizah and it is upon this corpus that the following description is based. The majority of the manuscripts of Hebrew poetry in the Cairo Genizah were copied in the Middle Ages between the 10th13th centuries. Many of these texts are informal copies. Recent studies of non-standard scribal use of Tiberian graphemes (see Sharvit 1986:119135; 1992:501518; Yeivin 1997:159178; Jefferson 2004) reveal that the copyists (due to the oral nature of the poetry) produced mixed patterns inuenced to a certain degree by each of the three main biblical reading traditons: Babylonian, Palestinian (its graphic variant: Palestino-Tiberian) and Tiberian. Commonly attested in medieval Hebrew poetic manuscripts are interchanges of the two a vowels ( qame and pata ) and of the two e vowels ( ere and seghol ), respectively. This phenomenon is also found in fragments of early piyyu im from the Genizah with Palestinian vocalization, reecting the redundancy of these vowels in the Palestinian pronunciation tradition (Yahalom 1997:1416). The most frequently occurring non-standard vocalization trait in the poems with Tiberian vocalization is the interchange of the qame and pata graphemes. This interchange is fairly random across the syllables, but it is especially common in nal syllables. Manuscript CUL T-S H6.15, for instance, has fteen examples including (standard Tiberian ni na ) was subdued (f.5.r.13), (standard Tiberian yi ra ) will grow (f.7.r.1) and (standard Tiberian y uppar ) will be atoned for (f.10.r.15). Interchanges of the seghol and ere graphemes also occur sporadically, mostly in initial syllables. For example, MS CUL T-S NS 299.98 has (standard Tiberian k < ) like grass (f.2.v.3) and (standard Tiberian l ) mourning (f.2.v.9). Manuscripts that exhibit a ere grapheme where seghol is normally required often attest such interchanges in nal syllables and segholate forms, including (standard Tiberian hy ) I will be (MS West. Coll. Lit.I, f.57.r.1) and (standard Tiberian m ) whisper (MS CUL T-S H6.15, f.9.r.10). Less frequent interchanges are recorded between the e and i vowels. ere is often found arbitrarily where standard Tiberian has ireq whereas ireq is found to occur more frequently in the sufxes of construct forms where ere is expected: (standard Tiberian sa : gr r ) the steady rain (MS CUL T-S H6.15, f.2.r.10). Interchanges of the o and u vowels are also recorded, with a greater number of olem vowel signs replacing the shureq vowel sign, e.g., (standard Tiberian m < uzz < m ) their stronghold (MS CUL T-S H6.15, f.6.r.7). Such substitutions are also typical of the Sephardi pronunciation of Hebrew (see the study of end rhymes in Garbell 1954:686690) and are widely attested in the Sephardi prayerbooks (see Dodi 2003:1518). The same traits are likewise found in the Ashkenazi reading tradition and the Ashkenazi prayerbooks of the 10th14th centuries (see the examples in Eldar 1978:1635). Another frequently attested vowel sign interchange in the medieval Hebrew poetry manuscripts is that of the pata and seghol graphemes. This may reect the inuence of the Babylonian pronunciation tradition, in which the corresponding vowels are believed to have had a similar if not identical quality (see Yeivin 1985). In poetic manuscripts vocalized in the Tiberian tradition the pata grapheme is often found where standard Tiberian vocalization requires seghol specically with guttural letters or in initial syllables (particularly after nun in nif al forms like (standard Tiberian n ( m < n ) he is trusted (MS CUL T-S H6.15, f.7.v.14), after he in hif il forms, like (standard Tiberian h ( r ) he crowned (MS JTS ENA 3015, f.3.v.7), and in segholates, J e f f e r s o n R J W t h e v o c a l i z a t i o n o f m e d i e v a l i n d d 1 Jefferson, R.J.W._the vocalization of medieval.indd 1 1 1 / 2 2 / 2 0 1 0 3 : 0 3 : 0 5 P M 11/22/2010 3:03:05 PM

PAGE 2

2 the vocalization of medieval hebrew poetrylike (standard Tiberian r ) prey (MS CUL T-S H15.139, f.1.v.7), but for the most no pattern can be discerned in its appearance. The seghol grapheme is also found randomly in places that normally require pata but it occurs with more regularlity in nal syllables. One noticeable pattern is the periodic vocalization of the interrogative pronoun what/how with seghol where standard Tiberian has pata ( ma ). A number of the deviations in the medieval Hebrew poetic manuscripts preserve genuine features of Tiberian pronunciation, but with a different notation. This is especially relevant in the case of shewa Masoretic literature reveals that standard Tiberian had three pronunciations of shewa : in general it was pronounced like pata ; before yod it was pronounced like ireq and before gutturals it acquired the quality of the vowel following the guttural (Yeivin 1980:281282 and Tiberian Reading Tradition). Copyists of poetic manuscripts occasionally reveal these pronunciation traits by placing pata where standard Tiberian vocalization requires shewa e.g., [gamo 1 l] (standard Tiberian g m l ) bestow! (MS JTS ENA 641, f.7.v.5 & f.8.r.4). Some of the manuscripts have ireq and not shewa before consonantal yod : [biy 1 o 1 ] (standard Tiberian b y < ) in his hand (MS CUL T-S H6.72, f.1.r.2), and some have ireq before a guttural with ireq where shewa is required in standard Tiberian: [ti inn 1e] (standard Tiberian t inn < t < ) your supplication (CUL MS T-S H10.38, f.1.v.1). Medieval Hebrew poetic manuscripts also attest the use of the shewa grapheme on initial and nal consonants that do not normally require a vowel sign, e.g., (standard Tiberian b an ) tested (Bodl. MS, f.49.v.6) and (standard Tiberian l < ha ; g ) he pronounced (MS CUL T-S NS 131.34, f.1.v.5)). This usage probably reects the inuence of Palestino-Tiberian vocalization, whose main trait was the use of diacritics and vowel signs to disambiguate as many of the letters as possible. (The term Palestino-Tiberian is just one of the names to describe the system that employs Tiberian graphemes to represent Palestinian pronunciation. On account of its distinctiveness and comparatively systematic representation in medieval manuscripts, it is often dened as a system in its own right (Dotan 1971:1466 72), although it is essentially a graphical variation of the Palestinian system and therefore not a separate pronunciation tradition.) Numerous medieval Hebrew poetic manuscripts exhibit other non-standard deviations from the Tiberian pronunciation tradition, such as the use of the shewa grapheme in place of the a ephim a graphic convention regularly found in medieval manuscripts. Many of the vocalizers appear to demonstrate some measure of confusion or carelessness in regard to the application of the a ephim and a number of poetic manuscripts have full vowels where a ephim are expected, e.g., (standard Tiberian n ( l < m < ) dumb (MS CUL T-S H10.12, f.1.r.7). The vocalization of pausal forms in medieval Hebrew poetry manuscripts is largely inconsistent. Some strophes are vocalized without the required pausal forms, see, e.g., MS JTS ENA 641, f.7.r.8: (standard Tiberian b ay ) with my debts and f.7.r.9: (standard Tiberian s ay ) around me; others have them where they are not expected, often as a result of reproducing a biblical phrase that occurs in pausal position in the original biblical verse, e.g., CUL MS T-S H6.15: nir < (standard Tiberian nir a ) shattered (f.6.r.15), y < < (standard Tiberian y < a ) advised (f.6.v.1), and w y < n (standard Tiberian w y an ) and is shown favour (f.6.v.4). Thus, where other vocalization traits show the inuence of non-biblical language traditions, the vocalization of pausal forms often betrays the strong inuence of a core biblical education upon these scribes and how readily it was recalled. Some of the morphological variants occurring in the Genizah copies of the medieval Hebrew poems betray the inuence of both the Palestinian and the Babylonian vocalization of Rabbinic Hebrew. For example, the vocalization of some of the segholate nouns reects the Palestinian vocalization of Rabbinic Hebrew, where the guttural consonants do not affect the surrounding vowels, e.g., (standard Tiberian p a ) transgression (MS CUL T-S H6.15, f.5.r.10, f.13.r.8), (standard Tiberian p : ga ) occurrence (MS JTS ENA 3015, f.3.r.9), and (standard Tiberian p sa ) Passover (MS West. Coll. Lit.I.26, f.1.r.9). This is particularly true of segholate J e f f e r s o n R J W t h e v o c a l i z a t i o n o f m e d i e v a l i n d d 2 Jefferson, R.J.W._the vocalization of medieval.indd 2 1 1 / 2 2 / 2 0 1 0 3 : 0 3 : 0 5 P M 11/22/2010 3:03:05 PM

PAGE 3

the vocalization of medieval hebrew poetry 3nouns with et as the third radical, which have seghol rather than pata preceding it, but also holds true of nouns ending in ayin (cf. Yahalom 1997:25). In spite of the aforementioned deviations from standard Tiberian vocalization, the rhythm and sound patterns in the copies of medieval Hebrew poems do not appear to have undergone great alteration. Rhyme would seemingly be most affected by interchanges of pata and seghol (for example, see MS CUL T-S NS 131.34: (standard Tiberian m yuqq r ) esteemed (f.1.r.9), which is supposed to rhyme with the words n q r investigated and l q < r at the entrance of the city (f.1.r.11). But, if in the scribes own pronunciation these graphemes represented one vowel sound, as in the Babylonian tradition, then, in fact, the rhyme remains intact. The scribe of MS CUL T-S H15.139 even maintains his non-standard pattern of vocalization in all of the rhyming syllables ending in (standard Tiberian ). Other non-standard features, such as those inuenced by the Palestinian tradition of pronunciation do not impact upon rhyme. The pata and qame graphemes were pronounced alike, as were the seghol and ere graphemes, and these vowels were already used in free variation in the rhyming syllables of Spanish poetry. Meter may have been impacted to some degree, yet, for the most part, the non-standard variations, such as the interchange of a ephim with full vowels, are merely graphic conventions that do not affect vowel quality and metrical value. Thus, many of the medieval copies of Hebrew poetry reveal a genuine mix of traditions. The qame / pata interchange of the Palestinian tradition together with the seghol / pata interchange of the Babylonian tradition, as well as standard Tiberian features. These mixed pronunciation traits, combined with many graphic deviations and genuine errors, all reect the orality inherent in the scribal process that produced the rendition of each text. Indeed, this type of vocalization indicates that, although the copyists main objective was to facilitate the reading of the poem by others, he did so in a manner that reected his own reading, adding vowel signs where he felt it necessary, and sometimes using those signs that were closest to his tradition of pronunciation. ReferencesPrimary Sources Mishna Codex Parma B De Rossi 497 (in Hebrew), ed. by M. Bar-Asher, Jerusalem. Secondary Sources Dodi A. 2003. Vocalized prayer books from Spain and its cultural sphere of inuence: Problems in manuscripts from the Christian era. Kenishta 2:920. Dotan A. 1971. Masorah. Encyclopedia Judaica 16:14011482 [supplementary article]. Eldar I. 1978. The Hebrew language tradition in medieval Ashkenaz (ca. 9501350 C. E.). Vol. I: Phonology and vocalization (in Hebrew). Jerusalem. Garbell I. 1954. The pronunciation of Hebrew in medieval Spain. Homenaje a Mills-Valliscrosa Vol. 1, 647696. Barcelona: Consejo Superior de Investigaciones Cientcas. Jefferson, R. J. W. 2004. Popular renditions of Hebrew hymns in the middle ages: Based on a selection of vocalized liturgical poems by Solomon Ibn Gabirol from the Cairo Genizah . PhD diss., University of Cambridge. Sharvit S. 1986. Niqqud bavli-tavrani bi-genizat qahir (in Hebrew). Massorot 2:119135. . 1992. iyyunho be-niqqudam ek katte tella mi-genizat qahir (in Hebrew). Mexqarim be-laon 5:509518. Yahalom J. 1997. Palestinian vocalised piyyu manuscripts in the Cambridge Genizah collections Cambridge: Cambridge University Press. Yeivin I. 1980. Introduction to the Tiberian Masorah Trans. E. J. Revell. Missoula, Montana: Scholars Press. . 1985. The Hebrew language tradition as reected in the Babylonian vocalization (in Hebrew). 2 vols. Jerusalem. . 1997. urot leon axmim be-xitve-yad el piyyu im (in Hebrew). Massorot 910:7789. Rebecca J. W. Jefferson (Price Library of Judaica, Gainsville) J e f f e r s o n R J W t h e v o c a l i z a t i o n o f m e d i e v a l i n d d 3 Jefferson, R.J.W._the vocalization of medieval.indd 3 1 1 / 2 2 / 2 0 1 0 3 : 0 3 : 0 5 P M 11/22/2010 3:03:05 PM