bi-hi, bi-him... fi-hu? Pronominal Suffix Harmonization Diversity in Some Vocalized Christian Arabic Gospel Manuscripts.

• Author: Stokes, Phillip W.

1. INTRODUCTION

   Christian Arabic, defined here as literature produced by Christians for communal purposes, (1) is an important source not only for the study of Christianity in the Islamic Near East, but also for the history of the Arabic language. In addition to the cultural and theological aspects of these texts, scholars in recent decades have become interested in the linguistic study of Christian Arabic manuscripts as well. (2) While many of these studies include references to linguistically relevant aspects of the manuscripts and texts, the linguistic framework in virtually all studies of the Christian corpora, as well as the methodology by which the data are presented, is that of Joshua Blau, and specifically of his three-volume grammar of Christian Arabic, which studies texts produced (primarily) in southern Palestine in the ninth and tenth centuries CE. (3)

   One important consequence of Blau's focus on the earliest datable manuscripts is that the manuscripts upon which his grammar is based are either unvocalized, or only sparsely vocalized. When remarks are made about aspects of the vocalization of a particular manuscript, they are often brief, and typically focus on saliently non-Classical Arabic (henceforth ClAr) features; (4) rarely are we presented a comprehensive picture of a feature, or whether it represents a consistent pattern in the manuscript instead of a one-off occurrence. Further, examples from multiple manuscripts are included together, increasing the impression of randomness and chaos. Subsequent studies, which again frequently rely on Blau's grammar almost exclusively, likewise often mention vocalizations primarily when they are deemed non-ClAr. (5) Consequently, despite progress in some areas of the linguistic study of Christian Arabic corpora, there remains a significant lacuna in our knowledge, namely, of the vocalization patterns attested in each manuscript, as well as any linguistic typology of those patterns.

   Vocalizations in any corpus constitute an extremely important source for the study of a range of linguistic and sociolinguistic topics. While the consonantal orthographic tradition was far from uniform, there was nevertheless a set of scribal traditions for how to write, originally developed for a Hijazi variety of Arabic, (6) that undoubtedly obscured much potential linguistic variation. As with the Quran, a more or less fixed consonantal text that can nevertheless be read in a wide range of manners, Christian manuscripts lacking vocalizations are in many ways opaque to linguists. While the ability of scribes to tailor the Hijazi-based scribal practices to express the nuances of different varieties of Arabic was originally rather limited, the proliferation of vocalization and other diacritic conventions expanded that ability significantly. Vocalizations provide a clearer window into the features of each text and how it was read. These vocalizations are being profitably studied in other corpora, especially Judaeo-Arabic, (7) but are virtually untapped in the linguistic study of Christian Arabic.

   This paper is a first attempt at documenting and analyzing one aspect of the reading traditions reflected in the vocalizations attested in a handful of medieval Christian manuscripts containing translations of the gospels. Specifically, I focus here on the realizations of the third person pronominal suffixes and the harmonization patterns, or lack thereof, reflected in the vocalizations of nine different manuscripts. I chose this feature for several reasons. First, pronominal suffixes are regularly vocalized in the Christian Arabic corpora, even in otherwise sparsely vocalized manuscripts. Second, there is significant variation in the patterns of pronominal suffix vocalization, both in the Christian and non-Christian corpora. Third and finally, it presents a useful heuristic case for the interpretation of oft-presumed sub- or nonstandard features and patterns in medieval texts.

   In what has become normative ClAr, third person pronominal suffixes assimilate to a preceding high vowel or the diphthong ay (Table 1).

   However, according to the grammarians, (8) this pattern was not the only one permitted in ClAr, at least for several centuries, nor was it ubiquitous among Quran reciters. Instead, as we will see, there were numerous patterns across corpora in the early Islamic period. Virtually none of the Christian manuscripts examined here follows the textbook ClAr pattern used today; rather, these Christian patterns find parallels--often identical--in earlier non-Christian corpora. Following a thorough documentation of the Christian data, I will contextualize attested pronominal forms and harmonization patterns in comparison with patterns attested in Islamic-era Arabic corpora, including those mentioned by the early grammarians, as well as various sources on the quranic readings. With only nine manuscripts included, the present study does not pretend at completeness or exhaustiveness; it is, as mentioned, a first foray into what I hope will become well-traveled ground. My goals in this paper are thus twofold. First, I aim to illustrate the value of fully documenting and studying the vocalization strata of each manuscript, taking it seriously as representative of a style or register in its own right. And second, I hope to offer a compelling case for broadening and deepening the linguistic background and contexts against and within which scholars approach the study of Middle Arabic texts.

2. PREVIOUS SCHOLARSHIP

   Because most of the linguistic study of Christian Arabic has been carried out on unvocalized or lightly vocalized texts, very little discussion of the shape of pronominal suffixes and their vocalization exists. In his grammar Blau reports only the 3ms suffix as -uh, regardless of syntax, (9) which in his analysis "demonstrates the loss of the cases." (10) Blau does not report the existence of any vocalized third plural pronominal suffix forms in the corpus he examines.

   Some scholars do not report on the form at all, presumably either because it is -uh, or because it was not considered significant. In a study of several manuscripts containing the book of Judges, (11) Bengt Knutsson does not mention the 3ms forms, with the only discussion of pronominal suffixes consisting of non-ClAr agreement, as when the dual is used where one would expect the plural, and when the feminine plural occurs where one might expect the masculine plural. (12) In Per Bengtsson's study of manuscripts attesting the book of Ruth, a brief note is included noting only that the "3rd p.masc.sing., [??], is -uh." (13) We may infer that the manuscripts that Knutsson and Bengtsson studied simply lack more than a handful of instances of vocalized pronominal suffixes, but we are not explicitly told how common these occurrences are in any given manuscript. Such discussion is vital, however, not only for our knowledge of this feature and its distribution, but also for the light these shed on other aspects of the language underlying a manuscript, including especially that of morpho-syntactic case marking. For example, following Blau, Bengtsson concludes without any hesitancy that the attestation of 3ms -uh in a syntactically genitive context constitutes an "example of the disappearance of the case vowels." (14) Yet such analysis is necessarily true if, and only if, the only possible case-bearing variety of Arabic looked like what has become normative ClAr, specifically with its pattern of third person pronominal suffix harmonization. As will be discussed further below in [section]4 (and see above, n. 4), however, this is contradicted by the evidence. But before discussing alternative interpretations, we first turn to a discussion of the manuscripts and the data.

3. MANUSCRIPTS AND DATA

   The manuscripts included in this study, with one exception, are from the monastery of St. Catherine in the Sinai, Egypt: Sinai Arabic (henceforth SAr.) 76, 80, 82, 89, 90, 91, 115, and 146. The lone exception, Leiden Or. 561, is from the Leiden University holdings. I chose these nine manuscripts for two reasons. The first was pragmatic--they have been digitized (15)--and the second was methodological--they represent a diversity of translation families, as argued by Hikmat Kashouh. (16) According to Kashouh, the manuscripts from St. Catherine are from three different subfamilies of a larger group of manuscripts (Kashouh's Family J) that descend from a translation of a Syriac Vorlage, but which were subsequently influenced by, and occasionally corrected against, Greek translations. According to Kashouh's classification, SAr. 115 and 146 belong to one subfamily (Ja), SAr. 80, 82, 89, 90 and 91 belong to another (Jb), and SAr. 76 to yet a third (Jc). (17) Finally, Leiden Or. 561 belongs to a completely different family (Family F), which consists of only three manuscripts, possibly translated from the Peshitta but which attest no Greek influence. (18) The bulk of the manuscripts date to the thirteenth century CE, while two date to the fifteenth (Table 2).

   I have included a brief description of each manuscript, with notes on relevant orthographic and vocalization details. Only one (SAr. 76) has to date been given detailed treatment. (19) Following each brief introduction, I list all pronominal suffix variants from the gospel of Matthew, which is attested in complete form in all the manuscripts included with the exception of SAr. 146, which begins with Matthew 18:43. Following the variants is a discussion of the pattern and variants found in each manuscript.

   Before proceeding to a discussion of each individual manuscript, it is necessary to offer a brief remark on the challenge posed by some of the vocalization patterns in the corpus of manuscripts selected for this study. In several manuscripts, non-ClAr vocalization patterns govern the distribution of certain diacritics, especially those of the damma and shadda. (20) This is, in fact...