Summary Tiberian Hebrew Pausal Forms University of Chicago Dissertation (1992) by Richard L. Goerwitz, III I. Motivation Since about 1987, when I first read Geoffrey Khan's "Vowel Length and Syllable Structure in the Tiberian Tradition of Biblical Hebrew" (Journal of Semitic Studies 32:1 [1987], pp. 23-82), I have become increasingly dissatisfied with virtually all modern treatments of biblical Hebrew phonology. Whether neogrammarian, descriptivist, structuralist, or generative, these treatments almost invariably assume the quasi-Sephardic pronunciation exemplified in the works of David Qimhi. Although it has been known for over a century now that Qimhi's was only one of several vowel systems known to medieval Jews, recent work on Karaite transcriptions found in the Cairo Geniza has provided us with clear proof that his was not the system actually used by Tiberian cantors and naqdanim. Though perhaps originally quantity-based, like Arabic, the system used by the medieval Tiberian scholars had become quality based. Vowels, in other words, were distinguished primarily on the basis of their timbre, and not on the basis of length. Put concretely, a patah could be either long or short, as could a qamets. Length simply did not constitute an intrinsic feature of these, or any other (except possibly the hatef), vowels. Trite as this observation might sound, it actually has some rather important consequences, and initially my goal was to assemble a new, ground-breaking reanalysis of Tiberian Hebrew phonology based on these notions. This task, however, proved far too ambitious for a dissertation. More suitable, I felt, would be a study of a specific phonological phenomenon - one which, though broadly attested, could be narrowed down to a manageable range of alternations and forms. The phenomenon I ultimately selected was pause. Through a study of Hebrew pause, I hoped to show how Khan's work ramified when applied systematically to a well-known (but in fact only rarely discussed) set of forms and alternations. II. What is Pause? The Hebraists' term pause (Hebrew hefseq) traditionally designates clause-final positions in the Tiberian Hebrew verse where we find regular changes in the vowel and/or stress patterns of certain words and word-classes. I have no quarrel with this definition. My quarrel is with scholars' habit of portraying the changes in question as predictable enough that they can be ignored, or else cast in terms of a few sweeping generalities. In reality, pausal vowel and stress changes come in many varieties, applying to everything from whole classes of inflective forms to single lexical items. Each of these varieties has its own quirks and idiosyncrasies. Each also offers us valuable insight into the nature and structure of the language. Though they are not without certain regularities, pausal forms are far more diverse and less predictable, than most textbooks and grammars care to admit. III. Failures in Previous Work The reason for the prevailing tendency to over-regularize analyses of pausal forms is that these forms have traditionally been conceived of as arising from low-level "automatic" lengthening and re-stressing processes. As noted above, however, recent work on medieval Karaite texts has shown that Hebrew had transformed itself from a length-based, syllable-timed system to a stress timed one in which quantity was of minimal phonological significance. During the transitional period, older length-based alternations were either lost or turned into qualitative shifts. To claim that pausal alternations in Tiberian Hebrew reflect automatic, stress/length-based processes is thus absurd, and reflects a deep misunderstanding of medieval Tiberian phonology. Although pausal alternations probably originated as such processes, they could not have been preserved as such within the new, quality-based medieval phonological matrix. Why, then, did they remain? IV. The Morphologization of Pause The reason pausal alternations remained in Tiberian Hebrew is that Tiberian Hebrew was not a natural language. It was, rather, a collection of fixed, liturgical corpora. Within this context, relatively meaningless phonetic quirks at one historical level were easily reincorporated as higher-level phenomena at the next. This is precisely what Tiberian Hebrew "speakers" did to the old pausal alternations. In effect, they re-learned the pausal forms as quasi-morphophonemic, or in some cases lexical, deviations from their nonpausal counterparts. Pausal forms became part of the liturgical reading process, much as dialectal v-syncope (o'er, e'er, e'en) has become part of English poetry and song. V. Conclusion Because pause cannot generally be viewed as a natural reflex of clause-final stressing or lengthing processes, and because it does not apply in a phonologically consistent manner, it must necessarily represent a higher-level, quasi-grammatical phenomenon - one that was memorized as part of the reading tradition. This one simple point essentially sums up the entire thrust of my dissertation. My aim in writing, though, has not been simply to cover this one little-studied area of Hebrew phonology. I hope this work will also will bring out that Tiberian Hebrew is quite different, on both the (morpho)phonemic and phonetic levels, from how most scholars typically conceive it. It is a fixed, liturgical dialect that shows many queer phenomena, particularly in the ways its pronunciation was handed down from generation to generation. Although most Jews in the Near East were probably speaking a syllable-timed language by the end of first millennium (Arabic), the Tiberian liturgical dialect still clung for a time to the older, quality-based system. This is the form in which the traditional "Masoretic" text is handed down to us. This is how it is also transcribed by contemporary Arabic speakers (so Khan). We will have little hope of understanding the dialect that underlies this text until these facts are recognized and the corpus is removed from the late medieval "Qimhian" grammatical sphere in which virtually all modern textbooks and grammars have evolved. Richard Goerwitz