skip to content

Standard Tiberian Pronunciation in a Non-Standard Form: T-S AS 64.206

Estara J. Arrant


Scholars of the Hebrew Bible typically present, as part of a general summary, the three vocalisation systems present in the most well-known manuscripts: the Tiberian, Palestinian, and Babylonian vocalisation traditions.  However there also are manuscripts which have variations in the use of Tiberian vocalisation and diacritical signs; these are thought to reflect a variety of linguistic phenomena and perhaps a diversity of reading traditions. Scholarship has grouped these together into an entity which is now known as Non-Standard Tiberian (hereafter, NST). Some of the most recognised features include the unexpected use of dagesh in alef and word- and syllable-initial sonorants, the use of shewa in place of furtive pataḥ or under particular consonants to reinforce their pronunciation, and the extensive interchange of nearly every different vowel sign with each other (but most commonly pataḥ/qameṣ and ṣere/segol, as well as pataḥ /ḥatef pataḥ and shewa/ ḥatef pataḥ).[1] The most well-studied form of NST is generally thought to reflect a Palestinian phonological influence on Tiberian vocalisation (and thus it is called 'Palestino-Tiberian'), especially in that it is thought to show a reduced vocalic inventory (from 7 vowels to 5) that is typical of Hebrew Bible and Targum fragments with Palestinian vocalisation.[2]

However, the manuscript featuring as April 2019's Fragment of the Month is a striking example of the use of NST vocalisation patterns to represent, or even to hypercorrect, a generally Standard Tiberian (hereafter, ST) pronunciation of the Hebrew Bible text. I came upon this fragment whilst conducting my ongoing PhD research on NST and 'common' Torahs in the Cairo Genizah.[3] My Ph.D. thesis constructs a typology of these Bibles both on a codicological- and vocalisation-oriented basis. Through my innovative methodology of comprehensive data collection, statistical clustering, and wide-ranging comparison with contact texts and languages, I am reconstructing the scene of Common/NST scribal culture and the language contact situation of the 10th-12th centuries in the Cairo Genizah. One of my findings is that it seems the naqdan would often vocalise the text in unexpected ways: there is an attempt to honour the burgeoning Tiberian Masoretic system, while simultaneously a motivation to adhere to his own community's practices. It is often inaccurate to label these unexpected ways as 'mistakes': as this month's fragment shows, an NST text can be quite thoughtfully sophisticated.

Through my recent comprehensive study of all complete fragments of 2- and 3-column parchment Torah codices in Cambridge's Genizah collection,[4] I was able to place T-S AS 64.206 in a codicological typology: it belongs to the group of the smallest Italian-Byzantine Torah codices. However, despite their diminutive size (T-S AS 64.206's full single page is 13.4 x 12.69 cm), the documents in this group are sophisticated. The scribal hands are all formal. They all mark the Palestinian triennial reading cycle, and typically have full Masoretic notes (Masorah Magna and Parva). They all have an Italian-Byzantine style of script. What makes T-S AS 64.206 special is that it systematically utilises multiple NST features that are particularly characteristic of Italian-Byzantine Torahs to represent a more or less ST text.


T-S Misc.27.2.e recto

T-S AS.64.206


T-S AS 64.206 is a small, 2-column parchment Bible with an Italian-Byzantine scribal script style. It has a full Masorah and 20 lines. Its text is Deuteronomy 4:47-5:8, 5:10-18, 13:7-14:1, and 14:5-8. It is not pricked but is still ruled carefully and has wide margins. It marks the Triennial Seder reading of the Palestinian tradition. It is cantillated and has no immediately evident corrections. Therefore, the scribe was clearly skilled. My research has shown that NST exists to a significant degree in manuscripts which have a sophisticated codicology, and which are written by skilled scribes and naqdanim. This implies that NST was inherent to general biblical copying culture in the Genizah.

Codicology is important for contextualising a manuscript's textual features. I have found that NST manuscripts with similar codicologies also tend to have similar vocalisation and diacritic features. This manuscript exhibits features which are shared mainly by other small-medium sized Italian-Byzantine 2-column manuscripts. These features are not so prevalent in other groups of manuscripts that have a distinctly different codicology and/or come from a different region of script style.[5]

Textual Features

In my research I was able to group manuscripts together based on shared NST features. The group to which T-S AS 64.206 belongs is the 2-column "Orthoëpic" group. This group includes manuscripts which use Tiberian vocalisation signs in unexpected ways to represent or even strengthen a standard Tiberian pronunciation. This is a very small group, and T-S AS 64.206 is one of the most extreme in terms of these features.[6] Below are the highlights of its textual features:

Lexically Specific Standard Vocalisation

Throughout T-S AS 64.206, shewa always substitutes for ḥatef pataḥ, a feature which is common in NST documents and which indicates the phonetic equivalence of vocalic shewa with ḥatef pataḥ. However, T-S AS64.206 has one word which uses the ḥatef pataḥ, every single time, as expected: אֲשֶׁ֖ר. As this is the only word in the manuscript using ḥatef pataḥ, it is an artificial feature reflecting not a difference in pronunciation but a learned habit of the scribe.  

Vocalic Equivalence 

The evidence that pataḥ, vocalic shewa, and ḥatef pataḥ were all pronounced with the same phonetic quality is strong: for every definite article the naqdan substitutes shewa for a pataḥ: MS: הְשּׁבָּ֖֨ת for L/BHS: הַשּׁבָּ֖֨ת (the word occurs two times with the definite article); MS: כְּדָּבָ֥ר for L/BHS: כַּדָּבָ֥ר (Deuteronomy 13:12).[7] This is universally applied throughout the whole text. Coupled with a consistent interchange of shewa for ḥatef pataḥ (MS: עְבָדִֽ֑ים for L/BHS: עֲבָדִֽ֑ים (twice in the document); MS: לְאֹהְבַ֖י for L/BHS: לְאֹהֲבַ֖י (Deuteronomy 5:10)) the whole picture indicates a 3-way phonetic equivalence of these vowels.

However, what makes this manuscript orthoëpic in nature is that, because of the prolific use of shewa to represent a vowel, the scribe took extra care to ensure that there was no confusing silent with vocalic shewa. Since shewa often represents a vowel (pataḥ or ḥatef pataḥ) in this document, the naqdan would sometimes omit the silent shewa, presumably to avoid confusion and to preserve an ST syllable division. One example is: MS: תִתאַוֶּ֜ה; L/BHS: תִתְאַוֶּ֜ה (Deuteronomy 5:21).

In this example, given the scribe's proclivity to use shewa vocalically, he presumably left the shewa off of the taw in order to ensure it was not accidentally pronounced vocalically, which would have added another syllable. This was not universally applied: the scribe only left off the silent shewa on places that he felt were potentially problematic.

Shewa on Quiescent Letters 

The naqdan would place silent shewa on matres lectionis and quiescent letters (usually alef) either before or after the vowel in order to 'bind' the letter to its corresponding vowel: for example: MS: מְלַאְכְתֶּֽךָ֒ for L/BHS: מְלַאכְתֶּֽךָ֒ (Deuteronomy 5:13); MS: וַיֹּ֣אְמֶר for L/BHS: וַיֹּ֣אמֶר (occurs twice in the manuscript). This is an orthoëpic/hypercorrective feature to ensure that the quiescent alef is not treated as a consonant by the reader, and to preserve the correct syllable structure. The shewa here is not placed on the alef to mark it as a glottal stop (in NST typically a dagesh in the alef would function as such), but again due to the prolific use of shewa to represent a vowel, its placement here binds the alef to its corresponding vowel and reinforces its status as quiescent.

To summarise, we have a grouping of features in T-S AS 64.206 which together demonstrate that while the scribe used pataḥ, shewa, and ḥatef pataḥ in unique ways, he took measures to ensure that they did not lead to confusion with regards to the silent shewa and quiescent letters. Therefore, the text was mainly still pronounced according to the rigours of the ST tradition. This represents a tidy compromise between the scribe's individualistic use of particular signs and the accepted Tiberian pronunciation of the text. This unique document reminds us that non-standard placement of vowel signs does not always result in a non-standard reading of the text.

[1] NST has been given many different names troughout the history of scholarship, such as "Non-standard Tiberian", "Palestino-Tiberian", "Extended Tiberian", and more. For an overview, see Shai Heijmans, “Vocalization, Palestino-Tiberian”, in: Encyclopedia of Hebrew Language and Linguistics, Edited by: Geoffrey Khan. For a more comprehensive study of NST manuscripts, see Samuel Blapp, 'The Non-Standard Tiberian Hebrew Language Tradition according to Bible Manuscripts from the Cairo Genizah', PhD thesis, University of Cambridge, 2017. One aspect of my Ph.D. thesis is to build upon Blapp's work and to develop a typology of subtypes of NST.

[2] Steven E. Fassberg, A Grammar of the Palestinian Targum Fragments from the Cairo Genizah. Harvard Semitic Studies 38, 1990, 27.

[3] For an excellent overview of 'Common' Bibles, see Ben Outhwaite's Fragment of the Month, January 2019

[4] Estara Arrant, "An Exploratory Typology of 'Near-Model' Torah Codices from the Cairo Genizah with Standard and Non-Standard Tiberian Vocalisation," publication forthcoming.

[5] Ibid., forthcoming.

[6] I present more comprehensive details of this manuscript in my article (see f.n. 4).

[7] This concurs with the current scholarly understanding of the Tiberian pronunciation of these vowel signs. See Geoffrey Khan, “Shewa: Pre-Modern Hebrew”, in: Encyclopedia of Hebrew Language and Linguistics, 2013.


If you enjoyed this Fragment of the Month, you can find others here.

Contact us:

The manuscripts in this article are included in the Cambridge Digital Library. To see these and other items visit:

The zoomable images are produced using Cloud Zoom, a jQueryimage zoom plugin:
Cloud Zoom, Copyright (c) 2010, R Cecco,
Licensed under the MIT License