OT Manuscripts

  • Old Testament Manuscripts
  • Some definitions: Witnesses, Versions, Manuscripts etc.
    • Qumran near the Dead Sea and other Judean caves
    • Most Hebrew manuscripts are based on the Masoretic text
    • Scribal notes in the Masoretic Margin
    • Geniza Manuscripts found in Cairo
    • Other Hebrew Manuscripts
    • Ancient Versions of the Old Testament
      • “Samaritan”
      • “Greek”
      • “Aramaic”
      • “Syriac”
      • “Latin”
  • Details that are available elsewhere
      1. Other ancient translations
      2. Individual manuscripts
      3. Support for the Standard text
  • The value of non-Standard Texts

Old Testament Manuscripts

by Dr. David Instone-Brewer

The books of the Old Testament were written over a period of many hundreds of years by a wide range of authors. The original text, had no vowels, like modern Hebrew. This did not cause any problems while it was a living language, but later scribes added vowels to remind readers how to pronounce it.

Although there are many hundreds of Hebrew manuscripts, the vast majority of these are virtually identical to each other, thanks to a standardization process which was started probably before the first century AD.  The text produced by this process is called the Standard text, and sometimes the Proto-Masoretic text because it was the forerunner of the medieval Masoretic text on which modern printed Bibles are based.

The only significant difference between the Standard text and the medieval manuscripts lies in the way that vowels were written. They were omitted or inserted as occasional letters in the Standard text, but in the medieval manuscripts they are represented as pointing - a full and complex system of dots and dashes under and over the letters. Modern Hebrew is normally written without vowels because they are unnecessary for understanding the meaning, though this produces occasional ambiguity, and most scholars assume that ancient Hebrew was also written without vowels. The vowels are therefore not part of the original text but were added to aid oral reading and to remove ambiguities.

Manuscripts which are older than the first century AD show much more variation. Translations of the Old Testament into other languages mirror these differences. This has often been regarded as an indication that these older manuscripts are closer to an older form of the Old Testament text than the Standard text, but recent studies suggest that the text chosen for the standardization process actually represent the oldest and most accurate form available, so it is closest to the original.

Some definitions: Witnesses, Versions, Manuscripts etc.

*  A “manuscript” (Ms[s]) is an individual physical piece of writing which has survived (e.g., 1QIsaa, Leningrad Codex);
* A "text" is the writing which is on one or many manuscripts;
* A “version” is an ancient translation of a text (e.g., Septuagint, Vulgate);
* A "reading" or "variant" is the text in a particular manuscript or group of manuscripts which differs from the majority;
* A "witness" is a version or a manuscript which agrees with a reading.
* An "edition" is a published text created from the text in one or more manuscripts; There are many published editions of the OT - see more here.
* A "critical edition" includes an "apparatus" which summarizes "variants" or "variant readings" (ie variations in the text) found in different manuscripts;
* An "eclectic edition" is compiled from multiple differing manuscripts by choosing the most likely original readings.

The following deals with groups of manuscripts. Some are grouped either by where they were found (Qumran, Geniza etc) and some are grouped by the version they represent (Vulgate, Masoretic etc).

Qumran near the Dead Sea and other Judean caves 

The scripture manuscripts preserved by the community at Qumran  (also known as the Dead Sea Scrolls), which was destroyed by the Romans in AD 68. They are the earliest Biblical manuscripts we have, dating from the second century BC to the early first century AD. Other fragments of Hebrew Bibles have been found at Masada and a few other sites in the Judean desert. The Jews at Masada were rebelling against the Romans, and barricaded themselves in this abandoned fortress. Their community was destroyed shortly after that of Qumran, during the Roman campaign which included the destruction of Jerusalem and the Temple in AD 70. This means that all the documents preserved at Qumran and Masada are definitely from the first century, and some are much older. The other sites where Hebrew Bible fragments have been found were destroyed in the early second century.

By the first century Jews were 'correcting' Hebrew Bibles towards a Standard Text. Many manuscripts at Qumran demonstrate this kind of correction, though many also preserve unchanged readings. The fragments of manuscripts found at Qumran and these other sites therefore tell us a great deal about the Hebrew Bibles used by Jews before the first century. There is a profound contrast between the manuscripts from these different sites. At Qumran there is a wide variety of Hebrew manuscripts, though a quarter of them are identical to the Standard text, and another quarter of them agree largely with the Standard text.[1] In contrast, all the Hebrew Bible manuscripts found outside Qumran agree with the Standard text. This Standard text is essentially identical to the Masoretic text.[2]

Most Hebrew manuscripts are based on the Masoretic text

Jewish scribes called the Masoretes in the 10th century were keen to preserve the Standard text, and the traditions about how to read it out loud. They used systems of ‘pointing’ or marks above, below and inside letters to indicate vowels and punctuation. The Masoretic pointing system became standard, and often preserves the most ancient ways of pronouncing words and proper names. They and earlier scribes also added marginal notes about rare words or apparent mis-spellings. These act as warnings to scribes who copy the manuscript, so that they are not tempted to ‘correct’ them. These notes even record the total number of letters and other checksums which help to make sure it was transmitted letter-perfect. We do not know when they started using such techniques, but their similarity to the Standard texts found at Qumran and elsewhere demonstrate their success in preventing changes to the text.

The most important manuscript is the Leningrad Codex, named after the city where it resides (now called St Petersburg). It was the final and best manuscript completed by the Ben Asher family of Masoretic scribes in 1008. Other Masoretic manuscripts include the Aleppo Codex (an exemplary copy of almost equal authority to the Leningrad codex) and the Petersburg Codex of the Prophets (the best example of Babylonian vowel pointing). These and other medieval manuscripts offer little additional information for the study of the ancient text because they all share an essentially identical consonantal text which mirrors the Standard text of the first century BC.

Scribal notes in the Masoretic Margin

As well as comments about rare words and the numbers of letters, the margins of many Masoretic manuscripts contain notes about potential differences in the text itself. The scribes considered the text too sacred to be altered, even when their notes indicate that they clearly thought the text contained an error.

These notes include about 350 ‘suppositions’ (sebirin) which highlight places where an editor might think the text is faulty (but it isn't) and about a few ‘scribal omissions’ (itturé sopherim) where a letter or word was known to be missing.

The scribes also recorded traditions in other documents about the eighteen ‘scribal emendations’ (tiqquné sopherim) where the text was known to have been changed in the ancient past to avoid anthropomorphisms and similar misunderstandings about God. For example, in Genesis 18:22 the original text read that “God stood before Abraham”, but in ancient idiom this could imply that “God served Abraham”, so it was changed to “Abraham stood before the Lord”.

The “Qere” readings form the largest group of notes about textual variants, totaling about 1300 (there are different numbers in different manuscripts) . These indicate the accepted way to pronounce the text (the qeré) when it was different to the way it was actually written in the body of the text (the Ketiv). Many of these preserve not only a different pronunciation, but a different meaning which was regarded as too important to lose when the official text was established (more details here).

Geniza Manuscripts found in Cairo

Bible manuscripts make up a large proportion of the thousands of fragments discovered in the 1880’s in a Cairo synagogue geniza. A geniza is a repository in a synagogue for manuscripts which are old or defective, but because they contain God’s name, they must be buried honorably rather than destroyed. This particular geniza was neglected for centuries and then walled in and forgotten, so an accumulation of manuscripts from the 5th to 10th centuries was accidentally preserved.

These Bibles often reflect traditions of the Karaite Jewish community who rejected many rabbinic innovations, so they preserved some manuscripts which had not been conformed to the Standard text and later to the Masoretic text.

Other Hebrew Manuscripts

Many other Hebrew Bibles have survived, many of which preserved texts which did not conform to the Standard text and were perhaps used for private devotions rather than public worship.

Another valuable source of non-Standard manuscripts is among the 1600 manuscripts collected in the 1800’s by Firkowitsch from Karaite synagogues. The age of these manuscripts is often difficult to ascertain especially as Firkowitsch has been suspected of forgery, and may have put false dates on them to justify the huge sale price to the Imperial Library of St Petersberg.

Ancient Versions of the Old Testament referred to in "Manuscripts and Meanings"


The Samaritan Pentateuch is a Hebrew text written in an alphabet derived from an archaic Hebrew script. Although the manuscripts are all relatively late (13th C and later), this version preserves some very ancient variants, as confirmed by the fact that many have been found in Qumran fragments. It has popularizing features such as modernizing of archaic phrases and simplification of sentences, which suggests that it dates back to at least a couple of centuries BC when Hebrew was still understood and read by ordinary Jews and, presumably, by Samaritans.

This can be compared in some ways to the Tyndale Bible – a venerable and exact older version which is difficult to read (particularly when the spelling is not modernized) but it is valuable for preserving some historical interpretations.


This refers to both the Old Greek and the Septuagint. The term “Old Greek” refers to the  original translation of the Hebrew text into Greek which has not survived, and “Septuagint” (abbreviated as “LXX”)  refers to the various revisions of it which have survived. This translation was made for Greek-speaking Jews in Egypt in the third to first CBC. It was named after the seventy two translators who, according to the ancient author Aristeas, wondrously agreed unanimously on their translation of the Books of Moses. The rest of the Old Testament was probably translated later, though it was all completed by the first century BC. It was later revised by Jews (especiallyAquila) and by Christians (most importantly Lucian and Origen) to conform more closely to the Standard text.

Most scholars[3] believe that all Septuagint versions were revisions of the Old Greek, except perhaps for Aquila’s revision which may be a fresh translation based on the Hebrew Standard Text. For this reason, all the Greek versions can be grouped together as witnesses of a single Hebrew text – though the actual Hebrew text which was used for different portions of the Old Testament may not be from the same manuscript.

Usually the translation is very literal, even to the extent of following the order of Hebrew words which results in very unusual Greek. However, when the original Hebrew is obscure or difficult it may use a paraphrase, and occasionally contemporary vocabulary is substituted for ancient terminology. For example, the warning against cult prostitutes in Deuteronomy 23:18 was translated as a warning against the Mystery religions which were the equivalent temptation at the time of these translators. Therefore, when it differs from the Hebrew text, this may be due to the translators, but it may also be due to differences in the Hebrew manuscript which the translators used. This is occasionally confirmed by Hebrew fragments found at Qumran which reflect Septuagint readings. The Septuagint therefore represents a very important source of variants which may represent Hebrew manuscripts which are mostly lost.

This version can be compared in some ways to the Good News translation – a scholarly phrase-by-phrase translation which nevertheless tries to use a consistent English word when the same original word occurs. Sometimes it changes the text slightly in order to make the meaning clear for modern readers.


The Targums which were Aramaic translations given orally after the Scripture readings in synagogues because Hebrew had ceased to be a living language. These translations were not written down till the late first or early second century and revised during several centuries. All the Targums are paraphrases to varying degrees except for Targum Onkelos which was an official translation written in the 2nd C and designed to conform literally to the Standard text. Other Targums often preserve variants from non-Standard manuscripts which are occasionally mirrored in Qumran Hebrew fragments.

This version can be compared in some ways to the original Living Bible – a paraphrase which was not afraid to add phrases or even sentences in order to help the reader understand the text.


The Syriac Peshitta was translated for the many Syrians who followed Judaism after the royal family of Adiabene converted in the mid first century. During the following two centuries the whole OT was completed and the NT was translated for Christians. Unfortunately this translation was often based on Targums rather than on Hebrew manuscripts so it helps to confirm the early date of variants in the Targums, but it is an uncertain source of variants in lost Hebrew manuscripts.

This version can be compared in some ways to the New Living Bible – it is based on the original Living Bible (like the Syriac is often based on the Targums) but it is much closer to the original text and is often a fresh translation.


This refers to both the Old Latin translations and Jerome's Vulgate which replaced them. The Old Latin versions were translated from the Septuagint and even though Jerome translated from the Standard Hebrew text, he often included readings from the Septuagint because many Christians regarded this as inspired.  Latin translations therefore help to confirm the early date of variants in the Septuagint, but they are of little value for establishing the Hebrew text.

This version can be compared in some ways to the New International Version – a translation which aims to produce a word-for-word translation but is willing to change the word order and add words in order to make the meaning clear, and tries to preserve older traditional translations in famous passages.

Details that are available elsewhere

A scholarly text of the Hebrew Old Testament such as the BHS[4] contains many more details about individual manuscripts, but these details are of relatively minor help in ascertaining the original text.  The details omitted from STEPBible are:

1) Other ancient translations

These include the Coptic, Georgian, Ethopic Armenian and Arabic. These are all based wholly or partially on the Greek Septuagint as well as on Standard texts. This means that they do not give us any information about any Hebrew manuscripts which may have perished or help to confirm variants from other sources. This is also largely true of the Latin and Syriac translations, but they are too important to ignore.

2) Individual manuscripts

Each manuscript the ancient versions and translations can have subtle or substantial differences. There is a great deal of variety between different manuscripts of the Greek Septuagint and the Aramaic Targums. Copiers and later editors of these translations recognized that they were often paraphrases of the original Hebrew, and felt free to ‘correct’ these translations or make them more consistent. The overall effect of these changes was almost always to make them conform closer to the Standard text. Consequently, if a manuscript of these versions agrees with the Standard Text, it is very unsafe to regard this as evidence in support of that reading in the original manuscript it was translated from. However, when any manuscript deviates from the Standard Text it is important to note this as it may be a witness to a reading in the original Hebrew manuscript which the translator had, and which is now lost.

This means it is very important to note variants within the manuscripts or versions of the Septuagint, Targums, Peshitta or Latin translations, but it is much less important to note exactly which manuscripts contain them and which do not.  It is sufficient merely to note the fact that one or more manuscripts of a particular version are different from the Standard text.

If a variant is recorded as having support from the Greek, Latin, Aramaic, Samaritan or Syriac versions, this means that one or more manuscripts in that tradition has this variant. It does not mean that all do, or that the majority do. Even when all the manuscripts except for one supports the Standard or Masoretic text, a single manuscript may preserve the original translation from the Hebrew text that was available at the time, against all the others which have been corrected to the Standard text which replaced that earlier Old Testament text.

3) Support for the Standard text

It would be unhelpful to list all the support for the traditional Hebrew text, because there is too much and because this evidence is not very useful. Editors and copyists tended to ‘correct’ ancient translations towards the Standard text, so it is impossible to know whether agreements with the Standard text represent the original ancient version. Hebrew manuscripts were similarly corrected towards the Standard text. Such corrections are already seen in Qumran manuscripts, so this process had started at least by the first century. This means that the agreement of a text or translation with the Standard text cannot be taken as an ancient verification of that text. This also means that when one versions of an ancient translation disagrees with the Standard text while another agrees with it, we can normally assume that the disagreement represents the earlier version. However, even when all surviving manuscripts of a particular translation agree with the Standard text, this cannot be cited as reliable evidence of support for that text, because they may all have been corrected towards the Standard text.

Therefore, as in most works on textual criticism of the Old Testament, STEPBible rarely cites refers to versions support  the Standard text, because this is unsafe as evidence in support of that version.

The value of non-Standard Texts

Occasionally the Standard text contains an error which has been faithfully copied by generations of scribes. As mentioned above, sometimes the scribal notes in the margin indicate that they know it is an error but they regard the text as too sacred to change.[5] On other occasions the only clue to the original text lies in the ancient versions and non-Standard manuscripts. When two otherwise unrelated ancient versions agree with each other against the Standard text, it is possible that they represent an older version of the text.

For example, in Gen. 10.2 the Masoretic text names a son of Japheth as “Meshek”, but the ancient Greek and Samaritan versions both agree that this person should be pronounced “Moshek”. The consonants for both versions are identical, so this does not represent a variant in the Standard text, but only in the vowel pointing which later scribes added to the text.

Another example in Gen.10.4 is an instance where the Standard text probably was corrupt. The Masoretic text says that one group descended from Javan are the “Dodanites” while the Greek and Samaritan versions both agree that this should read “Rodanites”, which also agrees with 1 Chronicles 1.7. The difference between the opening letters is very slight – the letter R is ר while the letter D is ד so it is an easy mistake for a scribe to make. The scribe also had a reason to believe that “Dodanites” was correct because these people were made famous by Homer – he recorded that their city was founded by Zeus and was destroyed in the Trojan wars (see e.g. Illiad 20.210-215) – whereas the Rodanites were a completely unknown group. Therefore it is not surprising that the Aramaic Targums and Syriac translation keep the reading “Dodanites”, though it is likely that the original text said “Rodanites”.

These examples concerned very minor matters, and this is typical of the type of corruptions which have been found in the Standard text. Variant readings marked as “Poss.” (i.e. 'possible') or “Prob.” (i.e. 'probable') to indicate the likelihood that they represent the original text. These designations are based on the considered opinion of a group of scholars assembled by the United Bible Society.[6] In the majority of cases the scholars decided that the Standard text contained the oldest wording.

The Standard text, represented by the Masoretic text, is therefore the form which almost always represents the oldest editing. But the non-Standard versions are important for an occasional corrective to the Standard text and for telling us how the text was interpreted in ancient times.

[1]  Emanuel Tov, in Scribal practices and approaches reflected in the texts found in the Judean desert (Studies on the texts of the desert of Judah (STDJ), 54; Leiden: Brill, 2004) pp. 332-335, analyses 124 Qurman Scripture manuscripts which are large enough have a recognizable text type. Of these, 63 follow the Standard text (though 20 of these have some Samaritan features and 13 of these have some Septuagint features). Of the 15 non-Qumran texts which are large enough to be analyzed, all of them follow the Standard text.

[2]  The only differences are the use of plene spelling – i.e. they insert some letters as vowels where later manuscript use pointing for vowels.

[3] This was convincingly argued by Paul de Lagarde against the rival theory of Paul Kahle that each version represents a fresh translations from the Hebrew text.

[4] K. Elliger. Biblia Hebraica Stuttgartensia (Stuttgart: Württembergische Bibelanstalt, 1968).

[5] For example, in Isa.9.6 a final mem occurs in the middle of the first word. The correction in the marginal Qere shows that the scribes were aware of this, and wished to continue copying it in this form.

[6] Dominique Barthélemy et al., Preliminary and interim report on the Hebrew Old Testament Text Project: Compte rendu préliminaire et provisoire sur le travail d´analyse textuelle de l´Ancien Testament hébreu, (London: United Bible Societies, 1970). Edited by a UBS committee, including D. Barthélemy, A. R. Hulst,N. Lohfink, W. D. McHardy, H. P. Rüger and J. A. Sanders, with A. Schenker and J. A. Thompson as secretaries to the committee.