ArticleConsonants and vowels in the Hebrew script
AuthorHelmut Richter
published / modified /
Article URL
containsentire article

Consonants and vowels in the Hebrew script

This is an introductory explanation how vowels are treated in the Hebrew writing system. It is intended for an audience that has no previous knowledge of the Hebrew language or script.

How biblical Hebrew is written

It is well-known that only consonants are written in the Hebrew script. This statement, however, is only partially true. While it is true that all characters of the Hebrew script are consonants, it is not so that a text in Hebrew contains as little information about the vowels as an English text from which all vowels were deleted. Basically, there are two features in the writing system that give additional information:


Now the whole world had one language and a common speech. (Gen.11:1, NIV)

In the original Hebrew language, but in Latin transcription (with “c” for “tz” as in “tzar” and “x” for “ch” as in “Loch Ness”; both in order not to render one Hebrew letter by more than one Latin letter), this sentence could for instance look like this:

wayehi kol haárec safa exat udvarim axadim

If we just left the vowels away, it would look like this:

wyh kl hrc sf xt wdvrm xdm

Actually, the Hebrew letters transliterated (with apostrophe for Alef) are these:

wyhy kl h'rc sfh 'xt wdvrym 'xdym

which is significantly more information about where which vowels are. In Hebrew script (which is written from right to left) it looks like this:

Gen.11:1 (unvocalised)

When one considers a script in which vowels are left away, at least sometimes, one would expect the most confusing effect when unrelated words are written alike, for instance in English: tan, ten, teen, tin, ton, tone, tun, tune. Already in English, cognates share more often their consonants than their vowels: in this example, “ten” and “teen”, “ton” and “tun”, “tone” and “tune” are pairs of cognates, but none of “ban”, “can”, “fan”, … , “tab”, “tad”, “tag”, … is a cognate of “tan”. In Hebrew, the same is true to a still larger extent: cognates are only recognised by their root, that is, by the consonant pattern which remains when all prefixes and suffixes are stripped off. Vowels come in to distinguish different words sharing the same root or different grammatical forms of one word; there is an example near the end of the article. This feature of the language may have contributed to treating consonants with higher priority than vowels in the script.

One should keep in mind that it is quite normal that not all features of a language are reflected in a writing system: in many languages, for instance in English, word stress and sentence stress are not marked, and most languages have much more phonemes than letters in their alphabet. It is sufficient when the script contains enough information to allow reading it without guessing. Experience throughout history has shown that the information contained in the Hebrew script is enough for readers knowing the language.

Additional aids for reading

Well, the last sentence has to be taken with a grain of salt. If it had been so, the script would not have developed as to include more information about vowels. In post-Biblical times it became customary to write the Yods and Waws for the vowels in more places than where they appear in the Bible, and consonantal Yod and Waw are often doubled to distinguish them from vowels. Although these changes are minor, they contribute largely to the readability. This is how modern Hebrew is usually written, and all Israeli newspaper readers demonstrate each morning that the information contained in this slightly enhanced script is absolutely sufficient for reading – again provided one knows the language.

The work of the Masoretes

At the end of the first millenium of the common era, after thousand years of Jewish diaspora, perfect mastery of the Hebrew language and profound knowledge of the Bible could no longer be expected from all members of the Jewish community who needed to read the Holy Scriptures. To aid them and to reduce the advent of multiple variant readings, a couple of Jewish scholars set out to standardise the texts and their reading. Part of this work was the annotation of the consonant text with marks for the vowels. The scheme shortly described in the previous section, albeit well introduced at that time, was not suitable for this purpose:

The project was called [MSWRH] (masora= tradition). This name emphasises that it understood itself as a record of how the Bible has always traditionally been read in the Jewish community and not as an introduction of new interpretations to a previously obscure text. It may well be that they encountered diverging traditions or opinions about how to read a certain passage – this is a normal process in text critics. But the goal was to fix what was the uninterrupted and living tradition.

Basically, their work consisted of the following tasks:

The text of the example above looks like this with the vocalisation marks:

Gen.11:1 (vocalised)

and with the additional cantillation marks:

Gen.11:1 (vocalised and accentuated)

The text as written (Ketiv) and the text as read (Qere)

In quite a number of cases, the problem arose that the written Bible text would make more sense if slightly modified. The tradition had then been not to “fix” the text but only to read it as if it were “fixed”. In other words, the reader is expected to read not exactly what is written. This is the same process which takes place when someone reads aloud from a book which contains spelling errors. In such cases, you have two forms for the word in question, usually differing only by one letter: one is the written form, Ketiv ( [KTYB] ) with consonants only, and the other is the form to be used when reading, Qere ( [QRY] ), with consonants and vowels. An obvious solution would be to include one of these forms in the running text and the other one in a footnote. However, it is done differently: the running text gets the Ketiv with the vowels of the Qere, and the footnote gets the consonants of the Qere. At first glance this looks odd, but it is useful:

As an example let us look at Jer.8:7. We read there in our Bibles:

Even the stork in the sky knows her appointed seasons, and the dove, the swift and the thrush observe the time of their migration. (Jer.8:7, NIV)

although the written text has [SW.S] (sus = horse) and not [SIYS] (sis = swift) among the animals that know the time for their return. The Masoretes decided that this is probably a copyist’s mistake so that the verse is to be read with the “swift” in it. Reflecting this correction, the Hebrew Bible contains now [SIWS] , i.e. the traditional Ketiv consonants with non-fitting vocalisation marks, equipped with a footnote telling the Qere consonants [SYS] .

Such cases where an existing word is replaced by another existing word with a different meaning are rather infrequent. The vast majority of corrections are for obvious mistakes such as letters missing altogether, or for illogical forms (e.g. Ex.22:4(5?): “If a man grazes her lifestock …” (instead of his). Very often a Yod replaces a Waw or vice versa, as these two frequent characters differ only in the length of the vertical stroke. In some cases the two versions are variants of the same word, and the correction was made even though both versions make sense, e.g. the first word of Ps.77:12(11?) [)ZKYR] (azkir = I’ll remind) was corrected to [)ZKWR] (ezkor = I’ll remember) which squares better with the context. (This is at the same time an example how cognates share their consonantal root, in this case [ZKR] .)

The frequency of such corrections varies a lot across the books of the Bible; the average is about one or two corrections per chapter. This number tells more about the reluctance of the Masoretes to modify the traditional consonant text (which had then already been standardised for about 850 years) than about the quality of text tradition: there are much more variants to the text across all manuscripts, but on the other hand many obvious errors could safely have been corrected without leaving a footnote. The actual process of text tradition is much more intricate, and is not the subject of this article which deals only with the writing system.

One case is a bit special: It is tradition not to pronounce the Divine Name in order to keep it holy. Instead, one of the words [):ADONFY] (adonai = my Lord) or [):ELOHIYM] (elohim = God) is to be read instead, the latter in the relatively few places where a form of [)FDOWN] (adon = lord, master) is adjacent to the Name: “the Lord our lord” would sound as weird in Hebrew as it sounds in English. Now, the way this is written in the Masoretic text is exactly the same as with other forms of Ketiv/Qere: the text contains the Ketiv, in this case the Tetragrammaton, with the vowels of the Qere. As this occurs so frequently that including a footnote each time would be very wasteful, this practice has led to confusion with non-Jewish readers who erroneously took the vowels of the Qere as belonging to the Ketiv, thus giving the Tetragrammaton a wrong vocalisation of “yehova” (or “yehovi”, resp.). The first of these has entered the Christian literature, mostly with an initial “J” (which is pronounced like a modern English “Y” anyway).

Note on the Latin transcription

The transcription of Hebrew words in this article is based on the modern Israeli pronunciation of Hebrew which is not demonstrably better or worse than any other tradition. The notation follows more or less the English pronunciation for the consonants; exceptions are explicitly noted. “k” and “q” are both pronounced like English “k”. The vowels, however, are not pronounced as in English, but roughly according to the International Phonetic Alphabet, or the Spanish or German language. This means: Depending on length, “a” is to be pronounced as in “but” or “father”, “e” as in “bet” or “bed”, “i” as in “fish” or “bee”, “o” is in “doll” or French “bistro”, and “u” as in “butcher” or “cool”. Most vowels are short, except “i” in the last syllable and all vowels in a word-final position. Stress is on the last syllable except when otherwise indicated by an acute accent.