According to the Academy of the Hebrew Language, in the 1880s (the time of the beginning of the Zionist movement and the Hebrew revival) there were three groups of Hebrew regional accents: Ashkenazi (Eastern European), Sephardi (Southern European), and Mizrahi (Middle Eastern, Iranian, and North African). Over time features of these systems of pronunciation merged, and nowadays we find two main pronunciations of colloquial – not liturgical – Hebrew: Oriental and Non-Oriental. Oriental Hebrew displays traits of an Arabic substrate. Oriental speakers tend to use an alveolar trill [r] rather than a uvular trill [ʀ] or a velar fricative [ɣ], preserve the pharyngeal consonants /ħ/ and (less commonly) /ʕ/, preserve gemination, and pronounce /e/ in some places where non-Oriental speakers do not have a vowel (the shva na). Some Oriental speakers maintain a few of the emphatic consonants also found in Arabic, such as /sˤ/ for Biblical /tsʼ/.
Non-Oriental (and General Israeli) pronunciation lost the emphatic and pharyngeal sounds of Biblical Hebrew under the influence of Indo-European languages (Germanic and Slavic for Ashkenazim and Romance for Sephardim). The pharyngeals /ħ/ and /ʕ/ are preserved by older Oriental speakers. Dialectically, Georgian Jews pronounce /ʕ/ as [qʼ], while Western European Sephardim and Dutch Ashkenazim traditionally pronounce it [ŋ], a pronunciation that can also be found in the Italian tradition and, historically, in south-west Germany. However, according to Sephardic and Ashkenazic authorities, such as the Mishneh Berurah and the Shulchan Aruch and Mishneh Torah, /ʕ/ is the proper pronunciation. Thus, it is still pronounced as such by some Sephardim and Ashkenazim.
The classical pronunciation associated with the consonant ר rêš /r/ was a flap [ɾ], and was grammatically ungeminable. In most dialects of Hebrew among the Jewish diaspora, it remained a flap or a trill [r]. However, in some Ashkenazi dialects of northern Europe it was a uvular rhotic, either a trill [ʀ] or a fricative [ʁ]. This was because most native dialects of Yiddish were spoken that way, and the liturgical Hebrew of these speakers carried the Yiddish pronunciation. Some Iraqi Jews also pronounce rêš as a guttural [ʀ], reflecting Baghdad Jewish Arabic. An apparently unrelated uvular rhotic is believed to have appeared in the Tiberian vocalization of Hebrew, where it is believed to have coexisted with additional non-guttural articulations of /r/ depending on circumstances. In Modern Israeli Hebrew, the dominant pronunciation is a velar fricative, [ɣ], which sounds similar to [ʁ].
Though an Ashkenazi Jew in the Russian Empire, the Zionist Eliezer Ben-Yehuda based his Standard Hebrew on Sephardi Hebrew originally spoken in Spain and therefore recommended an alveolar [r]. The first waves of Jews to resettle in the Holy Land were northern Ashkenazi, and Standard Hebrew would come to be spoken with their preferred uvular articulation as found in Yiddish or modern standard German. Nearly all Jews of Israel now speak Hebrew with a uvular r because of its modern prestige and historical elite status.
Many Jewish immigrants to Israel spoke a variety of Arabic in their countries of origin, and pronounced the Hebrew rhotic consonant /r/ as an alveolar trill, identical to Arabic ر rāʾ, and which followed the conventions of old Hebrew. Under pressure to assimilate, many of them began pronouncing their Hebrew rhotic as a voiced uvular or velar fricative, often identical to Arabic غ ġayn [ɣ~ʁ]. However, in modern Ashkenazi, Sephardi, and Mizrahi poetry and folk music, as well as in the standard (or "standardised") Hebrew used in the Israeli media, an alveolar rhotic is sometimes used.
The Hebrew word for consonants is ‘itsurim (עיצורים). The following table lists the consonant phonemes of Israeli Hebrew in IPA transcription:1
The pharyngeal consonants /ħ/
are only pronounced by older Mizrahi speakers. Most speakers replace them with /x~χ/
The glottal consonants are not usually pronounced, as other than in careful or formal speech they are normally elided in unstressed syllables, and sometimes in stressed syllables as well.
Commonly transcribed /r/
. This is usually pronounced as a velar fricative [ɣ]
, sometimes as a uvular fricative or approximant [ʁ]
, and sometimes as a uvular or alveolar trill, depending on the background of the speaker.
The phonemes /w, tʃ, dʒ, ʒ/
were introduced through borrowings.
Obstruents assimilate in voicing. Voiceless obstruents (stops/affricates /p, t, ts, tʃ, k/ and fricatives /f, s, ʃ, x/) become voiced ([b, d, dz, dʒ, ɡ, v, z, ʒ, ɣ]) when they appear immediately before voiced obstruents, and vice versa. For example:לִסְגֹּר /lisɡoɣ/ > [lizɡoɣ] ('to close'), /s/ > [z]
זְכוּת /zχut/ > [sχut] ('a right'), /z/ > [s]
חֶשְׁבּוֹן /χeʃbon/ > [χeʒbon] ('a bill'), /ʃ/ > [ʒ]
מַדְפֶּסֶת /madpeset/ > [matpeset] ('a printer'), /d/ > [t]
אַבְטָחָה /avtaχa/ > [aftaχa] ('security'), /v/ > [f]
/n/ is pronounced [ŋ] before velar consonants.
Standard Israeli Hebrew (SIH) phonology has undergone a number of splits and mergers in its development from Biblical Hebrew (BH).BH /t/ and /tˤ/ merged into SIH /t/.
BH /ʕ/ and /ʔ/ generally merge into SIH /ʔ/, but the distinction is maintained in the educated speech of many Sephardim and some Ashkenazim.
BH /p/ had two allophones, [p] and [f], which split into separate phonemes /p/ and /f/ in SIH.
BH /b/ had two allophones, [b] and [v]. The [v] allophone merged with /w/ into SIH /v/. A new phoneme /w/ was introduced in loanwords (see Hebrew vav as consonant), so SIH has phonemic /b, v, w/.
BH /k/ had two allophones, [k] and [x]. The [k] allophone merged with /q/ into SIH /k/, while the [x] allophone merged with /ħ/ into SIH /x/, though a distinction between /x/ and /ħ/ is maintained in the educated speech of many Sephardim and some Ashkenazim.
The consonant pairs /b/~/v/, /k/~/x/, and /p/~/f/ were historically allophonic, as a consequence of a phenomenon of spirantisation known as begadkefat. In Modern Hebrew, the six sounds are phonemic. Similar allophonic alternation of BH [t]~[θ], [d]~[ð] and [ɡ]~[ɣ] was lost, with the allophones merging into simple /t, d, ɡ/, though /ɣ/ has reappeared from BH /r/ in the accent of the majority of the population.
These phonemic changes were partly due to the mergers noted above, to the loss of consonant gemination, which had distinguished stops from their fricative allophones in intervocalic position, and the introduction of syllable-initial /f/ and non-syllable-initial /p/ and /b/ in loan words. Spirantization still occurs in verbal and nominal derivation, but now the alternations b~v, k~x, and p~f are phonemic rather than allophonic.
The Hebrew word for vowels is tnu'ot (תְּנוּעוֹת). Hebrew has nine vowel phonemes, five short and four long:
Long vowels occur where two identical vowels were historically separated by a pharyngeal or glottal consonant, and the first was stressed. (Where the second was stressed, the result is a sequence of two short vowels.) They also often occur when morphology brings two identical vowels together, but they are not predictable in that environment.
Any of the five short vowels may be realized as a schwa [ə] when far from lexical stress.
There are two diphthongs, /aj/ and /ej/.
In Biblical Hebrew, each vowel had three forms: short, long and interrupted (chataf). However, there is no audible distinction between the three in Modern Hebrew, except that /e/ is often pronounced [ej] as in Ashkenazi Hebrew.
Vowel length in Modern Hebrew is environmentally determined and not phonemic, it tends to be affected by the degree of stress, and pretonic lengthening may also occur, mostly in open syllables. When a glottal is lost, a two-vowel sequence arises, and they may be merged into a single long vowel:תַּעֲבֹד /taʔavod/ ('you will work') > [taːvod]
שְׁעוֹנִים /ʃeʔonim/ ('watches') > becomes [ʃoːnim]
Modern pronunciation does not follow traditional use of the niqqud (diacritic) "shva". In Modern Hebrew, words written with a shva may be pronounced with either /e/ or without any vowel (or sometimes as an actual schwa), and this does not correspond well to how the word was pronounced historically. For example, the first shva in the word קִמַּטְתְּ 'you (fem.) crumpled' is pronounced /e/ (/kiˈmatet/) though historically it was silent, whereas the shva in זְמַן ('time'), which was pronounced historically, is usually silent ([zman]). Orthographic shva is generally pronounced /e/ in prefixes such as ve- ('and') and be- ('in'), or when following another shva in grammatical patterns, as in /tilmedi/ ('you [f. sg.] will learn'). An epenthetic /e/ appears when necessary to avoid violating a phonological constraint, such as between two consonants that are identical or differ only in voicing (e.g. /lamadeti/ 'I learned', not */lamadti/) or when an impermissible initial cluster would result (e.g. */rC-/ or */Cʔ-/, where C stands for any consonant).
Stress is phonemic in Modern Hebrew. There are two frequent patterns of lexical stress, on the last syllable (milrá מִלְּרַע) and on the penultimate syllable (mil‘él מִלְּעֵיל). Final stress has traditionally been more frequent, but in the colloquial language many words are shifting to penultimate stress. Contrary to the prescribed standard, some words exhibit stress on the antepenultimate syllable or even further back. This often occurs in loanwords, e.g. פּוֹלִיטִיקָה /poˈlitika/ ('politics'), and sometimes in native colloquial compounds, e.g. אֵיכְשֶׁהוּ /ˈexʃehu/ ('somehow'). Colloquial stress has often shifted from the last syllable to the penultimate, e.g. כּוֹבַע 'hat', normative /koˈvaʕ/, colloquial /ˈkovaʕ/; שׁוֹבָךְ ('dovecote'), normative /ʃoˈvax/, colloquial /ˈʃovax/. This shift is common in the colloquial pronunciation of many personal names, for example דָּוִד ('David'), normative /daˈvid/, colloquial /ˈdavid/.
Historically, stress was predictable, depending on syllable weight (that is, vowel length and whether a syllable ended with a consonant). Because spoken Israeli Hebrew has lost gemination (a common source of syllable-final consonants) as well as the original distinction between long and short vowels, but the position of the stress often remained where it had been, stress has become phonemic, as the following table illustrates. Phonetically, the following word pairs differ only in the location of the stress; orthographically they differ also in the written representation of vowel length of the vowels (assuming the vowels are even written):
When a vowel falls beyond two syllables from the main stress of a word or phrase, it may be reduced or elided. For example:זֹאת אוֹמֶרֶת
('that is to say')
?אֵיךְ קוֹרְאִים לְךָ
/eχ koʁʔim leχa/
('what is your name?')
When /l/ follows an unstressed vowel, it is elided, sometimes with the surrounding vowels:אַבָּא שֶׁלָּכֶם
הוּא יִתֵּן לְךָ
/hu jiten leχa/
('he will give you')
Syllables /rV/ drop before /x/ except at the end of a prosodic unit:בְּדֶרֶךְ כְּלָל
but: הוּא בַּדֶּרֶךְ [u badeʁeχ] ('he is on his way') at the end of a prosodic unit.
Sequences of dental stops reduce to a single consonant, again except at the end of a prosodic unit:אֲנִי לָמַדְתִּי פַּעַם
/ani lamadəti paʔam/
('I once studied')
but: שֶׁלָּמַדְתִּי [ʃelamadəti] ('that I studied')