Varieties of Arabic - Variation - Phonetics

Phonetics

• CA /ʔ/ is mostly lost.
  • Depending on the exact phonetic environment, this either caused reduction of two vowels into a single long vowel or diphthong (when between two vowels), insertion of a homorganic glide /j/ or /w/ (when between two vowels, the first of which was short or long /i/ or /u/ and the second not the same), lengthening of a preceding short vowel (between a short vowel and a following non-vowel), or simple deletion (elsewhere). This resulted initially in a large number of complicated morphophonemic variations in verb paradigms.
  • In CA and Modern Standard Arabic (MSA), /ʔ/ is still pronounced.
  • However, because this change had already happened in Meccan Arabic at the time the Qur'an was written, it is reflected in the orthography of written Arabic, where a diacritic known as hamzah is inserted either above an ʾalif, wāw or yāʾ, or "on the line" (between characters); or in certain cases, a diacritic ʾalif maddah ("lengthened ʾalif") is inserted over an ʾalif. (As a result, proper spelling of words involving /ʔ/ is probably one of the most difficult issues in Arabic orthography. Furthermore, actual usage is inconsistent in many circumstances.)
  • Modern dialects have smoothed out the morphophonemic variations, typically by deleting the associated verbs or moving them into another paradigm (for example, /qaraʔ/ "read" becomes /qara/ or /ʔara/, a third-weak verb).
  • /ʔ/ has reappeared medially in various words due to borrowing from CA. (In addition, /q/ has become in many dialects, although the two are marginally distinguishable in Egyptian Arabic, since words beginning with original /ʔ/ can elide this sound, whereas words beginning with original /q/ cannot.)
• CA /q/ changes widely from variety to variety. In Bedouin dialects from Mauritania to Saudi Arabia, it is pronounced, as in most of Iraq. In the Levant and Egypt (except in Upper Egypt (the Sa'id), as well as some North African towns such as Tlemcen, it is pronounced as a glottal stop, apart from rural areas in the South West Levant where it becomes emphatic . In the Persian Gulf, it becomes in many words (adjacent to an original /i/), and is otherwise. Elsewhere, it is usually realized as uvular .
• CA varies widely. In some Arabian Bedouin dialects, and parts of the Sudan, it is still realized as the medieval Persian linguist Sibawayh described it, as a palatalized . In Egypt, parts of Yemen and parts of Oman, it is a plain . In most of the Levant and most of North Africa, apart from north Algeria, it is . In the Persian Gulf and southern Iraq, it often becomes . Elsewhere, it is usually .
• CA /k/ often becomes in the Persian Gulf, Iraq, some Rural Palestinian dialects and in some Bedouin dialects (adjacent to an original /i/, particularly in the second singular feminine enclitic pronoun, where replaces an classical /ik/ or /ki/). In a very few Moroccan varieties, it affricates to /k͡ʃ/. Elsewhere, it remains .
• CA /r/ is pronounced in a few areas: Mosul, for instance, and the Jewish variety in Algiers. In all northern Africa, a phonemic distinction has emerged between plain and emphatic, thanks to the merging of short vowels.
• CA /θ/, /ð/ become /t, d/ in Egypt and some regions in North Africa (including Malta), and become /s, z/ in the Levant (except for some words, in which they become /t, d/), but remain /θ/ and /ð/ in Iraqi, Yemenite, Tunisian, rural Palestinian, Eastern Libyan, and some rural Algerian dialects. In Arabic-speaking towns of Eastern Turkey, (Urfa, Siirt and Mardin) they respectively become /f, v/.
• CA /t/ (but not emphatic CA /tˤ/) is affricated to in Moroccan Arabic; this is still distinguishable from the sequence .
• CA /ʕ/) is pronounced in Iraqi Arabic and Kuwaiti Arabic with glottal closure: . In some varieties /ʕ/ is devoiced to before /h/, for some speakers of Cairene Arabic /bitaʕha/ → /bitaħħa/ (or /bitaʕ̞ħa/) "hers". The residue of this rule applies also in the Maltese language, where neither etymological /h/ nor /ʕ/ are pronounced as such, but give in this context: tagħha "hers".
• The nature of "emphasis" differs somewhat from variety to variety. It is usually described as a concomitant pharyngealization, but in most sedentary varieties it is actually velarization, or a combination of the two. (The phonetic effects of the two are only minimally different from each other.) Usually there is some associated lip rounding; in addition, the stop consonants /t/ and /d/ are dental and lightly aspirated when non-emphatic, but alveolar and completely unaspirated when emphatic.
• CA short vowels /a/, /i/ and /u/ suffer various changes.
  • Original final short vowels are mostly deleted.
  • Many Levantine Arabic dialects merge /i/ and /u/ into a phonemic /ə/ except when directly followed by a single consonant; this sound may appear allophonically as /i/ or /u/ in certain phonetic environments.
  • Maghreb dialects merge /a/ and /i/ into /ə/, which is deleted when unstressed. Tunisian maintains this distinction, but deletes these vowels in non-final open syllables.
  • Moroccan Arabic, under the strong influence of Berber, goes even further. Short /u/ is converted to labialization of an adjacent velar, or is merged with /ə/. This schwa then deletes everywhere except in certain words ending /-CCəC/.
    • The result is that there is no more distinction between short and long vowels; borrowings from CA have "long" vowels (now pronounced half-long) uniformly substituted for original short and long vowels.
    • This also results in consonant clusters of great length, which are (more or less) syllabified according to a sonority hierarchy. (For some subdialects, in practice, it is very difficult to tell where, if anywhere, there are syllabic peaks in long consonant clusters in a phrase such as /xsˤsˤk tktbi/ "you (fem.) must write". Other dialects, in the North, make a clear distinction; they say /xəssək təktəb/ "you want to write", but */xəssk ətkətb/ just won't do).
  • In Egyptian Arabic and Levantine Arabic, short /i/ and /u/ are elided in various circumstances in unstressed syllables (typically, in open syllables; for example, in Egyptian Arabic, this occurs only in the middle vowel of a VCVCV sequence, ignoring word boundaries). In Levantine, however, clusters of three consonants are almost never permitted. If such a cluster would occur, it is broken up through the insertion of /ə/ – between the second and third consonants in Egyptian Arabic, and between the first and second in Levantine Arabic.

• CA long vowels are shortened in some circumstances.
  • Original final long vowels are shortened in all dialects.
  • In Egyptian Arabic and Levantine Arabic, unstressed long vowels are shortened.
  • Egyptian Arabic also cannot tolerate long vowels followed by two consonants, and shorten them. (Such an occurrence was rare in CA, but often occurs in modern dialects as a result of elision of a short vowel.)

• In most dialects, particularly sedentary ones, CA /a/ and /aː/ have two strongly divergent allophones, depending on the phonetic context.
  • Adjacent to an emphatic consonant and to /q/ (but not usually to other sounds derived from this, such as /ɡ/ or /ʔ/), a back variant occurs; elsewhere, a strongly fronted variant ~ is used.
  • There is a tendency for emphatic consonants to cause non-adjacent low vowels to be backed, as well; this is known as emphasis spreading. The domain of emphasis spreading is potentially unbounded; in Egyptian Arabic, the entire word is usually affected, although in Levantine Arabic and some other varieties, it is blocked by an /i/ or /j/ (and sometimes /ʃ/).
  • The two allophones are in the process of splitting phonemically in some dialects, as occurs in some words (particularly foreign borrowings) even in the absence of any emphatic consonants anywhere in the word. (Some linguists have postulated additional emphatic phonemes in an attempt to handle these circumstances; in the extreme case, this requires assuming that every phoneme occurs doubled, in emphatic and non-emphatic varieties. Some have attempted to make the vowel allophones autonomous and eliminate the emphatic consonants as phonemes. Others have asserted that emphasis is actually a property of syllables or whole words rather than of individual vowels or consonants. None of these proposals seems particularly tenable, however, given the variable and unpredictable nature of emphasis spreading.)
  • CA /r/ is also in the process of splitting into emphatic and non-emphatic varieties, with the former causing emphasis spreading, just like other emphatic consonants. Originally, non-emphatic occurred before /i/ or between /i/ and a following consonant, while emphatic occurred mostly near .
    • To a large extent, Western Arabic dialects reflect this, while the situation is rather more complicated in Egyptian Arabic. (The allophonic distribution still exists to a large extent, although not in any predictable fashion; nor is one or the other variety used consistently in different words derived from the same root. Furthermore, although derivational suffixes (in particular, relational /-i/ and /-ijja/) affect a preceding /r/ in the expected fashion, inflectional suffixes do not.)
    • In Moroccan Arabic, short /a/ and /i/ have merged, obscuring the original distribution. In this dialect, the two varieties have completely split into separate phonemes, with one or the other used consistently across all words derived from a particular root except in a few situations.
  • In Moroccan Arabic, the allophonic effect of emphatic consonants is more pronounced than elsewhere.
    • Full /a/ is affected as above, but /i/ and /u/ are also affected, and are lowered to and, respectively.
    • In some varieties, such as in Marrakesh, the effects are even more extreme (and complex), where both high-mid and low-mid allophones exist ( and, and ), in addition to front-rounded allophones of original /u/ (, ), all depending on adjacent phonemes.
    • On the other hand, emphasis spreading in Moroccan Arabic is less pronounced than elsewhere; usually it only spreads to the nearest full vowel on either side, although with some additional complications.
  • Emphasis spreading also pharyngealizes consonants between the source consonant and affected vowels, although the effects are much less noticeable than for vowels, since the rise of emphasis spreading is associated with a concomitant decrease in the amount of pharyngealization of emphatic consonants.
    • Interestingly, emphasis spreading does not affect the affrication of non-emphatic /t/ in Moroccan Arabic, with the result that these two phonemes are always distinguishable regardless of the nearby presence of other emphatic phonemes.
  • Certain other consonants, depending on the dialect, also cause backing of adjacent sounds, although the effect is typically weaker than full emphasis spreading and usually has no effect on more distant vowels.
    • The /x/ and the uvular consonant /q/ often cause partial backing of adjacent /a/ (and lowering of /u/ and /i/ in Moroccan Arabic). For Moroccan Arabic, the effect is sometimes described as half as powerful as an emphatic consonant, as a vowel with uvular consonants on both sides is affected similarly to having an emphatic consonant on one side.
    • Interestingly, the pharyngeal consonants /ħ/ and /ʕ/ cause no emphasis spreading and may have little or no effect on adjacent vowels. In Egyptian Arabic, for example, an /a/ adjacent to either sound is a fully front . In other dialects, /ʕ/ is more likely to have an effect than /ħ/.
    • In some Gulf Arabic dialects, /w/ and/or /l/ causes backing.
    • In some dialects, words such as الله /aɫɫaː/ Allāh has backed 's and in some dialects also velarized /l/.
• CA diphthongs /aj/ and /aw/ have become and (but merge with original /iː/ and /uː/ in Maghreb dialects, which is probably a secondary development). The diphthongs are maintained in the Maltese language and some urban Tunisian dialects, particularly that of Sfax, while and also occur in some other Tunisian dialects, such as Monastir.
• The placement of the stress accent is extremely variable between varieties; nowhere is it phonemic.
  • Most commonly, it falls on the last syllable containing a long vowel, or a short vowel followed by two consonants; but never farther from the end than the third-to-last syllable. This maintains the presumed stress pattern in CA (although there is some disagreement over whether stress could move farther back than the third-to-last syllable), and is also used in Modern Standard Arabic (MSA).
    • In CA and MSA, stress cannot occur on a final long vowel; however, this does not result in different stress patterns on any words, because CA final long vowels are shortened in all modern dialects, and any current final long vowels are secondary developments from words containing a long vowel followed by a consonant.
  • In Egyptian Arabic, the rule is similar, but stress falls on the second-to-last syllable in words of the form ...VCCVCV, as in /makˈtaba/.
  • In Maghrebi Arabic, stress is final in words of the (original) form CaCaC, after which the first /a/ is elided. Hence جَبَل ǧabal "mountain" becomes .
  • In Moroccan Arabic, phonetic stress is often not recognizable.