Egyptian phonology

While the consonantal phonology of the Egyptian language may be reconstructed, its exact phonetics are unknown, and there are varying opinions on how to classify the individual phonemes. In addition, because Egyptian is also recorded over a full two millennia, the Archaic and Late stages being separated by the amount of time that separates Old Latin from modern Italian, it must be assumed that significant phonetic changes would have occurred over that time.

Phonologically, Egyptian contrasted labial, alveolar, palatal, velar, uvular, pharyngeal, and glottal consonants, in a distribution rather similar to that of Arabic. It also contrasted voiceless and emphatic consonants, as with other Afroasiatic languages, although exactly how the emphatic consonants were realized is not precisely known. Early research had assumed opposition in stops was one of voicing, but is now thought to either be one of tenuis and emphatic stops, as in many of the Semitic languages, or one of aspirated and ejective stops, as in many of the Cushitic languages.

Since vowels were not written, reconstructions of the Egyptian vowel system are much more uncertain, relying mainly on the evidence from Coptic and foreign transcriptions of Egyptian personal and place names. The vocalization of Egyptian is partially known, largely on the basis of reconstruction from Coptic, in which the vowels are written. Recordings of Egyptian words in other languages provide an additional source of evidence. Scribal errors provide evidence of changes in pronunciation over time. The actual pronunciations reconstructed by such means are used only by a few specialists in the language. For all other purposes the Egyptological pronunciation is used, which is, of course, artificial and often bears little resemblance to what is known of how Egyptian was spoken.

Consonants
The following consonant system is posited for Archaic (before 2600 BC) and Old Egyptian (2686–2181 BC), with IPA equivalents in square brackets where they differ from the usual transcription scheme:

* possibly unvoiced ejectives

The phoneme did not have an independent representation in the hieroglyphic orthography, and was frequently written with the sign for  or. The probable explanation is that the standard for written Egyptian was based on a dialect in which former had merged with other sonorants. was rare and also not indicated orthographically. The phoneme was written as $\langle\rangle$ in initial position ($\langle\rangle$ =  'father') and immediately after a stressed vowel ($\langle\rangle$ =  'bad'), as $\langle\rangle$ word-medially immediately before a stressed vowel ($\langle\rangle$ =  'you will appear'), and as null word-finally ($\langle\rangle$ =  'father').

In Middle Egyptian (2055–1650 BC), a number of consonantal shifts took place. By the beginning of the Middle Kingdom period, and  had merged, and the graphemes $\langle\rangle$ and $\langle\rangle$ were used interchangeably. In addition, had become  word-initially in an unstressed syllable (e.g. $\langle\rangle$  >  'color) and following a stressed vowel (e.g. $\langle\rangle$  >  '[the god] Apis').

In Late Egyptian (1069–700 BC), the following changes are present: the phonemes ' gradually merge with their counterparts ' ($⟨⟩$ > Akkadian transcription ti-ba-an 'dbn-weight'); ' often become, though they are retained in many lexemes; ' becomes ; and  become  at the end of a stressed syllable and eventually null word-finally (e.g. $⟨⟩$  > Akk. transcription -pi-ta 'bow').

More consonantal changes occurred in the first millennium BCE and the first centuries CE, leading to the Coptic language (1st–17th century AD). In Sahidic ' merged into ' (most often from ') and   (most often '). Bohairic and Akhmimic are more conservative, having also a velar fricative ( in Bohairic,  in Akhmimic). Pharyngeal * merged into glottal, after having affected the quality of surrounding vowels. is only indicated orthographically when following a stressed vowel, in which case it is marked by doubling the vowel letter (except in Bohairic), e.g. Akhmimic, Sahidic & Lycopolitan  ', Bohairic  ' 'to be' <  * 'has become'. The phoneme  probably was pronounced as a fricative, and became   after a stressed vowel in syllables which were closed in earlier Egyptian (compare  <  'gold' and  < * 'horn'). The phonemes are only found in Greek borrowings, with rare exceptions triggered by a proximate  (e.g.  <  'school').

Earlier * were preserved as ejective  in prevocalic position in Coptic. Despite the fact that these were written using the same graphemes as for the pulmonic stops ($\langle\rangle$), their existence may be inferred based on the following evidence: The stops $\langle\rangle$ were allophonically aspirated  before stressed vowels and sonorant consonants. In Bohairic these allophones were written with the special graphemes $\langle\rangle$, while other dialects did not mark aspiration, thus Sahidic vs. Bohairic  'the sun'. It then may be observed that Bohairic does not mark aspiration for reflexes of older *, e.g. Sahidic & Bohairic  'horn'. Similarly, the definite article is unaspirated when a word beginning with a glottal stop follows, e.g. Bohairic  'the account'.

The consonant system of Coptic is as follows: * various orthographic representations; see above

Vowels
The following is the vowel system posited for earlier Egyptian: Vowels were always short in unstressed syllables (e.g. $⟨⟩$ = * 'first'), long in open stressed syllables (e.g. $⟨⟩$ = * 'man'), and either short or long in closed stressed syllables (e.g. $⟨⟩$ = * 'we' vs. $⟨⟩$ = * 'to stay').

Late New Kingdom, after Ramses II i.e. c. 1200 BC: * > * (parallel to Canaanite vowel shift), e.g. $⟨⟩$ '(the god) Horus' * > * (Akkadian transcription: ). This provoked * > *, e.g. $⟨⟩$ 'tree' * > * (Akkadian transcription: ).

Early New Kingdom: short stressed * > *, e.g. $⟨⟩$ 'Menes' * > * (Akkadian transcription: ). Later, probably circa 1000–800 BC, short stressed * > *, e.g. $⟨⟩$ 'Tanis' * was borrowed into Hebrew as * but later transcribed as $⟨⟩$ in the Neo-Assyrian period.

Unstressed vowels, especially after the stress, became *, e.g. $⟨⟩$ 'good' * > * (Akkadian transcription ). * > * next to and, e.g. $⟨⟩$ 'soldier' * > * (earlier Akkadian transcription: , later: ).

In Sahidic and Bohairic Coptic, Late Egyptian stressed * becomes * and * becomes, while in the other dialects these are preserved, e.g. $⟨⟩$ * 'brother' > SB $\langle\rangle$, ALF $\langle\rangle$; $⟨⟩$ 'name' * > * > SB $\langle\rangle$, ALF $⟨⟩$. However, SB preserve * and Fayyumic renders it as $\langle\rangle$ in the presence of guttural fricatives, e.g. $⟨⟩$ '10000' * > SAL $\langle\rangle$, B $\langle\rangle$, F $\langle\rangle$. In Akhmimic and Lycopolitan, * becomes before etymological, e.g. $⟨⟩$ 'river' * >  > S $\langle\rangle$, B $\langle\rangle$, A $\langle\rangle$, F $\langle\rangle$. Similarly the diphthongs *, *, which normally have reflexes, in Sahidic and are preserved in other dialects, in Bohairic are written $\langle\rangle$ (in non-final position) and $⟨⟩$ respectively, e.g. "to me, to them" S $\langle\rangle$, AL $\langle\rangle$, F $\langle\rangle$, B $\langle\rangle$. Sahidic and Bohairic preserve * before (either etymological or from lenited  or tonic-syllable coda ), e.g. SB $\langle\rangle$  'to you (fem.)' < * < *. * may also have different reflexes before sonants, in proximity of sibilants, and in diphthongs.

Old */aː/ surfaces as /uː/ after nasals and occasionally other consonants, e.g. $\langlenṯr\rangle$ 'god' */ˈnaːcar/ > /ˈnuːte/ $\langlenoute\rangle$ /uː/ has acquired phonemic status, as evidenced by minimal pairs like 'to approach' $\langlehôn\rangle$ /hoːn/ < */ˈçaːnan/ ẖnn vs. 'inside' $⟨houn⟩$ /huːn/ < */ˈçaːnaw/ ẖnw. Etymological */uː/ > */eː/ often surfaces as /iː/ next to /r/ and after etymological pharyngeals, e.g. SL $\langlehir\rangle$ < */χuːr/ 'street' (Semitic loan).

Most Coptic dialects have two phonemic vowels in unstressed position. Unstressed vowels generally became /ə/, written as $\langlee\rangle$ or null ($\langlei\rangle$ in Bohairic and Fayyumic word-finally), but pretonic unstressed /a/ occurs as a reflex of earlier unstressed */e/ in proximity to an etymological pharyngeal, velar, or sonant (e.g. 'to become many' $\langleašai\rangle$ < ꜥšꜣ */ʕiˈʃiʀ/), or unstressed */a/. Pretonic [i] is underlyingly /əj/, e.g. S 'ibis' $\langlehibôi\rangle$ < h(j)bj.w */hijˈbaːj?w/.

Thus the following is the Sahidic vowel system c. 400 AD:

Phonotactics
Earlier Egyptian had syllable structure CV(C), where V was long in open, stressed syllables and short elsewhere. In addition, syllables of the type CV:C or CVCC could occur in word-final, stressed position. However CV:C only occurred in the infinitive of biconsonantal verbal roots, and CVCC only in some plurals. In later Egyptian stressed CV:C, CVCC, and CV became much more common because of the loss of final dentals and glides.

Stress
Earlier Egyptian: penultimate or ultimate. According to some scholars this is a development from a stage in proto-Egyptian where the antepenult could be stressed; this was lost as open posttonic syllables lost their vowels, e.g. * > * 'transformation'.

Egyptological pronunciation
As a convention, Egyptologists make use of an "Egyptological pronunciation" in English, in which the consonants are given fixed values and vowels are inserted in accordance with essentially arbitrary rules. Two consonants, alef and the ayin, are generally pronounced. The yodh is pronounced, and w. Between other consonants, is then inserted. Thus, for example, the Egyptian king whose name is most accurately transliterated as  is transcribed as "Ramesses", meaning "Ra has Fashioned (lit., "Borne") Him". In transcription, $\langle\rangle$, $\langle\rangle$, and $\langle\rangle$ all represent consonants; for example, the name Tutankhamen (1341–1323 BC) was written in Egyptian . Experts have assigned generic sounds to these values as a matter of convenience, but this artificial pronunciation should not be mistaken for how Egyptian was actually pronounced at any point in time. For example,  is conventionally pronounced  in English, but in his time was likely realized as something like *.