Phoenician language

Phoenician ( fə-NEE-shən) is an extinct Canaanite Semitic language originally spoken in the region surrounding the cities of Tyre and Sidon. Extensive Tyro-Sidonian trade and commercial dominance led to Phoenician becoming a lingua-franca of the maritime Mediterranean during the Iron Age. The Phoenician alphabet was spread to Greece during this period, where it became the source of all modern European scripts.

The area in which Phoenician was spoken includes Greater Syria and, at least as a prestige language, Anatolia, specifically the areas now including Lebanon, coastal Syria, coastal northern Palestine, parts of Cyprus and some adjacent areas of Turkey.[2] It was also spoken in the area of Phoenician colonization along the coasts of the southwestern Mediterranean Sea, including those of modern Tunisia, Morocco, Libya and Algeria as well as Malta, the west of Sicily, Sardinia, Corsica, the Balearic Islands and southernmost Spain.

In modern times, the language was first decoded by Jean-Jacques Barthélemy in 1758, who noted that the name "Phoenician" was first given to the language by Samuel Bochart.[3][4]

The Phoenicians were the first state-level society to make extensive use of the Semitic alphabet. The Phoenician alphabet is the oldest verified consonantal alphabet, or abjad.[5] It has become conventional to refer to the script as "Proto-Canaanite" until the mid-11th century BCE, when it is first attested on inscribed bronze arrowheads, and as "Phoenician" only after 1050 BCE.[6] The Phoenician phonetic alphabet is generally believed to be at least the partial ancestor of almost all modern alphabets.

The most important Phoenician trade routes and cities in the Mediterranean Basin

From a traditional linguistic perspective, Phoenician was composed of a variety of dialects.[7][8] According to some sources, Phoenician developed into distinct Tyro-Sidonian and Byblian dialects. By this account, the Tyro-Sidonian dialect, from which the Punic language eventually emerged, spread across the Mediterranean through trade and colonization, whereas the ancient dialect of Byblos, known from a corpus of only a few dozen extant inscriptions, played no expansionary role.[9] However, the very slight differences in language and the insufficient records of the time make it unclear whether Phoenician formed a separate and united dialect or was merely a superficially defined part of a broader language continuum. Through their maritime trade, the Phoenicians spread the use of the alphabet to Northwest Africa and Europe, where it was adopted by the Greeks. Later, the Etruscans adopted a modified version for their own use, which, in turn, was modified and adopted by the Romans and became the Latin alphabet.[10]

Punic colonisation spread Phoenician to the western Mediterranean, where the distinct Punic language developed. Punic also died out, but it seems to have survived far longer than Phoenician, until the 6th century, perhaps even into the 9th century CE.[11]

Phoenician was written with the Phoenician script, an abjad (consonantary) originating from the Proto-Canaanite alphabet that also became the basis for the Greek alphabet and, via an Etruscan adaptation, the Latin alphabet. The Punic form of the script gradually developed somewhat different and more cursive letter shapes; in the 3rd century BCE, it also began to exhibit a tendency to mark the presence of vowels, especially final vowels, with an aleph or sometimes an ayin. Furthermore, around the time of the Second Punic War, an even more cursive form began to develop,[12] which gave rise to a variety referred to as Neo-Punic and existed alongside the more conservative form and became predominant some time after the destruction of Carthage (c. 149 BCE).[13] Neo-Punic, in turn, tended to designate vowels with matres lectionis ("consonantal letters") more frequently than the previous systems had and also began to systematically use different letters for different vowels,[13] in the way explained in more detail below. Finally, a number of late inscriptions from what is now Constantine, Algeria dated to the first century BCE make use of the Greek alphabet to write Punic, and many inscriptions from Tripolitania, in the third and fourth centuries CE use the Latin alphabet for that purpose.[14]

In Phoenician writing, unlike that of abjads such as those of Aramaic, Biblical Hebrew and Arabic, even long vowels remained generally unexpressed, regardless of their origin (even if they originated from diphthongs, as in bt /beːt/ 'house'; Hebrew spelling has byt). Eventually, Punic writers began to implement systems of marking of vowels by means of matres lectionis. In the 3rd century BCE appeared the practice of using final 'ālep Phoenician aleph.svg to mark the presence of any final vowel and, occasionally, of yōd Phoenician yodh.svg to mark a final long [iː].

Later, mostly after the destruction of Carthage in the so-called "Neo-Punic" inscriptions, that was supplemented by a system in which wāw Phoenician waw.svg denoted [u], yōd Phoenician yodh.svg denoted [i], 'ālep Phoenician aleph.svg denoted [e] and [o], ʿayin Phoenician ayin.svg denoted [a][15] and Phoenician he.svg and ḥēt Phoenician heth.png could also be used to signify [a].[16] This latter system was used first with foreign words and was then extended to many native words as well.

A third practice reported in the literature is the use of the consonantal letters for vowels in the same way as had occurred in the original adaptation of the Phoenician alphabet to Greek and Latin, which was apparently still transparent to Punic writers: hē Phoenician he.svg for [e] and 'ālep Phoenician aleph.svg for [a].[17]

Later, Punic inscriptions began to be written in the Latin alphabet, which also indicated the vowels. Those later inscriptions, in addition with some inscriptions in Greek letters and transcriptions of Phoenician names into other languages, represent the main source of knowledge about Phoenician vowels.

The following table presents the consonant phonemes of the Phoenician language as represented in the Phoenician alphabet, alongside their standard Semiticist transliteration and reconstructed phonetic values in the International Phonetic Alphabet.[18][19]:

The system reflected in the abjad above is the product of several mergers. From Proto-Northwest Semitic to Canaanite, and have merged into , and *z have merged into *z, and *θʼ, *ɬʼ and *s’ have merged into (*s’) *śʿ. Next, from Canaanite to Phoenician, the sibilants and were merged as () *s, *ḫ and *ḥ were merged as () h, and *ʻ and *ġ were merged as *ʻ.[20][19][citation needed] These latter developments also occurred in Biblical Hebrew at one point or another.

The original value of the Proto-Semitic sibilants, and accordingly of their Phoenician counterparts, is disputed. Recent scholarship argues that š was [s], s was [ts], z was [dz], and was [tsʼ],[21] against the traditional sound values of [ʃ], [s], [z], and [sˤ] as reflected in the transcription.[22]

On the other hand, it is debated whether šīn Phoenician sin.svg and sāmek Phoenician samekh.svg, which are mostly well distinguished by the Phoenician orthography, also eventually merged at some point, either in Classical Phoenician or in Late Punic.[23] Krahmalkov suggests that *z may have been [dz] or even [zd] based on Latin transcriptions such as esde for the demonstrative 𐤅z.[19]

In later Punic, the laryngeals and pharyngeals seem to have been entirely lost. Neither these nor the emphatics could be adequately represented by the Latin alphabet, but there is also evidence to that effect from Punic script transcriptions.

There is no consensus on whether Phoenician-Punic ever underwent the lenition of stop consonants that happened in most other Northwest Semitic languages such as Biblical Hebrew and Aramaic (cf. Hackett[20] vs Segert[24] and Lyavdansky).[25] The consonant /p/ may have been generally transformed into /f/ in Punic and in late Phoenician, as it was in Proto-Arabic.[25] Certainly, Latin-script renditions of late Punic include many spirantized transcriptions with ph, th and kh in various positions (although the interpretation of these spellings is not entirely clear) as well as the letter f for the original *p.[26] However, in Neo-Punic, *b lenited to v contiguous to a following consonant, as in the Latin transcription lifnim for 𐤋𐤁𐤍𐤌‎ *lbnm "for his son".[19]

Knowledge of the vowel system is very imperfect because of the characteristics of the writing system. During most of its existence, Phoenician writing showed no vowels at all, and even as vowel notation systems did eventually arise late in its history, they never came to be applied consistently to native vocabulary. It is thought that Phoenician had the short vowels /a/, /i/, /u/ and the long vowels /aː/, /iː/, /uː/, /eː/, /oː/.[20][27] The Proto-Semitic diphthongs /aj/ and /aw/ are realized as /eː/ and /oː/. That must have happened earlier than in Biblical Hebrew since the resultant long vowels are not marked with the semivowel letters (bēt "house" was written 𐤁𐤕bt, in contrast to Biblical Hebrew ביתbyt).

The most conspicuous vocalic development in Phoenician is the so-called Canaanite shift, shared by Biblical Hebrew, but going further in Phoenician. The Proto-Northwest Semitic /aː/ and /aw/ became not merely /oː/ as in Tiberian Hebrew, but /uː/. Stressed Proto-Semitic /a/ became Tiberian Hebrew /ɔː/ (/aː/ in other traditions), but Phoenician /oː/. The shift is proved by Latin and Greek transcriptions like rūs for "head, cape" (Tiberian Hebrew rōš, ראש‎), samō for "he heard" (Tiberian Hebrew šāmāʻ, שמע‎); similarly the word for "eternity" is known from Greek transcriptions to have been ʻūlōm, corresponding to Biblical Hebrew ʻōlām and Proto-Semitic ʻālam. The letter Y used for words such as ys "which" and yth (definite accusative marker) in Greek and Latin alphabet inscriptions can be interpreted as denoting a reduced schwa vowel[17] that occurred in pre-stress syllables in verbs and two syllables before stress in nouns and adjectives,[28] while other instances of Y as in chyl and even chil for /kull/ "all" in Poenulus can be interpreted as a further stage in the vowel shift resulting in fronting ([y]) and even subsequent delabialization of /u/ and /uː/.[28][29] Short /*i/ in originally-open syllables was lowered to [e] and was also lengthened if it was accented.[28]

Stress-dependent vowel changes indicate that stress was probably mostly final, as in Biblical Hebrew.[30] Long vowels probably occurred only in open syllables.[31]

As is typical for the Semitic languages, Phoenician words are usually built around consonantal roots and vowel changes are used extensively to express morphological distinctions. However, unlike most Semitic languages, Phoenician preserved numerous uniconsonantal and biconsonantal roots inherited from Proto-Afro-Asiatic: compare the verbs kn "to be" vs Arabic kwn, mt "to die" vs Hebrew and Arabic mwt and sr "to remove" vs Hebrew srr.[32]

Nouns are marked for gender (masculine and feminine), number (singular, plural and vestiges of the dual) and state (absolute and construct, the latter being nouns that are followed by their possessors) and also have the category definiteness. There is some evidence for remains of the Proto-Semitic genitive grammatical case as well. While many of the endings coalesce in the standard orthography, inscriptions in the Latin and Greek alphabet permit the reconstruction of the noun endings, which are also the adjective endings, as follows:[33]

In late Punic, the final /-t/ of the feminine was apparently dropped: 𐤇𐤌𐤋𐤊𐤕ḥmlkt "son of the queen" or 𐤀𐤇𐤌𐤋𐤊𐤕ʼḥmlkt "brother of the queen" rendered in Latin as HIMILCO.[29][34] /n/ was also assimilated to following consonants: e.g. št "year" for earlier */sant/.[29]

The case endings in general must have been lost between the 9th century BCE and the 7th century BCE: the personal name rendered in Akkadian as ma-ti-nu-ba-ʼa-li "Gift of Baal", with the case endings -u and -i, was written ma-ta-an-baʼa-al two centuries later. However, evidence has been found for a retention of the genitive case in the form of the first-singular possessive suffix: ʼby /ʼabiya/ "of my father" vs ʼb /ʼabī/ "my father". If true, this may suggest that cases were still distinguished to some degree in other forms as well.

The written forms and the reconstructed pronunciations of the personal pronouns are as follows:[36]

1st: /ʼanōkī/ ʼnk (Punic sometimes ʼnky), also attested as /ʼanek/
2nd masc. /ʼatta(ː)/ ʼt
2nd fem. /ʼatti(ː)/ ʼt
3rd masc. /huʼa/ , also [hy] (?) hy and /huʼat/ hʼt
3rd fem. /hiʼa/

1st: /ʼanaḥnū/ ʼnḥn
2nd masc. /ʾattim/ ʼtm
2nd fem. unattested, perhaps /ʾattim/ ʼtm
3rd masc. /himūt/ hmt
3rd fem. /himūt/ hmt

Enclitic personal pronouns were added to nouns (to encode possession) and to prepositions, as shown below for "Standard Phoenician" (the predominant dialect, as distinct from the Byblian and the late Punic varieties). They appear in a slightly different form depending on whether or not they follow plural-form masculine nouns (and so are added after a vowel). The former is given in brackets with the abbreviation a.V.

1st: // , also y (a.V. /-ayy/ y)
2nd masc. /-ka(ː)/ k
2nd fem. /-ki(ː)/ k
3rd masc. /-oː/ , Punic ʼ, (a.V. /-ēyu(ː)/ y)
3rd fem. /-aː/ , Punic ʼ (a.V. /-ēya(ː)/ y)

1st: /-on/ n
2nd masc. /-kum/ km
2nd fem. unattested, perhaps /-kin/ kn
3rd masc. /-om/ m (a.V. /-nom/ nm)
3rd fem. /-am/ m (a.V. /-nam/ nm)

In addition, according to some research, the same written forms of the enclitics that are attested after vowels are also found after a singular noun in what must have been the genitive case (which ended in /-i/, whereas the plural version ended in /-ē/). Their pronunciation can then be reconstructed somewhat differently: first-person singular /-iya(ː)/ y, third-person singular masculine and feminine /-iyu(ː)/ y and /-iya(ː)/ y. The third-person plural singular and feminine must have pronounced the same in both cases, i.e. /-nōm/ nm and /-nēm/ nm.

These enclitic forms vary between the dialects. In the archaic Byblian dialect, the third person forms are h and w // for the masculine singular (a.V. w /-ēw/), h /-aha(ː)/ for the feminine singular and hm /-hum(ma)/ for the masculine plural. In late Punic, the 3rd masculine singular is usually /-im/ m.

The same enclitic pronouns are also attached to verbs to denote direct objects. In that function, some of them have slightly divergent forms: first singular /-nī/ n and probably first plural /-nu(ː)/.

The near demonstrative pronouns ("this") are written, in standard Phoenician, z for the singular and ʼl for the plural. Cypriot Phoenician displays ʼz instead of z. Byblian still distinguishes, in the singular, a masculine zn / z from a feminine zt / . There are also many variations in Punic, including st and zt for both genders in the singular. The far demonstrative pronouns ("that") are identical to the independent third-person pronouns. The interrogative pronouns are /miya/ or perhaps /mi/ my "who" and /mū/ m "what". An indefinite pronoun "anything" is written mnm. The relative pronoun is a š, either followed or preceded by a vowel.

The definite article was /ha-/, and the first consonant of the following word was doubled. It was written h but in late Punic also ʼ and ʻ because of the weakening and coalescence of the gutturals. Much as in Biblical Hebrew, the initial consonant of the article is dropped after the prepositions b-, l- and k; it could also be lost after various other particles and function words, such the direct object marker ʼyt and the conjunction w- "and".

Of the cardinal numerals from 1 to 10, 1 is an adjective, 2 is formally a noun in the dual and the rest are nouns in the singular. They all distinguish gender: ʼḥd, šnm/ʼšnm[37] (construct state šn/ʼšn), šlš, ʼrbʻ, ḥmš, šš, šbʻ, šmn(h), tšʻ, ʻsr/ʻšr[38][39] vs ʼḥt, unattested, šlšt, ʼrbʻt, ḥmšt, ššt, šbʻt, šmnt,[40] unattested, ʻšrt.[41] The tens are morphologically masculine plurals of the ones: ʻsrm/ʻšrm,[39][42] šlšm, ʼrbʻm, ḥmšm, ššm, šbʻm, šmnm, tšʻm. "One hundred" is mʼt, two hundred is its dual form mʼtm, whereas the rest are formed as in šlš mʼt (three hundred). One thousand is ʼlp. Ordinal numerals are formed by the addition of *iy -y.[43] Composite numerals are formed with w- "and", e.g. ʻsr w šnm for "twelve".

The verb inflects for person, number, gender, tense and mood. Like for other Semitic languages, Phoenician verbs have different "verbal patterns" or "stems", expressing manner of action, level of transitivity and voice. The perfect or suffix-conjugation, which expresses the past tense, is exemplified below with the root q-t-l "to kill" (a "neutral", G-stem).[45][36]

The imperfect or prefix-conjugation, which expresses the present and future tense (and which is not distinguishable from the descendant of the Proto-Semitic jussive expressing wishes), is exemplified below, again with the root q-t-l.

The imperative endings were presumably /-∅/, /-ī/ and /-ū/[46] for the second-person singular masculine, second-person singular feminine and second-person plural masculine respectively, but all three forms surface in the orthography as /qutul/ qtl: -∅. The old Semitic jussive, which originally differed slightly from the prefix conjugation, is no longer possible to separate from it in Phoenician with the present data.

The non-finite forms are the infinitive construct, the infinitive absolute and the active and passive participles. In the G-stem, the infinitive construct is usually combined with the preposition l- "to", as in /liqtul/ "to kill"; in contrast, the infinitive absolute (qatōl)[47] is mostly used to strengthen the meaning of a subsequent finite verb with the same root: ptḥ tptḥ "you will indeed open!",[46] accordingly /*qatōl tiqtul/ "you will indeed kill!".

The missing forms above can be inferred from the correspondences between the Proto-Northwest Semitic ancestral forms and the attested Phoenician counterparts: the PNWS participle forms are *.

/qātil-, qātilīma, qātil(a)t, qātilāt, qatūl, qatūlīm, qatult or qatūlat, qatūlāt/

Most of the stems apparently also had passive and reflexive counterparts, the former differing through vowels, the latter also through the infix -t-. The G stem passive is attested as qytl, /qytal/ < */qutal/.;[46] t-stems can be reconstructed as /yitqatil/ ytqtl (tG) and /yiqtattil/ (Dt) yqttl.[50]

Some prepositions are always prefixed to nouns, deleting, if present, the initial /h/ of the definite article: such are b- "in", l- "to, for", k- "as" and m- /min/ "from". They are sometimes found in forms extended through the addition of -n or -t. Other prepositions are not like that:ʻl "upon", .ʻd "until", ʼḥr "after", tḥt "under", b(y)n "between". New prepositions are formed with nouns: lpn "in front of", from l- "to" and pn "face". There is a special preposited marker of a definite object ʼyt (/ʼiyyūt/?), which, unlike Hebrew, is clearly distinct from the preposition ʼt (/ʼitt/).

The most common negative marker is bl (/bal/), negating verbs but sometimes also nouns; another one is ʼy (/ʼī/), expressing both nonexistence and the negation of verbs. Negative commands or prohibitions are expressed with ʼl (/ʼal/). "Lest" is lm. Some common conjunctions are w (originally perhaps /wa-?/, but certainly /u-/ in Late Punic), "and" ʼm (/ʼim/), "when", and k (/kī/), "that; because; when". There was also a conjunction (ʼ)p (/ʼap/"also". l- (/lū, li/) could (rarely) be used to introduce desiderative constructions ("may he do X!"). l- could also introduce vocatives. Both prepositions and conjunctions could form compounds.[51]

The basic word order is verb-subject-object. There is no verb "to be" in the present tense; in clauses that would have used a copula, the subject may come before the predicate. Nouns precede their modifiers, such as adjectives and possessors.

Most nouns are formed by a combination of consonantal roots and vocalic patterns, but they can formed also with prefixes (/m-/, expressing actions or their results, and rarely /t-/) and suffixes /-ūn/. Abstracts can be formed with the suffix -t (probably /-īt/, /-ūt/).[48] Adjectives can be formed following the familiar Semitic nisba suffix /-īy/ y (e.g. ṣdny "Sidonian").

Like the grammar, the vocabulary is very close to Biblical Hebrew, but some peculiarities attract attention. For example, the copula verb "to be" is kn (as in Arabic, as opposed to Hebrew and Aramaic hyh) and the verb "to do" is pʿl (as in Aramaic pʿl and Arabic fʿl, as opposed to Hebrew ʿśh, though in Hebrew pʿl has the similar meaning "to act").

The significantly divergent later form of the language that was spoken in the Tyrian Phoenician colony of Carthage is known as Punic and remained in use there for considerably longer than Phoenician did in Phoenicia itself by arguably surviving into Augustine of Hippo's time. Throughout its existence, Punic co-existed with the Berber languages, which were then native to Tunisia (including Carthage) and North Africa. Punic disappeared some time after the destruction of Carthage by the Romans and the Berbers. It is possible that Punic may have survived the Muslim conquest of the Maghreb in some small isolated area: the geographer al-Bakri describes a people speaking a language that was not Berber, Latin or Coptic in the city of Sirte in rural Ifriqiya, a region in which spoken Punic survived well past its written use.[55] However, it is likely that arabization of the Punics was facilitated by their language belonging to the same group (both being Semitic languages) as that of the conquerors and thus having many grammatical and lexical similarities. Most Punic speakers may have been linguistically Berberized and/or Latinized after the fall of Carthage.

The ancient Libyco-Berber alphabet that is still in irregular use by modern Berber groups such as the Tuareg is known by the native name Tifinagh, possibly a derived form of a cognate of the name "Punic".[citation needed] Still, a direct derivation from the Phoenician-Punic script is debated and far from established since the two writing systems are very different. As far as language (not the script) is concerned, some borrowings from Punic appear in modern Berber dialects: one interesting example is agadir "wall" from Punic gader.

Perhaps the most interesting case of Punic influence is that of the name of Hispania (the Iberian Peninsula, comprising Portugal and Spain), which, according to one of the theories, is derived from the Punic I-Shaphan meaning "coast of hyraxes", in turn a misidentification on the part of Phoenician explorers of its numerous rabbits as hyraxes.[citation needed] Another case is the name of a tribe of hostile "hairy people" that Hanno the Navigator found in the Gulf of Guinea. The name given to those people by Hanno the Navigator's interpreters was transmitted from Punic into Greek as gorillai and was applied in 1847 by Thomas S. Savage to the western gorilla.

Phoenician, together with Punic, is primarily known from approximately 10,000 surviving inscriptions,[56] supplemented by occasional glosses in books written in other languages. In addition to their many inscriptions, the Phoenicians are believed to have left numerous other types of written sources, but most have not survived.

The Phoenician alphabetic script was easy to write on papyrus or parchment sheets, and the use of these materials explains why virtually no Phoenician writings – no history, no trading records – have come down to us. In their cities by the sea, the air and soil were damp, and papyrus and leather moldered and rotted away. Thus disappeared the literature of the people who taught a large portion of the earth's population to write. The only written documents of Phoenicians and Carthaginians are monumental inscriptions on stone, a few ephemeral letters or notes on pieces of broken pottery, and three fragmentary papyri. Thus, no Tyrian primary sources dating from Hiram I's time are available.[57]

Roman authors, such as Sallust, allude to some books written in the Punic language, but none have survived except occasionally in translation (e.g., Mago's treatise) or in snippets (e.g., in Plautus' plays). The Cippi of Melqart, a bilingual inscription in Ancient Greek and Carthaginian discovered in Malta in 1694, was the key which allowed French scholar Jean-Jacques Barthélemy to decipher and reconstruct the alphabet in 1758.[58] Even as late as 1837 only 70 Phoenician inscriptions were known to scholars. These were compiled in Wilhelm Gesenius's Scripturae linguaeque Phoeniciae monumenta, which comprised all that was known of Phoenician by scholars at that time.

Basically, its core consists of the comprehensive edition, or re-edition of 70 Phoenician and some more non-Phoenician inscriptions... However, just to note the advances made in the nineteenth century, it is noteworthy that Gesenius’ precursor Hamaker, in his Miscellanea Phoenicia of 1828, had only 13 inscriptions at his disposal. On the other hand only 30 years later the amount of Phoenician inscribed monuments had grown so enormously that Schröder in his compendium of 1869 could state that Gesenius knew only a quarter of the material Schröder had at hand himself.[59]

Die phönizische Sprache. Entwurf einer Grammatik nebst Sprach- und Schriftproben

Since bilingual tablets with inscriptions in both Etruscan and Phoenician dating from around 500 BCE were found in 1964, more Etruscan has been deciphered through comparison to the more fully understood Phoenician.