User:Florian Blaschke/Romance languages

Logudorese Sardinian with its pentatonic vowel system /i ɛ a ɔ u/ is difficult to derive from the Proto-Romance and Classical Latin system. Rather, I suspect that its oldest layer descends directly from Old Latin of the 2nd century BC.

The following consideration is significant: Between about 250 and 150 BC, Old Latin had three distinct long front vowel phonemes: /iː/, /eː/ (from the Proto-Italic diphthong *ei), and /ɛː/. By around 150 BC, /eː/ had merged into /iː/. The result was the following system of late Old Latin:

Short vowels: /i ɛ a ɔ u/ (plus the sonus medius [y], which may not have been a phoneme)

Long vowels: /iː ɛː aː ɔː uː/

Diphthongs: /ae oe au ui eu ei/

The short vowels /i/ and /u/ may have already had a tendency towards [ɪ] and [ʊ].

Occasional spellings indicate that in the 2nd century BC, in some Old Latin dialects, especially rural dialects outside the city of Rome, the diphthong /ae/ was already monophthongised, with the result being /ɛː/ or perhaps /æː/, but these spellings are sporadic and Classical Latin retained a diphthong /ae/.

It is unclear to what extent popular spoken Latin of the first century BC reflected the diphthongal pronunciation or the innovative monophthongal one, which has plausibly suspected to have been promoted by interference of Umbrian, or Sabellic in general. From the Pompejan inscriptions, the monophthong was already common by the (late) first century AD.

As the reflex of /ae/ merges with that of short /ɛ/ and not long /eː/ in most of Romance, the diphthongisation must have prevailed completely in spoken Latin (early Romance) only relatively late – it is said to have been carried through fully only in late antiquity.

Decisively, in Classical Latin of the first century BC we find that not only have short /i/ and /u/ been lowered to what is most likely /ɪ/ and /ʊ/; moreover, /ɛː/ has been raised to /eː/ (taking the place of the Old Latin */eː/ that merged into /iː/) and symmetrically, /ɔː/ has been raised to /oː/.

As such, the pronunciation of Classical Latin vowels was very similar to that of German.

Linguistically, there are said to be clear indications (apart from the vowel system), especially lexically, that Sardinian had already started to diverge by the first century BC.

That means that the Classical Latin system can only be said to underlie the mainstream of Romance, but not all of Romance.

It is not inconceivable that the dialect underlying Sardinian already featured the monophthongisation /ae/ > /ɛː/ by the 2nd/1st century BC.

If so, the result merged directly with the long vowel /ɛː/ before merging into short /ɛ/ as a result of the loss of vowel lengths.

The same system is also found in southern Basilicata around Senise, as well as in the far south of Corsica, except that instead of merging with /a/, ancient /au/ behaved differently and apparently became /ɔ/ there – in southern Corsica it even remains distinct from /o/ (from Old Latin /ɔ/ and /ɔː/).

Moreover, African Romance apparently featured the same system.

As the pentatonic system of Sardinian is identical to that of Modern Greek, and almost identical to that of Koine Greek at the time of Classical Latin (it had a sixth vowel /y/ additionally), it is possible that the simple loss of quantity is connected with a Greek substratum, but a Punic substratum could also be the cause, or even both in concert.

As for the Italo-Western development (the symmetrical Quantitätenkollaps traditionally attributed to "Vulgar Latin" as a whole, even though even Romanian has a distinct system), Sabellic influence has been suggested as a cause.

By late antiquity, and arguably already in the first century BC, various dialects in Sardinia, Corsica and southern Italy were noteworthy for their conservativity and apparently direct descent from Old Latin.

As for the Sicilian vowel system, it looks like the result of the Italo-Western system overlaying a Sardinian-like substratum (just like in the Senise region); that means that it appears to be the last direct remnant of the substratum in question) with a pentatonic vowel system (as the result is a pentatonic system too, except that its origin is different, in that all of the high and mid-high vowels have merged into /i/ and /u/). Although the pentatonic system could also be due to an Italo-Greek sub- or adstratum, there are other traits Sicilian has in common with Sardinian, and African Romance, which was apparently similar to Sardinian, isn't far off, either.

There are, in fact, texts from the 13th century which have been adduced as evidence for an intermediate Sardinian-like pentatonic system underlying Sicilian (apparently as a substratum, as just posited).

The details are not altogether clear, but it appears that in late antiquity, conservative Sardinian-like dialects – apparently directly descended from Old Latin – were spoken all over southern Italy, overlaid by a mainstream Romance layer directly descended from Classical Latin.

When Dacia began to be colonised in the 2nd century AD, likely at least at a significant part of the colonists came from southern Italy. Either the "compromise" vowel system typical of Romanian had already formed in southern Italy, or it formed in the Balkans as a result of the encounter of colonists speaking dialects with different vowel systems.

Perhaps quantity loss had already happened in the conservative dialects at the time, while possibly still being absent from mainstream dialects.

(Septimius Severus is said to have retained his African accent until old age, implying that African Romance already began to be distinct in the 2nd century, presumably along with Sardinian.)