In my recent blogpost on the alif al-wiqāyah I argued that <wʔ> denotes final vowels /ū/ and /ō/ (or /aw/) in the orthography of the QCT. While final <w> denoted actual consonantal /w/. While it is fairly clear that this is the rule that is operating in the QCT orthography, I at the time did not have a good explanation how such an orthographic device arose, after all, if clarity was really the aim, you'd think there were more pressing matters in Arabic orthography than distinguishing /w/ from /ū, ō/, and it'd be unclear why the scribes did not use a similar orthography device to distinguish /y/ from /ī, ē/. yet there's no alif al-wiqāyah on the <y>.
But one of the readers of this blog, Jadhimah, provided a fascinating solution in a reply to that blogpost. I would like to work out his suggestion, provide some of my own insights and see if we can see some evidence for the suggestion.
In Jadhimah's suggestion, he suggests that the reason why the alif al-qiqāyah was introduced, at a time that wawation was still written on the vast majority of the Arabic nouns, but at at time that the wawation had become completely empty of phonetic realization. Wawation is the practice of an apparently otiose w, especially on the end of Nabatean Arabic names. This practice has almost completely died out, but is still present in the spelling of the name ʕamr which is spelled عمرو <ʕmrw>, which was probably retained, because it helps to distinguish it from ʕumar عمر <ʕmr> which would have otherwise been homographs. Wawation, had not yet completely disappeared in the Early Islamic times for writing names. In the earliest known Islamic Arabic-Greek bilingual papyrus PERF 558 (dated 22 AH, mere 10 years after the prophet's death!), the author of the papyrus Ibn Ḥadīd signs the papryus as ابن حديدو <ʔbn ḥdydw>, rather than the expected ابن حديد <ʔbn ḥdyd>.
Jadhimah suggests that originally, most of the Arabic names had wawation, and that as a result, the vast majority of the Arabic nouns ended on 'silent w's'. At this point, the alif al-wiqāyah would have been introduced to distinguish a 'pronounced' <w> from a silent <w>. The reason why words with pronounced w's like ʕafw did not have an alif al-wiqāyah, was because they were nouns, and were originally written with this <w> and the subsequent wawation *عفوو <ʕfww>. Later this otiose wawation was simply dropped off all nouns, but was retained for a while as a marker of names. This final practice (at least occasionally) lasted into the Islamic period, and ended up being retained in one name, in the Arabic orthography today.
This solution is incredibly elegant, and only suffers from the problem that we lack any texts that have both wawation on all nouns and a noun with an alif al-wiqāyah. This is hardly a surprise, as we have very few documents of Pre-Islamic Arabic written in the Nabatean Arabic script that would actually allow us to confirm this. We simply do not have any texts where we could reliably expect it to begin with.
The origins of Wawation
So, how did the orthographic practice of placing wawation on all nouns develop? The wawation certainly at some point in the history of Arabic Orthography had a phonetic function, as it has a very peculiar distribution in Arabic names within Nabatean Arabic: Names that are diptotic (these are nouns that don't take nunation in Classical Arabic) don't have wawation, Names that are triptotic do take wawation. This is the reason why we have the pair عمرو ʕamr versus عمرʕumar. In Classical Arabic, the former is a triptote nom. ʕamrun/gen.ʕamrin/acc.ʕamran while the latter is a diptote nom. ʕumaru/gen.-acc. ʕumara.
This can be best understood through the following developments from a Proto-Arabic situation:
- Loss of word-final vowels: ʕumaru/ʕumara becomes ʕumar, and subsequently is spelled <ʕmr>.
- Loss of nunation, the originalʕamrun comes to be pronounced ʕamru, and is spelled <ʕmrw> (the other cases, although mostly unattested would be <ʕmry> and <ʕmrʔ> (Nabatean did not have a length distinction for final vowels, it us to be expected that short vowels would simply be spelled as a final long vowel).
This is presumably the phase in which Nabatean Arabic was, when the names were first being written down in Nabatean script. The next phase, was the loss of final /u/ and final /i/, but not final /a/. This may, in fact, have coincided with a general syncope of /i, u/ in open syllables, throughout the word. Many modern Arabic dialects retain /a/ in open syllables, but lose /i/ and /u/: This gives rise to Nom.-Gen. /ʕamr/, Acc. /ʕamra/.
At this point, the <w> that appeared on most nouns would have become a purely orthographic device that had lost all function. This was then spread to all places where the noun had a zero ending. This is presumably also the phase of language, where in Nabatean Aramaic we start seeing spellings of diptotic nouns that take wawation, e.g. ʔabǧaru, as <ʔbgrw> instead of the more regular <ʔbgr>, there was nothing phonetically that distinguished nouns with final <w> from those without final <w>, and the only true difference was whether they had an indefinite accusative -a or not. This would be the phase of the language around which we start to see the first fully Arabic texts in the Nabatean script. Especially the En Avdat inscription, that we recently discussed in our blogpost.
<p ypʕl lʔ pdʔ w lʔ ʔtrʔ>
<p kn hnʔ ybʕnʔ ʔlmwtw lʔ ʔbʕh>
<p kn hnʔ ʔdd grḥw lʔ yrdnʔ>
Rather than Al-Jallad's transcription, which assumes that wawation in these nouns has a phonetic realization, it might be that those spellings might already be otiose (except, of course the accusative, which did mean someting). This inscription has gotten a lot of discussion, and not necessarily a very satisfying translation. Perhaps the insight that the wawation need not be a nominative, while the alif on <ʔtrʔ> has to be an accusative, might give new insights into the interpretation.
My adapted transcription, sticking with the interpretation of Kropp, and by extension that of Al-Jallad we get the following form:
fa-yafʕal lā fedā wa lā ʔoṯrā
fa-kon honā yabġe-nā ʔal-mawt lā ʔabġā-h
fa-kon honā ʔadāda gorḥ lā yorde-nā
Of course, this inscription gives us little reason to think that I'm right the <w> had become otiose in writing, but <ʔ> (for accusative) had not. Both cases seem to function perfectly. But this is not true for all pre-Islamic Arabic inscriptions. If we turn our attention to the JSNab 17 inscription we find a very different situation.
Below bold marks Aramaic words, Italics mark ambiguous words and normal type marks either Arabic names, or unambiguously Arabic words (note that outside of the dating formula, and the demonstrative almost none of the text is unambiguously Aramaic)..
<dnh qbrw ṣnʕh Kʕbw br
Ḥrtt l-Rqwš brt
ʕbdmnwtw ʔmhw hy
hlkt py ʔl ḥgrw
šnt mʔh w štyn
w tryn byrḥ tmwzw lʕn
mry ʔlmʔmn yšnʔʔl qbrw
d[ʾ]w mn yptḥhḥšy w
wldh w lʿn mn yqbr w {y}ʕly mnh>
"This is the tomb which was built by Kaʿb son of Ḥāriṯat for Raqōš daughter of ʕabdu-manawat ("the slave of manawat"), his mother. And she died in Al-Ḥijr, in the year one hundred and sixty-two in the month of Tammūz. And may the Lord of eternity curse anyone who alters this tomb or opens it except his children and may he curse anyone who buries and removes [a body] from it."
What is especially interesting it the wawation on ʕbdmnwtw, which is unexpected, as the feminine ending behaves as a diptote in Nabatean Arabic names.
The phrase py ʔl ḥgrw can only be read as Arabic, having both the typical Arabic preposition fī'in', and the Arabic definite article al. If the final short vowels were still around, this phrase would be read as fī (a)l-ḥiǧri. Instead, we find a final-w, which must certainly be otiose.
At this time the alif al-wiqāyah must have been introduced as a way of writing final -ū, but it is difficult to find unambiguous examples of that in the inscriptions that we have (and we might expect them to be found sooner in genuine Arabic writing, rather than the highly Arabized mixed Aramaic register we find in JSNab 17.
The motivation for using <wʔ> to distinguish it from the otiose <w> should probably also be found in Aramaic orthography. The third person independent pronounce hū in Aramaic, is spelled <hwʔ>, with an otiose <ʔ>. This sequence <wʔ> representing /ū/ for Aramaic, may have been felt by the Nabatean Arabic speakers as a way of writing /ū/ that was not ambiguously also read as zero.
The problem here is that the phase where the <wʔ> was used as a way of writing /ū/ outside of <hwʔ> is so far unattested in the epigraphy. Perhaps because that device was only used in written Arabic at a later stage, of which we have little written evidence, and the device wasn't used for Aramaic writing.
The final stage of written Arabic stopped using the otiose <w> on all nouns that did not have a vocalic ending, and only continued to use it for names. This stage can be found in the Namarah inscription: While the inscription has received many different interpretations, it is quite clear that only (place)names (and not even always) receive wawation, e.g.:
mdynt šmrw /madīnat šammar/ 'the town of Šammar'
mlk mʕdw /malik maʕadd/ 'the king of Maʕadd'
ʕmrw mlk ʔlʕrb /ʕamr malik al-ʕarab/ 'ʕamr, King of the Arabs'
The scenario presented above is still far from certain, but would be confirmed if we would find a Pre-Islamic inscription in the same language stage as JSNab 17, which unambiguously has a 3pl.m. verb with the ending <-wʔ>. No such inscription has so far been found.