The doings of American philologasters are, in truth, a curious study.

Archive for the month “February, 2013”

Some Preliminary Data

As the 18th chapter of the Doctrine of John is the only one thus far that contains a clear terminus post quem (that being 691 CE, the date of the completion of the Dome of the Rock), I thought that it might make for a useful “anchor” for my relative chronology. The first step was to collect data about the language of the text; the chart below illustrates this data, and an explanation of the data follows.

1) gṭal 145 0.52 0.91
2) gṭil l- 14 0.05 0.09
3) nigṭul 13 0.05 0.12
ḏ-nimar 11 0.04 0.85
ḏ-nimar 2 0.01 0.15 0.02
4) gṭul 7 0.03
5) migṭal 5 0.02
6) (qa)gaṭil 95 0.34 0.88 0.98
Total 279 1

The rows represent different verbal stems, starting with 1) the  inherited perfect (“suffix conjugation”), 2) the innovated periphrastic perfect, 3) the inherited imperfect (“prefix conjugation”), 4) the imperative, 5) the infinitive, and 6) the innovated participial present-future tense. As I previously mentioned, 2) and 6) came to replace 1) and 3), respectively, in most Neo-Aramaic dialects; my purpose in collecting this data is to determine how far along this process Classical Mandaic was at the time that this chapter was composed.

I immediately ran into some serious problems.  For starters, although the inherited imperfect (3) appears 13 times in the text, almost all of the examples are from a formula used to introduce direct speech (ḏ-nimarlun). Formulae of this sort are much more likely to retain archaic grammatical constructions than spontaneous speech, so I felt that the data had to reflect this in some way.

Column A reports the distribution of the 279 verbal forms that appear in the text.  As you can see, the preponderant majority were the inherited perfect and the innovated participial present-future tense, suggesting a relative proximity to the contemporary reflex of Mandaic. Column B contains a breakdown of the 13 instances of the imperfect, divided between formulae and all other forms.

Column C reports the relative proportion of these verbal forms. Column D reports a breakdown of the proportion of instances of the imperfect, divided between formulae and all other forms.

Column E reflects the proportion of inherited perfects to the innovated periphrastic perfect in the text.  The overwhelming majority (91%) of examples are of the inherited perfect.

Column F reflects the proportion of formulaic instances of the inherited imperfect to non-formulaic instances.  85% of the examples were formulaic.

Column G reflects the proportion of inherited imperfects to the innovated participial present-future tense in the text. Only 12% of the examples reflected the inherited imperfect, a number which decreases even further when you remove the formulae from the proportion (Column H); then the proportion becomes 98% to 2%, a situation very much like that found in Neo-Mandaic, in which the only vestiges of the inherited imperfect are found in formulae.

It will be interesting to compare the breakdown of the data in Chapter 18 to that of the other chapters that have been parsed.


How Can You Date a Text?

The question I keep asking myself as I read selections from the Doctrine of John is how these texts relate to one another and to the rest of the Mandaic corpus. I’d like to be able to express these relationships quantitatively, but for the time being I must stick to the qualitative differences that have been identified by previous researchers. These include:

  1. References to Islam or historical events that followed its advent;
  2. Orthographic conventions, such as historical or phonetic spellings;
  3. The presence of Arabic words, either in the form of loan words or names; and finally,
  4. “Modern” (Neo-)Mandaic forms such as the indicative particle qa- or the 3rd pl. personal morpheme -iun on the perfective;

All of these are found to some degree or another in Chapter 18 of the Doctrine of John.  The problem with 1) is that it is subjective and open to interpretation, as I have hopefully demonstrated elsewhere.  In the context of Chapter 18, there is a reference to a “Dome of the Priests,” qumba ḏ-kahnia, which appears to be a synonym for the Temple, but this would place the text (or at least this version of it) squarely in the Umayyad period, as the Second Temple was most certainly not domed, and was destroyed in 70 CE in any case; no dome would appear until 691 CE, when the Dome of the Rock was completed. This confusion between the Temple of Solomon and the present occupant of its former site is by no means restricted to the Mandaeans; at times, Muslims, Christians, and even Jews have confused the two. The incipit to the Gospel of Luke in the Harley Golden Gospels, which was composed sometime during the first quarter of the 9th century, features an illustration of the Annunciation to Zechariah (the very subject of Chapter 18) against the backdrop of the Dome of the Rock:

Incipit to Luke

2) is a slightly better indicator of the antiquity of a text, but still problematic, as these texts were copied and recopied constantly throughout the centuries, and with each copy came new opportunities to revise the spelling of each word (potentially to “modernize” or even “archaicize” it).  It is for this reason that the preponderant majority of the variants between the existing texts of the Doctrine of John are spelling variants.

With 3) we find ourselves on slightly better ground, but its utility as a metric for dating texts is still quite limited.  For starters, Arabic and Mandaic belong to the same Central branch of West Semitic, and share a common inherited vocabulary. Words inherited by both languages will be distinguished by their phonology; while Arabic retains the distinction between most Proto-Semitic phonemes, in Mandaic many of these same phonemes have been merged together.  Unfortunately, the Mandaic script reflects its phonology, with the result that Arabic phonemes not found in Mandaic are lumped together with their closest Mandaic equivalents, which are not infrequently reflexes of the same ancestral phonemes. Thus distinctions between Arabic and Mandaic words are occasionally invisibilized by the script. Furthermore, Arabic words are rare in this text, and the pedigree of a given word is not always unimpeachable.

4) is perhaps the best indicator of the relative antiquity of a text.  The presence of Neo-Mandaic forms, when they are absent from other classical texts, indicate that the text in question was composed more recently, relative to those texts.

To these diagnostics, I’d add another qualitative difference, and perhaps even a quantitative one:

  1. The presence of the periphrastic gṭil l- perfect conjugation, which is neither Classical nor, strictly speaking, Neo-Mandaic;
  2. The gradual replacement of the old Semitic imperfect or “prefix conjugation” by a participial present-future tense.

The latter requires some explanation. Between the period of Late Antiquity and the present day, the Eastern Aramaic dialects underwent a major restructuring of their verbal system. As a result of this restructuring, the ancestors of most surviving Aramaic dialects lost their inherited perfect and imperfect conjugations, which were replaced by periphrastic conjugations based upon the participles. These participial constructions were already present in earlier Aramaic dialects, in which they were used to supplement the inherited tenses. Construction 1), mentioned above, eventually came to replace the West Semitic perfect or “suffix conjugation” in most surviving dialects. It is conspicuously absent from Neo-Mandaic, but appears occasionally in the Doctrine of John.

The process by which these conjugations were replaced was gradual, and Classical Mandaic, especially insofar as it is represented by the language of the Doctrine of John, is a virtual “snapshot” of this process at work. Thus it is possible that this process could be used as a metric for a relative chronology of the texts; if the proportion of the innovated forms to the inherited forms can be observed within the corpus of texts to grow over time, then the relative age of individual texts could be potentially be gauged.

Arabic Loanword

I’ve encountered what could very well be the first genuine Arabic loanword I’ve yet seen in the text of the Doctrine of John. The word is hus, which appears at least twice, first on p. 4, ln. 15, and again on p. 76, ln. 14. Both contexts require something meaning the “source” of the Jordan, and have been translated by Lidzbarski as “reservoir.”

According to Nöldeke (Mandäische Grammatik, p. XXXIII, ln. 17), the word appears three times in the Doctrine of John, and comes from the Arabic حوض ḥawḍ- “basin, cistern.” On p. 670 of Book 1 of his Arabic-English Lexicon (London: Willams & Norgate, 1863), Lane identifies it with the root √ḥ-w-ḍ “to collect (water)” inter alia, which would seem to indicate that it is indeed a good Arabic root, and not potentially a loanword into Arabic from some other language. What is it related to?

Assuming no ad hoc sound changes, the PS root could only be *ħ-w-ɬʼ. This would give us Aramaic √ḥ-w-ʕ (or √ḤWQ), Hebrew √ḥ-w-ṣ, Akkadian √ʔ-w-ṣ, and so forth. The root √ḥ-w-ʕ is unattested in Aramaic; I likewise found nothing in Akkadian that would indicate the existence of this root (the expected G infinitive êṣum “to be(come) too little, small,” is actually derived through regular sound changes from Old Babylonian wiāṣum, according to the CDA, and the expected cognate to Arabic ḥawḍ-, ūṣum, only appears with the meaning “arrowhead,” which clearly comes from the unrelated PS root *ħ-ϑʼ-w).

Hebrew, on the other hand, has two roots that are potential candidates, חוץ I and II, the first of which is represented by the word חוץ ḥuṣ “outside” and the second by the word חיץ ḥáyiṣ “barrier, partition.” BDB relates the latter to Arabic root حوص √ḥ-w-ṣ, meaning “to sew” or “to contract,” but it seems to me that both could be related to the Akkadian root √ḫ-ṣ-ṣ (e.g. ḫaṣāṣum “I to snap off; II to erect (a reed hut)”), which would make them completely unrelated to PS *ħ-w-ɬʼ.

Quite apart from the question of the origin of the Arabic term is the question of how it came to be adopted into Mandaic.

  • Why would the Mandaeans, who live in a water-rich environment and have a correspondingly rich vocabulary for water features, need to borrow such a term from Arabic, which is not generally known for its aquatic vocabulary? It would represent a total inversion of the established principles by which borrowing operates, as if the Inuit were to borrow the English word for “snow,” or if New Jerseyans were to borrow the Inuit word for “hairspray.”
  • Why then, for that matter, would this stray Arabic word appear, isolated, in a text that is otherwise almost completely bare of other loans?
  • Why employ a borrowed word that means “basin,” i.e. a place where water is collected, in a context that clearly requires something like “source”? They’re not the same thing, and neither Nöldeke nor any of his followers have attempted to explain the semantic shift.
  • Finally, how can we explain the anomalous use of s to represent Arabic ض? Here is a spare list of words containing the same phoneme, borrowed into Mandaic:
Arabic Transliteration Mandaic Gloss
أرض ʔarḍ- arda earth
بياض bayāḍ- baiad whiteness
ضعيف ḍaʕīf- daeip weak
حوض ḥawḍ- hus basin

I could find no other words in Drower and Macuch, and at least one of these is dubious in the extreme, but the pattern is clear: the Arabic phoneme ḍ-, when borrowed into Mandaic, is represented by d, except for this one word.

Granted, one could easily counter that we’re not dealing with a lot of data here, and that the word hus, if it was indeed borrowed, may have been borrowed from Arabic when the articulation of ḍ- was quite different from what it was at a later date. Even so, there’s not much evidence that this was borrowed from Arabic, despite the superficial similarity between the two words, and as Mark Rosenfelder reminds us, chance resemblances between words in two unrelated languages are not at all uncommon. In the final analysis, it’s just not a good candidate for an Arabic loan word.

So, what does it mean, and where does it come from? I can’t find any good Middle Persian candidates, unless we assume it’s somehow related to xwaš “pleasant, sweet, nice,” which somehow entered nearly all the languages of Iraq, Semitic and non-Semitic, as xōš “good” and which precedes the noun it modifies in each of these. Unfortunately, this interpretation founders upon the same issue, that of the final sibilant, which is just not a match.

Post Navigation