A morphological innovation continues
Regular readers will recall that, just about a year ago, I found two young cousins using an innovative strategy to prevent consonant clusters in feminine nouns when vowel-initial possessive suffixes are added. I predicted that “Most probably, the next time I go to Dellys I'll find these two children using the normal forms and denying they ever spoke this way”. It turns out I was wrong: for the time being, at least, both of them are still using it, as confirmed by spontaneous data (quww-at-ək قوّاتك “your strength”, sənsl-at-ək سنسلاتك “your chain” rather than everyone else's quww-t-ək, sənsəl-t-ək.)
Elms between Europe and Arabia
A new word I learned lately is nəšma نشمة (pl. nšəm نشم) “elm tree”. Knowing that most of Arabia is desert, you might assume that this would be a prime candidate for a substratum word to borrow from Berber. In reality, however, it reflects Classical Arabic našamah نَشَمَة, a word used by the pre-Islamic poet 'Imru' ul-Qays and defined in the first Arabic dictionary, Kitab al-`Ayn, as “a tree from which bows are made” (even though the Modern Standard term appears to be dardār دَرْدَار). Clearly it would be a mistake to imagine the pre-Islamic Arabs as uniformly living in an isolated desert environment. At first sight, this word looks nothing like English elm, Latin ulmus, or Kabyle ulmu. However, in general Arabic š corresponds to Proto-Semitic *ɬ, so the original form would have been *naɬam-, which looks rather more similar. The mountains of the northern Middle East where the elm grows have been a zone of contact between Semitic and Indo-European for a long time, and given the tree's distribution, a borrowing into Semitic from IE would seem plausible a priori, especially since it doesn't seem to have cognates in Syriac or Hebrew; but the etymology would require more investigation than I can undertake on holiday. Within Indo-European, the form in question seems to be limited to European branches (Slavic, Germanic, Italic, Celtic), so how it would have reached Arabic is not obvious; coincidence is not to be excluded.
8 comments:
Where would the n- come from?
(I'm ignoring the vowels because I guess they could be explained through morphology or epenthesis somehow.)
Abu lxayr al icbili, umdat attabib [Mediëval Andalus]
- ulmus: awalmay, awalmi, ulmu, tizzaght
I often hear children conjugating the verb kla/kul as nkul, tkul... rather than nak°el, tak°el.... However, I don't understand their transition from this "wrong" (but regular) conjugation to the usual (but irregular) one.
I'm assuming the Kabyle /ulmu/, because of its vowels and absence of initial /n/, is a Latinism. Is it found elsewhere in Berber, and (if so) does its distribution indicate a word borrowed by Proto-Berber itself?
David: I don't have a good explanation for the n; Semitic-internally, one might envision a reinterpretation of a phrase with an early version of the definite article, like *han-'aɬam > *ha(n)-naɬam, but that's pretty speculative.
Anon: Thanks!
Emad: Interesting observation; I wonder at what age they correct themselves?
Etienne: The distribution of elms in North Africa is too limited to tell for certain whether the word could be Proto-Berber or not (it doesn't grow in Mauritania, for a start), but the medieval, non-Kabyle evidence cited by Anonymous is suggestive.
Unfortunately I didn't have the chance to get enough information about that. However, the inverse happened to me: I was using "ak°el" as the imperative of "kla" instead of "kul". I changed it at the age of 15 after my friends corrected me.
About elms: if elm wood was a sufficiently useful/widespread trade item, I could easily imagine that the loanword /ulmu/ could have existed in Proto-Berber even if elm trees did not grow in the Urheimat.
The quotation by Anonymous is suggestive, but it could also suggest that /ulmu/ was borrowed from Mozarabic Romance (phonologically, it could be late Latin or Medieval Southern Romance).
Wikt says that našam has no cognates among the Semitic languages, and so it may be a Wanderwort or have been borrowed from IE very long ago.
L1 learning often involves a two-way shift from regular to irregular back to regular. Anglophone children often begin with sing, sang and move, when they internalize the regular pattern for verbs, to sing, singed, only to correct themselves or be corrected back to sing, sang. Of course, many people do not internalize ox, oxen ever and stick with ox, oxes, which is why there are only about 20 irregular nouns (except for Latin and Greek loanwords) surviving in English. There are about 100 verbs of the strong conjugation (all synchronically irregular) surviving and about an equal number of irregular weak ones like sell, sold.
Post a Comment