Avestan: "naxtar" ~ "night"

Discussion in 'Etymology, History of languages and Linguistics (EHL)' started by Phosphorus, Aug 21, 2012.

  1. Phosphorus Senior Member

    Kurdish
    Greetings,

    I recently noticed Avestan "upa-naxtar" ~ "adjoining the night". It appears to me that something such as "naxta*" (with a probable Old Iranian form of "nakta*") existed in Avestan for "night"-along with "xshapa".

    I wonder is "naxta*" (or something like it) attested in the Avestan materials at hand? It would certainly share the same root with Sanskrit "naktam" ~ "at night" (as it is asseverated here: http://www.etymonline.com/index.php?term=night&allowed_in_frame=0).

    P.S. This Avestan entry will eventually explain the etymology of modern Kurdish "notek"/"nutek" ~ "sheer darkness"/"darkness of night" (< "nuxtak*" <? "naxuta-ka*").

    Thanks in advance for your reply.
     
  2. berndf Moderator

    Geneva
    German (Germany)
    The root nokt-/nekt- is really common Indo-European and can be traced through all through the entire IE group. Cf. Latin nox, oblique root noct-; Classical Greek νύξ [nÿkʰs], oblique root νυκτ- [nÿkt-]; German Nacht [naxt]; English (before the muting of "gh"): night [nɪçt] (in Germanic, [x] and [ç] were allophones of /h/, the reflex of PIE /k/), Lithuanian naktis...
     
  3. Phosphorus Senior Member

    Kurdish
    Yes seemingly a significant number of Indo-European languages share a similar root in this case. For modern Iranian languages, as far as I know, it has only survived in Kurdish in shape of "nutek". I could not initially figure out its exact origin: via an Old Iranian root or through the Hittite proper? But noticing Avestan "upa-naxtar" clarified its Iranian origin to me.

    I just wish to know what is the exact Avestan root ("naxta*"?) and is it attested independently in Avestan materials, namely not in combinations such as "upa-naxtar"?
     
  4. berndf Moderator

    Geneva
    German (Germany)
    Or directly inherited from PIE. It doesn't have to be a loan at all. That's why I showed you it is a very common IE word.
    I found a reference to naxtar=night outside the phrase upa-naxtar here., p.9, footnote 44.
     
  5. Phosphorus Senior Member

    Kurdish
    Yes I did not took it as borrowing from Avestan, I was trying to have an idea about a probable Old Iranian common root. And you are right one should also take into account the direct inheritance, thank you.

    Many thanks pal, I am grateful for the provided link.
     
  6. fdb Senior Member

    Cambridge, UK
    French (France)
    The adjective upa.naxtar occurs twice in the Avesta. Bartholomae, Altiranisches Wörterbuch, col. 391, glosses it as “an die Nacht angrenzend”. It implies a simplex naxtar- (n.), like Skt. nakt- (f.), but with an r extension, as in Greek νύκτωρ “at night” (adv.). The simplex does not occur in the Avesta, but the root survives in Wakhi (one of the Iranian languages in the Pamir mountains) as naγδ, a perfectly regular Eastern Iranian reflex of *naxt. Otherwise, Iranian has replaced the inherited word for “night” by reflexes of xšap- (f.), xšapan (f., n.), xšapā- (f.) etc., like Skt. kṣap- (f.), e.g. New Persian šab.

    The reference supplied by Bernd is not Avesta but Pahlavi (Dēnkard VIII). The passage is very problematic. I am not aware of any other occurrence of *naxtar in Middle Persian and would like to defer judgement on this passage until a critical edition of Dēnkard VIII becomes available. Due to the ambiguity of Pahlavi script nhtl could just as well be whtl, wʼtl, nʼtl, etc.

    I do not have an etymology for Kurdish notek/nutek, but a derivation from *naxtar- seems very difficult from the point of view of regular sound correspondences.
     
    Last edited: Aug 22, 2012
  7. Phosphorus Senior Member

    Kurdish
    Thanks a lot for the precise reference, I appreciate it.

    Thank you for the clarification. I just initially presumed that Pahlavi "naxtar" is only another religious term that has entered Middle Persian lexicon during, let's say, the revival of Zoroastrianism under the Sassanid pax. I believe "spenag" ~ "sacred" or "srishwadag" ~ "1/3" are potentially similar examples (I have not read it somewhere, but based on the typical sound changes in Avestan and Persian, I assume these to be very later borrowings).

    At the first sight it appears so, but Kurdish in some cases has already developed "-x-" into "-w-". This development happens via "x" turning into "gh" or "g" and then this sound results in "w" ("-g/-" > "-w/-" shift is typical in Central Kurdish, e.g. "agir" > "awir" ~ "fire", "cirag" > "ciraw" ~ "lamp", "bedbext" > "bewbext"/"bobext" ~ "unlucky"). An at hand example, for "x" > "w" shift, might be common Kurdish "nexsh" (< Ar. "نقش") which has a variant such as "newsh" (< "negsh*" < "nexsh"; also one "shewtan" ~ "to burn" may suggest a similar development, cf. Av. "saoxt*", Per. "suxtan"). I presume the same thing may have happened to a supposed common Iranian root of "naxta*" in Kurdish: "naxta*" > "naxt-aka*" > "nawtak*" > "notak" or as in Kurdish alphabet "notek".

    Thanks to your mention to Wakhi "naghd", I found this link (http://en.wiktionary.org/wiki/Appendix:Proto-Indo-European/nókʷts). There also states Kurdish "nixte" to be of the same origin (although its sense might not properly fit the significance of darkness or night-I am afraid it may be only representing an outward resemblance and nothing more). But if considered of the same root then "nixte" is the Northern (Kurmanji) Kurdish cognate of Central (Sorani) "notek"/"nutek". However retention of "x" before "t" in Kurdish appears a little bit far-fetched to me, unless in case of loans.
     
  8. fdb Senior Member

    Cambridge, UK
    French (France)
    Your explanation of the Kurdish word looks good. I was not aware of the wiktionary entry. The Avestan script does not show up on my screen, so I cannot comment on its correctness, but the Wakhi word should have a fricative -δ (as correctly in Bartholomae), not -d.
     
  9. Phosphorus Senior Member

    Kurdish
    The Avestan script does not appear on my screen either. Yes it is most likely a spelling error putting "naγd" instead of the original "naγδ". Thanks again for your help pal.

    I wonder if you could help me out with Persian "sir" ~ "saturated" here in this thread: "http://forum.wordreference.com/showthread.php?t=2476993". I will add some updates in there.
     
  10. Dhira Simha Senior Member

    UK
    Russian
    There is no doubt that all these words are cognates, however, postulating a root nokt-/nekt, obtained by a simple mechanical application of regular correspondences, may not provide the correct picture. If we have a closer look at the Vedic nakta (n.) we can see that there are also forms like náktā (f) and also nakti (f.) all meaning "night". Such gender endings are typical for verbal adjectives formed by the Past Passive Participle ending -ta. In fact, nakta by its internal structure is a typical Past Passive Participle: compare akta "driven" (from aj "to drive, propel"), ukta " said" (from vac "to speak"), yukta "united" (from yuj "unite, join") and dozens of others. This is actually acknowledged by C. Watikins in his dictionary: where he, following Pokorny's nekʷ-(t-), nokʷ-t-s, has the IE root as *nekw-t- accompanied by a note "Probably from a verbal root "neg"-, to be dark, be night." I could say a lot more but, I am afraid, it will be off-topic.
     
  11. Phosphorus Senior Member

    Kurdish
    Very instructive notes Dhira Simha, thanks for you contribution.
     
  12. Dhira Simha Senior Member

    UK
    Russian
    I thought that it was important because it it would be interesting to establish whether a cognate for "night" in a given branch or language was inhered as a noun, which had already lost its connection with the verbal root *nVg/nVj, or was it formed within it by ways of a P. P. P. or some other verbal form. This is, of course, subject to the acceptance of the existence of such hypothetical root. In any case, such possibility should not be discounted. One element of proof in favour of this root is the rather obscure Vedic t-less nak ind. (g. %{svar-Adi} , as nomin. RV. vii , 71 , 1) night. I shall find the passage in Rig-Veda and analyse it.

    As for your
    You probably are aware of this, in case somebody does not know, the final -k(a) is a highly prolific comparative/diminutive suffix used in forming adjectives and nouns to express diminution, deterioration, or similarity e.g. putraka, a little son; aśvaka, a bad horse or like a horse. (see Edgerton, F. The K-Suffixes of Indo-Iranian, John Hopkins University, 1911).

    So note-k may be interpreted as "similar to night i.e dark" or "dark as night". Loss of /k/ before /t/ is not unusual. Cp It. notte, however the explanation of fdb through /w/ is more fitting here, of course.
     
    Last edited: Aug 24, 2012
  13. Phosphorus Senior Member

    Kurdish
    Your suggestion about the exact etymological origin of "night" is an intriguing idea.

    Thanks for your notes. You are right "-ka" suffix proposes a semantic change in case of Kurdish "notek" and yes loss of "k" before "t" is unusual neither in Kurdish (e.g. "wakt-*" > "waxt" > "wet", "raekt*" > "rext" > "re(h)t", "pakt-*" > "paxt*" > "pe(h)t"). But what makes me uphold another development, rather than "naxtaka*" > "naxutaka*" > "nuxtak*" > "notek", for the Kurdish entry is presence of "-o-" as well as a cognate such as "naghdh" attested in another Iranian language (Wakhi).

    That is to say here I speculate of an earlier form of "naghtak", most likely existed in the specific Scythian speech spoken in today Kurdish speaking areas, which can be safely followed by modern Kurdish "notek" through "naghtak" > "nawtak" > "notek"/"nutek". Kurdish "shewtan" also suggests a similar development.
     
  14. Dhira Simha Senior Member

    UK
    Russian
    Sorry, this is not my kettle of fish. All I can say that the transition /u - o/ is most natural and quite common across many languages. The medial /d/ in the cognate "naghdh" might be a simple voicing of /x/ (perhaps via /ɣ/). I am sure, fdb can comment much better and more professionally. I have looked up the Rig Veda RV 7,71.1.jpg and it does have the word nak "night" in the context (without sandhi) apa svasuḥ uṣaḥ nak jihīte... which I would translate as "From (her) sister Ushas (with)draws Night [nak]..." but with sandhi it reads nagjihīte. I do not know if it helps you.
     
    Last edited: Aug 24, 2012
  15. Phosphorus Senior Member

    Kurdish
    It is ok pal, your considerations in regard with the P.I.E root of night have already been helpful. I appreciate it. :)

    P.S. As far as I know turning middle "x", specifically before "t", into "gh" frequently happens in Eastern Iranian languages.
     
  16. Dhira Simha Senior Member

    UK
    Russian
    I could not find naxtar in the vocabulary of A.V. W. Jackson, however, I would like to note that in the text you suggested it is translated as "night, starry night". This is significant. It may mean that it does not directly relate to Skr. nakta(m) but is actually a compound nak - tāra where nak is "night" and tāra - star. This explains the otherwise problematic /r/. Could you give the reference to this text. I could not find any source.
     
    Last edited: Aug 24, 2012
  17. Phosphorus Senior Member

    Kurdish
    Is "tāra" itself attested in Avestan besides "stara*" for "star"? By the way I wonder under which rule a supposed Avestan "nak-tāra" could give "naxtar"?

    I believe Wakhi "naghdh" and Kurdish "notek" confirm a common Old Iranian word such as "naxt(a)*", but I due to lacking an inclusive Avestan glossary cannot suggest whether a suffix is responsible here or not (perhaps one similar to modern Kurdish "-al" < "-ar*", e.g. "mert" > "mirtal" ~ "carrion; slim", "kend" > "kendal" ~ "canal").
     
  18. Dhira Simha Senior Member

    UK
    Russian
    Only star, stara m. star (http://www.avesta.org/avdict/avdict.htm#dctt) appears to be attested. However, compared to Vedic, only a small part of Avestan vocabulary is attested due to the restricted volume of preserved texts. I would like to point out, that in Skr. the semantics of tāra is uncertain. It has a general meaning mfn. "carrying across, a saviour, protector") which is a semantic blending of two separate roots: tṛṛ "to carry through or over, save" and trai "to protect, preserve" . The meaning "star" is usually derived from the root stṛ "to spread, spread out or about, strew, scatter" :
    tṛ́ [p= 453,1] [L=86521] n. (= stṛ́) nom. pl. tā́ras , the stars.

    So, as you see, it is a bit complicated. As for your second question, it is known that /s/ tended to become debuccalised to /x/ (in my humble opinion, due to the retraction of the general articulatory setting under the influence of admixture with surrounding languages - Semitic?) in Avestan. The /k/ in nak could have coalesced with the following /s/ producing /kh/ - /x/. Usually, /st/ did not change to /xt/ but the added influence of the preceding velar in nak could have triggered the change. Or, we could postulate the existence of the non-attested parallel form *tara in Iranian but this is more speculative.

    Finally, what is the other explanation of the final /r/ in naxtar? Why is it also translated as "starry night"?
     
    Last edited: Aug 25, 2012
  19. Dhira Simha Senior Member

    UK
    Russian
    I do not know, if you hoticed the other intersting Avestan word: naxturu n. "nocturnal"
     
  20. fdb Senior Member

    Cambridge, UK
    French (France)
    This is not another word, it is an erroneous citation of the same word. In its two Avestan attestations we in fact have the locative plural written upa.naxturušu, for an expected *upa.naxtərəšu (thus Bartholomae).

    The -r- has been explaned in no. 6.
     
  21. Dhira Simha Senior Member

    UK
    Russian
    Thank you! With Sanskrit I can check the source directly but I have to trust dictionaries here. I took it from http://www.avesta.org/avdict/avdict.htm. I mostly use Jackson's vocabulary and came across this site today. It appeared quite credible. At least this is clear. Sorry, I missed the bit about -r- but I still cannot understand what you mean by "extension". Is it a suffix? What does the Greek have to do here? I would understand Sanskrit but the Greek evidence, which does look similar, can hardly be taken as a straightforward explanation. I am just curious. Are there any other Avestan words with this "extension"? Does it have an analogue in Sanskrit? There is a primary suffix -ra which is added to make adjectives i.e. ug-ra " mighty", aji-ra "swift". Generally, ra (like -ka quoted in 12) has its own meaning "acquiring, possessing" which is quite suitable to create adjectives but it does not convey any locative aspect as in Greek νύκτωρ “at night”.
     
  22. fdb Senior Member

    Cambridge, UK
    French (France)
  23. Dhira Simha Senior Member

    UK
    Russian
    I can see that the electronic dictionary is crap. But I could not find the answer to the origin of -r in naxtar
     
  24. fdb Senior Member

    Cambridge, UK
    French (France)
    In Indo-European we often find synonymous stems formed from the same root with different extensions. E.g. in Avestan we have in the meaning “strength”:

    aog-ar- (n.) IE. *-or/n- (heteroclite)
    aog-ah- (n.) IE. *-os-
    aoj-ah- (n.) IE. *-es- cf. Skt ojas-

    Similarly:

    xšap- (f.) ‘night’
    xšap-an- (f. or n.) ‘night’
    xšuuaš.xšap-ar- (n). ‘period of six nights’.

    And so forth.
     
    Last edited: Aug 26, 2012
  25. Dhira Simha Senior Member

    UK
    Russian
    Thank you! I am not sure for IE, which is not attested, but I agree with you that it may be the case in Indo-Aryan. I still think that it is of the same nature as in Skr. ug-ra " mighty", aji-ra "swift" etc.
    I do not think I have more to say. Obviously, the "starry-night" hypothesis is incredible and I only proposed in view of the the Pahlavi naxtar "starry night" which, as you say, is doubtful. I am inclined to support you that most probably the poorly attested Avestan *naxtar/naxturu cannot be considered as the prototype for the Kurdish words. Still, I think, the P.P.P. origin of nakta and a hypothetical root *nVg > *nek/nak seems to be an interesting venue to explore. Thank you!
     
    Last edited: Aug 26, 2012
  26. Phosphorus Senior Member

    Kurdish
    You are right about the probable Avestan words that we are possibly unaware of, but in respect with our present ken "tāra" is certainly not attested in Avestan. I always had the idea that classical Persian "tārā" ~ "star" (in modern Persian only as a female first name) is a loan of Indian origin.


    As far as I know "k" and "g" are supposed to give "x" when they precede "s" in Avestan, this shift is attested-but "s" always remains intact. So based on your proposed pattern for now we have "nak*" + "star" > "naxstar*". Now the question is whether such a development is confirmed in Avestan to change, or maybe to loose, "s" in a "-xst-" combination? I also have no idea whether a middle Old Iranian "s" is ever to turn into "x" in Avestan (I think the famous development of P.I.E "s" into "h" and later "x" in Avestan and Iranian languages is rather out of context in this case).

    Yes I too believe that, concerning our current confirmed knowledge in regard with Avestan and-generally-Iranian languages, such a speculation is not safe.
     
  27. Phosphorus Senior Member

    Kurdish
    Interesting citation, thank you. If I am not mistaken "-ar" in "xšap-ar" may eventually propose a solution for the exact Avestan root in "naxtar": "naxt-*" + "-ar" (specifically when new Iranian entries such as "naghdh" and "notek" or their Indo-European cognates are taken into account)
     
  28. Phosphorus Senior Member

    Kurdish
    I wonder if this "-ar" in Avesta has anything to do with New Persian "-ār" (e.g. in "mordār") or Kurdish "-al" (e.g. in "mirtal", "kendal", "gutal", etc.)?
     
  29. Dhira Simha Senior Member

    UK
    Russian
    Sorry, what does Persian mordār means? I hope you do not confuse it with the prolific "agent ending" like Skr. -tṛ in mātṛ.
     
  30. fdb Senior Member

    Cambridge, UK
    French (France)
    Yes, exactly.
     
  31. fdb Senior Member

    Cambridge, UK
    French (France)
    No, the suffix –ār (with long ā) is a semi-productive morpheme in Middle Persian to form nouns from the past stem of verbs: murd-ār ‘carrion’, dīd-ār ‘sight’, guft-ār ‘speaker’ etc.

    By contrast, the stem-extension –r- or –ar- is operative only at a pre-historic (Indo-European or Indo-Iranian) level.
     
  32. Phosphorus Senior Member

    Kurdish
    New Persian "mordār" means "carrion". The suffix used used here is "-ār", which I do not know whether it has anything to do with Sanskrit "-tṛ" or not.
     
  33. Phosphorus Senior Member

    Kurdish
    That is cool. Thank you.

    I see. Well I assumed a connection because Kurdish "-al"/"-ar" shares evidently the same root with Persian "-ār", but surprisingly the Kurdish suffix attaches nouns as well-as opposed to the Persian proper: "gutal"/"guwal" ~ "crap" < "gu(t)" ~ "feces", "roal" ~ "reddish" < "ro" ~ "red", "sewzal" ~ "tan" < "sewz" ~ "green".

    Do you have any idea what might be the etymology of Modern Iranian "-ār/-al"?
     
  34. Phosphorus Senior Member

    Kurdish
    I wonder does Wakhi "gh" (γ) usually result directly from "x" or it comes through "g" < "x"-as the famous "g" > "gh" change in Eastern Iranian languages?
     
  35. fdb Senior Member

    Cambridge, UK
    French (France)
    xt > γδ or γd
    ft > βδ or βd
    This is the general rule in Eastern Iranian.
     
  36. Phosphorus Senior Member

    Kurdish
    I see, it is clear now. Many thanks pal.
     

Share This Page