25 Mar 2012

Learning, the unending battle against paradox

I use my own personal models in order to understand things rather than merely adopting the exact model of another (ie. blind rote learning). I believe it's only by allowing oneself to freely explore and question within one's personal models of things that one can learn more rapidly and gain a deeper understanding of things. After an interesting discussion with a commenter about my model of Egyptian vocalism, we uncovered some small inconsistencies in my theory. Of course, there were inconsistencies in his account of things too. Despite being a zealous Peust fanboy, Peust's theories apparently weren't enough for him to reconstruct the Egyptian word for 'man' with any confidence (whether *zī*zij or *zijV; I'd wager *ziˀ or *zaˀ). While he seems to insist there's some sort of "consensus" out there for Egyptian reconstruction, the undetermined vowels in Woodard and Loprieno's representation of Egyptian shows that there are still large gaps in the field. Let's not be deluded.

So since not all is written on this subject and we all have something to learn here, this online discussion helps us get our bearings straight on where we each need to improve. I can certainly take a few things from this exchange, myself. I've been inspired as of late to begin keeping stricter notes on the Coptic language in my personal offline database. It can't hurt, and it's already helped me gain a clearer picture.

One of the inconsistencies that popped up in the discussion was my understanding of the pattern of Egyptian noun-plus-possessive forms such as nb.f 'his lord' and rn.f 'his name'. I now perceive that Middle Egyptian nouns with long vowels must have drawn stress to themselves in these possessives, away from a penultimate default. So, while *rin 'name' with short *i leads to *ranífa 'his name' (nb. the preservation of Pre-Egyptian oblique case marker *-i correlating with that in Proto-Semitic), a word like *nība 'lord' must lead to *nībafa 'his lord' with accent on the long vowel of the first syllable (nb. the original vowel quality of the petrifact oblique marker is reduced in unstressed syllables). This seems to work well. For example, I account for the cuneiform-rendered personal name Bukurninip with *Boˀka-n-Ranífa [ˈbɔˀkn̩ɾəˌnɪfə] 'Servant-of-his-Name' (from an earlier *Bāˀka-na-Ranífa).

Another question I came face to face with as we were talking was: How precisely do vowels evolve from a Loprieno-derived model like mine to Coptic proper. After sitting on this for a few days while looking at a number of Coptic examples of my sound changes at work, I see that just a few minor modifications can fix things.

Let's say, as before, that by the 1st millennium BCE only Middle Egyptian long  leads to Proto-Coptic *o (nb. no phonemic length anymore) mirroring the contemporaneous Canaanite Shift in North-West Semitic languages. I maintain that short *a simply must have remained a in Sahidic in at least some cases, judging by how MEgy *sanáwi 'two' (rendered directly in cuneiform as ši-na-ah-wu in EA 368) became Sahidic snau. The AA cognates of this numeral only add to this likelihood. With  becoming a Late Egyptian *o, the selection of either omikron or omega in Sahidic (a matter of phonemic quality, not phonemic length) should depend only on whether the syllable was closed or open at the time, respectively. The presence of nasal stops just adds a slight twist by raising *o further to ou /u/, as in Sahidic noute 'god' < *note and moui 'cat' < *moya.

Likewise, as long  and  merge to *e (with loss of length contrast) this vowel must have similarly split into either eta (/e/) or iota (/i/) based on the openness of the syllable. However, in contradiction to my previous version of things in the commentbox, I now realize that I have it backwards. Sahidic i appears to correlate with *closed* syllables while ē (a front-high /e/ without phonemic length) matches best with open syllables in a later stage of the language[1]. This then necessitates some interesting tweaks.

For Proto-Coptic *CéC (closed syllable), we then have:
  • MEgy *pasīj 'nine' > PCop *pset > Sahidic psis
    (Note EA 368 pi-ši-iṭ in cuneiform.)
  • MEgy  *maḥīt 'north' > PCop *mxet > Sahidic mxit 
For Proto-Coptic *Cé (open syllable):
  • MEgy *mūˁat 'truth' >  PCop *méˀe > Sahidic mēē
  • MEgy *rīˁa 'sun' >  PCop *réˀa > Sahidic  
  • MEgy *mūḏa 'ten' > PCop *méta > Sahidic mēt
Now this seems nicer and more regular (cross fingers). Just the way I'd like to keep it - facts willing! Of course this makes me think up new questions about the exact processes of the language. My understanding is, as always, a work in progress. One way or another, however, I'm determined to figure out those undetermined vowels because I've always loathed wildcard symbols in reconstructions. It's the principle of it all, you understand.

NOTES
[1] I botched that up again! So sorry. I wrote "Sahidic i appears to correlate with *open* syllables". Please read the opposite. Sigh. Of all the typing mistakes, I make the most confusing one. LOL!

17 comments:

  1. How do you explain the i/a alternations in Coptic (see Peust p. 175 for examples)? They seem to be quite analogous to the o/ō alternations, and there are cases like mise < *mst "to give birth" where it seems quite clear that i has been in an open syllable at least since Late Egyptian.

    ReplyDelete
  2. Michael Hahn: "How do you explain the i/a alternations in Coptic [...]"

    From Middle Egyptian /*i.

    "[...] and there are cases like mise < *mst(sic) 'to give birth'"

    It's spelled msy and I await proof of your claim that the syllable is assuredly open "since at least Late Egyptian".

    What facts mitigate against the following development: MEgy *mīsiya > PCop *mesya > Sahidic mise?

    ReplyDelete
  3. It's spelled msy and I await proof of your claim that the syllable is assuredly open "since at least Late Egyptian".

    The Coptic "infinitive" is derived from the old infinitive (if you look at the development of verbal morphology, you see that the Coptic infinitive appears where Late Eyptian has an infinitive), and the infinitive of msj was mst, of course. The feminine ending had disappeared by Late Egyptian in pronunciation, so its pronunciation must have been close to *misV > mise.

    From Middle Egyptian *ī/*i.

    Okay, so you assume that the pronominal state could involve vowel shortening (as in mise/mastf "giving birth to him"), right?.

    ReplyDelete
  4. I get grumpy when some stray from the necessarily strict rules of rational debate. I'm a stickler but it's for good reason. I dare say that you've manipulated the argument in at least three ways:

    1. You neglect to prove your statement that the first syllable was open in msy "since at least Late Egyptian". Or do you mean its derivative msy.t? What exactly are you trying to prove exactly?

    2. Despite lack of proof, you doubly fail to admit confusing opinion with fact and retract. (One retracts errors out of respect for debate and the surrounding witnesses of that debate.)

    3. You fail to state only *relevant* facts and instead distract with already-agreed-upon minutia (ie. *-at [feminine] > Sahidic -e and deverbal nouns using this feminine) that is completely non sequitur to your initial objections that remain as yet unjustified.

    I'm certain you're aware "to give birth" is a triliteral msy, not biliteral. So its deverbal noun in -t is evidently msy.t originally. Later we see it reduced to ms.t, but this is already self-evident by Coptic mise. It's quite natural for a sequence -VsyV- to reduce to -VsV- over time.

    Now will you prove with clarity when this reduction occurred?

    "Okay, so you assume that the pronominal state could involve vowel shortening (as in mise/mastf 'giving birth to him'), right?."

    Yes. But notice that the accent remains on the first syllable in mastf. So the opposition is, as I've always said, restricted to stressed environments.

    ReplyDelete
  5. I'm certain you're aware "to give birth" is a triliteral msy, not biliteral. So its deverbal noun in -t is evidently msy.t originally.

    You should go to the grammar literature to see that the hieroglyphic writings consistently show that the -j of weak verbs like msj only appeared (even in Old Egyptian) in certain conjugational forms where it can just as well be analyzed as a suffix. As long as you don't show why the Egyptians never wrote the -j- in infinitives like your *msy.t, but chose to write it in other conjugational forms, ms.t is preferable.

    ReplyDelete
  6. MEgy *sanáwi 'two' (rendered directly in cuneiform as ši-na-ah-wu in EA 368)

    How do you explain the -h-? The interpretation in the book you linked (šnʿ.wj) seems a bit more plausible, given the -h-.

    ReplyDelete
  7. "As long as you don't show why the Egyptians never wrote the -j- in infinitives like your *msy.t [...]"

    And why should I prove against a falsehood? Try again.

    ReplyDelete
  8. "How do you explain the -h-? The interpretation in the book you linked (šnʿ.wj) seems a bit more plausible, given the -h-"

    Yes, it would have secondarily acquired an ahistorical ayin by association with "one". No big whoop. The ayin isn't present in the Semitic cognate as far as I know though.

    ReplyDelete
  9. And why should I prove against a falsehood? Try again.

    Nice try, but msj.t is a feminine participle in this example.

    Yes, it would have secondarily acquired an ahistorical ayin by association with "one". No big whoop.

    As you certainly know, šnʿ was a unit of mass, and it seems quite natural that šinah and šunuh are transcriptions of šnʿ and its broken plural. Do you really think the list should read "one šnʿ-weight", "two", "three šnʿ-weights", "four šnʿ-weights" etc., when šinahwu could more easily be interpreted as the dual of šnʿ?

    ReplyDelete
  10. Michael Hahn: "Nice try, but msj.t is a feminine participle in this example."

    Show how the two forms are separate words despite the identical feminine suffix. Active participles and infinitives aren't distinguished in all languages (ie. Mandarin, Etruscan) and they carry similar semantics afterall.

    "As you certainly know, šnʿ was a unit of mass, [...]"

    I must admit, it sure seems that your only purpose here is to bicker about ANY alternative provided that doesn't fit your narrow Peust model which is riddled with lazy wildcard "v"'s like Loprieno et alia.

    And if this is a misunderstanding about its semantic value, it's a very common misunderstanding, if indeed.

    But ironically, saying "šinah and šunuh are transcriptions of šnʿ" would admit to the very schwa in the unstressed position that you say can't exist.

    ReplyDelete
    Replies
    1. And if this is a misunderstanding about its semantic value, it's a very common misunderstanding, if indeed.

      My understanding is taken from the book you linked in your post (just look at the Dutch translation in the snippet!). At least in the scholarly literature, your interpretation does not seem so widespread.

      Delete
    2. Show how the two forms are separate words despite the identical feminine suffix.

      Haha you got to be kidding :-)

      Delete
  11. Obviously I'm not but your aim is clear: to condescend. You're a dime a dozen online.

    ReplyDelete
  12. For other would-be commenters here, let me make something perfectly clear. I expect commenters to tell me *directly* without snarky overtones how and why something is in error. I'm willing to accept that I (and apparently others) have misinterpreted an inscription. This subject demands a great deal of attention afterall but it's another thing to make offensive remarks like "you got to be kidding" and other remarks that speak of low self-esteem on the part of the participant.

    Got it? Good. Let's move on.

    ReplyDelete
  13. Moving on now, none of this snarky nonsense speaks against *sanáwi > Sahidic snau which shows no sign of an ayin anyway (particularly impossible in the feminine) and the Semitic data still shows *a (Hebrew shnayim שניים). What then?

    Callender's got *sínway for "two", Loprieno's got *sinuwwvj, and Woodard's got *sinuwwaj. This raises important questions: how does Semitic *a correlate with a *u in Egyptian at all? And is Callender an idiot, or is Michael Hahn being unnecessarily hostile towards asking questions, having no fascinating blog of his own to commit to? I believe the latter is true. These questions are legitimate. I have a right to ask them. I also have a right to not be talked down to for doing so because other readers might *also* be wondering the same thing. Let's end the snark and start debating for real (minus the ego). I put myself on the line online every day. Others need to step up to the plate with me too.

    ReplyDelete
  14. My apologies for the "got to be kidding" comment, that was inappropriate. The comment should have been: There is a correlation between syntactic use and the writing of -j, namely that -j correlates with participial use. The same correlation can be seen in the verbs II gem., where reduplication correlates with infinitival use. So I'd say that there is good reason to believe that (perfective, the imperfective form used reduplication) feminine participles and the infinitive were separate forms.

    ReplyDelete
  15. Thanks for your apology and for the otherwise intelligent conversation. It's too bad you won't answer other questions I posed and I get the strong impression that you have some sort of hidden axe to grind with me. This isn't a reasonable debate if only one side feels willing to answer questions.

    So we'll resume one day when you're feeling better. Best wishes.

    ReplyDelete