10 Nov 2011

The reconstruction of the Pre-Egyptian case system

Antonio Loprieno states something confusing to me on page 55 of Ancient Egyptian: A linguistic introduction (1995):
"Also, the ending *-u is still preserved, although functionally reinterpreted, in the forms of some singular patterns as well: when the original stem ended in a vowel, for example *u in *ḥāruw '(the god) Horus,' *-a in *upraw 'form,' or *-i in *masḏiw 'enemy,' the ending was maintained as a glide, often written in good orthography as <-w> in the case of *-aw as opposed to <-ø> in the case of *-iw or *-uw: <ḫprw> =: *ḫupraw 'form,' <ḥfȝw> =: *ḥaf3aw 'snake.'"
Stated more directly, he's claiming that the *w in *upraw was written by scribes according to "good orthography" while strangely ignored in *masḏiw and *ḥāruw despite being present in all these words. It's hard to understand why that would be so. It's rather as if we have *upraw with *w but *masḏi and *ḥāru without. But then this would be inconsistent with what he's stated on the development of the case system from Pre-Egyptian into Old Egyptian.

So it seems that either I'm missing something here or his theory needs a few tweaks. If I ventured an attempt at revisal, perhaps we could try Pre-Egyptian nominatives *ḫaprúwu, *másḏiyu and *ḥārawu. After reduction of unstressed vowels, this becomes *ḫaprūwa /xəpʰˈɾəwə/, *masḏi /'masɟi/ and *ḥāru /'ħaːɾu/ before the case ending was omitted altogether: *ḫaprū, *masḏi and *ḥāru. I contend that only the first word ever motivated writing w. I question its existence altogether in the pronunciation of the second. In the third, 'hawk', I suspect the word was built on the notion of 'that which is above', consisting of *ḥar 'above, upon' and an ancient masculine suffix *-aw, becoming therefore *-u. As such, it couldn't have consonantal w during literate times either since we have only a short vowel. This then explains Loprieno's "good orthography" which now reflects a transparent, underlying reality. No more arcane scribal rules on whether or not to write the trailing semivowel. No more wildcard symbols either, as I've shook my fist at beforehand.


  1. Loprieno's formulation is a bit unclear. Egyptian scribes were a bit inconsistent with respect to -w and -j. All forms msḏj (from masḏiw by assimilation) and msḏ, ḥrw and ḥr, ḫprw and ḫpr are attested. But the theory Loprieno bases his idea on claims that the frequency of the -w version is determined by the vowel. This conclusion is not mere speculation, it is supported by words whose final vowel we can reconstruct from other evidence.

  2. What I'm saying effectively reinterprets in different terms this frequency that Loprieno notes. Loprieno's *-aw/*-iw/*-uw here become *-u/*-i/*-ū in my view. As I've said, the reason behind Loprieno's claimed scribal rule is obscure given his reconstruction, but it becomes clear when we simply accept that the "w" is signalling consonantal w where present. This way, we can reconcile Loprieno's observation with a more natural reason for why scribes are not writing -w where expected. If *masḏi and *ḥāru, then consonantal w is simply absent already by Old Egyptian and what Loprieno is reconstructing is in effect Proto-Egyptian before the Old Egyptian stage.

  3. Okay, so you would assume the pronunciation was ḥaruw in those instances where the -w was written and ḥarū elsewhere, right? This might be possible, as we know next to nothing about how the pronunciation of words was afected by their syntactic environment. Of course, one could go to the texts and check whether the -w version of these words (Schenkel's monographs on "Nominalbildung" and "Pluralbildung" from 1983, references are given in Loprieno's book, have extensive lists of reconstructed nouns) can be associated with certain syntactic contexts.

  4. Sorry, I think I've only managed to confuse things in my last comment. Loprieno's *-aw/*-iw/*-uw should rather become *-ū/*-i/*-u (ie. the order I gave in my previous comment was wrong).

    It's his *-aw in *ḫupraw which I reinterpret as a long *-ū, thereby motivating -w in writing quite naturally. Given *ḥāru with short -u however, one would expect scribes to write the final vowel with less frequency. And what happens when unstressed Old Egyptian *-a, *-i and *-u in turn fall to *-a (schwa)? Then there's doubly little motivation to write a -w in that word because, simply put, it's not there.

  5. Okay, I see, so you assume that, when written, -w in *ḥāru was rather a mater lectionis and did not represent a consonant, right?

  6. Yes, it seems to me that there's room for this kind of variation but on the whole I'd say that it's more probable for a scribe to represent long with w rather than short *u. This would especially be true if long vowels coincide with stress accent.