From: Piotr Gasiorowski
Message: 290
Date: 1999-11-18
----- Original Message -----From: Mark OdegardSent: Thursday, November 18, 1999 9:58 PMSubject: [cybalist] Lenis and Fortis.Piotr writes:
Proto-Indo-European, as reconstructed with the help of the comparative method (I mean the most recent common ancestor of all the IE languages including Anatolian), was a stress language in which syllable strength was chiefly a matter of pitch differences and, presumably, of intensity (loudness).
At this point I think I need another lesson regarding lenis and fortis. Fortis seems to describe volume (loudness), or the intensity of a puff of air. Lenis is the opposite of this, describing a reduction in volume or the intensity of the air flow. If I've gotten this right, then one example of this in English is the difference between an unexploded word-terminal d or t. If you say 'bad', and keep your tongue in place after articulating the d, the effect is quite different from what you do when you remove your tongue and let out a puff of air, particularly if you are being emphatic, in which case the puff of air takes a vowel-like quality. From what I've read, fortis and lenis is phonemic in one or more of the Celtic languages, at least historically (by this I mean you have minimal pair of words, distinguished solely by lenition and fortition).
The contrast between the two pronunciations of bad is between released and unreleased. If you want to hear the distinction between fortis and lenis, say ten and den. English /d/ may be considerably devoiced (also in word-initial position), but it is inherently weaker than /t/. the difference is easiest to hear in the onset of a stressed syllable, where the fortis character of /t/ (extra air-flow) manifests itself as aspiration or as the devoicing of the following sonorant (as in tree, price, clock or queen). It follows that also in English (as well as in German) the fortis/lenis distinction may be contrastive. Polish also has /t/ and /d/, but in their case the only distinctive feature is voicing, /t/ being never aspirated (to an English or American ear it might sound like /d/ ins ome positions).There's a second question I want to ask, but I'm having a very difficult time formulating it. Mostly, it's what Beekes says, in section 11.4.8. I'm speaking of what has traditionally been referred to as the set of 'voiced aspirate' stops. Gamkrelidze et al. are said to reconstruct this as "voiceless aspirated stops, voiceless ejective stops and voiced aspirated stops" (EIEC, 'Proto-Indo-European', p. 461). The second school, lead by Beekes, "have proposed a system of *p, *p’, and *bh " (id.) Beekes himself says "the last two were 'lenis' while the first was 'fortis'" (p.133), with the first being glottalized.
What I'm asking is if lenis/fortis was phonemic in PIE (and so far as I understand all this, it would seem to be). I also wonder if the fortis/lenis distinction needs to be interpreted as part of the PIE stress system.
I can only offer a guess here. As (late) PIE doesn't seem to have had strong expiratory stress, it is a priory unlikely that it correlated stress with consonant strength. I prefer to be agnostic about the minutiae of PIE phonetics and to stick to the traditional transcription system, whatever its detailed interpretation. I admit complete inability to make sense of Gamkrelidze and Ivanov's system (as laid out in their opus magnum), and in particular of positing variable aspiration in the first and third series of stops. Beekes' system is a well-known alternative to analysing the middle series as voiced stops. To be sure, can't see how ejectives (or glottalised voiceless stops) can be regarded as lenis. An ejective is pronounced with the simultaneuos glottal and oral closure. Then the closed glottis moves up and compresses the air between the two closures to produce a sharp puff upon the release of the stop. In terms of the fortis/lenis distinction this is the very epitome of fortis, like the aspiration of voiceless stops.The fortis/lenis contrast became important in all the branches which introduced emphatic stress contrasts (Germanic, Celtic, early Italic); such languages also typically show a preference for initial stress). But Classical Greek and Sanskrit, despite having voiceless /th/ etc., depended on voicing and aspiration as distinctive features. Unlike English, they had no strength contrast between /t/ and /d/. In my opinion, PIE contrasts were of a similar kind.
My third point here is speculative, and may be based on my own ignorance of the facts. Elsewhere in the same article in EIEC, Douglas Adams speaks of PIE's use of grammatical gender. He says
the primary use of gender would seem to be demarcative. The agreement of adjectives an other modifiers with their head nouns delimits the scope of a particular noun phrase. When agreement stops, the speaker is inter alia signaling the end of a noun phrase. (p. 465)
Am I reading him right? Gender would seem to have developed out of a system devised in part to delimit a noun phrase from any other adjacent unit. Once such a feature developed, it's not difficult to see how it would be pressed into other duties, not the least of which is lexical. Here I am engaging in a flight of fancy, taking the thought wherever it leads me. The earliest distinction between masculine and feminine would be an alternation between the O-vowels and A-vowels found in the daughter languages. It would have started as an allophonic variation, something coming out of existing stress patterns, and maybe even one of fortition and/or lenition. In this context, it's almost easy to envision the stages whereby the masc/fem distinction became 'grammaticalized', mostly because you would have had a choice about what variation you would start with. It might have included the distinction found in other languages between women's speech and men's speech.
Can't answer this one now. Maybe a little later.Piotr