17 December 2015

Sex, Greek, and Rix’s Law

A recent comment by David Marjanović made me reflect on a Sanskrit word, yábhati ‘have sexual intercourse’ (as the Monnier-Williams dictionary tactfully puts it). The verb is of special interest to speakers of Slavic languages, because its exact cognate – Proto-Slavic *jebe/o- (with a host of Slavic derivatives) – remains one of the most favourite obscenities in all the languages belonging to that branch of Indo-European. Interestingly, the verb is only very sparsely attested in Iranian and seems to be completely absent from Baltic. In Modern Indo-Aryan its reflexes are quite numerous, though hard to recognise after more than two millennia of sound change, sometimes combined with euphemistic deformation.

By comparing Indo-Iranian and Slavic cognates, we arrive at the stem *jébʰ-e/o- (3sg. *jébʰeti, 3pl. *jébʰonti) as the most parsimonious reconstruction of their ancestral form. It’s a so-called “simple thematic present” – an imperfective stem built to the root *jébʰ-, with the vowel *e in the root and the “buffer vowel” *-e/o- added before personal endings. If the verb has a deeper origin in Indo-European, its oldest form must have been different. Simple thematic presents occur in large numbers in most of the branches of the family; for example, they accout for much of the third conjugation in Latin. However, they are absent from the most outlying lineage of Indo-European (the Anatolian languages), and their low number in Tocharian, the next group that split off before the divergence of the modern branches, shows that they evolved gradually in post-Proto-Indo-European times. Further speculation about the origin of *jébʰ-e/o- via internal reconstruction is difficult because simple thematics have more than one historical sources.

I hope it is not all Greek to you.
[Source: Wikipedia]
When we run out of exact cognates, we can focus on next best thing – plausibly related words with a different morphological structure. Everybody agrees that Ancient Greek oípʰō (with the same meaning) must be a relative of yábhati. Pre-Greek *jébʰ-e/o-, however, would have produced Gk. ˣzepʰō (here, ˣ, not to be confused with the asterisk, marks an unattested, incorrectly predicted form), so the origin of oípʰō must be different. Since the Greek reflex of the root morpheme (oípʰ-) contains an unexpected o, it is justifiable to suspect that one of the Proto-Indo-European “laryngeal” consonants, the one conventionally written *h₃ (probably a voiced pharyngeal fricative [ʕ], if you prefer phonetic symbols) is lurking about. This consonant was vocalised in Greek as o in some positions; it could also (already in PIE) change an adjacent *e into *o. This is why the root we are discussing is often reconstructed as *h₃jébʰ- to accommodate the o-colouring fricative. Unfortunately, most sources just put the laryngeal there and don’t attempt to explain the Greek form in detail.

The trouble is that oípʰō can’t be derived from *h₃jébʰ-e/o- either. According to recent work on PIE syllable structure (Byrd 2015; see also here), the sequence *h₃j- was simplified to *j- in word-initial positions very early in the history of Indo-European, so in this case too we should expect Gk. ˣzepʰō, just as if the *h₃ weren’t there. Some authors propose that *h₃jébʰ-e/o- had a metathetic byform *h₃óibʰ-e/o-, in which *j and *e had swapped places, which caused the latter to get coloured to *o by the preceding *h₃. Such a solution, however, is desperately ad hoc. There is no morphological or phonological motivation for the metathesis, and the wish to see the desired output is not enough.

Another ad hoc solution is adopted in the Lexicon der indogermanischen Verben (Lexicon of Indo-European Verbs, LIV), where the root is listed as *jebʰ-, and its Greek reflex is reconstructed as a present stem with the zero grade of the root and the prefix *o-, that is, *o-ibʰ-e/o-. The problem is that such an alleged verb prefix is vanishingly rare in Greek (so rare that its very reality is questionable), and its function (if any) is unspecified. Solving one mystery by creating another is not sound etymological practice.

A more ingenious suggestion was made by Johnny Cheung in his Etymological Dictionary of the Iranian Verb (2007). Cheung proposes that the Greek present was reduplicated. Grammatical reduplication in PIE involves copying the initial consonant, extending it with the vowel *e or *i, and pasting it back onto the root as a prefix. There are several classes of Indo-European verb stems formed in this way. Following Cheung’s suggestion, we should reconstruct *h₃e-h₃ibʰ-e/o-, which after the laryngeal colouring of the first *e yields *h₃oh₃ibʰ-e/o- and – hey presto – Gk. oípʰō.

Alas, the formation of reduplicated presents is something we understand rather well – well enough to see a couple of problems with this reconstruction. First, although *e may appear as redupllication vowel in IE present stems, it does so only in so-called “athematic” ones (without the *-e/o- suffix). In thematic presents, i-reduplication occurs instead, as in *si-sd-e/o- ‘sit’ > *sizde/o- > Gk. hízō. Secondly, even in athematics, *e seems to have alternated with *i. The details of the alternation are still debated, but one thing is sure: Greek generalised i-reduplication thoroughly in this class, so that we find it in Ancient Greek present stems (thematic and athematic alike) to the complete exclusion of e-reduplication. Therefore, *h₃e-h₃ibʰ-e/o- just won’t float – not in Greek waters.

There remains another possibility, also considered by Cheung but qualified as less likely than the reduplicated root: a zero-grade thematic present, *h₃ibʰ-é/ó-. Such a stem structure is also well-known; one typical example is *gʷr̥h₃-é/ó- ‘devour, swallow’ (Sanskrit giráti, Slavic *žьreti). Both *h₃ibʰ-é/ó- and *(h₃)jébʰ-e/o- (with an early loss of *h₃) could be independently derived from a still older common prototype, most probably a root verb without any suffixes. Why, then, should *h₃e-h₃ibʰ-e/o- be “more likely” than *h₃ibʰ-é/ó- as the source of oípʰō?

The problem here is that we aren’t really sure what happened to initial *h₃i- in the transition from PIE to Ancient Greek. There was a pre-Greek sound change, known as Rix’s Law, which changed any initial *HR̥- into Greek VR-. In these formulae, R stands for any liquid or nasal (l, r, mn), is its syllabic variant, H is any of the three PIE laryngeals, and V is a vowel whose quality matches the phonetic “colour” of the laryngeal (e, a, o for, respectively, *h₁, *h₂, *h₃). To what extent the sequences *Hi- and *Hu- were also affected by Rix’s Law has been a matter of some dispute. PIE *i, *u can be regarded as syllabic variants of the corresponding glides *j and*w; therefore, it is at least thinkable that Rix’s Law could apply to them as well.

As for the sequence *Hi-, however, it can be demonstrated with good examples that no initial vowel developed if the laryngeal was *h₁. It has furthermore been suggested that the outcome could be Gk. hi- (with an initial aspirate) rather than simply *i- (Bozzone 2013). For *h₂ and *h₃ the evidence is inconclusive (no unambiguous examples). But there is no clear counterevidence either to rule out *h₂i- > Gk. ai- or *h₃i- > Gk. oi- (pace Peters 1980¹, who argues for *Hi- > Gk. *i- across the board). As for *Hu-, we have several convincing cases showing that *h₂u- > Gk. au-, one or two possible cases of *h₃u- > Gk. ou-, but no examples at all of *h₁u- > Gk. ˣeu-. This may mean that the Greek reflexes of *h₁u- are indistinguishable from *u- since both merged as Gk. hu-, while the other two laryngeals followed the pattern of Rix’s Law. It is therefore possible that *Hi- and *Hu- developed in parallel, and that the expected outcome of *h₃i- is Gk. oi-.

This insight has far-reaching consequences for our understanding of the various combinations of *i/*j and *u/*w with the laryngeals in the prehistory of Greek, but I can only skim the surface of the topic in a blog post. It’s getting too long anyway, so it’s time for the moral. The hero of this little essay is a swear-word so obscene that some old ladies in my country might faint if they saw it printed in a newspaper. On the other hand, you can hear it all the time in the street, adorned with modifying prefixes, converted into derived nouns, adjectives and adverbs, and spawning lots of specialised meanings. It has functioned like that literally for millennia – taboo or no taboo. Living the merry life of an outlaw, it has become a respectable archaism, almost a living fossil, with an impeccable pedigree and aristocratic Vedic connections. Together with its equally naughty Ancient Greek cousin, it may provide a precious piece of crucial evidence needed to solve a vexing problem in Greek historical phonology. Not bad for a dirty little word.

———

¹ Martin Peters. 1980. Untersuchungen zur Vertretung der indogermanischen Laryngale im Griechischen. Vienna: Verlag der Österreichischen Akademie der Wissenschaften.

[See also: REDUPLICATION]

102 comments:

  1. On *h₃e-h₃ibʰ-e/o-: One could argue that the reduplication vowel wasn't replaced by /i/ in this word because due to the colouring effect, this formation had already become opaque when the regularisation of the reduplicated presents happened. On the other hand, wouldn't one expect a long /o:/ due to the second laryngeal? Or am I missing something?

    ReplyDelete
  2. The *o would remain short because the following laryngeal is not tautosyllabic. The contraction of disyllabic *-o.i- to Greek oi- is not a big problem. But i-reduplication is original in thematic presents, so *h₃e-h₃ibʰ-e/o- is aberrant either way. Some old reduplicated thematics became "de-thematised" in Greek on the analogy of the τίθημι type, but not the other way round.

    ReplyDelete
  3. "But there is no clear counterevidence either to rule out *h₂i- > Gk. ai- or *h₃i- > Gk. oi- "

    These from the Bozzone paper look pretty good, no?-

    • Gk. ἰθαρός ‘cheerful, bright’ < PIE *h2eydh- ‘kindle’, i.e., PIE *h2idh- > Proto-Greek *ith-. Compare the full grade in αἰθήρ ‘clear sky’.

    • Cypriot Gk. ἰκμαμένος ‘wounded’ (ICS 217.3, spelled i-ki-ma-me-no-se) < PIE *h2eyḱ- ‘pierce’. Compare the full grade in αἰχμή ‘spear point, spear’ (§1.3.2).

    Is it primarily the semantics that you find less than convincing?

    ReplyDelete
    Replies
    1. Not very convincing. For *h₁i- we have unambiguous conjugational forms of 'go' and 'throw' (I would also add zero-grade forms of *h₁eish₂-). For the other laryngeals we only have potential root-equations. And how do we know that αἰθήρ (note the suffixal accentuation) shows a zero-grade? What if it reflects *h₂idʰḗr?

      Delete
    2. Oops, I wanted to write "shows a full grade".

      Delete
    3. This is only tenuously connected to the topic, in that it relates to laryngeals and reduplication, so feel free to ignore, but...

      I was reading your “Meaning of Life” paper - a wonderfully compelling hypothesis, and the apparent relics of pre-PIE verb of motion morphology are particularly exciting. I have a few questions:

      1. What do you think would motivate the change of *gwi-gwu̯-ó- > *gwi-ɣu̯-ó- ? You don’t spend much time talking about this. Compared to e.g. *gwi-gwáh2- › Ved. jigā́ti, one obvious difference is the presence of the -u̯-. One could imagine de-labialising of /gw/ in that environment, as in Latin, but I’m not sure how that would help. Is this kind of dissimilating lenition observed elsewhere in PIE?

      On the other hand, those who don’t accept your theory have to motivate ‘laryngeal hardening’ in Gmc. *kwikwa- and Latin visor, which seems like an exactly equivalent problem!

      2. When talking about root aorists, you say in a footnote "Lat. perf. vixī and p.p. victus may owe their velar stop to the influence of forms preserving their -g-…” But how could this develop from a root aorist of *gweu̯- ? I don’t see where the second /g/ is coming from if not from the reduplicated form? I'm certain I'm being dumb here, but I can't work out how...

      3. What is the mechanism for Latin retaining reflexes of both the *gwi-gwu̯-ó- and the *gwi-ɣu̯-ó- forms? With Germanic it’s easy to imagine because it only (I think?) has the former. But for Latin to have both forms, would we need them both to survive post-PIE, but then one or the other die out in all branches except Italic? Or are you sceptical of the notion of singular proto-daughter-languages, along the lines of Andrew Garrett's work?

      4. WTF (that’s my effort to stay relevant to the post topic) is going on with the Greek reflexes? I can see why you’d get ζ- < *gwí-, with or without vowel breaking, and β- from *gw- with a non-front vowel as in βοῦς (or in fact with any vowel except /i/, judging by the Homeric βέ- forms). But then why βίος < *gwiu̯-ó- (or < *gwíh3-o-s)? And then /g/ strolling all nonchalant into the picture in ὑγιής < *h2i̯u-gwih3ēs ???

      5. Re the archaic correspondences - don’t we also have *stah2- (stand) | *stem- (push) | *steu- (push) - semantically less pleasing, admittedly, but on the other hand in Slavic (or at least Russian) the stand/sit words do pattern with the verbs of motion...

      6. Finally, has the paper had much traction yet? Do you get the sense that others in the field are persuaded by the *gwi-ɣu̯-ó- idea and its link to cow and motion words?

      Delete
    4. Sorry "visor" was meant to be "vigor" in para 4

      Delete
    5. Lotsa questions there :). I don't mind answering them all, but it's 1 a.m. in my time zone, so let me get some sleep first.

      Delete
    6. Let me begin with (4). The dialectal development of old labiovelars in Greek is messy. It is well known that they generally changed into labials in Aeolic also before front vowels (though even Aeolic has τίς, τε). It is less known that in Attic-Ionic the regular development of the voiced labiovelars *gʷ, gʷʰ is β, φ, not δ, θ, although the latter is regular before short and long *e (β in this position is due to analogy or interdialectal borrowing). Therefore, βίος is regular. Palatalisation did take place before the palatal glide *j, which is why we get *gʷih₃wós > *gʷjōwós > ζωός. In ὑγιής (no matter if the first element is *h₂ju- or *h₁su-, the initial * of the second element was preceded by *u, which caused it to delabialise in accordance with the so-called βουκόλος rule (*gʷou-kʷolh₁o-).

      Delete
    7. the regular development of the voiced labiovelars *gʷ, gʷʰ is β, φ, not δ, θ,

      I mean, before the high vowel /i/.

      Delete
    8. Thanks!

      That's a bit odd, palatalising before /e/ but not /i/...

      Delete
    9. Or rather not palatalising, but fronting or whatever you call it.

      Delete
    10. Palatalisation was certainly the first stage before they got fronted still further (affrication and then de-affrication probably also happened in the process). It's odd, too, that labiovelars were affected while plain velars weren't (as in Albanian), but that's what we see in the data. There's a lot of variation from dialect to dialect. Anyway, the failure of the fronting of voiced labiovelars before /i/ and /iː/ in Attic-Ionic is supported by several sure examples beside the 'live' word-family. Where * is not affected and analogy can be ruled out, as in Homeric πίσυρες 'four', Aeolic is usually blamed, though, OTOH, I don't think πίσυρες is directly attested in any Aeolic dialect.

      Delete
    11. OK, some other questions:

      1. When *gʷig(ʷ)- came to be treated as a "neo-root" rather than a reduplication, its structure was perceived as aberrant (normal IE roots don't contain two mediae), so it was "repaired" by replacing the second stop with the phonetically closest voiced fricative available in the system. The change of [g] to [ʕ] or a similar sound is common cross-linguistically (Ukrainian has it, for example).

      6. It takes time. There hasn't been much reaction so far, but the article has been downloaded hundreds of times from the repositories where I placed it, so I hope some of my colleagues have read with it and in due time will find it worth their time comment on the idea. My etymology was cited by Guus Kroonen in his etymological dictionary of Germanic with partial approval. Guus's own analysis is slightly different, but at least he bought the idea of a reduplicated present.

      Delete
    12. 2. I mentioned those forms before the putative root aorist. My suggestion is that they are innovations based on *gʷig(ʷ) (no longer understood as a reduplicated present but as a unitary root).

      3. It's normal (and inevitable) for innovations to coexist with their predecessors. Competition between related variants may last a long time, and there is no reason why a daughter language should not have inherited both variants. The result is incomplete sorting of ancestral variation. Some descendant languages may eliminate the older form, some may retain it and eliminate the alternatoin, and some may keep both, especially if they manage to specialise by developing different meanings.

      5. An interesting suggestion. I haven't given the matter much thought since I published that article. Perhaps I should have another look at the data.

      Delete
    13. Got it, and the explanations are very much appreciated.

      Delete
    14. alternatoin

      Why the heck did I post that? I meant innovation :(

      Delete
    15. That's a bit odd, palatalising before /e/ but not /i/...

      Could /e/ have become /je/ before /j/ was abolished?

      Ukrainian has it, for example

      AFAIK, Ukrainian (and everything from there to Upper Sorbian) has [ɦ], derived from earlier [ɣ] that is preserved in Belorussian and southern Russian (and only recently died out, in favor of [g], in the Russian prestige accent). Is there any evidence that an intermediate step involved retraction of the back of the tongue and was then undone?

      Delete
    16. Ukrainian г is often described as a voiced pharyngeal fricative. I have never seen X-ray images of its articulation. The sound of friction is weak (unlike, say, Arabic /ʕ/) but audible, so the sound could probably be described as a lax, breathy-voiced pharyngeal fricative (but not quite an approximant or a purely glottal sound): Ганна.

      Delete
    17. If I'm not mistaken, Dutch and Galician had similar developments, although into voiceless fricatives.

      Delete
    18. I'll try to listen tomorrow!

      I have no idea about Galician. In Dutch, I have heard g before a vowel as [ɣ], [x] and [χ] from different people, likely in a south-to-north sequence – the voiceless versions result from a merger into ch.

      Delete
    19. And to correct myself: I've recently heard some Czech – while [ɦ] exists, other people use a short [x] instead, even between vowels (Praha). That was a surprise.

      Delete
    20. Ганна.

      Finally had an opportunity to listen! I think this begins with [ʔɦ] with delayed voice onset... there is [ɑ]-coloring, but glottal consonants do that to some extent... I lack the equipment to cut the annoying [ʔ] off and listen again without it. I sometimes insert [ʔ] myself in front of utterances that otherwise begin with [h] (or, actually, the preaspirated version of this thing).

      My experience with pharyngeal consonants is admittedly very limited. Here is a bunch of examples where I expected a voiced pharyngeal fricative and basically couldn't hear it at all; it turned out to be an epiglottal approximant...

      Delete
    21. Again, we are dealing with a continuum in three dimensions: place of articulation (pharyngeal -- epiglottal -- glottal), degree of opening (fricative -- approximant) and phonation (breathy -- modal). An instrumental study of the Ukrainian fricative (with a decent sample of native speakers) would be useful, but none seems to have been done so far. I know some laboratory phoneticians who might be interested in conducting such a study.

      Delete
    22. Epiglottal is different, though. Epiglottal consonants pull vowels towards [æ], not towards [ɑ]; sound files here.

      Delete

    23. The proverb ἄριστα χωλὸς οἰφεῖ (cf. "dumm fickt gut"), if its accent has been transmitted correctly in comic fragments, belongs to οἰφέω rather than οἴφω. The form OIΠEI from the Law Code of Gortyn (2:3), with Π standing for Φ, although edited as οἴφει or οἴπει, could equally well represent οἰφεῖ, and the infinitive OIΠEN (ib. 2:17) could represent οἰφεῖν. The only basis for presuming οἴφω appears to be the imperative οἶφε in Plutarch (Pyrrhus 28:3), which is not above suspicion of being analogical. After his heroic actions defending Sparta against Pyrrhus, Acrotatus was encouraged by some of the elders to go and copulate with his woman Chilonis. Plutarch quotes them as shouting "οἶχε, Ἀκρότατε, καὶ οἶφε τὰν Χιλωνίδα". Perhaps "οἶχε καὶ οἶφε" i.e. 'go and fuck' was a somewhat vulgar collocation which acquired rhyme like "rough and tough" (these adjectives don't rhyme in OE or Chaucer) or "cows and sows" (these rhyme in East Norse, independently of English, but not generally in Gmc.).

      If οἰφέω is the historically correct form, it could stand to the missing *ζέφω as zero-grade βδέω 'I fart' (*bzd-) stands to full-grade *pezd- (Latin pēdō). This of course would require that PIE *h₃i- becomes Greek οἰ-, and presumably also *h₂i- > αἰ-, despite *h₁i- becoming ἰ- not εἰ-. I see no inherent problem with such an asymmetry in Rix's Law. Robert Woodhouse has commented on a similar asymmetry in Francis' Law (https://www.academia.edu/14184890), see esp. note 15. PIE *h₁ was the most recessive laryngeal, while *i and *u were the most vocalic resonants. It's not surprising that *h₁ should lose its laryngeal force in these environments before *h₂ and *h₃ did.

      Delete
    24. I'd rather say that *h₁ didn't become a vowel; the "reflex" ε is epenthetic, and no epenthetic vowel was needed in *h₁i-. *h₁ really must have been [ʔ], which at some point wasn't perceived as a phoneme any longer, or perhaps [h].

      Delete

    25. I see no advantage to positing epenthesis in Greek, or regarding *h₁ as a stop. If a Greek word began with too many consonants, the cluster was simplified, not epenthesized. Thus *pk̂t- > *kt- in κτείς, *bzd- > *bd- in βδέω. Πτάρνυμαι provides no good evidence that *pst- > *pt-, since it could have been remodelled after πτύω. But στήνιον corresponds to Av. fštāna- and illustrates *pst- > *st- in Greek. If laryngeals were stops, one would expect *h₂st- to be similarly simplified, not vocalized to *ast- in ἀστήρ. Insisting on epenthesis of a neutral vowel anyway, to be colored by the laryngeal before the latter vanished like the Cheshire Cat (i.e. *h₂st- > *h₂əst- > *h₂ast- > *ast-), only complicates the situation unnecessarily. Evidently the laryngeals stood between resonants and stops on the scale of sonority, placing them in fricative territory (generally speaking, in the range from lax approximants to affricates). If *h₁ alone among the laryngeals were a stop, one would not expect θεός from *dʱh₁s-ós, but simplification of the cluster *dʱh₁s-. Also the fact that *h₁ alone among the laryngeals fails to color *e points to sub-oral, that is glottal, articulation. Had *h₁ been the glottal stop [ʔ], one would not expect PIE *h₁jós 'which' to become Greek ὅς. No plausible mechanism would devoice the cluster *ʔj-. But *sjo- does become ὅ-, so *h₁ was presumably an unvoiced fricative like [s], namely the glottal fricative [h].

      Delete
    26. Sorry about the confusion – the idea isn't that the epenthesis happened on the way to Greek, but that it was already present (just not phonemic) in PIE.

      If laryngeals were stops

      *h₁ may have been, the others clearly were not.

      Had *h₁ been the glottal stop [ʔ], one would not expect PIE *h₁jós 'which' to become Greek ὅς. No plausible mechanism would devoice the cluster *ʔj-. But *sjo- does become ὅ-, so *h₁ was presumably an unvoiced fricative like [s], namely the glottal fricative [h].

      This is a good argument that *h₁ was [h] shortly before it was lost in Greek. However, that doesn't tell us if it was [ʔ] in PIE. Such a sound shift from [ʔ] to [h] has happened in several modern descendants of Classical Nāhuatl.

      Delete
    27. I sometimes insert [ʔ] myself in front of utterances that otherwise begin with [h]

      And so did Hillary Clinton when she said she always carried "hot sauce" in her pocket. I think it's an aborted attempt to say "...uh,".

      Delete
  4. Is it archaism in Polish? I have heard other words in this meaning, e.g. in the pun " rapier doła go zarzucił", but I was not sure which one is the basic word. BTW, is there some deeper etymology of the word? Later euphemisms often derive from verbs meaning 'strike,hit'.

    ReplyDelete
    Replies
    1. OK, I misused the term "archaism" for rhetorical effect. I just mean it has a very long history and retains the original form and meaning very faithfully. Of course it's neither antiquated nor obsolescent. Quite the opposite; it has perhaps never been as vigorous as it is now, judging from the number of recent derivatives (a few have been formed within living memory and are already extremely popular, e.g. pojeb 'an irritating nutcase' or zajebisty 'cool, fabulous'.

      There probably is a deeper etymology, and reflexes of a possible ancestral form can be seen in Tocharian, where they mean 'enter, penetrate'. It's a one-way semantic path, since 'fuck' > 'enter' (in innocent literal senses) is an unlikely development, so the Tocharian verb is probably closer to the common ancestor.

      Delete
    2. I see..."Power" verbs obiously can develop figurative meaning in almost any direction. Russians also exclaim zajebis' for somenthing cool, while in Bulgarian the same prefix forms the meaning 'quit, abandon'.
      I tried to find etymology in Starostin's database /http://starling.rinet.ru/ - in vain.
      Thanks anyway

      Delete
    3. It can be found there. Try this link:

      Link

      WARNING: They use an outdated reconstruction (from Walde-Pokorny's IE dictionary), ignoring the laryngeal theory. Some comparative material, also taken from that old dictionary, shouldn't be there at all.

      Vasmer's Russian etymological dictionary, which is also included in the database, omits ebat' ~ et' (the old form of the infinitive) just like the original edition of the Oxford English Dictionary ignored fuck.

      Delete
    4. Thanks! It wouldn't occur to me that the "English" key word is futuere.

      Delete
    5. 'fuck' > 'enter' (in innocent literal senses) is an unlikely development

      Occupy, no?

      Delete
    6. reflexes of a possible ancestral form can be seen in Tocharian, where they mean 'enter, penetrate'

      I was quite disappointed to find out that the most vigorous obscenity of the whole language family once was a euphemism.

      It wouldn't occur to me that the "English" key word is futuere.

      Turns out the "Russian meaning" is futuere, too... There are languages in the Caucasus where the words for genitals and what to do with them are only known to science from 19th-century wordlists written by German explorers (again with meanings given only in Latin), because the Soviet linguists/anthropologists/ethnographers were too prude to record them at all.

      Delete
    7. Occupy meant 'fuck' in EModE, but now has lost all trace of that sense in favor of "innocent" senses. Florio's 1598 Italian-English dictionary defines fottere as "To iape, to sard, to fucke, to swive, to occupy". But on investigation (meaning looking in the OED), the obscene sense is about contemporary with most of the others (the sense 'employ' is half a century older), around the end of the 14C or beginning of the 15C, allowing for the fact that it's less likely to be found in print.

      Ah well, mammals and dinosaurs arose around the same time too.

      Delete
    8. Tabooed in China (for a different reason): link.

      As Doll Tearsheet remarked in Henry IV,

      A captain! God's light, these villains will make the word as odious as the word 'occupy'; which was an excellent good word before it was ill-sorted.

      She was of course a good wench, one that occupied freely, as Florio put it.

      Delete
    9. BTW, it makes "Occupy Wall Street" an unintended historical pun for the erudite.

      Delete
  5. But i-reduplication is original in thematic presents, so *h₃e-h₃ibʰ-e/o- is aberrant either way. Some old reduplicated thematics became "de-thematised" in Greek on the analogy of the τίθημι type, but not the other way round.
    Hmmm... my opinion on this is that basically all reduplicated thematic verbs ultimately go back to reduplicated athematic verbs; some of them became thematic already in PIE, others only in the daughter languages. How do we know that with the Greek examples the athematic forms aren't retentions? On *h₃e-h₃ibʰ-e/o- specifically, my idea was that it was formed as an athematic verb already in PIE and then became thematic at one stage either in late PIE or on the way to Greek. I also assume that the neat distribution of the reduplication vowel (/i/ in the present, /e/ in the perfect) is a late development that started in PIE but was finalized only in the daughter languages, a development in which *h₃e-h₃ibʰ- didn't participate because, due to colouring, it had already become *h₃o-h₃ibʰ- and wasn't seen by the speakers as belonging to the reduplicated type any more (a bit like post-laryngeal *pibe/o-, where many daughter languages didn't "correct" the irregular reduplication consonants).

    ReplyDelete
    Replies
    1. I would have to lay out my opinion about the origin of i-reduplication to reply in full. I'll probably do that in one of my future blog post. In a nutshell, I agree that reduplicated thematics are secondary with regard to athematics. Still, unlike simple thematics and "Brugmannian" perfects, they are well represented in Hittite (and the reduplication vowel is invariably /i/ also there), so they must have existed already in PIE. IMO, the reduplication vowel *i developed from earlier *e through regular sound change in pre-PIE. The change is arguably older than laryngeal colouring, because there are presents like *h₂í-h₂ǵ-e/o- (Ved. ī́jate 'go'). In any case, it seems that the meaning 'copulate' developed out of 'enter' in the "Inner IE" group after the separation of both Anatolian and Tocharian. Even if an extremely conservative form like *h₃o-h₃ibʰ-e/o- had somehow survived, the semantic change would have had to take place both in the reduplicated stem and in the Inner IE simple thematic *jébʰ-e/o-, ultimately cognate but neither derived from, not transparently related to the reduplication (in synchronic terms).

      I find it easier to believe that the common ancestor was a root verb (probably an aorist) *(h₃)jébʰ-/*h₃ibʰ-, which developed the secondary meaning of 'penetrate (you know what)' in Inner IE and gave rise to productively formed alternative thematic presents, exactly like Vedic tiráti and tárati, both from the old root aorist *térh₂/*tr̥h₂- 'get through'. *h₃ibʰ-é/ó- survived in Greek and *jébʰ-e/o- in the nuclear Satem group.

      This, at any rate, is the best I can come up with as a plausible scenario.

      Delete
    2. Vasmer's Russian etymological dictionary, which is also included in the database, omits ebat' ~ et' (the old form of the infinitive) just like the original edition of the Oxford English Dictionary ignored fuck.
      Exactly. The word is also excluded from the Oxford Dictionary of English Etymology, first published in 1966, which I came upon it recently.

      Delete
    3. Dictionary editors should support swearwords:
      http://languagehat.com/people-who-curse-have-better-vocabularies/

      Delete
    4. Somebody there mentioned the Baltic thunder-god Perkūnas derives from *perkʷu- 'oak', but in my opinion this is only a homonymy.

      Delete
    5. I hope I'm not flogging a dead *H1ek'wos here...
      I am with you that i-reduplication in the present is common IE, but I assume that it's just the deictic "hic et nunc" /i/ that we also have in the primary endings. (I'm also extremely wary of attempts to explain /i/ and /u/ as old stress variants of the /e/ /o/ series - for my taste, this gets too close to "everything is possible".:-) )
      We have evidence for /e/ and /i/ being used for reduplication in the present tense in PIE, and *h₃e-h₃ibʰ- may have been formed when that process was still productive, so, in my view, the only problem with deriving oípʰō from *h₃e-h₃ibʰ- is that an e-reduplicated present of *H3yebʰ- could only be an old PIE formation, but is attested only in Greek.

      Delete
    6. I am going to devote the next post to those reduplications (and perhaps a few furter posts to reduplication in general, not only in PIE). I'll explain my opinion there. Sorry if it takes some time, but Christmas is coming and I have to do my part of house-cleaning, cooking, etc.

      Delete
    7. No need to apologise! Wesołych Świąt!

      Delete
    8. The change is arguably older than laryngeal colouring

      I don't think laryngeal coloring happened at any particular time. AFAIK, all languages that have epiglottal, pharyngeal or uvular consonants and also have small vowel systems have vowel coloring. While PIE didn't have a particularly small system of unreduced vowels, there was only one reduced vowel, *e...

      Anyway, the reduced vowel of Classical Latin was i, and the reduced vowel of my German dialect is usually [ɐ], but [ɪ] between two /s/; (further) reduction of *e (colored or not) to *i wouldn't be an outright "crazy rule".

      (Languages with larger vowel systems and uvular consonants adapt the latter to the former. In Lakhota, /ʁi/ comes out as [ʀi].)

      Delete
  6. Indicentally, I think the use of the Greek alphabet for transcribing Ancient Greek could be legitimate in Classics but in IE studes is old-fashioned and rather snobbish.

    ReplyDelete
    Replies
    1. I don't do that in the OP, but copying a Greek word from a dictionary is less trouble, so excuse my old-fashioned snobbery in the comments.

      Delete
  7. would have produced Gk. ˣzepʰō

    Indeed no such verb seems to be attested, but the expected root zepʰ is – with a meaning that fits pretty well.

    Anyway: why do you think *h₃ was pharyngeal rather than uvular?

    ReplyDelete
    Replies
    1. I'm really agnostic about that. Something like [ʁʷ] would work as well. Likewise for *h₂. A pharyngeal *h₂ would more naturally account for a-colouring, but [χ] is also thinkable. It's also quite possible that the pronunciation varied in time and space. Even in living modern languages "guttural" consonants can confuse phoneticians.

      Delete
    2. It's also quite possible that the pronunciation varied in time and space.

      Oh, certainly.

      A pharyngeal *h₂ would more naturally account for a-colouring, but [χ] is also thinkable.

      Uvular consonants cause a-coloring all over Semitic, Quechua and Eskimo-Aleut, and show up as the allophones of velar consonants in back-vowel words all over central Eurasian vowel harmony, so this isn't a reason to favor one interpretation over the other. Indeed, I've seen the fact that PIE *a seems less rare than usual next to plain velars used an argument by those who interpret them as uvulars (something I disagree with for other reasons).

      The arguments for a specifically uvular articulation are all weak, but there are some:
      1) ...I was sure I had read in a Kümmel paper that uvulars are typologically more likely than pharyngeals, but now I can't find it. (He certainly assumes uvulars in several works, but without explanation.) Anyway, this can't be a particularly strong tendency.
      2) In the alphabetic Anatolian languages, the reflexes of *h₂ and *h₃ were written with letters for velar plosives, derived from κ, Ϟ (q), χ. That points toward [x] or [χ], rather not [ħ] for which I'd expect η (h) at least sometimes. In the cuneiform ones, they're uniformly written with syllables with , which, if taken literally, should mean [χ] or maybe [x], but there wasn't any other obvious way of representing [ħ]. Anyway, none of these languages was PIE, and Proto-Anatolian wasn't PIE either.
      3) There are some hints that *h₂ comes from a Proto-Nostratic (or thereabouts) *q. I've posted on Languagehat about PIE *(h₁)éǵoh₂. Another comparison I find intriguing is the sun: PIE *sah₂w-l/n-, Proto-Altaic *si̯ògu-n-lV, Proto-Eskimo *ciqi-nəʁ. In a sad case of irony, the Moscow School has barely noticed this, because they still get their PIE roots from Pokorny; they're aware that laryngeals should be reconstructed, but they seem unsure of how many to reconstruct and where to put them! Mudrak (2008), my source for the comparison of PA and PE, raised the possibility that the rare correspondence of PA *g to PE *q could indicate an earlier *q, but didn't mention the PIE form at all, and G. Starostin (in a more recent paper which I can't find right now) added the PIE form as "*sa(H)wel", IIRC, without drawing any conclusions from it... Anyway, if *h₂ comes from a [q], that doesn't guarantee it was still uvular in PIE, although that's certainly the most parsimonious hypothesis.

      Delete
    3. Oh, I forgot to talk about *h₃. Certainly it's tempting to consider it labialized, but the other labialized consonants didn't cause o-coloring in PIE, "or the queen's wedding would be the quoon's wadding" as someone put it. But it's not like I have a better idea to offer. Perhaps the choice of [ɑ]-coloring for *h₂ vs. [ɒ]-coloring for *h₃ was ultimately arbitrary... :-/

      Delete
    4. Yes, I agree that the "plain velars" were actually uvulars if not in PIE proper, then at least in its not-too-distant past. I would actually argue that they also had a colouring effect on short *e in some positions.

      Delete
    5. Anyway, none of these languages was PIE, and Proto-Anatolian wasn't PIE either.I'm afraid "PIE" didn't exist at all, at least in the way it's commonly thought. :-)

      Delete
    6. Oh, I forgot to talk about *h₃. Certainly it's tempting to consider it labialized, but the other labialized consonants didn't cause o-coloring in PIE, [...]. Perhaps the choice of [ɑ]-coloring for *h₂ vs. [ɒ]-coloring for *h₃ was ultimately arbitrary...
      I don't think so. For example, IE *h₂arh₃- 'to plough' corresponds to Semitic *ħaruθ- 'to till, to plough'. Hence we've got IE *h₃ ~ Semitic *u.

      Delete
    7. Yes, I agree that the "plain velars" were actually uvulars if not in PIE proper, then at least in its not-too-distant past.

      I really don't think so. The reasons I know that have been given for this are:
      1) Several proponents (and indeed several other people) seem to believe that the "palatovelars" are traditionally thought to have been actual palatals. Palatal plosives are globally rare, so they'd be an extraordinary hypothesis that would require extraordinary evidence, and the kentum merger would mean that palatals become velars, which is rarer still (...apparently there's a branch of Samoyedic where palatalization ran backwards so that *si became ki). Both problems evaporate when palatalized velars are assumed instead.
      2) A system with palatalized, plain and labialized velars is also a bit crowded, though. IIRC, some seem to believe this is impossible. It is found, however, in Hausa and several West Caucasian languages – and PIE is widely acknowledged to show contact effects with West Caucasian.
      3) Within such a system, the plain velars should be the least marked ones. Yet, they're much rarer than the other two series. I simply deny that this is a problem: in the ancestry of Ubykh, the plain velars were so rare that they merged into the palatalized velars, not the other way around! (The gaps were then filled by loanword phonemes from Turkish and Adyghe, BTW.) This rarity makes perfect sense if such systems arise from transfers of frontness and roundedness from vowels to consonants. The globally most common number of vowels is five, of which only one is neither front nor rounded.
      4) The attraction of PIE *a to plain velars can be immediately explained as vowel coloring if they were uvulars. But that makes it very hard to understand why the coloring didn't happen in all environments. Coloring by [q] in attested languages is not so picky.

      Positing that the plain velars were uvulars also creates additional problems that weren't there before:

      5) If the "palatalized velars" were plain velars, and the "plain velars" were uvulars, were the "labialized velars" velar or uvular? They seem to differ from the "plain velars" in only one feature, not two, but a system with plain and labialized uvulars but only plain velars would be quite strange.
      6) [q] is rare globally, but not terribly so. [ɢ], on the other hand, is terribly rare (at least as an independent phoneme); and [ɢʱ] doesn't appear to be attested at all. The implosive [ʛ], incidentally, is attested from one language known to Wikipedia.
      7) When uvular plosives are lost as such, they sometimes become velar plosives. At least as often, however, [q] becomes [ʔ], and [qʰ] and especially [ɢ] become fricatives. If PIE had uvular plosives, why have they remained plosives in every single branch? I can find only one possible partial exception, namely Germanic (Vennemann has jumped on this), where *[q] > *[qʰ] would have become *x (Grimm)/ (Verner) and *[ɢʱ] would have become , but even there the high-maintenance [ɢ] would have become *k across the board in all environments. (And we'd have to assume that the kentum merger happened after Grimm – which is possible according to none lesser than Ringe, but not very parsimonious.) Why aren't there mergers of PIE *k with *h₁ and/or *h₂ and/or zero all over the place? Why aren't there mergers of *g with *h₃ or zero? Why aren't there mergers of *gʰ with *h₃, or rather, why can it be reconstructed at all? – Proposing that the uvular plosives became velars already in Pre-PIE would solve that problem, but internal reconstruction is unconstrained; it is hardly testable as long as external cognates or sufficiently early loans remain unidentified.

      Delete
    8. I forgot: I need to explain the occurrences of *a next to plain velars as retentions of an older *a that presumably became *e in some but not all environments and/or by whatever processes of analogy after the transfers of frontness and roundedness were completed. Alternatively, I'd need to assume a second vowel that was neither front nor rounded and merged into *e.

      I'm afraid "PIE" didn't exist at all, at least in the way it's commonly thought. :-)

      I know, and I disagree, because there's just no way contact could result in all these intricate and systematic regular correspondences between the sound systems and the grammars.

      For example, IE *h₂arh₃- 'to plough' corresponds to Semitic *ħaruθ- 'to till, to plough'. Hence we've got IE *h₃ ~ Semitic *u.

      Perhaps. But:
      1) In a loan from Semitic to IE, **h₂arw- would be a superior option in every respect, so why wasn't it taken? In a loan from IE to Semitic, why weren't the three consonants of the root identified as such? You could posit an unknown third party that loaned the root to both, but there you'd risk ending up with too many degrees of freedom to render the hypothesis testable unless you can identify that source. And if the root is actually cognate, you'd need to explain why it developed from an appropriate Paleolithic meaning ("sharp stick", "scratch"?) to "plough" in both families; identifying potential cognates with such meanings in other families would help.
      2) Alternatively, the *h₃ could correspond to the . The chain of inference for that would have to be a bit tenuous, though: *[θ] > *[ð] > *[ɰ]...?
      3) Finally, perhaps only *ħar- was borrowed around, and both IE and Semitic added a root extension independently. I think this is the most likely option, and of course Semitic is famous for extending almost all roots from 2 to 3 consonants... but the meanings/functions of root extensions in both families are unknown, so this is a case of obscurum per obscurius at least for now. Is *h₃ at least a known root extension in IE?

      Delete
    9. I wrote earlier:

      There are some hints that *h₂ comes from a Proto-Nostratic (or thereabouts) *q.

      In some cases, that is. Most likely *h₂ also had other sources.

      Delete
    10. I see your points. Still, the colouring effect of "plain velars" seems real enough. I have my own theory on how it was constrained, but since this is something I'm preparing for publication, I wouldn't like to say too much. Anyway, velar/uvular is a continuum of articulations, not a bipolar contrast; so is palatal/velar. Fronted velar vs. retracted velar (a.k.a. prevelar/postvelar) is a perfectly possible contrast (with the labiovelars not specified for backness).

      Delete
    11. 1) In a loan from Semitic to IE, **h₂arw- would be a superior option in every respect, so why wasn't it taken?
      Good point. Possibly the labiovelar glide /w/ was hardened into a voiced labiovelar approximant /ɣʷ/ like in Welsh and some Romance languages.

      Delete
    12. And if the root is actually cognate, you'd need to explain why it developed from an appropriate Paleolithic meaning ("sharp stick", "scratch"?) to "plough" in both families; identifying potential cognates with such meanings in other families would help.This is a neat case of a Neolithic Wanderwort belonging to agricultural lexicon.

      and PIE is widely acknowledged to show contact effects with West Caucasian.I won't call it "PIE" but "Kurganic", one of the layers which makes up the IE family.


      Delete
    13. Still, the colouring effect of "plain velars" seems real enough.

      Fair enough, I'm looking forward to your publication.

      Fronted velar vs. retracted velar (a.k.a. prevelar/postvelar) is a perfectly possible contrast

      Sure; however, a contrast between front velars (not palatalized) and back velars (not uvular) seems to be attested from approximately one language...

      Possibly the labiovelar glide /w/ was hardened into a voiced labiovelar approximant /ɣʷ/ like in Welsh and some Romance languages.

      Then why is *w so extremely common in PIE reconstructions? (Some even think it's so suspiciously common that this may have been where the missing *b went, though others think it may have merged into *m instead, which is likewise unexpectedly common... whatever.) Can you point to a phonetic environment that would have made this more probable? (Is *rw suspiciously uncommon? Do *h₂ and *h₃ occur in the same root much more often than *h₂ and *w?)

      one of the layers which makes up the IE family

      Have you written this idea up somewhere? Right now I can't see how IE coulc consist of layers, but I have no idea what evidence you have in mind.

      Delete
    14. Sure; however, a contrast between front velars (not palatalized) and back velars (not uvular) seems to be attested from approximately one language...

      This may be a terminological artefact. The obstruents described as "uvular" in many languages of North America are actually back velars, hardly more uvular than English /k/ in cool. The same languages have "plain velars" which are phonetically front velars. Both types can be labialised.

      Delete
    15. Then why is *w so extremely common in PIE reconstructions? (Some even think it's so suspiciously common that this may have been where the missing *b went,
      According to my own research, at least at word-initial, the answer is yes!

      though others think it may have merged into *m instead, which is likewise unexpectedly common... whatever.
      Actually, it looks like in words the labial nasal at word-initial underwent denasalized as *bh in traditional reconstructions. I'm very reluctant to write *bʱ (breathed voiced) and **bʰ (voiced aspirated) is utterly incorrect.

      Can you point to a phonetic environment that would have made this more probable? (Is *rw suspiciously uncommon?
      Aragonese Romance (now largely replaced by a regional variety of Spanish) has rw > rɣʷ as in e.g. Teruel /terɣʷél/.

      Have you written this idea up somewhere?
      I'ma fraid not, but IE-ists such as Rodríguez Adrados propose a "split" PIE model which is somewhat closer to my own idea.

      Right now I can't see how IE coulc consist of layers, but I have no idea what evidence you have in mind.
      I've already posted a few snipets right here. Remember lexical isoglosses such as *sa(n)k- vs. *yag´- or Latin sanguis vs. *yak(k)-? This should ring a bell.

      Delete
    16. This may be a terminological artefact. The obstruents described as "uvular" in many languages of North America are actually back velars, hardly more uvular than English /k/ in cool. The same languages have "plain velars" which are phonetically front velars.

      ...I shouldn't be surprised.

      *bh in traditional reconstructions. I'm very reluctant to write *bʱ (breathed voiced) and **bʰ (voiced aspirated) is utterly incorrect.

      All three are meant to be the exact same thing in phonetic terms. BTW, the IPA spelling for breathy voice without aspiration is [b̤].

      This should ring a bell.

      Well, these were other cases where you barely hinted at the very idea instead of explaining anything...

      Delete
    17. To cut a long story short, the IE family would be the result of a (possibly complex) series of expansions and replacement processes over many millenia, so there wasn't a single protolanguage involved (i.e. traditional "PIE") but several of them.

      Delete
    18. The first one of those that has any living descendants, as such and not just as a few substrate features, would then be PIE.

      Delete
    19. By "the first one" do you mean chronologically older?

      Delete
    20. In other words, you seem to be saying that the reconstruction of PIE is massively wrong because much of it actually belongs to later languages, and some may belong to older substrata; but you're not in fact saying that PIE didn't exist.

      Delete
    21. I see what you're hinting to. Then the actual PIE would be the ancestor of Anatolian and possibly other less known languages, while most of the lexicon commonly attributed to "PIE" would actually belong to either "Kurganic" and its branches or older substrata.

      Delete
    22. Yes; and this sounds like "Kurganic" is what our esteemed host would call "Proto-Core IE" (last common ancestor of IE except Anatolian) or "Proto-Neo-IE" (last common ancestor of the IE crown-group – IE without Anatolian and Tocharian).

      Delete
    23. I'm afraid this is too simplistic. "Kurganic" would be roughly equivalent to Rodríguez Adrados' IE III A (polythematic) and thus the direct ancestor of Indo-Iranian and the Greek-Armenian-Phrygian group, although Karl Horst Schmidt and Villar would include Celtic as well.

      Delete
    24. Vowel coloring by plain velars in Kabardian!

      ...However, the plain velars are loanword phonemes. Perhaps the speakers chose the vowel allophones that best approximate the Russian central [ä] or the Turkish back [ɑ] and not-quite-[ɯ]... or there's something wrong with the table, which currently mentions "palatalized palatovelars", "labialized palatovelars" and several "pharyngeals" while the consonant table only mentions [ħ]. I'll have to read the papers the article cites.

      Delete
    25. In fact, the real PIE would be more alike to "Proto-Nostratic". Of course, the real Nostratic (in the sense hinted to by Krens) is quite another thing.

      Delete
    26. As regarding morphology, "Kurganic" would be roughly similar to Adrados' IE III A, with features such as the augment and the aorist. Phonetically, "Kurganic" has reversed the so-called "thorn" clusters and plenty of words with initial *y- (likely IPA [ɟ] or [ʝ]), including the relative prounoun *yo-.

      Although Celtic has got Kurganic lexicon, its morphology is not, reflecting its strong Italic substrate/adstrate, which some scholars mistake for an Italo-Celtic node.

      Delete
  8. Whatever the phonetics of the three dorsal series, isn't it most natural for them to contain the laryngeals? Years ago I convinced myself that if so, you could explain the mystery of PIE /b/ via laryngeals. 

    My story went as follows: If the laryngeals fell neatly into the dorsal series, there would have been a 'hole' in the phonological system for a labial fricative - all other series would have a fricative member. So speakers started pronouncing /b/ as some sort of fricative when it occurred in positions where the phonotactics would allow it; eventually almost all /b/s went that way; then the new labial fricative merged with /w/.

    There are certainly a lot of reconstructed /w/'s in PIE roots - it's the most common phoneme, at 10.6% in LIV. If you assume that original /b/ had a similar frequency to /p/ and /d/ (both 4.2%), and subtract this from the /w/'s, you get fairly close to the frequency of /y/ (6.8%) - the phoneme that /w/ seems to pattern most closely with.

    I imagine there are a multitude of reasons why this doesn't hold up. Not being a scientist I never bothered looking for them, just sat in the college bar eating crisps. But I guess one prediction would be that /w/ should appear in some phonotactic contexts where other resonants don't but plosives do.

    ReplyDelete
    Replies
    1. But I guess one prediction would be that /w/ should appear in some phonotactic contexts where other resonants don't but plosives do.

      Well, it forms word-initial clusters with *r, *l, *j, but it shares this combinability with *m.

      Delete
    2. Maybe then /b/ got eaten by both /w/ and /m/?

      Delete
  9. And actually if you posit that reconstructed /b/'s are the remnant of the originals - rather than the result of /m/+/l/, /p/+laryngeal, loans etc. - and add them back into the /w/ score, the numbers for /y/ and /w/ come out exactly equal!

    ReplyDelete
  10. If the laryngeals fell neatly into the dorsal series

    It has indeed been proposed that *h₁, *h₂, *h₃ were *[xʲ], *[x], *[xʷ]; Illich-Svitych in the 1960s comes to mind. One problem is that *h₁ doesn't seem to have had any fronting effects on neighboring vowels or consonants.

    Speaking of Illich-Svitych – the Moscow School Nostraticists derive PIE *p, *b, *bʰ from Proto-Nostratic *pʼ, *p, *b by a somewhat strange chain-shift. If the PN plosive system was of the geographically Caucasian type, where the "plain voiceless" plosives are aspirated and the ejectives are the least marked members of the system, the scarcity of PIE *b suddenly makes perfect sense. It goes without saying, though, that all this needs a lot more work.

    ReplyDelete
    Replies
    1. One problem is that *h₁ doesn't seem to have had any fronting effects on neighboring vowels or consonants.The thing is *h₁ doesn't seem to represent a single but several phonemes. In some instances it would be a glottalic stop [ʔ] but a voiced dental/alveolar fricative [z] in others.

      I also disagree with the Moscow School's view of the 3 stops series in IE.

      Delete
    2. The thing is *h₁ doesn't seem to represent a single but several phonemes. In some instances it would be a glottalic stop [ʔ] but a voiced dental/alveolar fricative [z] in others.

      That's particularly interesting to me because the Moscow School has Proto-Nostratic *z- disappearing in PIE. :-þ

      What's your evidence for [z]?

      Delete
    3. The example I've got in mind is Kartvelian *zisx-L- 'blood' ~ IE *h₁esh₂r/n- 'flowing blood'. According to Gamkrelidze-Ivanov, the Kartvelian form would be a reduplicated one.

      Delete
    4. That's indeed intriguing. But if these two stems are cognates, that may still not tell us if *z corresponds to *h₁ or rather to zero: some think that PIE didn't allow words to begin with a vowel phoneme and automatically put the dummy consonant *h₁ in front of all that would have.

      Kloekhorst in "Towards a Hittite historical grammar" interprets the spelling e-eš-ḫar as ʔe-es-ḫar, phonemically /ˈʔesχr̩/ or thereabouts.

      Anyway, if *z corresponds to *h₁, that doesn't mean that any occurrences of *h₁ actually were [z] in any kind of IE.

      Delete
    5. This comment has been removed by the author.

      Delete
    6. That's right. That consonant must have disappeared at an early data without leaving any trace, so a dummy *h₁ occupies its place in reconstructions. However, I think we should replace it with *z in those cases where external data tells us to do so.

      Delete
    7. @ David

      The way you've laid it out above suggests that it's PIE /p/ that corresponds to /p'/. I thought in the Glottalic Theory it was the traditional /d/ series that was supposed to have been voiceless ejective, thus accounting for the missing /b/?

      Accounting for /b/ is the most attractive feature of the GT. Among its many problems as I recall is that speakers of languages with ejectives generally treat those series as the LEAST marked; and that the theory posits voicing of the ejective series happening independently in most of the IE daughter languages.

      If the Nostratic version you're talking about makes the ejectives correspond to the PIE /t/ series instead, that avoids these problems, but I'm not sure it helps with the /b/ mystery, which is all the GT really had going for it.

      Delete
    8. This comment has been removed by the author.

      Delete
    9. However, I think we should replace it with *z in those cases where external data tells us to do so.

      That's anachronistic.

      The way you've laid it out above suggests that it's PIE /p/ that corresponds to /p'/. I thought in the Glottalic Theory it was the traditional /d/ series that was supposed to have been voiceless ejective, thus accounting for the missing /b/?

      Yes; the trick is that the Moscow School people aren't glottalists.

      Among its many problems as I recall is that speakers of languages with ejectives generally treat those series as the LEAST marked;

      That depends. It's true in the Caucasus area and maybe in the Afroasiatic languages; it's not true in Navajo or the like, where the ejectives are produced with even higher air pressure and an even longer delay in voice-onset time than the aspirates, which are already aspirated really hard (think Mandarin, not English).

      If the Nostratic version you're talking about makes the ejectives correspond to the PIE /t/ series instead, that avoids these problems, but I'm not sure it helps with the /b/ mystery, which is all the GT really had going for it.

      It does if we assume a Caucasus-style system where the "plain voiceless" plosives (and affricates...) were aspirated.

      Of course here we get into an area where evidence is really thin on the ground. "Plain" /t/ and /k/ are aspirated in Arabic, and they were in Hebrew since the earliest Greek transcriptions, but they weren't in Phoenician (or the Greek letters wouldn't have developed as they did)...

      Delete
    10. (South Africa is another area where ejectives are lenes.)

      Delete
    11. Sorry if I'm being dumb - I still don't get it. How does aspirating the /t/ series help get rid of /b/? The advantage of ejectives is that they often omit /p'/ in the series... but isn't what you're suggesting that "PN" /ph/ for some reason disappears during the chain-shift while /p'/ survives as PIE /p/? If so, what is the relevance of ejectives to the /b/ mystery?

      Delete
    12. However, I think we should replace it with *z in those cases where external data tells us to do so.

      That's anachronistic.
      Maybe so, but very convenient in my opinion. For example, the late García Calvo reconstructed the so-called "augment" prefix as *ze-, if I'm not mistaken.

      Delete
    13. I still don't get it. How does aspirating the /t/ series help get rid of /b/?

      It helps because [pʰ] is harder to maintain than [tʰ] or [kʰ]; it is quite prone to developing > [f] > [h] > 0. If this happened before the voicing chain-shift, that would neatly explain why PIE *b is absent or nearly so.

      If so, what is the relevance of ejectives to the /b/ mystery?

      I suppose deglottalization might have triggered that chain shift. But the Moscow School doesn't reconstruct ejectives for Proto-Nostratic for purely IE reasons; it reconstructs them because they're preserved as such in Kartvelian and (largely) Afro-Asiatic.

      The advantage of ejectives is that they often omit /p'/ in the series...

      Well. First, this never seems to happen in the Caucasus area, where [pʼ] is not even rare. In Hausa and the Mayan languages, */pʼ/ has shifted to the implosive [ɓ] (while /kʼ/ remains as such and /tʼ/ sometimes does, showing up as [ɗ] the rest of the time). /pʼ/ is indeed absent in the Na-Dene languages, but those also lack the equally expected /pʰ/ (with onomatopoetic exceptions in a few languages), and many even lack the plain /p/. The only languages where specifically */pʼ/ is absent, to the best of my heterogeneous knowledge, are the Semitic ones (...the Ethiosemitic branch excepted, where /pʼ/ seems to be a loanword phoneme, or perhaps not).

      Delete
  11. a zero-grade thematic present

    The topic of my dissertation! If I'd thought of including this verb, I might even have finished it...

    ReplyDelete
  12. > In Modern Indo-Aryan its reflexes are quite numerous, though hard to recognise after more than two millennia of sound change, sometimes combined with euphemistic deformation.

    Do you have a list of these? I checked Turner, who does have a few, though none that I've come across before.

    ReplyDelete
    Replies
    1. I have never tried to compile a full list. As you can guess, Turner's dictionary was what I checked in the first place to find some modern reflexes:

      Searching the entire dictionary for yabh*.

      Delete