Kartvelian and the Issue of Glottalization in Nostratic Arnaud Fournet

Abstract: The article deals with the issue of glottalization in Nostratic with a special focus on
Kartvelian. Several hypotheses of sound correspondences are tested between Kartvelian and ProtoIndo-European. It is concluded that Kartvelian glottalization is not inherited but is probably a feature originally of prosodic nature. Pre-Proto-Kartvelian only had a distinction between voiced and voiceless phonemes.

Keywords: Nostratic, Kartvelian, Glottalization.

1. The consonants of Proto-Kartvelian The Proto-Kartvelian phonological system may be reconstructed as follows (Cf. Gamkrelidze 1967: 709; Gamkrelidze—Mačavariani 1982: 25—61; Schmidt 1962: 60; Bomhard 2008: 142): p Obstruents pˀ b t tˀ d c cˀ dz s z Resonants Vowels m n l č čˀ dž š ž r k kˀ g x γ y/i w/u q qˀ (G) h

o, ō a, ā e, ē Table1: The proto-phonemes of Kartvelian

It can be noted that (1) The voiceless stops and affricates were aspirated (*ph, *th, *ch, *čh, *kh, *q ). The aspiration is usually considered phonemically non-distinctive as glottalization is supposed to be the “real” phonemic feature, (2) The reconstruction of a voiced postvelar stop *G in ProtoKartvelian is controversial. In Georgian, the glottalized postvelar *qˀ was preserved, while the voiceless (aspirated) and voiced postvelar stops merged with x and γ respectively.

2. The issue of glottalization in Nostratic Two main approaches of the glottalized feature can be distinguished: the “American” school, most prominently represented by Bomhard, and the “Moscovite” school, historically initiated by Illič-Svityč and Dolgopolsky. Both Kartvelian and PIE have a three-way opposition between stops. The phonetic nature of Kartvelian features is coherent with the set that the Glottalic Theory of PIE proposes: voiced, voiceless and glottalized. For that matter Bomhard (2008) who adheres to the Glottalic Theory of PIE compares directly Kartvelian t d tˀ with PIE t dh d (=tˀ). The sound correspondences according to Bomhard (2008) are therefore as follows: Kartvelian Traditional PIE Glottalic Theory *t *t *t *tˀ *d *tˀ *d *dh *d Table2: Sound correspondences in the Glottalic Theory of PIE

The Macro-Comparative Journal

Vol.1 No. 1

Clearly. These constraints turn out to be a simple voicing agreement rule with the corollary that two glottalics cannot cooccur in a root. a glottalic is to be reconstructed for ProtoNostratic. in H[amito-]S[emitic] as glottalized or plain voiceless (the distribution being probably due to prosodic factors). they assume that. in IE (in its traditional interpretation) as voiceless. another is that the glottalized phonemes of Kartvelian are inherited as well. in U[ralic] (in the intervocalic position) as geminated voiceless stops. One is that the three-way opposition is inherited. whenever there is a voiceless stop in the Proto-Indo-European examples they cite. The common denominator of their K. One cannot determine the original phonetic realization of this additional effort (glottalization.] Illič-Svityč and Dolgopolsky posit glottalics for Proto-Nostratic on the basis of a small number of seemingly solid examples in which glottalics in Proto-Afrasian and/or Proto-Kartvelian appear to correspond to traditional plain voiceless stops in Proto-IndoEuropean. The conclusion reached by Bomhard (2008: 23) is that the system of sound correspondences proposed by the “Moscovite” school cannot be correct and should be entirely dismissed. The conclusions described above seem compelling at first sight. [.Kartvelian and the Issue of Glottalization in Nostratic Dolgopolsky (2008: 8) and the “Moscovite” school had a different approach: The emphatic stops are represented in K[artvelian] as glottalized.1 No. It can also be noted that in the Glottalic framework the simultaneous existence of a threeway contrast in PIE. fortis articulation?). as in Bomhard (2008: 23): The mistake that Illič-Svityč and Dolgopolsky made was in trying to equate the glottalized stops of Proto-Kartvelian and Proto-Afrasian with the traditional plain voiceless stops of ProtoIndo-European. HS. On the basis of these examples. The gist of the argument is that the voiceless phonemes traditional PIE are the most frequent and least marked and they should only correspond with phonemes displaying the lowest level of markedness in the other languages.. The conclusion is valid only if these additional premises are accepted. But there are a number of hidden premises in the reasoning. even when there are no glottalics in the corresponding Kartvelian and Afrasian forms! This means that the Proto-Nostratic glottalics have the same frequency distribution as the Proto-Indo-European plain voiceless stops. This is for example the point of view in Bomhard (2008:55-56): For the first time. Kartvelian and Afrasian is considered to be a kind of implicit “proof” that this three-way contrast should be ascribed to their common ancestor but in my opinion there are serious reasons to doubt that this three-way contrast is very ancient in any of these sub-families. I prefer to denote them as "emphatic" and to use the traditional Orientalistic underdot as their symbol. this cannot be correct. and Quechua as examples of natural languages exhibiting a similar constraint against the cooccurrence of two glottalics. The Macro-Comparative Journal Vol. U and A reflexes is an additional effort (if compared to the reflexes of N[ostratic] plain voiceless stops). Another argument for the Glottalic Theory of PIE is that it proposes an apparently “natural” explanation for the gaps in the distribution of phonemes in Proto-Indo-European. Akkadian may be added to this list as well if we take Geers’ Law to be a manifestation of such a constraint.. in A[ltaic] as fortes. 1 2 . If the glottalized phonemes of Kartvelian were originally plain voiceless then the correspondence between PIE *t and Kartvelian *tˀ is possible. aspiration. the root structure constraint laws can be credibly explained. Yucatec Mayan. Hopper (1973:160) cites Hausa. Kartvelian Traditional Nostratic Afrasian PIE Dolgopolsky Dolgopolsky *tˀ *t *tˀ *tˀ *t *d *t *t *d *dh *d *d Table3: Sound correspondences in the “Moscovite” approach The “Moscovite” approach has been criticized by the promoters of the Glottalic Theory for several reasons.

*ḳuć -̣ ‘to crush. *ḳeṗ. cut off’.‘to blow’ betk. *c ̣ḳwerṭ. in the other case Kartvelian tˀ is equated with PIE (trad. čkal. Which works best with Kartvelian lexical data? The following examples are based on my own survey of Kartvelian as compared to PIE (as reconstructed in Klimov 1998). c ̣ḳeṗ.‘to strike’. 1 3 . Bomhard proposed)? 3. Kartvelian. Gamkrelidze. č[ ̣ ḳ]ar. ceck. hit’ ~ PIE 112 bhau-d/t. which somehow acts as an intellectual substrate for the Glottalic Theory.‘to cut’ K.‘to get wet’. K. small’.‘to cut in little pieces.‘to cut in little pieces’ ~ PIE 929 sek. It can also be noted that Kartvelian accepts roots with two (or more) glottalized phonemes: *c ̣ḳeṗ.) t. cut’ K. cut off’ ~ PIE 923 skel. Latin battuere The Macro-Comparative Journal Vol.‘to break. *č e ̣ č ̣ḳ. ḳap. etc. creep’ K./ partx. Should Kartvelian glottalized be compared with PIE traditional voiceless (as Dolgopolsky or Illič-Svityč proposed) or with voiced phonemes (as Hopper.‘to blow’ ~ PIE 120 bhel. če ̣ č ̣ḳ. ḳeṗ.‘to cut’ ~ PIE 923 skel. *ḳreč-̣ / *ḳrič̣ ‘to shear.‘to cut into small pieces’ ~ PIE 931 skep.‘to scythe’ ~ PIE 911 ser-p.‘to cut’ K. par. chop’ ~ PIE 931 s-kep ‘to scrape.‘plank.‘to roll up’.) d.‘to cut’ K. kwec.‘prop. stake’ ~ PIE 123 bhel-g/k./ cick.‘to tie. band.Arnaud Fournet The article is dedicated to an assessment of these two premises: 1. ḳreč-̣ / ḳrič-̣ ‘to shear. 4. More examples will be listed below. plait’ ~ PIE 127 bhendh. ṗerpel ‘butterfly’ ~ Latin papilio ‘butterfly’ These items undermine the diachronic status of the contrast between voiceless and glottalized phonemes in Proto-Kartvelian. K.‘to cut in little pieces’ ~ PIE 931 s-kep ‘to scrape.‘to crawl. peel’.‘to beat. birch tree’ ~ PIE 139 bherHk ‘birch tree’ barž. to cut off’ K.‘to cut’ K.‘to bind’ bal ‘merry tree.‘to cut. cic ̣ il ̣ ‘snake’ ~ PIE 912 serp.‘to fly’ ~ PIE 835 pleuk ‘to fly’ K. tear’ ~ PIE 929 sek.‘to cut’ K.‘to cut’ K. sker. Testing different patterns of sound correspondences for Kartvelian lexical data As regards the second premise it is possible to compare and test the lexical yield of two hypotheses: in the first case Kartvelian tˀ is equated with PIE (trad. Examples which work in both hypotheses: b ~ bh K./ kuc ‘to cut’ ~ PIE 586 kes ‘to cut’ K.‘to lop. hook’. *ḳaḳ ‘to bend. cut off’ ~ PIE 586 kes. How much certainty is there that the three-way opposition is really inherited? 2. Critical assessment As regards the three-way opposition a number of variants suggest that Kartvelian glottalization is acquired and that Kartvelian voiceless and glottalized both correspond with PIE voiceless: K. *cum ̣ ṗ. *c ̣ḳenṭ / *c ̣ḳinṭil ‘(bird) faeces’. K. cal-/cel-/cil. *ṗrṭq̇ el‘flat’ . cel. cut’ K.‘to tear apart’ ~ PIE 895 sek-.1 No.‘to cut’ K. K. ḳwec. supports neither the impossibility or rarity of *ṗ nor the impossibility of two (or more!) glottalized phonemes occuring in a row in the same root./ ṗar.‘sickle.‘to cut into small pieces’.‘to cut in little pieces’. bad ‘net’. beam’ ber.‘to tear away.

glottalized “correspond” with PIE (trad.‘to give a good beating’ ~ PIE 170 bhreus. to go’ ~ PIE 418 ğhē ‘to release. K. K. The Macro-Comparative Journal Vol.‘to shine’ buṭḳ ‘inflorescence.‘to howl’ dag. K.‘to go’ Examples where K. salty’ ~ PIE 1039 sūro ‘sour’ cic ̣ il ̣ ‘snake’ ~ PIE 912 serp. K. wrap’ ~ PIE 1114 webh ‘to weave’ ɣor. *g > ɣ. ghi.‘to sing’ ~ PIE 123 bhel. K. K.‘this’ gargal. K.‘to injure’ ć q ̣ ̇ al. die away’ ded ‘mother’.‘to shout’ blanc ‘fern’ ~ PIE 108 bhar-s ‘bristle.infix and suffix -g-) bez. strike’ ~ PIE 160 bhlīg. projection’ brdɣwen. K.‘to breast-feed’ deg-/dg.‘fishing pole.‘wolf’ ~ PIE 493 ğhwer ‘wild beast’ txow. K. bex. cook’ ~ PIE 240 dhegwh. roll’ ɣrɣad ‘goose’ ~ PIE 30 al-bho ‘white’ ɣul. K.‘to proceed (solemnly)’ ~ PIE 456 ghredh ‘to walk’ gz(a). dew-/dw ‘to lie.‘to deceive’ ~ PIE 1140 wel. drip. K.‘to stand.‘to glitter’ ~ PIE 120 bhel-. smear’ gog. attach’ ɣwer. K. K. ask for’ ~ PIE 1109 wadh.‘to roll up’ ~ PIE 639 kwel.‘to set.> dumb g./ ɣriḳ. sknt-) ‘excrement’ c ̣ḳeṗ. to scold’ ~ PIE 115 bheH2. K. K./ c ̣ḳur. K. K. K. K. wink’ ~ PIE 928 skel. to disappear’ ~ PIE 249 dhen. K.‘to burn’ dagr. K. ghrēi ‘to rub. hen’ ~ PIE 241 dhē. K. K. K. wink’ (-w.‘to shout’ c ̣ḳenṭ / c ̣ḳinṭil.‘to flow.‘to pour. gim ‘earth’ ~ PIE 414 dhğhem. K.‘to deceive’ ɣreḳ.‘pledge’ wed. K.‘to curse’ ~ PIE 897 sekw. duɣ.‘to plait’.‘to break. xwe(w).‘to speak’ bir.‘to turn. to grow (plant)’ čxaṗ. ache’ ~ PIE 912 skēth.1 No.‘to perish. K.‘to ask’ (with assimilation) ~ PIE dheH1 ‘to say’ wed. carry’ ~ PIE 75 aw > Germanic audaz ‘property’ ɣwed. lay’ ~ PIE 235 dhē. sew’ ̣ ć ar ̣ ‘sour.‘to wound’ ~ PIE 912 skēth.‘to cut’ c ̣ḳwar.‘road.‘to speak. hook’ baċ ‘rope’ ~ PIE 111 bhas-ko ‘band.‘to close the eyes./ c ̣ḳr. dura ‘deaf’ ~ PIE dheu. crooked’ ~ PIE 1152 wer ‘to turn. K. put’.‘to ring.‘to cut into small pieces’ ~ PIE 931 skep.‘this’ ~ PIE 417 ghe-.‘to revolve’ cob/ cuc ̣ / coẉ ̣ ̣ ‘to suck’ ~ PIE 912 seu.‘to crawl. K./ ć q ̣ ̇ l.‘to drink’ cq ̣ ̇ . K. put’ diɣom.‘bent. K.‘to speak. let go’ ɣob.‘to be silent’.‘to spin.) voiceless K.‘to scatter’.‘to wish. K. K. K. twist’ ~ PIE 1140 wel. K. K. make noise’ ~ PIE 548 kel. 993 s-per ‘to strew’ ć (a)n‘to plait’ ~ PIE 973 snē. 261 dhwen ‘to disappear’ duɣ.‘to smear’ ~ PIE 457 gher-. Cf.‘to flow’. pound’ (with -r-) bežɣ. K. to make noise’ ~ PIE 428 ghel.(< *digom ?)1 ‘humus’. ḳund ‘(bird) faeces’ ~ PIE 947 sker.‘to beat.‘to be. leaf’ ~ PIE 146 bheu-t. blisk./ ć q ̣ ̇ ar.‘to die’ ~ PIE 487 ghdhei.‘to annoy. creep’ cip ̣ ‘birch’ ~ PIE 55 apsā ‘aspen’ c ̣ḳar.‘to burn’ dum. K.‘ground’ dn.‘to splash’ ~ PIE 992 s-phereg. K. to say’ Possibly because of the back vowel -o-./ wid.‘to blink. 45 ank.!) c ̣ḳwerṭ. bundle’ berċq̇ . K.‘to brand’ ~ PIE 240 dhegwh.‘strap’ ~ PIE 1116 wedh ‘to bind. dedal ‘female.‘to have.‘to growl’ ~ PIE 160 bhlē.(Gen. K. K.‘to bend. K. K. 1 4 . bend’ ɣun. K.‘to entwine. soak’ ~ PIE 1145 wel-k ‘wet’ mgel. 1 anḳes ~ PIE 2.‘to go’ ~ PIE 1109 wādh.Kartvelian and the Issue of Glottalization in Nostratic K. K. to scream.‘to strike’ (with -l. K.‘to call’ glas.‘to bend’ ~ PIE 1148 wen ‘to bend. curve’ ɣ(w).(< *dug ?) ‘to boil.‘to injure’ ć q ̣ ̇ (ew).‘to get angry. K.

‘to cackle. K./ ḳiw.‘to revolve’ (-w.‘to purify. K. K. cut off’ ~ PIE 923 skel.‘to strike.‘eye’ mčax‘very sour’ ~ PIE 1039 sūro ‘sour’ ̣ ṗerc ̣ḳ ‘to split. K.‘to eat.‘tail’ ~ Latin cauda ‘tail’ ḳwam-/ḳwm. paste’ ḳiw. K. K.‘to split. K.!) ḳać. ḳrčx.‘to pound’ ~ PIE 985 spel.‘spongy . pierce’ ̣ ć wet‘(water) drop’ ~ PIE 1050 sweid. K. prune’ ṗuṭ-wn. roast./ ḳwarćx ‘leg’ ~ PIE 928 skel.‘to mix in’ ~ PIE 632 kwet. cut off’ ~ PIE 586 ǩes.‘to smoke’ ~ PIE 595 kēu./ ḳrap.‘to shout’ ḳmin-. K. ɣač ̣ ‘jaw’ ~ PIE 784 ous. son’ ć w ̣ ‘to burn./ rḳi-.‘to ferment. strike.‘to give birth. hack’ ~ PIE 827 peuə.‘flat’ ~ PIE 805 pelə-. K.‘to look. 1046 swem curbel ‘leech’ ~ PIE 1045 swel. aus ‘mouth’ (-w. źeź.‘to shake.‘to cut.‘to twist. K. K. K. K. ring’ ~ PIE 639 kwel.‘male child’ ~ PIE 913 seu. K.‘to split’ ṗerpel ‘butterfly’ ~ Latin papilio ‘butterfly’ ṗox. K.‘to hum’ ḳo. K.‘to cut in little pieces’ ~ PIE 931 s-kep ‘to scrape.‘to lop.!) ḳwad. thorn’ ~ PIE 18 aǩ ‘sharp’ (-w. small’ ~ PIE 586 ǩes. K. K. short’ ḳalmax ‘fish’ ~ PIE 958 s-kwalos ‘big fish’ (wanderwort ?) ḳap.‘to tear’ ~ PIE 918 skhed.1 No.Arnaud Fournet K. K.‘clothes’ ~ PIE 951 s-keu.‘to cut.‘to bind’ ~ PIE 565 kenk. K. K. limb’ (-w. K.!) ḳurćx.‘to like. hold’ ~ PIE 888 seğh. K. leg.‘to shout’ ~ PIE 548 kel. K. K.‘to cut’ če ̣ č ̣ḳ.‘to cut’ ~ PIE 923 skel. wonder at’ ~ PIE 775 okw.!) ḳrox.‘to speak’ ~ PIE 996 sp(r)eg. cut’ ḳar. K.‘to split’ (-w. 1 5 . be sour’ ḳwec. skull’ ~ PIE 529 kap-ut ‘head’ ḳeṗ. 506 yes. K.‘to get wet’ ~ PIE 1052 swomb(h)o.‘to burn’ ḳwart. make fire’ ~ PIE 68 as.!) cum ̣ ṗ. K. cleanse’ ṭq̇ ub ‘twins’ ~ PIE 642 kwetwer ‘four’ ṭq̇ w. K. K. beat’ ~ PIE 545 kel. gather’ ~ PIE 938 kerp.‘to cut’ (-w.‘unripe. K. drink’ ̣ ć u ̣ ź.‘to moan (quietly). break off’ The Macro-Comparative Journal Vol. brood-hen’ ~ PIE 548 kel-ə. desire’ ḳon.‘to hold’ č[ ̣ ḳ]ar./ ḳl.‘to revolve’ ḳwerćx.‘to sweat’ ̣ č[ ̣ ḳ]ar.‘to cut. K. K. swamp’.‘crooked.‘to wish’ ~ PIE 515 kā. cut’ ḳer ‘fireplace’ ~ PIE 551 ǩel ‘warm’ ḳet. K.‘to cut’ (-w./ *ć q ̣ ̇ wd. coil. scone’ ~ PIE 639 kwel. K.‘to shout’ ḳuć -̣ ‘to crush. chop’ ~ PIE 931 s-kep ‘to scrape. K.‘to gird.‘to speak’ (with -l-) zaṗ.‘to cover’ ḳwax.‘not to suffice. K. K.‘to pluck’ ḳreč-̣ / ḳrič-̣ ‘to shear.‘to cut’ č[ ̣ ḳ]ir.!) ḳwer ‘round. howl’ ~ PIE 556 kem.‘to be under an obligation’ ɣać -̣ ‘cheek’. K. K.‘fat’ ~ PIE 793 peyə ‘fat’ ṗrṭq̇ el. K.‘to split’ ~ PIE 923 skel./ ḳud.‘to hit./ ḳr.‘to pluck (fowl)’ ~ PIE 827 peuə.‘to speak’ zeṗ-. K. K. K.‘to burn’. K. short’ ~ PIE 938 s-ker ‘to cut. cut’ ḳep ‘back of the head./ *ć q ̣ ̇ wid. K.‘to talk’ ~ PIE 1088 tolkw.!) hweḳ ‘sharp end.‘crooked. chop’ ~ PIE 586 ǩes ‘to cut’ ḳal. crack’ ~ PIE 988 s-pel.‘to cut.‘to boil’ ć wer‘to stump (wood)’ ~ PIE 1050 swer. limb’ ć q ̣ ̇ wed. sour’ ~ PIE 627 kwath.‘to cut.‘to grasp.‘to collect. ḳum. K. K. leg.!) ḳwir. ć q ̣ ̇ rta ‘elbow’ ~ PIE 928 skel. bind’ ḳrab.‘to need. 833 plat ‘flat’ ṗu. K.‘to cut’ ḳrḳ. K.‘to cut’ (-w.‘to cut in little pieces’ ~ PIE 929 sek. have a need’ ~ PIE 927 skel.

‘to urinate’ ~ Latin pissāre ‘id.‘big hoof.‘to die’ ~ PIE 466 gwedh.‘to shake’ ~ PIE 801 pel ‘to thrust. K.‘to burn’.‘sickle./ cr.‘to sweep.’ pšw-en. pig’ ~ PIE 1038 sū ‘pig’ kas. recite’. K.‘to fly’ ~ PIE 835 pleuk ‘to fly’ per. cxir ‘to sift’ ~ PIE 889 sē.‘many’ ~ PIE 798 pel.‘to cut’ čkep.‘to expand. etc.‘to mince. K.‘to push. K. fleece’ par.‘to run away’ berq̇ en ‘wild pear’ ~ PIE 173 bhrūg ‘fruit’ buq̇ wn.‘to say. K. K.) voiced K. voiceless “correspond” with PIE (trad. K./ partx. 1 6 . čem ‘me’ ~ PIE 1114 we ‘we.‘acorn’ (-w.‘to break. a lot’ prec-̣ / pric-̣ ‘to tear’ ~ PIE 835 plēǩ. pestle’ pu. K. K.‘bright’ ḳrḳo.‘gray’ ~ PIE 804 pel ‘pale’ pertx. K. K.‘to boil’ ćxi(n)ḳ. K.‘to cut’ cem.‘to flow’ čwem ‘we.‘to flow’ (with assimilation) ~ PIE 175 deH2 Examples where K. K.‘to flow’ ~ PIE 893 seikw. to cut off’ car. voiceless “correspond” with PIE (trad.‘(a) little’ ~ PIE 842 pou.‘to swell’ puc. strike’ pir(ś)ṭw.‘to marry’ kwer. deiğ tx. K. tear’ ~ PIE 929 sek. read. to rise’. K. cave’ papar. wheeze’ kor ‘house’ ~ PIE 639 kwel > Latin col.’ keš.‘to pant. speak’ ḳr.(?).‘mane’ ~ PIE 803 pel ‘skin.‘to laugh’ (Georgian si-c-il-) ~ PIE 1040 sward. K. K.‘acorn’ ~ PIE 472 gwel. ati ~ (?) PIE 191 deǩm ‘ten’ karb ‘belly’ ~ PIE 1145 gwelbh ‘womb’ kmar ‘husband’ ~ PIE 369 ğemə. to smear’ ćx ‘heat’ ~ PIE 68 as.) voiced K. K. K. us’. K.!) te ‘light’ ~ PIE 183 dei(w) titi. Latin pilus./ pšwn. K. K. K. ca ‘sky’ ~ (?) PIE 48 ansu ‘god’ cal-/cel-/cil. sk.‘to suffocate’ ~ PIE 931 kwes. K. 506 yes. scale.‘to scythe’ ~ PIE 911 ser-p. K. K. ḳwet ‘to promise’ ~ PIE 480 gwet. hide’.Kartvelian and the Issue of Glottalization in Nostratic Examples where K. foot’ ~ PIE 823 per-sna ‘heel’ po(ś)tel. husk. kurč ‘bark. few’ px. 506 yes. K.!) ḳwed-/ḳwd. K./ cick.‘to tear away.‘to sift’ cc.‘to ask.‘to trample’ ~ PIE 116 bhegw. K. crumble’ ~ PIE 796 peis.‘to inhabit’ kwab ‘cave’ ~ PIE 592 kew ‘hole. husk’ ~ PIE 923 s-kel. ḳit (?) ‘finger’ ~ PIE 188 deiǩ.‘to gobble up’ ~ PIE 170 bhağ.> Greek phagein ‘to eat’ ḳitx. K./ ṗar. 702 me ‘me’ ešw ‘boar. K.‘lungs’ ~ PIE 835 pleu ‘lung’ pol.‘to boil’ The Macro-Comparative Journal Vol. beḳ.1 No.‘full. K. us’. K. K.‘to glitter’ ~ PIE 366 ğel. K. dandruf’.‘to cool down’ ~ PIE 365 gel ‘cold’ (-w. K. glottalized “correspond” with PIE (trad.‘little. injure’ Examples where K. peel’.‘to smear’ ~ PIE 970 smeru. K.‘to spread’ pr.‘to crush. K.‘to scrape’ ker ‘scale.‘fire’ ʒećx ‘fire’ ~ PIE 68 as. clean up’ ~ PIE 585 kes.‘shell. 838 pleus ‘feather.‘grease. K.‘leaf of plant’ ~ PIE 824 pet. K.‘to burn’.‘to tear apart’ ~ PIE 895 sek-. K. cel. K. pupul ‘boil’ ~ PIE 847 pu.‘warm’ ~ PIE 828 pūr.‘to sneeze’ ~ PIE 971 snē ‘imitative words involving the nose’ čkal.‘to tear’ ps.> Germanic skal.) voiceless K.‘to laugh’ ceck. K.

!) ḳwer.‘to feed’ ḳaḳ ‘to bend.‘to melt’ (irregular -l-) ~ PIE 669 leikw.‘to lie.‘to break’2 cag ‘thorn’. 165 bhreg-. butter’ ʒ.‘breast’ (irregular -r.‘to drag.‘salt’ zrk-el. voiceless 33 (21%) K. K.‘strong. sturdy’ (irregular -g-) ~ PIE 708 meg. K.‘heart’ ʒ́el.‘to boil’ Contradictory examples K. K. to attach’ zɣwar. deiğ ‘finger’ mṭḳaw ‘five fingers’ ~ PIE 188 deiǩ.‘to line up. K. hook’ ~ PIE 537 keg ‘hook’ ḳaṗ ‘stick.‘great’ mḳerd. ʒ́egw ‘thorny bush’ ~ Latin sagitta ‘arrow’ čxartw ‘magpie’. The number of cases where Kartvelian glottalized corresponds with PIE voiced probably reflects the potential for chance coincidence. hammer’ (irregular -g-) ~ PIE 165 bhreğ. chin’ ~ PIE 423 ghebhel ‘head’ ḳep ‘roof’ ~ PIE 423 ghebhel ‘head. K.‘post.‘to strike. to be’ ~ PIE 340 es. K. 154 bhlag. ǯiǯɣ./ q̇ wer ‘crow’ ~ PIE 383 ger-.‘to gnash the teeth’ (onomatopeic ?)~ PIE 26 al. K. K. K. yard’. titi ‘finger’ ~ PIE 188 deiǩ. K. to break’. to sit’ Potential loanwords and look-alikes K.‘to limit. seal’ zisxl. cover’ (-w. ɣrč-̣ / ḳrč.‘to put’ (irregular -l-) ~ PIE 660 legh.‘to burn’. fence’ ~ PIE 898 swel. breg. trap’ ~ PIE 911 ser. voiced 7 7 Note the infixal paradigm: PIE 115 bheg-.‘to hide. pull.‘limit.‘to knock.Arnaud Fournet K.and -d-) ~ PIE 579 ǩerd. K. zɣw-ed. K.‘liquid’ loḳ./ lqw.‘blood’ ziz. K. K.‘blood’ ~ PIE 343 es. conceal’ (a loanword ?) ~ PIE 553 ǩel. deiğ ‘finger’ zalw.‘to be. voiceless voiced aspirate K.‘to hide.‘to cry hoarsely (bird)’ (onomatopeic) q̇ el ‘neck’ (irregular -l-) ~ PIE 639 kwel. zašw ‘blackbird’ ~ PIE 1096 trozd ‘thrush’ (onomatopeic ?) ekśw ‘six’ (a loanword ?) ~ PIE 1044 s(w)eks ‘six’ ḳwel.‘to lick’ ma(n)g.‘to be satiated’ ~ PIE 876 sā ‘to be satisfied. lay’ leqw. K.‘to lie. ʒeɣ. K. cane’ ~ PIE 409 ghabholo ‘branch’ ḳaṗ ‘jaw. voiced 8 38 (22%) Table4: Quantitative “yield” of the different “correspondences” The conclusion is that an overwhelming majority of Kartvelian glottalized phonemes correspond with PIE voiceless . 567 ker.) voiced ~ Kartvelian glottalized.‘fat’ ~ PIE 901 selp.‘to lick’ (irregular -l-) ~ PIE 668 leiğh. doorsill’ zirṭ./zɣw-d. roof’ ḳit. The Macro-Comparative Journal Vol.‘fetters. ʒeša ‘firewood’ ~ PIE 68 as. K. K. Results of the tests The results can be summarized in the following table: PIE trad. In fact the correct conclusion is that Kartvelian glottalized cannot correspond with PIE plain voiceless and 2 PIE trad. K. It is therefore unsurprising that Bomhard (2008) contains very few instances of the (erroneous) correspondence: PIE (trad.‘fat.‘to turn. neck’ lag. 1 7 . K. K. PIE trad.‘to (over)fill’. K. satiated’ zoɣw ‘sea’ ~ PIE 878 sal. the “Moscovite” school proposed .1 No. K. K. 506 yes.and not with PIE voiced.‘tree’ ~ PIE 879 salik ‘willow’ 5. K. glottalized 70 (44%) K.‘to glide’ ~ PIE 901 selk.

(3) *speg-: OHG spehhan. A similar example is K.1 No. K. K. In this approach the gaps in PIE roots are the trace of a phonological split. no **ṗ-) *ć. K.‘to pour’ ~ (no **p-. Another case.‘to press. string . K. chatter’ (with final *-k).Kartvelian and the Issue of Glottalization in Nostratic Afrasian emphatics at the same time. *anḳes ~ PIE 2. even though it has already been noted that there are numerous instances of glottalized sequences. no **š) *c-̣ ‘to dip’ ~ *z. These examples tend to show that Kartvelian glottalization is correlated with PIE accent: inherited voiceless phonemes became glottalized in (pre-)Proto-Kartvelian when they were followed by the accented vowel.‘to wish’ ~ *r.‘to curse’ ~ PIE 897 *sekw. MHG spaht ‘chatter.Kartvelian *t > secondary split into *th (unaccented) and *tˀ (accented) It is probable that the glottal split of *C > Cˀ / Ch in Kartvelian has a relationship with the position of accent in Proto-Kartvelian or in a deeper stage. bring’ ~ *n. which may be a borrowing.‘to thread.‘to pound’ ~ PIE 985 spel. (2) *spreg-: Albanese shpreh ‘I speak out’ (< *spreg-sk-). OHG sprāhha ‘language’.‘to exceed’ ~ (no **ś-. OE spaec ‘speech’. *b. 1 8 . *zeṗ. OE spraec.‘to arrive. The Indo-European comparanda show that neither -r.‘to carry. hook’: Indic aṇkáh. This would explain why Kartvelian generally admits only one glottalized per root as there is only one accented syllable. OE sprecan. The Indo-European comparanda are: (1) *sprek-: Welsh ffreg ‘talk.‘to speak.Kartvelian *d . which is therefore the unaccented zero grade of √z_p. break off’. Another indication that glottalization is linked with prosody is that voiceless phonemes are conspicuously rare in Kartvelian monoliteral verbs: K.‘to give . no **ḳ-) *i. voiceless and voiced. to wait’ ~ *m. win’ ~ (no **k-.‘to speak’.‘to split. to acquire. no **t-. *zaṗ. arrive’ ~ *ž. 6. K.nor the final -g / -k can be part of the root. to say’: Greek and Germanic indicate that this word was not accented on the root. which is listed as having s-mobile but nearly all comparanda do have s-. An interesting example is K. Cf. to beat’ ~ *ć -̣ ‘to reach. English to speak. which is also coherent with an unaccented root √z_p > *sp-. There is little doubt that PIE voiceless corresponds with both Kartvelian glottalized and voiceless. speech. loud song’. OHG sprehhan ‘to speak’. no **ṭ-) *g.‘to decrease . Greek and Germanic < *ankón are coherent with an unaccented root. Kartvelian glottalization and accent In the analysis which I propose here Kartvelian first lost glottalization then reacquired glottalization and aspiration as described below: . reach’ *l. is K. PIE 1000 splei ‘to splice. split’. K. spehhen ‘to chatter’. One more example is K. OE specan ‘to speak’. to save’ The Macro-Comparative Journal Vol. This situation leads to the following scenario: Proto-Kartvelian and Pre-PIE had only two series. *ć q ̣ ̇ (ew). Traditional PIE *tˀ = *th *t = *d *d *dh Table5: Modified Sound correspondences Kartvelian This means that the comparanda for the specific relationship between PIE and Kartvelian in Dolgopolsky are probably better than those in Bomhard (2008) on the whole. as the glottalized phonemes of Kartvelian do not come from originally glottalized phonemes. 45 *ank‘fishing pole.‘to win . squeeze’ ~ (no **s) (no **d-.

to let go. It can also be noted that dental stops are quite unexpectedly very poorly attested in these two types of verbs.‘to meet’ ~ *ɣ(w). p pˀ b t tˀ d c cˀ dz s z m č čˀ dž š ž k kˀ g x γ q qˀ (G) h n l r y w Table6: Phonemes attested in monoliteral roots t tˀ d c cˀ dz s z č čˀ dž š ž k kˀ g x γ q qˀ (G) h p pˀ b m n l r y w Table7: Phonemes attested in Cw-type roots In monoliteral verbs. *x. voiced and glottalized phonemes in Kartvelian does not appear to be inherited.‘to soften. possess’ K. roast’ ~ *źw.‘to get used to’ *cw. no **q̇ ) K. to insert’ ~ *św.‘to buzz’ ~ *sw. K.1 No.‘to expand. to eliminate’ *zu. lay’ *gu. K.‘to select. K.‘to lie. soft’ ~ *čw. The Macro-Comparative Journal Vol.‘old’ ~ *ččw.‘to smear . to string on’ *dew-/d(w). Conclusion The three-way opposition between voiceless. to rise (dough)’ *ć w ̣ . carry. where voiceless phonemes are incredibly rare. Glottalization is an innovative and possibly recent feature of Kartvelian which was correlated with accentuation. K. to stop. where voiceless phonemes are well represented and glottalized phonemes are rare.‘to defecate’ *xw. be late’ *qw. to ressemble .‘to put on.‘to feed . possess’ There is nearly a complete distributional symmetry between monoliteral verbs. 6. K. (no **q-. 1 9 . *dž. carry.‘to have. to have . K. Originally (Pre-)Proto-Kartvelian only had two series: voiced and voiceless. hinder.‘to touch’ ~ *ɣ(w).‘to burn. where accent must have immediately followed the only consonant of the root.‘to get used to’ ~ *gw. K. and Cw-type verbs.Arnaud Fournet K. K. voiceless consonants are nearly absent.‘to drink’ *šw. The rarity or near absence of the least marked phonemes in the class of monoliteral verbs is in all cases a typological oddity. leave’ ~ *žw.‘to have.‘to be (lying)’ ~ (no **č-) Another set of verbs corresponding to the zero grade of *C_w involves many more voiceless phonemes and nearly no glottalized: K. *pu(w).‘to give birth .

d dh K c cˀ dz s z PIE K č. 875pp) & 2 (976pp). Berlin. Comparative Phonology. Volumes 1 (xxi. ć ̣ PIE K k kˀ PIE k.Kartvelian and the Issue of Glottalization in Nostratic Annexe: Sound correspondences of the comparanda K p pˀ b PIE p. and Vocabulary. 2008 Reconstructing Proto-Nostratic. New York: Mouton de Gruyter. Klimov. The Macro-Comparative Journal Vol. w y/i K q qˀ (G) PIE k. g gh H1 H 2. g / K PIE s dž š ž s g x γ h / m m n n l r r l y/i w/u w/u Table8: Sound correspondences References Bomhard. Allan R.1 No. ć čˀ. 1 10 . Boston: Brill. Leiden. Georgij Andrejevich 1998 Etymological dictionary of the Kartvelian languages. Morphology. b bh K t tˀ d PIE t.