Posts Tagged ‘kinship terms’

Dene-Caucasian Kinship and Dene-Caucasian Kinship Terms

Friday, June 26th, 2015

John Bengtson and George Starostin have posted a synopsis of the current status of the Dene-Caucasian, or Dene-Sino-Caucasian hypothesis. While any long-range linguistic proposal faces considerable challenges gaining acceptance among mainstream historical linguists who specialize in the reconstruction of first-order language families, Dene-Caucasian, according to Bengtson & Starostin, is a hypothesis that “offers the most logical, simple, and systemic explanation to a set of stunning similarities that manifest themselves as exclusive links between a number of linguistic taxa.” Kinship terms rarely furnish good material to test long-range linguistic connections but Dene-Caucasian is somewhat special in this regard because there is a number of intriguing similarities not only between the grammatical and lexical properties of Dene-Caucasian kinship terms from Na-Dene all the way west to Basque but also between the developmental trajectories inferred for Dene-Caucasian grammatical and kinship terminological structures. Bengtson & Starostin observe that

“the original stage [in the linguistic evolution of Dene-Caucasian.-G.D.] was characterized by significant complexity of phonology and a system of noun classification that permeated the whole morphological structure.”

And this is the precisely the conclusion I reached when comparing Dene-Caucasian sibling terminologies (The Genius of Kinship, 324-5). Yeniseian and North Caucasian sibling terms sets are radically simplified compared to Proto-Na-Dene sets, while Sino-Tibetan, Burushaski and Basque sets are intermediate (with a couple of Sino-Tibetan languages showing apparent structural archaisms placing them closer to Proto-Na-Dene systems). All sibling terminologies progressively undergo simplification but the dramatic drop in the internal diversification of the sibling set between Na-Dene, on the one hand, and Yeniseian and North Caucasian, on the other, is remarkable. Similarly, grandparental terminological complexity reconstructible for Proto-Na-Dene must have undergone simplification in the other Dene-Caucasian branches. Both diachronic tendencies fit the above description by Bengtson & Starostin of the grammatical changes affecting the putative Dene-Caucasian phylum.

Bengtson & Starostin go on to write:

“As the Na-Dene family developed separately in what is now Alaska, the overt class marking of nouns diminished, while the marking of verbs remained fully developed and, perhaps, even extended. The level of phonetic complexity in Na-Dene remained relatively high, probably because most of the non-DSC Native American languages, with whom speakers of Na-Dene had to have areal interaction, were fairly complex in that regard themselves.”

This assessment of Na-Dene grammatical and phonological conservatism again finds confirmation in Na-Dene kinship terminologies. Although kinship terminologies in all Na-Dene branches underwent noticeable changes, our best perspective on the original Dene-Caucasian kinship terminological system comes from Proto-Na-Dene.

Of special interest to the students of comparative kinship terms in Eurasia is the proposed reconstruction of Dene-noun prefixes (see below, Table 7).

KinshipStudies-DeneCaucasianNounPrefixes copyThe origin of Basque osaba ‘uncle’ and iseba ‘aunt’ is clarified by comparing the initial formants o– and i– to the pervasive and productive North Caucasian noun prefixes such as Avar was ‘son’ vs. yas ‘daughter’, wac ‘brother’ vs. yac ‘sister’. Notably, in Na-Dene the potential cognate (e.g., y– in Eyak yahs ‘woman’s child’ or in Birket-Smith & De Laguna’s [The Eyak Indians of the Copper River Delta, Alaska, 1938, 566] notation siac, with a possessive prefix s-) indexes Ego-Gender and not Alter-Gender. Harry Hoijer (“Athabascan Kinship Systems,” 1956, 330, no. 50) reconstructs –yaze ‘woman’s son’ for Proto-Athabascan but a number of other forms for children (nos. 48-52) show a similar prefix (see below).

KinshipStudies-AthabascanHoijer copy


This is very much in agreement with the presence of Ego-Gender in Proto-Na-Dene sibling terminologies – precisely the feature that gives the only New World branch of Dene-Caucasian its archaic complexity, which is barely found in the Sino-Caucasian languages of the Old World (see above).

Note also how Alter-Gender is neutralized in both Eyak yahs, Avar –as ‘child’ and Basque –saba ‘parent’s sibling’ (without the noun prefixes), a semantic development that likely reflects an archaic emphasis on Ego-Gender vs. Alter-Gender.

Basque –saba ‘parent’s sibling’ (with fossilized possessive s-?) is intriguing in the light of Navaho shibizhi ‘father’s sibling; step-father’. Navaho and Western Apache are unique among Na-Dene languages in neutralizing Alter-Gender in the terms for ‘parent’s sibling’ (see Opler, “The Kinship Systems of the Southern Athabaskan-Speaking Tribes,” 1936) and this is precisely the development observed in Basque. Nav shibizhi was extended to ‘father’s brother’ from the original concept ‘father’s sister’ (Proto-Athabascan *ma-, *me-, *metce ‘father’s sister’ in Hoijer 1956 (see below, although in the light of Basque –seba one may need to reconstruct *ba-, *be-, *betce).

KinshipStudies-AthabascanHoijerFZ copy

It’s likely that the Proto-Na-Dene term for ‘father’s sister’ also covered its reciprocal, namely ‘woman’s brother’s children’ (as in Navaho and Western Apache) – an ancient semantic pattern altogethaer missing from Basque. Bengtson & Starostin lament that Na-Dene internal reconstruction is still work-in-progress and this fully applies to Proto-Na-Dene kinship systems. This is what Kroeber (“Athabascan Kin Term Systems,” 1937) had to say about the cognates of Nav shibizhi:


Finally, a note on the possible connection between Dene-Caucasian and Kartvelian sibling sets. I have already pointed out (also The Genius of Kinship, 325-6) that the structural similarity between Basque, Burushaski and Svan (the most divergent of Kartvelian languages) sibling sets is remarkable, while nothing even remotely reminiscent of the Svan sibling set is found among the so-called “Nostratic” languages. One of Svan sibling terms, udil, widil ‘woman’s sister’ (< *udild, *widild) is composed of a fossilized possessive prefix u-, root –d– and a diminutive –ild. The possessive prefix can be compared with the Dene-Caucasian noun prefix u-. It’s quite likely that a generalized possessive prefix evolved from a more specific Ego-Gender marker (‘my, the woman’s, sister’).



Indo-European Labiovelars: A New Look

Wednesday, July 23rd, 2014

The standard theory behind Indo-European (IE) labiovelars (gwh, gw, kw) is that they were “primitive” IE phonemes that were lost in virtually all IE dialects and yielded velar-only reflexes in most languages outside of Greek. Presumably only Greek shows a triple split into dental (usually before e/i), velar (usually before u) and labial (usually before o and a) reflexes of PIE labiovelars. Some of the well-known examples include Gk theínoo “I strike, kill” < *gwhen-yoo vs. phónos “slaughter” < *gwhón-osdelphús “womb” < *gwelbh– (Skrt garbha-) vs. bíos “life” < *gwih3wos (Goth qius “alive”), tís “who?” < *kwis (Lat quis). A dissenting voice came from A. G. E. Speirs, The Proto-Indo-European Labiovelars (Amsterdam 1978). He argued that the Greek pattern of a triple-split of the PIE labiovelar phoneme depending on the quality of the following vowel was in fact a proto-Indo-European phenomenon and hence dental and labial reflexes of labiovelars can be found in all IE dialects, not just in Greek. Speirs takes three well-known IE roots, *bhudh– ‘ground, bottom’, *dheg’h-n-/*dheg’h-m– ‘earth’ and *pekw– ‘bake, cook’. He notices that in all the three cases some IE dialects show developments that look like an irregular metathesis. But, instead, Speirs suggests, *bhudh– ‘ground, bottom’, *dhegh-n-/*dhegh-m– ‘earth’ and *pekw– ‘bake, cook’ should be reconstructed as roots with PIE labiovelars.


Speirs collected a number of examples in which labial and dental reflexes of labiovelars seem to occur in IE branches other than Greek. However, his effort was dismissed. Most recently, Jouna Pyysalo (System PIE: The Primary Phoneme Inventory and Sound Law System for Proto-Indo-European, 355-6), in an otherwise very thought-provoking and open minded dissertation, summarizes the objections to Speirs’s argument in the following way:

“the underlying superphonemes – allegedly yielding velars, labials and dentals – would violate the principle of the regularity of sound change. On the contrary, it must be concluded that the places of articulation PIE *k p t are irreducible and the oppositions are distinctive. Any attempt to derive these items from other places of articulation is doomed to failure.”

Meanwhile, the velar, dental and labial split of a primitive labiovelar phoneme is a well attested and universally accepted phonetic development in ancient Greek, and there can be no doubt that it does not contradict the “principle of the regularity of sound change.” The counterargument based on a general principle is therefore invalid. It’s all the matter of how well the etymological material from IE languages supports the triple split of labiovelars at the deeper, PIE level. In the following I will show that Speirs was basically right but he didn’t go far enough.

PIE *kw

1. PIE *wlkwo– ‘wolf’ (Gk lukos, Lat lupus, Skrt vrka, Toch B walkwe, Slav *vliku, Lith vilkas, OHG wulf, OEng wolf, ONorse ulfr ‘wolf’, ylgr ‘she-wolf’) ~ Slav *lutyji ‘crazy, vicious, angry’. It has long been observed that the Germanic forms going back to *wulhwaz show a labial reflex of the PIE *kw attested in Latin lupus, and hence labial reflexes of PIE labiovelars are indeed attested outside of Greek and Latin. Slav *lutyji completes the picture and proves that PIE *kw yielded velar, labial and dental reflexes across IE languages. Gk lussa ‘madness’ presently linked to *lukos agrees with the Slavic form semantically and may represent either *luk(w)iyo– or *lutiyo-.

2. PIE *penkwe– ‘five’: Skrt panca, Avest panca, Arm hing, Gk pente, Lat quiinque, OIr coic, Goth, OHG fimf, Toch pan, B pis, Lith penki, Slav penti. This well-known set shows a similar phonetic pattern to the WOLF set. Germanic yields –f– as a reflex of PIE *kw, while Slavic is aligned with Greek in having a dental reflex.

3. PIE *kwekw– ‘bake’: IE *pekw– ‘bake’ (Skrt pacati ‘cooks, bakes, roasts, boils’, Gk pesso ‘cook’, peptos ‘cooked’,  Toch A pak, Toch B pak ‘cook, boil, ripen’, papaksu ‘cooked’, Lat coquo ‘cook’, Welsh poeth ‘baked, roasted, hot’, pobi ‘bake’ (p– < *kw-), Lith kepu, kepti ‘bake’, Alb pjek ‘I bake’, Slav *peku, *pekti ‘bake, roast, oven’) ~ Slav *potu ‘sweat’, *poteti ‘to sweat’. It’s presently assumed that Slav *potu derives from *pokto, with –kt- developing into –t– before a front vowel. A simpler solution, however, postulates that Slav *poteti ‘to sweat’ (comp. Skrt pacati) stems directly from *pokwe-. Speirs above compared IE *pekw– ‘cook, bake, heat’ (with a suggestive “metathesis” seen in Latin coquo, Lith kepu ‘roast, bake’ and Gk –kopos in arto-kopos ‘bread-baker’) and IE *tep– (Skrt tapati ‘be hot, heat, burn’, etc.), with the resulting PIE *kwekw-. This completes the paradigm of labial, dental and velar reflexes of PIE *kwekw– ‘bake’.

4. PIE *yekwr(t)-/*yekwn(t)– ‘liver, innards’: IE *yekwr(t)-/*yekwn(t)– ‘liver’ (Skrt yakrt, Gen. yaknas, Gk heepar, Gen. heepatos, Lat iecur, OHG lebara, OEng lifer, ONorse lifr, Arm leard, Lith jeknos ‘liver’, Slav *(j)ikra ‘calf (of leg), fish roe’) ~ IE *enter– ‘innards’ (Skrt antra ‘entrails’, Gk enteron ‘innards’, Lat interior ‘internal’, Arm enderk’ ‘innards’, ONorse ithrar ‘innards’, Slav *jentro ‘liver, entrails’, *wentro ‘womb, liver, entrails’). Germanic LIVER forms have always been puzzling. Now, the medial labial stops being an anomaly but regularly reflects PIE *kw which is further supported by the medial –t– in the INNARDS set. The nasal infix in the INNARDS set is likely a secondary innovation originally derived from the heteroclitic *yekwn– form (*yekwn– > *yenkwn) and later distributed across the whole paradigm. The INNARDS is morphologically derived as it lost the ancestral heteroclitic paradigm. The initial l– in Germanic and Armenian forms (comp. also Hitt lissi ‘liver’) remains problematic, although it’s reminiscent of the n– in Slav *nentro ‘inside’. One possible explanation is that it dissimilated from *rekwor (> *lekwor) and the latter was formed by the same process of assimilation (*ekwor > *rekwor) as the forms in the INNARDS group (*yekwn– > *yenkwn). For a close morphological parallel see OHG zahar and trahan ‘tear’ (< *dakru-/*daknu-).

5. PIE *leikw– ‘stick, leave behind, leave’: IE *leikwo– ‘leave’ (Skrt rinakti, Gk leipoo ‘I leave’, Lat linquoo ‘I leave, abandon, forsake’, Arm lk’anem ‘to leave’, Goth leihvan, OHG liihan ‘to lend’, Lith likti ‘to stay’, at-laikas ‘remains’, liekas ‘that which is left over’, OCS ot-leeku ‘remains’) ~ IE *lep– ‘to stay, to stick, to leave’ (Skrt lepayati ‘applies ointment’, lipyate ‘glues on, sticks’, Gk lipos ‘grease’, Lat lippus ‘having watery eyes’, Toch lip– ‘stay’, Goth aflifnan ‘stay’, bilaibjan ‘leave’, Lith lipti ‘stick’, lipus ‘sticky’, Latv lipinat ‘stick’, laipns ‘friendly’, Slav *lep– ‘stick, form from clay’, *lipkiji ‘sticky’).

6. PIE *kwel– ‘revolve, turn, move’: IE *kwel– ‘revolve, turn, move’ (Gk kuklos ‘circle’, tellomai ‘turn around in circles’, pelomai ‘am in motion, go’, poleoo ‘go around, range, haunt (intrans.); turn up the soil (trans.)’, bou-kolos ‘cowherder’, ai-polos ‘goatherder’, amphi-polos ‘female servant’, polos ‘axis’, teleoo ‘finish’, telos ‘end’, Lat colere ‘cultivate, move around, inhabit’, Skrt carati ‘he moves, wanders’, cakra ‘wheel’, Toch kukal ‘wagon’, OEng hweohl ‘wheel’, Lith kelias ‘road, way’, kaklas ‘neck’, OPruss kelan ‘wheel’, maluna-kelan ‘miller’s wheel’, Slav *kolo ‘wheel’) ~ IE *pel- ‘flour, dust, ashes’ (Lat pollen ‘finely milled flour’, pulvis ‘dust’, Gk palee ‘finely milled flour, dust’, Skrt palalam ‘ground seeds’, Lith pelenai ‘ashes’, pelene ‘hearth’, Slav *poleeti ‘burn’, *polmen ‘flame’, *pepelu ‘ashes’). The highly productive PIE root with an abstract meaning ‘move in a circle’ could naturally yield forms with the meaning ‘flour, dust’ (the outcome circular movement of millstones). The meaning ‘ashes’ attested only in Balto-Slavic could either evolve by analogy with ‘flour’ or represent an independent development from the underlying notion of circular movement via the agricultural process of slashing and burning (comp. Lat colere ‘cultivate’) or via an existential understanding of ashes as the end (comp. Gk telos) of a cycle of life. The perfect morphological match between Slav *pepelu ‘ashes’, on the one hand, and Gk kuklos ‘cycle’, Skrt cakra ‘wheel’, Toch kukal ‘wagon’ and OEng hweohl ‘wheel’ is a further proof that the two cognate sets are related.

7. IE *kwel– ‘full, fertile’: IE *kwel– ‘crowd’ (Skrt kulam ‘herd, lineage’, OIr clan, cland ‘offspring, lineage, clan’, Lith kiltis ‘clan’, Gk telos ‘crowd’, Slav *celedi ‘serfs, servants’) ~ IE *pel– ‘full, offspring, multitude’: Gk pleerees ‘full’, pleethos ‘multitude’, Skrt puurnas ‘full’, Lat pleoo ‘fill up’, pleenus ‘full’, pleebees ‘crowd’, OIr lan ‘full’, Goth full ‘full’, Lith pilnas ‘full’, Slav *pulnu ‘full’, *plemen < *pled-men– ‘tribe’, *plodu ‘fruit, offspring’. There’s a complete semantic and morphological alignment between Slav *celedi ‘serfs, servants’, OIr cland ‘clan’, on the one hand, and Lat pleebees ‘crowd’ and Slav *pled-men ‘tribe’.

8. PIE *skwel- ‘split, tear, rip’: IE *skel- ‘split’ (Hitt iskallaa(i)– ‘break, smash’, Lith skeliu, skelti ‘split’, skilti, ski?lu ‘crack’, ski?ltis ‘cut piece’, skyle ‘hole’, Gk skallo ‘dig, chop’ , ONorse skilj? ‘divide’, Goth skilj? ‘butcher’, Arm celum ‘split’ [< *sk-] ~ IE *(s)pel– ‘skin’ (Gk spolia ‘fine wolf plucked from the legs of sheep’, pelma ‘sole of the foot’, Lat pellis ‘animal skin, hide’, spolium ‘animal skin, hide’, OHG fel, OEng fell ‘animal skin, hide, pelt’, filmen ‘film, membrane, foreskin’, Lith plene ‘skin (on milk), scab,’ OPruss pleynis ‘meninges’, Slav *pleva, *plena ‘membrane’). It has been observed (EIEC 269) that the underlying verb *(s)pel– ‘to tear’ is implied by the ‘skin’ nouns comprising the *(s)pel- set but is not directly attested in IE dialects. Now it can be seen that this is because this verb is widely attested in the more basic skel– form. The semantic alignment between the two sets is perfect as evidenced by Goth skilj? ‘butcher’ that naturally sits with both groups of forms. Finally, both sets feature s-mobile.

9. PIE *kweH2ur ‘fire’: IE *peH2ur– ‘fire’ (Hitt pahhur ‘fire’, Gk puur ‘fire’, Toch A por, Toch B puwar ‘fire’, Arm hur ‘fire’, hn-oc’ ‘oven’, Goth foon, OHG fiur ‘fire’, ONorse fuurr ‘fire’ (poet.), OPruss panno ‘fire’, Czech pyrzi ‘burning coal, firebrand’) ~ IE *kur– ‘smoke’ (Goth hauri ‘coal’, ONorse hyrr ‘fire’, Lith kurti ‘stoke fire’, Slav *kuriti ‘to smoke, to stoke fire’, Bulg chur ‘smoke’, churia ‘I smoke’.

10. PIE *aukw– ‘seeing faculty’: IE *okw– ‘eye as a physical organ’ (Skrt aksi ‘eyes’, Gk osse ‘eyes’, omma ‘eye’, opsomai ‘I shall see, perceive, inwardly appreciate’, ossomai ‘seeing with the mind’s eye, have a foreboding’, opsis ‘sight’, Lat oculus ‘eye’, Goth augo ‘eye’, Toch A ak, Toch B ek ‘eye’, Arm akn ‘eye’, Lith akis ‘eye’, Slav *oko ‘eye’ ~ Balto-Slav *aupmen ‘intelligence’ (Lith aumuo ‘mind’, aumenis ‘memory, omenis ‘sense, consciousness’, Slav *umu ‘mind, intelligence’). The morphology of the Balto-Slavic root is unmistakably the same as the morphology of Gk omma (< *opmn) and Arm akn (< *akw-mn, according to Olsen, Birgit A. The Noun in Biblical Armenian. Berlin and New York: De Gruyter, 1991, 171), hence we must postulate a labial in Balto-Slav *aupmen corresponding to a velar in Slav *oko and Lith akis. The loss of –p– before –m– would be an expected outcome of the cluster. The semantic alignment between Gk opsomai, ossomai and the Balto-Slavic forms is perfect (on the non-physical meanings of the Greek reflexes of IE *okw-, see Gonda, Jan. “Reflections on the Indo-European Medium II.” In Selected Studies. Vol. 1. Indo-European Linguistics. Leiden: Brill, 1975. Pp. 179-80). Without the Balto-Slavic forms the non-physical semantics of the Greek words would be an exception considering that in all other IE dialects (and IE *okw– is widely attested) the reflexes refer exclusively to the eye as a physical organ. Finally, the inclusion of the Balto-Slavic group derived from *aupmen explains *au– in Germanic forms for ‘eye’ (Goth augo, OHG ouga, OEng eeage). Hittite evidence confirms the proposed equation between IE *okw– ‘eye’ and Balto-Slav *aupmen ‘mind, intelligence’. It has 1sg. pres. uhhi ‘I see’, 2sg. pres. autti, 1pl. pres. aumeni, 2pl. pres. autteni, which implies *aukwi-, *aut-ti, *aup-meni with the same distribution of reflexes between velar (“laryngeal”), dental and labial places of articulation. The words denoting ‘seeing faculty’ must ultimately derive, by means of a –kw– affix, from PIE *au– seen in Slav *aviti ‘show oneself’, Lith ovyje, ovyties ‘come to someone in a dream’, Ved aavih, Avest aavis ‘obvious, revealed’, Lat audio ‘I hear’, Gk aioo ‘I perceive’ (< *aFioo), aisthanomai ‘I perceive, I realize’ (< *aFisthanomai). The IE forms for ‘ear’ (Gk ous [< *ousos], Lat auris, Goth ausoo, OHG oora, OEng eeare, Lith ausis, Arm unkn, Slav *uxo, Dual *ushi ‘ears, intellect’, Alb vesh ‘ear’) seem to belong here as well but they are formed from the basic root *au– by means of the affix –s– and not –kw-. This solution makes it unnecessary to explain the diphthong in the Germanic forms for ‘eye’ as contamination by the Germanic forms for ‘ear’ or the shape of Arm form for ‘ear’ (unkn < *ukn) as modeled on the form for ‘eye’ (akn). One remaining phonological challenge of this combined cognate set is the initial alternation between o– (H3-) and au– (H2ew-, H2u-). The same alternation is seen in Skrt ukha ‘cooking pot’, Lat aulla ‘pot’ (< *auksla), Goth auhns ‘oven’ next to Arm akut’ ‘hearth’, OHG ofan ‘oven’, OEng ofen ‘furnace’ (< *H2/H3ukw– in EIEC 443 but this doesn’t account for Gk ipnos ‘oven’, Myc i-po-no ‘cooking bowl’). The onset of the IE terms for ‘ear’ is uncertain (Nom. *Hous, Gen. H2eus, H2us are postulated in EIEC 173). Hitt uhhi shows that there was no initial laryngeal in this group (comp. Hitt huhhas but Lat auus ‘grandfather’ < *H2euH2o-) but it also, surprisingly, establishes identity between a laryngeal and a labiovelar.

11. PIE *kwerp– ‘body’: IE *kwerp– ‘body’ (Lat corpus ‘body’, Skrt krpaa ‘beautiful appearance’, Avest kerefs ‘form’, Gk prapis, prapides ‘midriff, diaphragm, understanding, mind’, OHG href, OEng hrif ‘womb, belly’ ~ IE *turp-/*trup– ‘dead body’ (Slav *trupu ‘body, dead body, tree trunk’, Lith trupus ‘breakable, fragile’, trupeti ‘break into small pieces’, Gk trupee ‘hole’, trupaoo ‘I bore’ (Vasmer, Max. Etimologicheskii slovar’ russkogo iazyka. Vol. 4, 109). Pokorny (620) mentions an opinion by Vendryes and Specht that the *kwerp- form could be metathesized *perkw– found in Goth fairhwus ‘life, world’. This fits the tendency noted above whereby roots with two labiovelars may develop mirror-like reflexes that look like products of a metathesis or dissimilation.

12. PIE *kwlokw-/*tlokw– ‘speak’: IE *tlokw– ‘speak’ (Lat loquor ‘say, speak, talk, declare’, Slav *tulku ‘make sense, interpret’, Skrt tarkas ‘suggestion’, tarkayati ‘contemplate’, OIr ad-tluch ‘thank’, totluch ‘ask’ ~ IE *lep-/*lop– ‘talk, mumble’ (Skrt lapati ‘he mumbles, moans, talks’, Pamir loowam, lewam ‘I speak’, Welsh llef ‘voice’ (< *lepmo-), Slav *lepetu, *lopotati ‘babble, talk’, Alb laperdi ‘dirty talk’). Although IE *lep-/*lop– comes across as onomatopoeic (and this may have caused the loss of the initial voiceless stop), but the second consonant is faithful to the underlying sound law. The initial consonant in *kwlokw-/*tlokw– is uncertain. The root very well may be the one with two labiovelars, although only t– forms have survived in the daughter languages.

13. PIE *anekw-/*H2nekw– ‘grandchild, sister’s son, sister’s daughter’: Slav *wnenku-/*wnuku- ‘grandchild’, OCS netiji ‘nephew’, ORus nestera ‘niece’, OIr necht ‘niece’ ~ IE *H2nepoot ‘grandchild’ (Skrt napaat– ‘grandson’, Avest napaat– ‘grandson’, Gk anepsios ‘cousin’, nepodes ‘descendant’, Lat nepoos ‘grandchild’, Lith nepuotis ‘grandson’, nepte ‘granddaughter’, OIr nia ‘sister’s son’, OHG nefo ‘sister’s son’, OEng nefa ‘grandson, sister’s son’, Alb nip ‘grandson, nephew’, mbese ‘granddaughter, niece’ (< *nepotiya). There’s a complete agreement among scholars that  Slav *wnenku/*wnuku ‘grandchild’ is related to the root of IE *H2nepoot- but it’s assumed that the Slavic form has a different affix representing an ancient *-ko-. In the light of the examples above showing that ancient labiovelar phonemes split into dental, velar and labial reflexes already in PIE times it’s possible to postulate that Slav *wnenku-/*wnuku- maintained a velar reflex of the underlying PIE *kw, which yielded /p/ in the majority of other IE dialects. Correspondingly, OCS netiji ‘nephew’ and ORuss nestera ‘niece’ may represent *nekwt– or even *nekwi– > *neti and not *nepti-, as presently assumed. The phonetic reconsideration of the PIE term for ‘grandchild’ as *H2nekw– instead of *H2nepoot– allows one to connect it to Hitt nekna, nikna ‘brother’, nika, neka, nega ‘sister’. Semantically Hitt neka/nekna will be closest to Gk anepsios ‘cousin’. For a parallel semantic development connecting kin categories in Gen 2 with kin categories in Gen0, comp. Lat avus ‘grandfather’, avunculus ‘mother’s brother’ next to Alb vella ‘brother’ (< *awentlo or *awenklo ‘mother’s brother’s son, cousin, brother’).

14. PIE *wekw-/*wokw– ‘give vocal utterance’: IE *wekw– (Gk epos ‘word’, eipon (aor.) ‘spoke, said’, ops ‘voice’, Skrt vakti ‘speaks, says’, vacas ‘word’, vaak ‘voice, sound’, Avest vac– ‘speak, say’, vaxs ‘voice, sound’, Lat vox ‘voice, sound, utterance, cry, call’, Toch wek ‘voice’, OHG giwahan ‘mention’, Arm gocem ‘cry, roar’, OPruss wackis ‘outcry’) ~ IE *wep-/wop– ‘cry out, yell’ (Slav *upeti/*vopiti ‘cry out, yell’, *vopli ‘outcry’, vypu ‘Ardea stellaris’, Latv uupet, uupeju ‘I cry’ (in reference to owls and wild pigeons), upuot ‘cry, call loudly’, upis ‘eagle owl’, Lith upas ‘echo’, OHG ufo, uvo, ONorse ufr ‘owl’, Avest ufyeimi ‘invoco’). OPruss wackis, Arm gocem andAvest ufyeimi secure a strong semantic link between the forms with a velar and the forms with a labial. The semantic integrity of this extended cognate set can be supported by a quote from Carl Buck (“Words of Speaking and Saying in the Indo-European Languages.” American Journal of Philology 36, no. 1 (1915), 16-17): “It is only in Indo-Iranian and Greek that the root has furnished the regular verb of ‘speaking, saying’. The more wide-spread noun, Skt. vak, Lat. vox, etc., means primarily ‘voice’, and the use of the other forms which occur outside of Indo-Iranian and Greek indicate for the parent speech a general application to the voice and to its product, speaking, calling, crying, etc. The more precise semantic source is hidden in the remote past, but it can hardly be doubted that it belongs somewhere under the general head of ‘sound’.”

15. PIE *sekw-/*swekw– ‘sap, juice, syrup’: IE *sekw-/*swekw– ‘juice, sap’ (Gk opos ‘juice’, Lith sakai ‘tree sap’, OPruss sackis ‘tree sap’, Latv svek?i ‘sap, tar’, Alb gjak ‘blood’, Slav *soku ‘juice, sap’) ~ IE *sap- ‘sap’ (OEng saep, ONorse safi ‘tree sap’, Mid Low Germ sabben ‘spit out’, sabbelen ‘make dirty’, Lat s??? ‘syrup, must’, Arm h?m (< *s??m?-) ‘juice’, Avest v?-š???– ‘having poisoned juices’, Slav *sopuli ‘snivel’.

16. PIE *sekw– ‘follow’: IE *sekw– ‘follow’ (Gk hepomai ‘follow’, Lat sequor ‘to follow’, socius ‘companion’, Skrt sacate ‘follows’, sakha ‘friend, companion’, OIr sechithir ‘follows’, Lith seku ‘follow’, OEng secg ‘follower’) ~ IE *sekwtm ‘seven’ (Gk hepta, Lat septem, Skrt sapta, OIr secht, Lith septyni ‘seven’ but sekmas ‘seventh’, OEng seofon, Goth sibun, Toch A spat, Toch B sukt). The IE numeral ‘7’ can be seen as a formation parallel to a well-known Latin derivative of the *sekw– root, namely secundus ‘second’ (“the one that follows”) suggesting that *sekwtm originally meant ‘number that comes after the set of numbers from 1 to 6’. The consolidation of the two sets is facilitated by the presence of forms with –k– in the SEVEN cluster (OIr secht, Toch B sukt, Lith sekmas) and, of course, the presence of –p– forms in the FOLLOW cluster (Gk hepomai). The new cognate set eliminates the need to postulate contamination with okt ‘8’ to explain the anomalous Toch B sukt ‘7’. Uralic is believed (see, e.g., here) to have borrowed several of its forms for ‘7’ from at least two different IE branches (Finno-Permic from Balto-Slavic and Ugrian and Samoyedic from Tocharian B). It will be interesting to see if the reinterpretation of the IE numeral ‘7’ as originally *sekwtm clarifies the prehistory of the Uralic forms in any way.

17. PIE *kwer– ‘oak’: IE *perkwo– ‘oak’ (Lat quercus, Goth fairguni ‘mountain chain’, ONorse fjor ‘tree’, OHG fereh-eih ‘oak-tree’) ~ IE *kwresno– ‘oak, brush’ (Gk prinos ‘holm-oak’, OHG hurst, horst ‘wood, wooded eminence’, OEng hyrst ‘hillock, height, wood, wooded eminence’, Welsh prys ‘woods’, Slav *xvorstu ‘oak, brush’). The semantic alignment between the two subsets is perfect. Under the new reconstruction, Lat quercus does not need to be explained through an (irregular) assimilation from *perkwos as *kw– is original in this superset. The meaning ‘wooded eminence’ for OHG hurst, horst and OEng hyrst is a good match for Goth fairguni ‘mountain-chain’, while the n-affix found in Gk prinos is the same as that of Goth fairguni. Slav *xvorstu likely derives from *skwer. In the light of the possibility that PIE had a complex palatolabiovelar phoneme k’w (by analogy with an aspirated labiovelar gwh), –s– in *kwresno– could represent a palatal reflex of PIE *k’w, while the –kw– of *perkwo– its labial outcome, so that a more complex morphological entity *kwerk’wo– could be assigned to the PIE level.

18. PIE *kwe– ‘father, father’s brother; older male relative’: IE *pH2ter ‘father’ (Toch A paacar, Toch B paacer, Skrt pita, Gk pateer, Lat pater, Arm hayr, Goth fadar, OIr athir) ~ IE *te– ‘father’ (Lith tevas, tetis ‘father’, OPruss taaws, towis ‘father’, thewis ‘father’s brother’, Skrt tata ‘father’ (r.), ‘any male relative or acquaintance’ (a.) [Karve 1953, 38]). Currently one can often read in literature that the laryngeal in *pH2ter does not directly translate into Goth or Arm –a– because in those languages laryngeals are regularly lost and not vocalized (e.g., Goth dauhtar, Arm dustr next to Skrt duhita, Gk thugater). So, for Gothic, it’s assumed that an interconsonantal laryngeal was first lost and then an anaptyctic vowel emerged, which later merged with a. For Armenian, a very short vowel is postulated to occur before or after an interconsonantal laryngeal (Ajello, Roberto. “Armenian.” In The Indo-European Languages, edited by Anna G. Ramat and Paolo Ramat. Pp. 197-227. Taylor & Francis, 1998, 203). For Tocharian, long aa is considered to be analogical with maacer ‘mother’ but this explanation does not work in light of Toch A ckaacar, Toch B tkaacer ‘daughter’. These piecemeal and ad hoc explanations can be dropped if we reconstruct IE *kweH2ter yielding *paH2ter and regularly leading to Toch *pacer, Germ. *fater and Arm *hathir. IE *paH2ter was in complementary distribution with *pH2ter. The latter form prevailed in Sanskrit, Greek and Latin. (For the original CeC- shape of this root see Kloekhorst, Alwin. “Indo-European Nominal Ablaut Patterns: The Anatolian Evidence,” in Indo-European Accent and Ablaut, edited by G. Keydana, Paul Widmer and Thomas Olander. Pp. 107-28. Copenhagen: Museum Tusculanum, 2013, 117.) Szemerenyi (1977, 6-7) suspected that the two sets are related and postulated *pte-wis to explain the Baltic group. The proposed new phonetic law simplifies the solution. The difficult Slav *stryju ‘father’s brother’ receives a satisfactory explanation as derived from *tetrujos > *ttrujos > *struju (morphologically identical to Skrt pitrvya and Lat patruus ‘father’s brother’ and in terms of the onset similar to OPruss thewis ‘father’s brother’). Compared to the phonetic chain proposed in (Vey 1931), this new one drops the most difficult first step that required the conversion of pt into tt. (See above for the origin of ORuss nestera ‘niece’ from *netetera < *nekw– instead of *neptera.) Similarly, the otherwise-puzzling Arm yawray ‘step-father’ can now be seen as regularly derived from *tatros (comp. Gk patroos). Arm hayr (< *kweH2ter), just like Arm hur ‘fire’ (< *kweH2ur) above show that the Armenian velar spirant h does not constitute a delabialized reflex of PIE *p (as in p > f > h) but descends more seamlessly from the labiovelar without the change in the place of articulation. It’s likely that Hitt atta, Slav *otici, Goth atta ‘father’ are also derived from *akwa-, and hence there was only one PIE term for ‘father’ (*kwe-, *akwe-) and not two (*pH2ter and *atta) as presently believed. Interestingly, in Gothic, the form fadar occurs only once, with atta being the main term for ‘father’. If the two forms are linked phonetically, the puzzle of why a PIE term for ‘father’ fell out of favor in Gothic simply disappears. Proto-Germanic *fathoo ‘father’s sister’ (OEng fathe, OFris fethe) is a perfect morphological match for Slav *teta ‘father’s sister’.

Unlike *pHter and *te-, which are isolated in a Eurasian perspective, PIE *kwe-, *akwe ‘father’ find plenty of potential cognates in the so-called “Nostratic” or “Eurasiatic” languages (comp. Nostr eka ‘older male relative’ such as ‘father’, ‘father’s brother’, ‘older brother’, ‘grandfather’, etc.).

PIE *gw

18. PIE *wegw– ‘water’, *wegwr ‘water creature’: IE *ngwr- ‘water snake, eel’ (Slav *ongri/*ugri ‘eel’, Lith ungurys, OPrus angurgis ‘eel’, Gk (Hes.) imbeeris ‘water snake’, Lat anguis ‘snake’, anguilla ‘eel’, angis ‘snake, adder’, OHG unc ‘adder’, OIr esc-ung ‘eel, water snake’) ~ IE *udro– ‘water snake, water animal, otter’ (Lith uudra, Slav *wydra ‘otter’, Skrt udras ‘water animal’, Avest udra ‘otter’, Gk hydros, hydra ‘hydra, water snake’, OHG ottar ‘otter’). The first root can be traced back to IE *wegw-/*ugw– ‘wet’ (Gk ugros ‘liquid, fluid’, ONorse vokr ‘wet, moist’ [EIEC 639, with further possibilities). The second root is transparently related to a more basic IE root *wed– ‘water’ (Hitt watar, Toch war, Gk hudoor, Skrt udan, Goth watoo, OHG wazzar, Slav *woda, Lith vanduo) which now needs to be reinterpreted as *wegw-. The nasal infix in the EEL set is secondary and intrusive. It sporadically emerges in the WATER set as well (Lith vanduo ‘water’, Lat unda ‘wave’).

19. PIE *ngwen- ‘bulge, testicle’: *IE *ngwen– (Lat inguen ‘bulge in the genital region, genitals’, Gk adeen ‘gland’) ~ Skrt andas ‘egg’, andam ‘testicle’, Slav *jendro ’round core, kernel’, *jendra pl. ‘testicles’.

21. PIE *gweHw-/*gweHi– ‘be, become, grow, live’: IE *gwiHwo– ‘live’ (Skrt jivati ‘he lives’, jiiva ‘living’, Toch A so, Toch B saw ‘live’, Gk zoooo ‘live’, bios ‘life’, Lat viivo ‘live’, viita ‘life’, Goth qius ‘living’, OHG quek, OEng cwic, ONorse kvikr ‘quick’, OIr beeo, Welsh byw ‘alive’, Lith gyvas ‘living’, gyju ‘become healthy’, Slav *zivu ‘living’, *ziti ‘live’) ~ *IE *bhewH– ‘be, become, grow’ (*bhewHiyo– [pres.]: Lat fiioo ‘become’, OIr biiu ‘become’, Welsh byddaf ‘be’, OEng beeo ‘am’; *bhewH-: Skrt bhavati ‘is’, Gk phuomai ‘grow, become’, phuo ‘beget’, Arm busanim ‘sprout up’, OHG buuan ‘live’, Lith buuti ‘be’, Slav *byti ‘be’, Alb buj ‘lodge, stay’; *bhewH-t-, *bhewH-d– [nomin.]: Skrt bhuuti ‘being’, Gk phyton ‘plant’, physis ‘nature’, phylee ‘tribe’, phytla ‘nature, species’, Lat futurus ‘yet to be, going to be’, OIr both ‘hut’, Welsh bod ‘dwelling’, Lith buutas ‘house’, buukla ‘residence’, Slav *bydlo ‘dwelling, residence; cattle’; *bhewH-s-: Lith busiont ‘future’, Slav *bystru ‘quick’. Semantic alignment between the two cognate sets is compelling (comp., especially, OHG buuan ‘live’ as well as OEng cwic and Slav *bystru). The key phonetic barrier separating them is aspiration in the onset of PIE *bhewH-. However, this can be attributed to the subsequent laryngeal, just like in IE *dhugH2ter a voiced aspirated stop comes from a combination “plain voiced stop + laryngeal” (see below). The variation in the root vocalism of IE *gwiHwo– and IE *bhewH– can be accounted for as a metathesis in the ancient affixed forms *gweHi-wo vs. *gweHw-yo. IE *gwous– ‘cow, cattle’ (Skrt gau, Avest gaus, Toch A ko, Toch B keu, Myc qo-u ‘cow’, Gk bous, Lat boos, Umbr bum, Arm kov, OIr boo, OHG chuo, OEng cuu, Latv guovs ‘cow’, Slav *govedo ‘bull, cattle’, Osset qug, gog ‘cow’), which currently does not have an etymology, seem to belong here as well. The general semantic fit is perfect (cattle was a source of life and a foundation of being for cattle-growing pastoralists). In addition, Slav *byd-lo ‘cattle’ (in Polish and eastern Slavic languages) provides an outstanding formal and semantic parallel to Slav *goved-o ‘cattle’.

PIE *gwh

22. PIE *ghwegwh-: ‘earth, ground, bottom’): IE *dheg’h-/*dheg’h– ‘earth’ (Hitt tekan, Toch A tkam, Toch B, kam, Gk khthoon, Lat humus, Skrt ksam) ~ IE *bhudh-/*dhubh– (Skrt budhna, Gk puthm?nLat fundus, OHG bodam/boden, Slav *du(b)no) ‘bottom’, Lith dubus ‘deep’, Goth diups ‘deep’).

23. PIE *sneghw– ‘sky, snow’: IE *sneigwh– (OPruss snaygis, Lith sniegas, Goth snaiws ‘snow’, OHG sniwit ‘it snows’, Gk neiphei ‘it snows’, nipha ‘snow’, Lat niivit, ninguitnix ‘snow’, OIr snigid ‘it rains, it snows“, Skrt snihyati “it gets wet”) ~ IE *Cnebho– (Hitt nepiš, Luw tappas ‘sky’, Lith debesis ‘cloud’, Gk nephos ‘cloud’, dnophos, gnophos ‘darkness’, Skrt nabhas ‘fog, sky’, OHG nebul ‘fog’, Slav *nebo ‘sky’). The onset remains somewhat enigmatic due to the unexpected d– in Luwian, Lithuanian and Greek. The consolidation of IE *sneigwh– and *Cnebho– cognate sets into one opens possibilities to resolve the puzzling onset but since the SKY group is always reconstructed as *nebho– the connection between the two sets can be easily established on the assumption of an s-mobile in the SNOW group. The diphthong in the sneigwh– group adds a nuance of difference to this comparison. It can be explained as the product of metathesis of an affixal vowel into the root, so that *snegwhyo– > *sneigwho– (see also *Hledhwero > *Hleudhero below).

24.  PIE *gwhe– ‘burn, smoke’: IE *gwher– ‘ burn’ (Skrt ghrnoti ‘shines, burns’, gharmas ‘summer heat’, haras ‘heat’, Gk theromai ‘am burning hot’, theros ‘summer, harvest’, thermos ‘warm’, Arm jer ‘warm, warmth’,OIr gorim, guirim ‘am warming up’, Lat formus ‘warm’, Alb zjarm ‘heat’, Slav *goreti ‘burn’, Lith gariu, gareti ‘burn, get enflamed with anger’, OPruss gorme ‘heat’) ~ IE *dhuH2mo– ‘smoke’ (Hitt tuhhuis, Skrt dhumas, Lat fuumus,Gk thuumos ‘spirit’ > Mod Gk ‘anger’, OHG toum ‘smoke, fog, steam’, Lith dumai, Slav *dymu). The original root *gwhe– is enlarged with –r-, –m– and –H2– determinants, or a combination thereof.

25. PIE *gwhei– ‘strike, slay, beat’: IE *gwhe(n)– ‘strike’ (Hitt kuenzi ‘he/she/it kills’, Gk theino ‘I strike’, phonos ‘manslaughter’, Skrt hanti ‘he/she/it strikes, kills’, Lith genu ‘I hunt, drive’, OCS *goniti ‘hunt, drive’ (< ‘beat repeatedly’) ~ IE *bhei– ‘beat’ (Gk phitros ‘log, tree trunk’, OHG biihal ‘axe’, Arm bir ‘club’, OIr benim ‘I cut, I strike’, Slav *biti ‘beat, strike’, *boju ‘fight’. The diphthong in Gk theino may come not from *thenyo– but represent the original diphthong also reflected in phitros. Alternatively, the bright vowel in phitros may parallel the same “anomaly” seen in Gk bios ‘life’ (< IE *gwiiwo-) instead of expected **dios [see Meier-Brugger 2003, 135].

26. PIE *gwher-/*gwhor– ‘door, court, garden’: IE *dhwer– ‘door’ (Lith durys ‘door’, dvaras ‘court’, Goth daur ‘gates’, OHG turi ‘doors’, Gk thura ‘door’, Lat forees ‘double-sided door’, foris ‘door’, forum ‘court’, Skrt dvaras ‘doors’, Alb dere ‘door’, Slav *dveri ‘door’, dvoru ‘court’) ~ IE *ghordho-/*ghorto– ‘garden, town’ (Skrt grhas ‘house’, Lat hortus ‘garden’, Alb garth ‘fence’, Goth gards ‘house’, Toch B kerciye ‘palace’ [< *ghordhiiom], Lith gardas ‘fence’, Slav *gordu ‘town’).

27. PIE *gwhegwh– ‘yearn, long for, be greedy, ask for’: IE *gwhedh– ‘yearn, desire’ (Skrt thessasthai ‘pray for, beseech’ [< *gwedhiyo-], pothos ‘desire, need, longing’, Avest jaidyat ‘he asked’, OIr guidim ‘I beseech’, Lith pasigendu ‘feel a lack of something’, gedauju ‘wish, yearn, strive’, godas ‘greed’, godus ‘greedy, stingy’, godziuos ‘wish, desire’, Slav *zendati ‘desire, be thirsty’, *zadnyji ‘greedy’) ~ IE *bhogho– ‘devour, part, lot’ (Gk phagein ‘swallow, eat up, devour, consume’ [< *phagFein?), Skrt bhajati ‘endows, allots’, Skrt bhagas ‘lot, happiness’, also ‘bestower’ as an epithet of gods, Avest baxsaiti ‘partakes’, baga– ‘lot, part’, Slav *bogu ‘god’). The semantic connection between the cognate subsets reveals a complex, reciprocal notion of yearning for a missing part (by people) and bestowing it (by gods upon people). The strength of the semantic link becomes evident if one observes that both cognate subsets preserve both the original religious meanings (e.g., Gk thessasthai and Slav *bogu) as well as the reinterpretation of this religious concept as a physical state of being thirsty, hungry (Slav *zendati) or of the act of satisfying hunger (Gk phagein). Interestingly, the second PIE labiovelar reflected as plain g in Gk phagein behaves like labiovelars are known to behave in satem languages (comp. the reflex of the first labiovelar in Lith gedauju) suggesting that the centum-satem division in the treatment of the labial component of a labiovelar (the loss of w in satem languages and the merger of labiovelars with velars) reflects a positional feature of certain roots in both centum and satem languages and not a unique phenomenon of satem languages.

27. PIE *H1legwh– ‘grow freely’: IE *H1legwhu-/*H1legwhro– ‘light, quick’ (Lat levis ‘light’, Gk elakhus ‘low’, elaphros ‘quick, light’, Ved raghu ‘quick, swift’, laghu ‘light, low’, OHG lungar ‘fast’, Goth leihts [with a nasal infix > *H1lngwhu-/*H1lngwhro-], OIr laigiu ‘less, worse’, Slav *liguku ‘light, easy’, Latv liegs ‘light’) ~ IE *H1leudhero– ‘free’ (Skrt rodhati ‘he/she/it grows’, Goth liudan ‘grow’, Gk eleutheros ‘freeman’, Lat liiber ‘freeman’, liiberi ‘children’, OHG liut ‘people’, Lith liaudis ‘people’, Latv laudis ‘people’, Slav *liudu ‘people’. In IE *H1leudhero– ‘free’ the diphthong is likely secondary and emerged from an earlier *H1ledwhero-. The semantic match is perfect as the inclusion of the LIGHT-QUICK set fills is the logical gap between the meaning ‘grow’ and the meaning ‘free’ in the FREE set. The presence of the meaning ‘low’ in the LIGHT-QUICK set (Gk elakhus, Ved laghu) fits the meaning ‘grow’ found in the FREE set. The meaning ‘children, offspring’ recorded in Latin does not look aberrant (as it seems to be among other social terms such as OHG liut, Lat liiber and Gk eleutheros) but ties back to the broader PIE notion of ‘quick, early growth’ supported by both sets. The emergence of the social meaning ‘freeman’ > ‘people’ clearly postdates the split of this cognate set into the LIGHT-QUICK and the FREE groups, as it’s not present in the LIGHT-QUICK group, while both groups show the original organic and pre-social semantic core.

28. IE *(H1)neghwro– ‘kidney’ (Gk nephros ‘kidney’, OHG nioro ‘kidney’, Lat nefrones ‘testicles’) ~ Slav *needro, pl. needra ‘chest, internal area in the body, breasts’. The Slavic form is considered difficult from an etymological perspective. H1 in the KIDNEY set accounts for the long vowel in the Slavic form.

29. PIE *(s)neghw-/*nogwh– ‘dusk, night, cloud, mist’: IE *negwh-/*nogwh– ‘night’ (Hitt nekuz, Toch A nakcu, Toch B nekc, Gk nuks, Lat nox, Skrt nakti, Lith naktis, Goth nahts,  Alb nate, Slav *not’i) ~ IE *sneudh– ‘dusk, cloud, mist’ (Gk nuthon ‘dusk’, Lat nuubees ‘cloud, mist’, Avest snaoda ‘cloud’, Welsh nudd ‘mist’). The NIGHT set is usually reconstructed as *nekw-/*nokw– but the medial stops in all the IE languages is equally compatible with PIE *gwh. Before t labiality, voice and aspiration would all disappear.

Cases with secondary aspiration

As I argued elsewhere (see, e.g., on this blog), some IE voiced aspirates can be explained as a combination of voiced stop + laryngeal, in which a laryngeal is found not adjoining the voiced stop but in the subsequent segment of the form. This rule can explain such a well-known IE word usually reconstructed as PIE *bhreH2ter (> Lat frater, Gk phrater, etc.) as representing earlier mr– (also found in Lat maritus ‘husband’, Gk meiraks ‘boy, girl’, Germ *brudi– ‘bride’, Latv marsa ‘brother’s wife’, etc.). The aspiration inferred from Lat frater and Gk phrater is the result of a feature throwback from the medial laryngeal in the underlying form *mreH2ter. This insight leads to the following comparanda featuring voiced labiovelars, voiced aspirates and medial laryngeals.

30. PIE *gwer-/*gwor– ‘devour; mouth, throat’ > IE *gwer– ‘devour, throat’ (Lith gerti, geriu ‘drink’, gurklys ‘throat, crop, craw’, Latv dzert, dzeriu ‘drink’, OPruss gurcle ‘throat’, Skrt girati, grnaati ‘he devours’, garas ‘potion’, Avest gar– ‘devour’ (in compounds), Arm keri ‘I ate’, Gk bora ‘fodder’, bibroosko ‘I eat, devour’, barathron ‘gulf, pit, muzzle’, Arcad dzerethron, Ion berethron, Lat vooro, voraare ‘I devour’ , Slav *zreti ‘devour’, *gordlo ‘throat’ ~ IE *bhardheH2– ‘beard’ (Lat barba, OHG bart, OEng beard, Lith barzda, Latv barzda, baarda, OPruss bordus, Slav *brada ‘beard, chin’). There is distributional synergy between the two cognate sets: Greek forms are missing from the BEARD set but they are abundantly present in the DEVOUR set. The semantic match is very good: beard grows in the same area of the human head where mouth and throat are located. The consolidation of these two IE cognate sets provides a clear etymological path for the BEARD forms. The morphology of Gk barathron and Slav *gordlo shows the same voiced aspirate *dh as the BEARD forms. In Lat barbab– is not aspirated (comp. Gk barathron), which has always been a problem. Regressive assimilation from –b– of the second syllable (*farba > barba) or a Latin version of Grassmann’s Law (*bhardha– > *bardha– > barba) are the two main explanations. Importantly, both explanations invoke an intersegmental process. For the BEARD forms a laryngeal is usually reconstructed (e.g., EIEC 251) to account for the last –a. It’s precisely this laryngeal that could have added aspiration to the originally unaspirated root *gwer-/*gwor-, so that *gwerdeH2– yielded *bhard-, *bardh– (Lat barba, Gk barathron) and *bhardh– (Goth bart) forms.

31. PIE *gwrew– ‘scruff, brow’: IE *bhrewH– ‘eyebrow’ (Toch A parwaam, Toch B parwaane [dual] ‘brows’, Skrt bhruu-‘ brow’, Avest brvat [dual] ‘brows’, Gk ophruus ‘brows’, OEng bruu ‘brow, eyelash, eyelid’, Eng brow ‘eyebrow, forehead’, ONorse bruun ‘brow’, OIr forbruu ‘brows’, Lith bruvis ‘brow’, Slav *brovi ‘brows’) ~ *gwrewio– ‘scruff, mane’ (> *gwreiwo-) (Skrt griiva ‘back of the head, scruff’, Avest griiva ‘back of the head’, Gk deree ‘neck, back of the head, throat’, Ion deiree ‘neck, throat’, Latv griva ‘mouth of river’, Slav *griva ‘mane’). The semantic connection between eyebrow, eyelash and eyelid, on the one hand, and scruff and mane, on the other, is motivated by both meanings referring to secondary forms of head hair. Both subsets show semantic development from hair to the underlying part of the head: mane > back of the head, eyebrow > forehead. (Notably, in the DEVOUR-BEARD set forms for ‘beard’ show a similar semantic evolution from beard to chin.) The SCRUFF-BROW cognate superset is similar to the preceding DEVOUR-BEARD superset and they may ultimately be related. But for the time being it’s better to keep them separate to avoid excessive semantic latitude. In addition, the DEVOUR-BEARD set and the SCRUFF-BROW set are distinct morphologically. PIE *gwrew– ‘scruff, brow’ is united by the distinctive –w– enlargement with pan-Indo-European distribution. But the emergence of a voiced aspirate from a sequence ‘voiced stop + laryngeal’ can clearly be in seen in both supersets. PIE *gwrewH- and *gwerdeH2- yielded, respectively, *bhrewH– and *b(h)ard(h)H2-, in both cases with a laryngeal throwback.

32. PIE *gwhe-/*gwe– ‘daughter, wife, junior female relative’: IE *dhugH2ter ‘daughter’ (Lyc cbatra, Luw tuwatari, Toch A tkacer, Toch B ckacer, Skrt duhita, Gk thugater, Osc fuutrei (, Goth dauhtar, Lith dukte, Arm dustr, Slav *dutji (OCS dušti)) ~ IE *gweneH2/*gwenH2 ‘wife, woman’ (Gk gunee, Goth qino, Slav *zena, Arm kin, Skrt janih, Toch B sana). The two sets are etymologically linked at the root *gwhe-/*gwe– enlarged by the gH2– + ter suffixes in the DAUGHTER forms and by the nH2/eH2 suffix in the WIFE forms. Lyc cbatra (Kloekhorst reconstructs *dwetr– and this reconstruction secures –w– in this form) likely stems directly from *gwatra and not from *twatra as it’s currently assumed. Aspirated dh in *dhugH2ter may be explained as a throwback from the following laryngeal, so that *gwe– > *du– > *dugH2– > *dhugH2– > *dhugH2ter. As I argued earlier, the same process likely accounted for the origin of IE *bhreH2ter ‘brother’ from the original PIE *mreH2ter < *mer– ‘affine’. Italic represented by Osc fuutir shows that the correspondences PIE *dh ~ Italic f and PIE *gwh ~ Italic f are in fact one and the same correspondence PIE *gwh ~ Italic f. The ending –nH2/-neH2 in the WIFE forms may be a hypocoristic.

Latin doesn’t have either a DAUGHTER or a WIFE reflex of the PIE *gwhe-/*gwe– root suggesting that the form got lost in the history of Italic at the time when it still had the undifferentiated DAUGHTER-WIFE meaning. And indeed Osc fuutrei refers to an epithet of a goddess from Ceres’s circle and not specifically to ‘daughter’. The semantic connection between ‘daughter’ to ‘woman, wife’ can be illustrated in the light of the peculiarities of Roman marriage. There were two kinds of legal marriages (justae nuptiae, justum matrimonium, legitimum matrimonium) in Rome: cum conventione in manum and sine in manum conventione. In the former case, a woman severed her ties with her natal family and became materfamilias in the family of her husband. She was entitled to a share of inheritance in her new family. In the latter case, a married woman remained a member of her father’s family. Uxor was a woman who entered into a connubium on the sine in manus conventione basis and was therefore a wife of her husband and a daughter of her father at the same time. In Cicero’s (Top. 3) words, “uxor is a genus of which there are two species; one is materfamilias, ‘quae in manum convenit;’ the other is uxor only.” (Materfamilias was a kind of daughter to her husband as well, and jurist Gaius in Institutes (I, 3; I, 118; II, 159) literally says just that – Usu in manu conveniebat quae anno continui nupta perseverabat: nam velut annua possessione usucapiebatur, in familia viri transibat filiaeque locum obtinebat.) (L.S. 1870, 740; Bierkan et al. 1907, 310). A 19th century German legal scholar Rudolph Sohm (2002, 365-366) echoes him saying that “An uxor in manu (materfamilias) stands legally, by virtue of the manus, ‘filiaefamilias loco’. The relations between her and her husband – both as regards her person and her property – are governed by the same rules of law as apply to the case of a child… The relations between a wife in manu and her children are governed by the same rules as apply between brother and sisters.” From an early period, however, this transfer of a woman from her father’s manus to her husband’s manus began to be avoided, at least in wealthy families, thus retaining the woman, along with rights to her dowry and inheritance in her familia of origin. Manus was rare by the end of the Republic and virtually extinct by Gaius’s time (Gardner & Wiedemann 1991, 6, n. 1).

The newly reconstructed onset for the PIE term for ‘daughter; wife’ *gwe– casts light on the enigmatic Lat uxoor ‘wife’. The Latin form can now be easily derived from *weksoor < *gweksoor (PIE *gw regularly gives v in Latin as in viivos ‘living’ < *gwiivo-) < *gwegH2soor (by analogy with *sosoor ‘sister’) < *gwegH2ter. Apparently in proto-Italic there were two complementary roots – *gwegH2– (> Lat uxoor) and *gwheg– (Osc fuutir). The unaspirated onset of *gwegH2– finds support in IE *gweneH2 ‘wife’ and in Skrt duhita ‘daughter’. Skrt duhita is usually taken as an example of Grassmann’s Law, which presumably operated independently in Indic, Greek, Tocharian and Latin, but considering that the onset of *gweneH2 is unaspirated either it’s more likely that *dh reconstructed on the basis of Gk thugater, Arm dustr and Goth dauhtar represents a secondary development caused by “aspiration throwback,” with the aspiration feature derived from medial H2. This means that only one of two consonants could carry aspiration in PIE: Gk thugater comes from *dhugH2ter, while Arm dustr and Goth dauhtar from *dhugter (and not *dhughter).

33. PIE *gwheighw– ‘build from stone or clay’: Slav *zidati build, erect, create (especially from stone or clay)’, Lith žiedžiu ‘form, shape from clay’, žaidas ‘oven’, židinys ‘hearth’, Goth deigan ‘form out of clay, smear’, OHG teig ‘dough’, Skrt dehmi ‘I smear’, dehi ‘damb’,  Gk teikhos ‘stone wall’, toikhos ‘wall’, Lat fingoo, fictus ‘I form, smear’. This is a well established cognate set that shows a root alternation suggestive of an original pair of labiovelars: Balto-Slavic points to *zeid, while all the other dialects *dheigh-. These forms can be linked to an otherwise isolated Greek-Slavic isogloss – Slav *ziduku ‘liquid, watery’, *ziza ‘slop’, Gk deisa ‘sludge’ (< *gweidhya-) showing that the original protoform must have been *gweigwH-, with the subsequent collapse of gw and H into gwh and a laryngeal throwback seen in Goth deigan.

34. PIE *gwor– ‘move quickly’: IE *gworH– (Gen. *gwr-H-) ‘mountain, forest, wind’ (Skrt giris? ‘mountain’, Avest gairi– ‘mountain’, Gk (Hom.) boreees,  (Att.) boreaas, boraas ‘northwind’, OPruss garian ‘tree’, Lith giria ‘forest’, Latv dzira ‘forest’, Slav *gora ‘mountain’,  Alb gur ‘rock’) ~ IE *bhur– ‘move quickly’ (Skrt bhurati ‘moves quickly, twitches, fidgets’, Lat furoo ‘rage’, Latv bauruot ‘to moo’, OIr buriud ‘mooing’, burach ‘tear the ground’, ONorse byrr ‘tailwind’, MidLowGerm borelos ‘windless’, Lith paburmai ‘furiously, rapidly’, Slav *bura ‘storm’). The laryngeal is reconstructed for the first set but not for the second but aspiration shows up in the second set. But the semantic link carried by Gk boreees, boreaas, boraas ‘northwind’ and ONorse byrr ‘tailwind’, MidLowGerm borelos ‘windless’, Slav *bura ‘storm’ is strong. Apparently, the abstract concept of ‘quick, violent movement’ stands behind such diverse natural referents as ‘mountain’ (product of violent natural forces), ‘forest’, ‘wind’, etc.

35. PIE *gwer-H– ‘carry; heavy’: IE *bher– ‘carry’ (Skrt bharati, bibharti ‘he carries, brings’, bhr?tis? ‘carrying, content, reward’, Avest baraiti ‘he carries’, Gk phero ‘I carry’, Arm berem ‘I carry’, Alb bie ‘I bring’, Lat fero ‘I carry’, fors ‘occurrence’, Goth baira ‘I carry’, baurþei ‘burden’, OEng byrðen ‘load, weight, charge, duty’, OIr biru ‘I carry’, brith ‘birth’, Slav *brati ‘to take’) ~ IE *gwreHu-/*gwrHu– ‘heavy’ (Toch A kraamaarts ‘heavy’, B kraamaar ‘weight, heaviness’, kramartse ‘heavy’, Skrt guru– ‘heavy’, Avest gouru– ‘heavy’, Lat gravis ‘heavy’, Gk barus ‘heavy’, Welsh bryw ‘live, vigorous, strong’, Latv gruuts ‘heavy’), Alb zor ‘heaviness, trouble’. The Tocharian forms (< *gwremr-) are especially close morphologically to Gk pherma ‘foetus’, Skrt bharma ‘care, preservation’, bhariman ‘carrying, preservation’, Avest bareman ‘carrying, preservation’, Slav *brema ‘burden, pregnancy’ (< *bhermn). If in the HEAVY set the laryngeal is reflected as a in Greek and Latin, in the CARRY set it must have turned an original voiced stop attested in the HEAVY set into a voiced aspirate. The meaning transformation went from the notions of ‘carry’ and ‘burdensome’ to the notion of ‘heavy’.

A Greek-Slavic isogloss *gwrebh– represented by Gk brephos ‘foetus, new-born, baby’ and OCS zrebe ‘foal’ seem to be of the same root *gwerH– enlarged with suffix –bho– (< –bo-, see below, no. 36).

36. PIE *gweH-l– ‘to mother’ > ‘to nurse, to give birth’: IE *gwelbho– ‘womb’ (Gk delphus ‘uterus’, adelpheos ‘brother’ [<*sm-gwelbheyos ‘from the same womb’], dolphos (Hes.) ‘womb’, Skrt garbha ‘uterus’, Avest gerebus ‘new-born animal’, garewa ‘uterus’, Goth kalbo ‘calf’, OHG chalb, chalp ‘calf’, kilbur ‘ewe lamb’, OEng clifor lamb ‘ewe lamb’, cealf ‘calf’, ONorse kalfr ‘calf’) ~ IE *dhe-l– ‘to suckle, to nurse, to milk, to mother’ (Gk theelee ‘mother’s breast’, theelus ‘female (animal), feminine’, Lat feelo ‘suck’, feelix ‘fertile, happy’, Arm dal ‘colostrum’, OHG tila ‘woman’s breast’, OEng delu ‘nipple’, ONorse dilkr ‘lamb’, Lith pirm-dele ‘cow which bears a calf for the first time’, Latv dels ‘son’, Alb dele ‘sheep’, MIr del ‘nipple’, delech ‘milk cow’. The new principle of grouping of cognates brings together such semantically and phonetically perfect matches as Gk a-delpheos ‘brother’ and Latv dels ‘son’, ONorse kalf ‘calf’/OHG kilbur ‘ewe lamb’ and ONorse dilkr ‘lamb’.

The PIE root *gweH– also spawned a variety of forms with different affixation: Gk *theesthai ‘milk a cow’, theesato ‘he sucked’, theenion ‘milk’, Skrt dhayati ‘suckles’, dhatri ‘nurse’, dadhi ‘yogurt, sour milk’, dheena ‘milk cow’, Avest daeenu ‘female quadruped’,  Lat feetus ‘offspring, pregnancy’, feemina ‘woman’, OIr dinu ‘lamb’, Goth daddjan ‘to nurse’, OHG diien ‘to nurse’, Arm diem ‘suck’, Latv deju ‘suck’, OPruss dadan ‘milk’, OCS dojiti ‘give breast’, deeva ‘maiden’ (the ending is like that of Gk theelus ‘female (animal), feminine’ [<*dhelFos?]), deeti ‘child’. Although there is a clear semantic specialization around the notion of ‘nursing’, ‘milking’ and ‘suckling’ the more neutral meaning of ‘mothering’ (whether through birth or through nursing) is seen in such form as Lat feetus.

Phonetically, aspiration initially carried by a laryngeal (prior to its loss with compensatory vowel lengthening) was redistributed across the consonants of the morph: e.g., in Gk delphus (< PIE *gweHlbo-) the second consonant became aspirated, while the first one stayed plain, while in Gk theelus (< PIE *gweHlwo-) the first consonant got aspirated (in the absence of the second one). The same pattern can be seen in Gk pherma ‘foetus’ and brephos ‘foetus’ (see above, no. 35).

37. PIE *gwhegwho– ‘limb, run’: IE *bhaghus/*bheH2ghus ‘(fore)arm, foreleg’ (Toch poke ‘arm’, Toch B pokai ‘arm, limb’, Skrt baahu ‘forearm, arm, forefoot of the animal’, Avest baazu ‘arm, foreleg’, Gk peekhus ‘elbow, forearm’, OHG buog ‘shoulder’, OEng buog ‘shoulder, arm, bough’, ONorse boogr ‘arm, shoulder’) ~ IE *gwhegw– ‘run’ (Gk phebomai ‘I run’, phobos ‘stampede, fear’, Lith begti ‘run’, Latv begt ‘run’, OCS beezati ‘run’). Notably, Balto-Slavic *beegeeti ‘run’ attests a long vowel that’s prominent in the forms for FOREARM-FORELEG but is missing in Gk phebomai. IE *bhaghus shows –u– suggestive of an earlier labiovelar. The original meaning likely referred to the animal, rather than human body part and its action. The contrast between Gk peekhus/Skrt baahu and Gk phebomai parallels the contrast between Skrt duhita (expected Gk tukhater**) and attested Gk thugater.

Complex cases

There are cognate sets in which the sound correspondences characteristic of the descendants of PIE labiovelars have been clouded by seemingly irregular processes involving sonorants.

38. PIE *gwhn-/*gwhngwhe– ‘mouth, cheek, jaw, tongue’: IE *dng’huH2-/*dngwheH2– ‘tongue’ (Lat lingua, OLat dingua, Osc fangvam, Goth tuggoo, Skrt jihvaa, Avest hizva, Slav *jenzyku, Lith liezuwis, Arm lezu, Toch A kantu, Toch B kantwo, OIr tengae) ~ IE *g’enu-/*g’endho– ‘cheek, chin, jaw’ (Gk genus ‘chin, jaw’, gnathos ‘jaw’, Skrt hanus ‘jaw’, Avest zaanu ‘jaw’, Lat gena ‘cheek’, Goth kinnus ‘cheek’, OIr glun‘ mouth’, Arm cnaut ‘jaw, cheek’, Toch A sanwem ‘jaw’, Lith zandas ‘jaw, cheek’). These are two classic IE roots widely attested across IE dialects and subjected to a myriad of phonological analyses. The semantics of the two sets holds no barriers for comparison, as both sets refer to the physical area of the mouth. The formal properties of the members of the TONGUE set have been most difficult to reconcile phonologically. While it’s generally agreed that the IE words for ‘tongue’ go back to *dnghuH2-/*dngwheH2-, the Osc fangvam points to *dhenghwen, Skrt jihvaa and Avest hizva to *gighwaa, Slav *jenzyku and Lith liezuwis to Balto-Slav *ingwu-. The CHEEK-CHIN-JAW set contains one “anomaly,” namely Skrt hanus which points to *g’henu. The TONGUE set contains a familiar “metathesis” (Toch A kantu, Toch B kantwo) that we’ve also encountered in the EARTH and BUILD sets and that seems to represent a “signature” of sets containing two labiovelars. This “metathesis” reveals the formal similarity between the two sets (Toch A kantu, Toch B kantwo next to Toch A sanwem). Skrt jihvaa doesn’t look like a product of assimilation anymore, but is rather a regular outcome of PIE *gwhngwhe-. Osc fangwam, too, now logically continues *gwhngwen. OIr glun ‘mouth’ (transparently from *gnun) may hold a key to the origin of the obscure l-onset in Lith liezuwis and Arm lezu. While it’s commonly assumed that the l– in the Lithuanian and Armenian forms suggests two independent cases of contamination with the IE verb ‘to lick’, it’s more likely that liezuwis and lezu comes from *dningwu– or, better, *gningwu- (comp. Slav *gneezdo ‘nest’ next to Lith lizdas) followed by *glingwu- and *lingwu. Gk gnathos and Lith zandas contain a clear affix *-dh-, which corresponds to the medial *-g’wh– in the TONGUE set morphologically but to the initial *d(h)– in the TONGUE set phonologically. This suggests that the medial *-g’wh– in the TONGUE set is a secondary enlargement of the original shorter root *dn-/*gwhn-. With this analysis, the d-forms found in the TONGUE set (OLat dingua, Goth tuggoo, OIr tengae) increasingly look marginal and secondary (in the –e/-n environment?) to the g– forms, which find full support in the entirely d-less CHEEK-CHIN-JAW forms. Gk gloossa ‘tongue’, which is sometimes considered as another anomalous member of the TONGUE set, now fits the larger MOUTH-CHEEK-JAW-TONGUE set more organically as it displays an original velar in the onset and the intrusive –l-, which can now be explained, following the lead of OIr glun ‘mouth’, Lith liezuwis and Arm lezu ‘tongue’, as stemming from *gnundhya > *glundhya > *gloossa. Alb gjuhë (< *glusaa [Orel 1998: 138]) must have undergone a similar development to the Greek form.

39. PIE *gw(h)egw(h)no-/*gw(h)egw(h)ro– ‘smooth, hairless, naked’: IE *negwno-/*negwro– ‘naked, hairless’ (Hitt nekumant, Gk gymnos, Skrt nagnas, Avest magna, Arm merk, Lat nuudus (< *nogwodos), Goth naqaths, OIr nocht, Lith nuogas, Slav *nagu) ~ IE *ghladh– ‘smooth’ (Lat glaber ‘smooth, hairless, bald’, OHG glat ‘smooth, shiny’, OEng glad ‘smooth’, Lith glodus, glodnas, glodnus ‘smoothly lying upon’, Slav *gladuku ‘smooth’. The suffixal morphology of the two sets is fully compatible: the *-no-/*-ro– ending is present in both cases. The semantic match is perfect, especially considering that IE *negwno– is thought to mean ‘naked’ in the sense of ‘hairless’, not just ‘undressed’ (EIEC 45). The NAKED set is widely considered irregular due to the unexpected sonorants in Avest magna, Arm merk, Gk gumnos and a “metathesis-like” structure of Gk gymnos (assumed to be from *nogwno– > *gwon-no– > gymnos). The presence of –l– in IE *ghladh– seems to be an insurmountable barrier for the consolidation of the two cognate sets. However, the existence of Arm lerk ‘bald, hairless, soft’ and olork ‘smooth, polished’ (both likely related but without a good etymology [Falileyev, Alexander, and Petr Kocharov. “Celtic, Armenian and Eastern Indo-European Languages: Comments on a Recent Hypothesis.” In Ireland and Armenia: Studies in Language, History and Narrative, edited by Maxim Fomin, Alvard Jivanyan and Seamus Mac Mathuna. Washington, 2012,  72-3]) overcomes this difficulty. Representing *legwro-, they form a bridge between such a form as Lat glaber ‘smooth’ and Arm merk ‘naked’ (< *megwro-). A plausible protoform that can account for the –l– of Arm lerk, olork and the –l– of IE *ghladh-) is *gwnegw(h)no-/*gwned(h)no-> *gwlegwno-/*gwled(h)no-. From *gwlegwno-/*gwled(h)no- the –l– spread to their heteroclitic counterpart *gwnegwro-/*gwled(h)ro-. Another challenging aspect of the proposed etymology is the presence of voiced aspirates in the SMOOTH set. Germanic forms (OHG glat, OEng glad) indicate that (at least from the point of view of the classic model of PIE phonology) both stops were aspirated in PIE. But PIE *gh is regularly lost before l in Latin, so the expected form is **laber, not *glaber. It’s assumed, therefore, that Latin underwent the dissimilation of aspirates (*ghladhro– > *gladhro-) akin to the process known as Grassmann’s Law described for Greek and Sanskrit but independent from it. But the –b– in glaber does not directly suggest aspiration either (it’s assumed that PIE *dh reflects as f word-initially but as b medially but the fact remains). Under the new interpretation proposed herein, Gk gymnos is expected to go back to *gugw(h)nos (comp. Gk amnos ‘lamb’ < *abnos > *agw(h)nos). The rest of the IE NAKED forms must therefore represent *gwhegwno– assimilated to *gwhnegwno– with the subsequent loss of initial gwh– in a cluster environment.


On the Dual Reflexes of Indo-European Laryngeals: A Note on Jouna Pyysalo’s “System PIE: The Primary Phoneme Inventory and Sound Law System for Proto-Indo-European”

Sunday, December 22nd, 2013

Jouna Pyysalo has written a very bold and thought-provoking dissertation that revisits some fundamentals of Indo-European historical phonology. He attempts to bring Indo-European linguistics, which has been plagued in the 20th century by fanciful multilaryngeal reconstructions, back to its frugal Neogrammarian virtues. Following Zgusta, Szemerenyi, Tischler and Burrows (but strangely missing Russian monolaryngealists such as V. Dybo, S. Starostin, A.Kassian), he rejects the complex combinatorics of H1+e, H2+e, H3+e, H4+e frequently postulated by Indo-Europeanists in favor of a single laryngeal /H/ firmly attested in the so-called “Old Anatolian languages” (Hittite, Luwian, Hieroglyphic Luwian, Cuneiform Palaic and Cappadocian) but lost elsewhere. The intriguing part of the dissertation is the hypothesis that Old Anatolian /H/ covered a voiceless and a voiced variants inherited from the Proto-Indo-European (PIE) stage and that they had no coloring effect on neighboring vowels (in the languages outside of the Old Anatolian cluster) and no syllabic (vocal) allophones in PIE. To account for the statistical regularity of association between Old Anatolian h and Narrow Indo-European a, Pyysalo postulates for PIE diphonemic sets *Ha and *aH: while laryngeals as obstruents get lost in Narrow Indo-European and New Anatolian languages (Lycian, Lydian, etc.), the associated vowel survives (or changes into other vowels, according to the phonetic rules of individual IE branches). Along the way, Pyysalo also simplified a portion of the Neogrammarian legacy by deconstructing a) labiovelars – hypothetical phonemes poorly unattested in actual IE languages – as a combination of velars and /w/; b) palatovelars as combinations of velars and /y/; c) voiced aspirates as combinations of voiceless stops and the voiced laryngeal, by analogy with the more widely accepted interpretation of Indic voiceless aspirates as combinations of voiceless stops and the (voiceless) laryngeal. Pyysalo’s c) has been my conviction since I’ve realized that *bhr– in the IE cognate set BROTHER (PIE *bhreH2ter > Lat frater, Gk phrater, etc.) derives from earlier mr– (also found in Lat maritus ‘husband’, Gk meiraks ‘boy, girl’, Germ *brudi– ‘bride’, Latv marsa ‘brother’s wife’, etc.) and that the aspiration of *bh– is the result of a feature throwback from the medial laryngeal in *mreH2ter) (see Dziebel G. V.2006. “Reconstructing ‘our’ kinship terminology: Comments on the Indo-European material in A. V. Dybo’s and S. V. Kullanda’s The Nostratic terminology of kinship and affinity,” in Algebra rodstva 11, 67-68). This etymology supports Szemerenyi’s (“Introduction to Indo-European Linguistics,” 1996, 144) contention (quoted by Pyysalo, p. 398):

“Since according to our conclusions the ‘laryngeal’ was a glottal spirant h, it is also clear that the unvoiced and voiced aspirates originally represented the combinations unvoiced stop+h and voiced stop+h, which in Indo-European counted as monophonematic.”

Pyysalo ends up with the following PIE phonemic inventory:


Pyysalo (p. 58) puts Indo-European linguistics on a hot seat by observing that


While it’s rewarding to see an Indo-Europeanist openly talking about the problems with Indo-European reconstructions (instead of just praising the remarkable achievements and the scientific rigor of comparative Indo-European studies), the root of these problems, as I have extensively wrote on the pages of the Algebra rodstva almanac from 2000 on, is not in the inability of Indo-Europeanists to come to terms with the Old Anatolian data. It’s the methodological weakness of the traditional comparativist method that is holding the Indo-European historical phonology back. This weakness stems from the logical contradiction inherent in the definition of a cognate set. Sound laws are supposed to be derived from the material assembled in the least law-like manner – by simple visual inspection in search of similarities in form and meaning. This approach can detect obvious cognates but it’s not well suited to detect cognates subject to more dramatic transformations of form and meaning due to accelerated language change, great time depth or other factors. Under the current approach to linguistic comparativism, formal similarity dominates over semantic relationships. Only once a protoform is reconstructed, it’s subjected to semantic interpretation. As a result, sounds laws are constrained by a linguist’s naive perception of “similarity” in the lexical material, while proposed etymologies are often dubious on semantic and anthropological grounds.

A gignetic approach to the reconstruction of kinship terminological systems and to linguistic comparativism broadly focuses on this initial stage of cognate set composition and uses semantic variation as an important counterpart to formal variation. Semantics is not an afterthought but an essential criterion actively contributing to a more in-depth definition of a cognate set. A more complexly defined cognate set – in reality a combination of two or more one-dimensional cognate sets – should yield a more sophisticated, complete and sound laws of linguistic change in a particular family of languages.

Pyysalo (p. 80) feels himself on a firm ground when he declares:


It’s precisely when the doubts seem to recede completely that the traditional comparativist method shows its limitations, while the gignetic method creates a breakthrough. The first cognate set listed by Pyysalo contains forms with the initial velar (Slav *kosti ‘bone’, Lat costa ‘rib’) (conveniently omitted by Pyysalo), while the second one cannot be divorced from PIE *pek’– (Skrt pášu, Avest pasu ‘cattle’, Lat pecu ‘cattle’, pec?nia ‘money’, Goth faíhu ‘money, movable goods’, OHG fihu, OEng feoh, ONorse f? ‘livestock, property, money’, Lith pekus ‘cattle’). The perfect semantic fit between the two sets can be clearly observed in such forms from the first group as Gk ??? ‘flock of sheep’, ???µ?? ‘guardian, herder’.

Having the alternation such as *peH-/*pek’– in mind, we can observe the same correspondence between, for example, Hitt haluga ‘message’ (a word with obscure etymology under the known sound laws) and PIE *k’lewo– (Gk kleFos ‘glory’, klu? ‘I hear’, Skrt srávas, Old Arm lulsem, Slav *slovo ‘word’, etc.). Similarly, IE *H3osdo– ‘branch’ (Hitt hasduir, Arm ost ‘Ast, Zweig’, Gk o?zdos ‘Ast, Zweig’, Goth asts ‘branch, palm branch, leafy branch’, etc.) shows parallel forms with an initial velar (possibly, labiovelar, which would be consistent with the labial component of /H3/ inferred from such forms as Gk o?zdos): OHG questa ‘bunch’, ONorse kuistr ‘branch’, Slav *gvozd ‘forest’ (> Russ gvozd’ ‘nail’). Also, Hitt huidar, HLuw huidar ‘wild animals, fauna’, Palaic huidumar ‘Lebe, Lebenwesen’ can be linked to PIE *gwiwo– ‘life’ in addition to forms such as ONorse vitni ‘creature’ (referenced in Pyysalo, p. 79-80) resulting in the identification of /H3/ with labiovelar.

Pyysalo (pp. 91-92) uses another well-known IE cognate set to demonstrate the lack of connection between vowel quantity and vowel coloring in PIE. Hitt mehhur/n ‘time, noon’ corresponds to Skrt m?ti, mím?te ‘to measure’, m?tr? ‘measure’, Lat m?tior ‘I measure’, m?ni ‘in the morning’,Goth mitan ‘to measure’, m?la ‘measure of grain’, OHG mezzan ‘measure’, Lith metai ‘year’, Alb mat, mas ‘I measure’, mot ‘year, weather’, etc. Again, this set finds its counterpart in PIE *meg’h– usually glossed as ‘great’ (Hitt megi– ‘big’, Gk megas ‘great, large, mighty’, Lat magnus, Gothic mikils, Old High German mihhil, Skrt mahan, etc. Pyysalo treats this set separately from the meH-set on p. 408. But both sets neatly derive from PIE *meg’h– or *meH– ‘to measure’; some forms chose the semantic path of TIME’, others the semantic path of SIZE and STRENGTH. The Gothic forms show a perfect morphological match between the me-l and the miki-l extensions reinforcing the phonetic similarity. Whether we postulate a law of velar weakening in some to-be-determined phonetic environments (potentially in non-accented syllables), a combination of a plain velar and a laryngeal (voiced in Pyysalo’s interpretation or voiceless in Szemerenyi’s interpretation) yielding the Indo-European palatovelars and labiovelars, or an entirely different ancient phoneme, Hittite (mehhur/n) doesn’t appear any more archaic in its phonetic development from the PIE root than Sanskrit (mahan) or Old High German (mihhil). Old Anatolian uniqueness is reduced to providing supporting evidence that, the TIME subset, too, contains the trace of an ancient consonant.

From a formal perspective one can conclude that the Indo-European material shows alternation between –ek’-/-eg’-/-ek– and –a-/-o– in correspondence sets that include Old Anatolian /H/. The “loss of the laryngeal” is only one path taken by non-Anatolian languages, the other one being a velar reflex. Both Puhvel and Kurylowicz are correct.

Importantly, one doesn’t need Anatolian evidence to establish a connection between IE *pa– ‘protect’ and IE *pek’– ‘livestock’. Overall, PIE *meg’h– or *meH– ‘to measure’ supports Pyysalo’s contention that Old Anatolian /H/ stood for both voiceless and voiced variants, but it undermines his belief that the advancement of the Neogrammarian agenda of the scientific study of Indo-European historical phonology can be achieved by tackling Old Anatolian data. Contra Pyysalo, the key to progress in Indo-European linguistics lies not in the preferred analysis of Anatolian data (obviously, it needs to be further studied but it shouldn’t command any priority over any other IE branches) but in the improvement of comparativist methodology itself.

Pyysalo comes face-to-face with the situation of a dual reflex of PIE “laryngeals” in languages outside of Old Anatolian when he correctly observes that Lycian sometimes shows /0/ where Old Anatolian has /H/ and sometimes a velar. Examples without the loss of a laryngeal include

Pyysalo-LycianThey are contrasted with the examples where Old Anatolian /H/ corresponds to Lyc /x/ (phonetically a voiceless velar, not a fricative [Melchert, Anatolian Historical Phonology, 40]) such as Lyc xuga next to Hitt huhhas ‘grandfather’ and Latin avus and Lyc xawa next to CLuw haui– ‘sheep’.

Pyysalo’s solution is to ignore the unmistakable similarity in form and meaning between the Old Anatolian and Lycian words and to make Lyc xuga and xawa unrelated to any other standard IE terms for ‘grandfather’ and ‘sheep’ but instead connected to obscure Hesychian forms. In the case of xuga, he relates it to

Pyysalo-Xugaand, in the case of xawa, he clusters it with

Pyysalo-xawa This solution is hardly convincing and contradicts the opinion of the majority of Indo-Europeanists. On the other hand, Lyc /x/ as a parallel to Anatolian /H/ remains tenuous unless we can show that the correspondence of the laryngeal to a velar is systematic in these cognate sets. And we indeed find that IE cognate sets GRANDFATHER and SHEEP are likely incomplete. PIE *Howi– (Skrt a?vi-‘sheep’, Arm hoviw (*ou?i-p?-) ‘shepherd’, Gk ???, ???, Lat ovis, OEng ewi, OHG ouwi, Lith avi?s, a?vinas, Slav *ovica ‘sheep’, etc.) can be compared with PIE *k’uwon-/*k’un– ‘dog’ (Skrt s?v??, s?(u)v??, Arm šun, Gk ????, ?????, Goth hunds, OEng hund, Lith šuo? (Gen. šun?s), Toch A ku, Toch B kunder. The morphological derivation of the DOG forms from the SHEEP forms by means of an n-extension is straightforward, while the semantic link is natural considering that proto-Indo-European pastoralists, with all certainty, used their dogs as shepherds and modern Europeans continue to call several dog breeds ‘shepherds’.

The situation with Lyc xuga ‘grandfather’ is more complicated. While the cognation of Hitt huhhas with Lat avus, Goth awo ‘grandfather’, OIrish (h)aue ‘grandchild’ is clear, a number of other forms, including Lyc xuga (< *xuxa, with voicing developing in an intervocalic position [Kloekhorst A. “Studies in Lycian and Carian Phonology and Morphology,” Kadmos 47 (2008), 125]), are similar enough to warrant serious consideration as belonging to the same group. Skrt susa ‘grandmother, grandfather, progenitor’ (from PIE *suHsiya), IE *suH– ‘son’ and Alb gjysh (from PIE *suHsos) ‘grandfather’ provide a case in point.

Importantly, Hitt huhhas is a reduplicative (comp. Neapolit vava ‘grandmother’ [Zimmermann 1922, 150] or USorb wowa ‘grandmother’ [Schuster-Šewc 1961; 1984, 796-797; 1987, 1682-1683; Machek 1968, 40] and hence the original form does not have to be *HewHo-. It could be *weHo-/*uHo- or even *CweHo-/*CuHo– assimilated into *HweHo-/*HuHo-. Skrt susa and Alb gjysh may, therefore, have preserved the original shape of the PIE GRANDFATHER root, while Hitt huhhas (and its direct cognates in other IE branches), so iconic in the context of the laryngeal theory, stems from earlier *suhhas. On the strength of OIr (h)aue ‘grandchild’ (Gen. aui ‘grandson’ [*awios < *awyios, with –i– developing into –e– under the influence of the following o [Maille 1910, 49-50]), úe ‘granddaughter’ (Meyer 1912, 183) and, now, IE *suH– ‘son’ (Toch we can confidently postulate a self-reciprocal meaning for PIE *suHo-/*sweHo– ‘grandfather; grandchild’. In many IE languages the GRANDFATHER-GRANDCHILD root spawned a host of derivatives including Lat avunculus ‘mother’s brother’, OIr amnair ‘uncle’ (through the assimilatory nasalization from *abn-air < *awn-air [McCone 1996, 49, 86] < PCelt *awen-tro), Welsh ewythr and OCorn eviter, Lith avýnas ‘mother’s brother,” OLith avà ‘mother’s sister, uncle’s wife’, OPrus awis (< *awio-s) “mother’s brother”, Slav *uj? ‘mother’s brother’, Alb *vella ‘brother’ (< *awnlada-< *awentlo– [Huld 1984, 128-129]). Morphologically, forms such as OIr (h)aue, OPrus awis and Slav *uj? have the same extensions as Toch A se, B soy, Gk huiús ‘(< *huios) son’ supporting the overall connection between the forms with s– and the forms with Old Anatolian h– and its accepted Narrow-IE counterparts.

The above interpretation of the PIE form for GRANDFATHER-GRANDCHILD yields to the clarification of the etymology of a widely-spread IE affinal term *swék’uro- ‘husband’s father’ (Skrt švašura ‘father-in-law’, Avest xvasura ‘father-in-law’, Prasun ?üj? [< *?u? < *?va?r < *sva?r-], Kati ??styü? [Buddruss 1976, 29-31], Gk ‘?????? ‘husband’s father’, Lat socer ‘father-in-law’, Goth swaihra ‘father-in-law’, OHG swehur ‘father-in-law’, OEng sw?or ‘father-in-law’, ONorse sv?r, OLith šešuras ‘husband’s father’, Slav *svek’?r? ‘husband’s father’, Alb vjerr, vjehër ‘father-in-law’, Arm. skesrayr ‘husband’s father’, Welsh chwegrwn, OCorn hwigeren ‘father-in-law’) and *swekr?s ‘husband’s mother’ (Skrt švašr?– ‘mother-in-law’, Avest xusr?, Waigali ??tr [dissimilated from *???r] [Buddruss 1976, 29-30], Gk ‘????? ‘husband’s mother’, Lat socrus, Goth swaihr?, OHG swigar, OEng sweger, ONorse sv?ra, Welsh chwegr, OCorn hweger  ‘mother-in-law’, Slav *svekry ‘husband’s mother’, Alb vjehërrë ‘mother-in-law’, Arm skesur ‘husband’s mother’.

The morphology of Lat avunculus (< *avonculus), PIrish *awentro, Welsh ewythr and OCorn eviter, PAlb *awentlo– parallels that of *swék’uro-. The forms with –t– may in fact be analogical with the respective terms for FATHER (PIE *pHter) and MOTHER (PIE *meHter, Alb motre ‘sister’), while –c– in Latin avunculus is identical to –c– in socrus (PIE *k’). The –n– in also likely intrusive and represents a secondary “nasal infix.” We therefore can align PIE *swék’uro-  ‘husband’s father’ with the Italo-Celtic-Albanian isogloss *awekro ‘mother’s brother’. Outside the Latin, Celtic and Albanian worlds, there are a few other indications that the IE terms for GRANDFATHER>MOTHER’S BROTHER and HUSBAND’S FATHER are cognates. Skrt švašura and OLith šešuras ‘husband’s father’ are assimilated from, respectively, *svašura and *sešuras just like Hitt huhhas and its counterparts in Narrow-IE languages assimilated from *suhhas (< *suHo-). Arm skesur ‘husband’s mother’ (judging by other IE forms, originally ‘husband’s father’) can be compared with k’eri ‘mother’s brother’. Until now, k’eri (< *sweriyos) was an isolate among IE terms for MOTHER’S BROTHER but in the light of the proposed reconstruction it’s naturally connected to skesur. Skesur shows the same assimilation of the earlier *k’esur into skesur as Skrt švašura and OLith šešuras.

Now that the various IE forms for GRANDFATHER, MOTHER’S BROTHER and HUSBAND’S FATHER have been shown to be aligned morphologically and subject to the same assimilatory processes, it becomes clear that PIE *k’ in *swék’uro corresponds to hh in PAnatolian *suhho-. The difficult Albanian –h– in vjehër ‘father-in-law’ (sometimes interpreted as a reflex of *H4 by multilaryngealists) matches the Old Anatolian hh nicely. The present analysis furnishes a better interpretation to another puzzling form. OHG ?heim, OEng eám ‘mother’s brother’ (< *eaham, with a syncope [Schhoof 1900, 232]) (Grimm & Grimm 1889, VII, 1198; De Vreese & Boekenoogen 1910, XI, 16; Lendinara 1990, 298) undoubtably contain the formant *au(n)– ‘grandfather’, but the final segment –heim causes problems. Gothic does not show this form but, according to the Grimms, it should be *áuhaims. The latest Etymologisches Wörterbuch des Deutschen (1993, II, 830, 915, 945) seems to take the Old English form as the basis and to explain the OHG voiceless fricative as phonologically conditioned in the intervocalic position. The majority of scholars, however, tend to consider ?heim an original form and to treat it as a compound going back to *au(n)-haimaz. The latter, however, is an ad hoc form without any parallels in other IE branches. The present analysis suggests that –h– in ?heim is part of the root and corresponds to –hh- in Hitt huhhas, -g- (< *-x-) in Lyc xuga, -h- in Alb vjeher, c in Lat avunculus and the reflexes of IE *k’ in the various HUSBAND’S FATHER terms in the satem and centum languages. Finally, an aberrant isogloss with Arm ustr ‘son’ corresponding to OEng suhterga ‘brother’s son’ becomes more understandable: semantically it belongs with  Toch A se, B soy, Gk huiús ‘(< *huios) ‘son’ but morphologically with the IE terms for HUSBAND’S FATHER (*swek’ro-/*suk’ro-).

To conclude, Pyysalo is correct in stressing the obstruent nature of the PIE “laryngeal,” the existence of voiced and voiceless laryngeal variants in Hittite and the polyphonemic nature of IE labiovelars and palatovelars. He is also justified in postulating diphonemic sets *Ha and *aH to explain the wide-spread correlation between Old Anatolian h and Narrow-Indo-European a. The evidence presented above, however, refutes a) Pyysalo’s (and others) conviction that “laryngeals” were lost outside of (Old) Anatolian and b) monolaryngealists’ reduction of all “laryngeals” to just one phoneme. In fact, we can provisionally postulate the following diphonemic sets for PIE:

1. *H2a/*aH2~ *k’e/*ek’                   *H2a/*aH2 ~ *g’e/*eg’

2. *H3o/*oH3 ~ *kwe/*ekw              *H3o/*oH3 ~ *gwe/*egw

3. *H1e/*eH1 ~ *ke/*ek                     *H1e/*eH1 ~ *ge/*eg

While the latter set is given here only for completeness (I can’t adduce any etymological material to support it), the former two are intimated in the correspondences laid out above. What they seem to be telling us is that PIE palatovelars and labiovelars were identical to, respectively, PIE H2 and H3. Both pairs were obstruents, both sets covary with their respective vowels (a and o) and both are “extinct” phonemes.

An example from IE numerals can illustrate the alternation between palatovelars and labiovelars in one root form:

PIE *ok’to– /*H3ek’to ‘eight’ and PIE *kwetwor-/ *kwetur– ‘four’ can be related to each other via PIE *H3eH2t-. The otherwise obscure medial a-vocalism of Lat quattuor ‘four’ fits well with –k’- in *H3ek’to, while the initial o-vocalism of *ok’to– corresponds to the labiovelar in *kwetwor-. The known Hittite corpus does not have reflexes of either roots. A connection between the numerals ‘2’ and ‘4’ in IE was noticed before (see Kassian A. “Anatolian *meyu– ‘4, four’ and Its Cognates,” Journal of Language Relationship 2 (2009): 65-78). Kassian argued that PIE *ok’to– /*H3ek’to- used to mean ‘four’ (comp. Iran *ašti– ‘(breadth of) four fingers’ (measure of length)) and he called the *kwetwor-/ *kwetur– root “enigmatic.” The enigma disappears if we accept the formal evidence presented above that IE palatovelars and labiovelars were initially found in complementary distribution with IE vowels a/aa and o/oo. Intriguingly, PIE *H3eH2t- ‘four; eight’ shows a good formal and semantic fit with Uralic *kekta ‘two’. Whether the connection between the two represents common descent between IE and Uralic or a borrowing from Uralic to IE or from IE to Uralic, it documents the progressive complexification of a numeral system with numeral ‘2’ serving as a generator of high-order numbers ‘4’ and ‘8’.

Indo-European and North Caucasian: Linguistic Typology, Kinship Terms and Autosomal Genetics

Saturday, July 7th, 2012

Ranko Matasovic presented rich typological evidence (consonant-to-vowel ratio, tonal accent, number suppletion in personal pronouns, the presence of gender and the morphological optative and, possibly, the presence of glottalized consonants and ergativity) in favor of areal contacts between Proto-Indo-European (PIE) and North Caucasian (NC) in the eastern part of the Pontic steppe. He writes (pp. 306-307):

“The adduced typological parallels between PIE and Caucasian languages make it likely that PIE was, indeed, in contact with languages of the northern Caucasus. However, these contacts could also have been of indirect nature, since there are no demonstrable loanwords from North Caucasian languages in PIE, or vice versa. If such loanwords exist, their number is certainly not high. If direct contacts did exist, we cannot determine their nature: both long-term bilingualism due to exogamy and trade networks, as well as rather rapid language shift appear equally possible.”

Matasovic (p. 288-289) mentions one of such loanwords, namely the PIE kinship term *snuso– (Gr. nuos, Arm. nu, OCS snuxa, etc. ‘daughter-in-law’), apparently borrowed into many North Caucasian languages (Chechen nus, Avar nus, Akhvakh nusa, Kabardian nesa) as well as into Megrelian (nosa). But he fails to mention the cases of 1) PIE *swesor ‘sister’ and Chechen sesag ‘wife’, Lak sus, Lezgin swas, Ubykh sasa ‘bride’ and 2) PIE *geme– ‘son-in-law’  (Skrt jamatar, Avest zamatar, Gk gambros, etc.) and Chechen zam-o, Ingush sam-e ‘best man’, Lezgin c:am, Agul zam (Nikolayev S. L., and S. A. Starostin. ? North Caucasian Etymological Dictionary. Moscow, 1994). The combination of the these three apparent loanwords, all referring to marriage and affinity, favor the long-term exogamy hypothesis between PIE speakers and the natives of the Caucasus.

Naturally, if long-term marriage exchange was indeed practiced between PIE speakers and the natives of the Caucasus, we may expect to find its genetic traces in mitochondrial, Y-DNA and/or autosomal DNA. The amateur genome blogger Dienekes Pontikos recently reported the elevated frequencies of a “Caucasus” autosomal component among modern Indo-Europeans and its absence among Basques and low frequencies among Finns. While he labeled it “West Asian” and misconstrued it as suggesting a Near Eastern origin for Indo-Europeans (in contrast to the mainstream Kurgan hypothesis), it seems to be the best available evidence for genetic exchange between Indo-European speakers and the speakers of North Caucasian languages in the Pontic steppe and North Caucasus areas. The Chechens whose language contains all three kinterm loans from PIE, show this component at 54.6%. Matasovic focuses on the Maykop culture (3700-2500 BC) as providing the best archaeological correlate to the contact zone between PIE and North Caucasian languages. The Maykop culture is centered in Adygeia, which shows 52.5% of that “Caucasus” autosomal component. We have, therefore, a very strong fit between linguistics, kinship studies, population genetics and archaeology in this case.

It remains to be seen if this fit is real or spurious. It’s noteworthy that all NC societies are strictly patrilocal and patrilineal. PIE society is also reconstructed as patrilineal and patrilocal (see The Encyclopedia of Indo-European Cultures, with the original proposal in Friedrich, Paul, “Proto-Indo-European Kinship,” Ethnology 5 (1966), pp. 1-36). This commonality would facilitate marriage exchange between the two populations in both directions, as PIE women marrying NC men would go live with their husbands without violating the rules of either society. The same works for NC women. At the same time, PIE men and NC men would rarely end up as son-in-laws in foreign households, unless they had been first taken as prisoners. We don’t know yet if PIE speakers contributed any genes to the NC speech community. Neither do we know if genetic admixture between PIE and NC speakers manifested in mtDNA and Y-DNA. If it did, then we may be able to infer from the pattern of this admixture whether this gene exchange was sex-biased and whether it took place under peaceful or violent circumstances.

Burushaski and Indo-European Kinship Terms: Burushaski suffix -taro and IE suffix *-ter.

Sunday, June 24th, 2012

Ilija Cašule, of Macquarie University in Australia, is attracting more and more attention with his Indo-European-Burushaski hypothesis. As part of his thoroughgoing comparison between Indo-European and Burushaski, Cašule has looked at kinship terms (see p. 12 in the attached). Although, at first glance (see here), Burushaski kin terms support the Dene-Caucasian placement of Burushaski (if one is willing to entertain any long-range proposals at all), Cašule brings up the Burushaski plural suffix –taro used almost exclusively with kinship terms, which does strike one as similar to the ubiquitous Indo-European kin term ending –ter. As I argued at length elsewhere (here and here), IE kinship terms are puzzling in a lot of respects. One of them is the fact that many key kinship terms are already complex morphological structures at the Proto-Indo-European level, and the daughter languages don’t shed any light on the function of those widely-spread IE suffixes. Other language families typically maintain a clear trace of the semantic and pragmatic functions of kinterm-specific grammatical forms, but in IE languages such grammatical forms are obscure. Hittite data has so far contributed nothing to our understanding of the meaning of PIE –ter (or *Hter in some laryngeal reconstructions). Nostratic kinterm reconstructions are of very poor quality and they contain nothing approximating IE *-ter (see Dziebel, G. V. “Reconstructing ‘our’ kinship terminology: Comments on the Indo-European material in A. V. Dybo’s and S. V. Kullanda’s The Nostratic terminology of kinship and affinity.” Kinship Algebra, No. 10 (2006)). This suggests that the most divergent IE language hasn’t been found yet. And, who knows, maybe Cašule found it and it’s Burushaski.

What’s in a Kin Term? Observations on Errol Morris’s What’s in a Name?

Monday, May 7th, 2012

Errol Morris has a three-part opinion page in The New York Times on the nature of personal names and photographs. Errol anchors his exploration in the semantics of naming in the recent scandal around a Rockefeller impostor, the former German student in the U.S., Christian Karl Gerhartsreiter. Gerhartsreiter, now universally remembered as Clark Rockefeller, adopted the Rockefeller name in the early 1990s to oil his way into the cream-of-the-crop social circles of Boston and New York. Gerhartsreiter’s self-inflicted identity fraud had remained unnoticed, even by Gerhartsreiter’s super-intelligent wife and the sole bread-winner in their upscale Boston household, Sandra L. Boss (a Harvard Business School Graduate and a senior McKinsey executive who happened to be a one-time boss of my former boss at Arnold Worldwide), until after he divorced her, kidnapped their daughter and was hunted down by the police.

Any story of personal names is a detective story. Morris’s narrative manifests the thrill that he experienced while researching and pondering the nature of personal names, human identity and photographic imagery. Borderline mystical phenomena such as coincidences, ironies, paradoxical role reversals and shocking revelations seem to be part and parcel of any empirical study of naming. The above paragraph already contains a string of such coincidences and ironies: Gerhartsreiter’s wife’s family name is Boss and she lived up to the literal meaning of her name by being a) a senior executive at an iconic company that consults major businesses in the U.S. and worldwide, and b) the sole bread-winner in her marriage with Clark Rockefeller. She was also an ex-boss of my own ex-boss. All of this took place in Boston. One of the role reversals of the story is that, by being a stay-home dad and teaching their daughter how to write by the age of 2, Clark Rockefeller acted like a nurturing mother, while Sandra Boss, by being the sole bread-winner, acted like a father. One of the ironies of the story is that Sandra Boss’s self-effacing and street-smart ex-husband exposed what may very well be the Achilles heel of armchair, academic intelligence: its inability to call real things with their proper names. A brilliant-at-work Sandra turned out to be woefully naive in real life.

Or compare the a real modern descendant of the legendary American industrialist, David Rockefeller, namely his grandson Clayton Rockefeller, with the impostor Clark Rockefeller. They share the first three letters of their first names (Clark may be a portmanteau of Christian and Karl, Gerhartsreiter’s real first and middle names), obviously their last name is the same but both of them also love to wear horn-rimmed eyeglasses. The last fact makes the two look uncannily alike. The irony, as Morris notes, comes from the fact that the real Clayton Rockefeller never made it past the page 15 of the Weddings/Celebrations section of Sunday Styles, while the fake Clark Rockefeller commanded the front page of some major East Coast venues such as the Boston Globe. This is precisely where Morris’s fascination with naming comes from: a name is not simply a baggage tag on a person, it carries powerful associations that, under certain circumstances, can propel a completely random person into a national spotlight. But the irony is again right around the corner: Morris praises and adopts Saul Kripke’s model of personal names as “rigid designators,” but his choice of an opening case-study seems to furnish a paradoxical refutation of Kripke’s model: a name is inalienably attached to the thing in all possible counterfactual worlds (i.e. worlds in which the thing is stripped of all its associations) with the exception of those worlds in which the name is assumed to be a rigid designator of the thing. Gerhartsreiter’s fraud became possible precisely because people around him believed that the name Rockefeller is always the name of a real Rockefeller. Within the social circle of his friends and his wife’s family and friends he became a celebrity by claiming to be a Rockefeller. He bolstered his personal myth by publicly ordering Oysters Rockefeller at restaurants and buying properties exclusively from Cushman & Wakefield — the managers of Rockefeller Center. Outside of the primary circle in which Gerhartsreiter’s Rockefeller myth lived he was not known at all – neither as a Gerhartsreiter, nor as a Rockefeller. His natal family in Germany had lost track of him, while those who knew the real Rockefellers did not know Clark. When Gerhartsreiter entered the national spotlight, he became famous not because he was Gerhartsreiter or Rockefeller but because he was a fake Rockefeller or a non-Rockefeller. But in any case it was the name Rockefeller that continued to define Gerhartsreiter in the public eye, as if he was re-baptized as Nonrockefeller by virtue of the fact that the real Rockefellers confirmed that Gerhartsreiter was not related to them. Kripke erred in his belief that names are rigid designators of things – the enigma surrounding names stems from the fact that they are things. Note that names are rarely translated from one language to another – they are usually reproduced in another language as faithfully as phonetics allows. And as sound things they can be manipulated and appropriated. Gerhartsreiter literally stole Rockefeller from the Rockefellers, as if it was a family relic, and proudly wore it around as if it was an expensive piece of clothing. But they are not simply things – they are things conditionally related to essences. And every culture has its own native theory of those essences.

Morris chose the example of Clark Rockefeller without fully exploring the fact that Rockefeller is not just a name but a family name. For his purposes, a family name is just another proper name to be looked at through a Kripkean prism. He noted, in passing, the connection between the Rockefeller genealogy and Gerhartsreiter’s deception:

“When we use the family name “Rockefeller” or the proper name “Clark Rockefeller,” what are we doing? We are calling up a rich set of associations. And we are also claiming provenance: a matriarchal or patriarchal link from the original Rockefeller, John D. Rockefeller, to his progeny — his son, John D. Rockefeller, Jr.; his grandsons, Laurance, Winthrop, Nelson, David, etc. All these Rockefellers are genetically related to the original Rockefeller and presumably heirs to part of his vast fortune. And then the fourth-, fifth- and sixth-generation Rockefellers. They are all part of the Rockefeller family tree with John D. Rockefeller as the trunk. When we use the name “Rockefeller” we imagine we are picking out some individual related to the 19th-century patriarch…Isn’t there a remote possibility that Gerhartsreiter really is a member of the Rockefeller family? Or that they had common ancestors? By this account, John D. Rockefeller’s roots can be traced back to Germany in the 1600s. Perhaps they had common ancestors.” 

And in the accompanying note 14:

“This kind of claim is complicated by uncertainties about John D. Rockefeller’s two half-sisters, Clorinda and Cornelia. The Rockefeller family line veers surprisingly from the straight and narrow during the life of John D. Rockefeller’s father, William Avery Rockefeller. William, or Big Bill as he became known, was from a young age a drifter, schemer and occasional con man — perhaps a spiritual if not actual progenitor of Clark Rockefeller. He once rode into a small town in upstate New York posing as a deaf-mute selling novelties, hoping to gain sales through sympathy. He also discovered that with this “deaf and dumb” identity he could overhear or extract town gossip he could use to his advantage. Months later, when a woman who’d befriended the supposed mute ran into him chatting at a social gathering, she was so impressed by the man’s “progress” that she said to him, “I see that you can talk better than when I saw you last.” Without missing a beat, Big Bill replied, “Yes, I’m somewhat improved.” Bill was equally unorthodox in his home life. During his travels, he met and married Eliza Davison. Some suggested it was her father’s farm and modest wealth that Bill found most attractive, considering he already had a mistress, Nancy Brown. And not long after bringing Eliza to the tiny farmhouse he’d constructed, he also brought Nancy to live there, in what essentially was a ménage à trois. In two years, in the same house, Bill had four children: Eliza had Lucy, and a few months later Nancy had Clorinda. The next year, Eliza gave birth to John Davison Rockefeller, and in that same year Nancy had another daughter, Cornelia. Eventually, Nancy and her two daughters went to live with her parents one town over. Clorinda died as a child, but Cornelia grew up and married, becoming Cornelia Saxton, and most of her neighbors never knew she was the half-sister of the richest man in America. The Saxton line, which without the name still bore the traces of Rockefeller lineage, continues off in another direction, save for a few exceptions in John D. Rockefeller’s records when they collide again to ask for money; these requests were turned down by John D.’s secretaries, and it remains unclear if he knew he had two illegitimate half-sisters.”

Christian Gerhartsreiter claimed that the name Rockefeller had been given to him by a man named “Harry Copeland,” his godfather from New York.

“My godfather gave it to me. He insisted that is what my name is.”

Christian Gerhartsreiter claimed descent from a member of the Rockefeller clan, George Percy Rockefeller, and his wife Mary Roberts who came from the privileged upper neck of Virginia. Gerhartsreiter’s imaginary parents died in a car accident when he was 18. Harry Copeland, his presumed godfather, died in the 1990s. By the time Gerhartsreiter was put to trial for kidnapping his daughter, his real father Simon Gerhartsreiter, a painter, was dead, too. His mother Irmgard, a homemaker, and brother Alexander recognized him. Gerhartsreiter’s pattern of deception matches the neurotic fantasy of noble origin described by Freud in his rarely quoted work “Der Familienroman der Neurotiker.”

“There are only too many occasions on which a child is slighted, or at least feels he has been slighted, on which he feels he is not receiving the whole of his parents’ love, and most of all, on which he feels regrets at having to share it with brothers and sisters. His sense that his own affection is not being fully reciprocated then finds a vent in the idea…of being a step-child or an adopted child…The later stage in the development of the neurotic’s estrangement from his parents, begun in this manner, might be described as ‘the neurotic’s family romance’… [T]he child’s imagination becomes engaged in the task of getting free from the parents of whom he now has a low opinion and of replacing them by others, who, as a rule, are of higher social standing” (Freud, Sigmund. “Family Romances,” in The Standard Edition of the Complete Psychological Works of Sigmund Freud. Vol. 9. Pp. 236-241. London, 1959, 237-239).

If Christian Gerhartsreiter chose to become Clark Gerhartsreiter, it wouldn’t make much of a difference. He was clearly going after the essence of being a Rockefeller, the richest man in America. People who are passionate about genealogical research are searching for the timeless essence of their real family name(s). Christian Gerhartsreiter abandoned his real family name, just as he left his native Germany, in favor of an American one, with possible deeper German roots, that already had a noble, albeit shallow, genealogy associated with it.

In “The Genius of Kinship” (2007) I touched upon the relationship between names and kin terms. Since then, I’ve been thoroughly fascinated with the opportunities contained in the cross-disciplinary analysis of names and kinship terms (and pronouns, for this matter). The boundary between kin terms, pronouns, names and natural kind terms can easily blur. Consider a few examples:

1. My dad is coming.

2. Yesterday I became a dad.

3. Dad, when are you coming?

4. (Mother to son) Go, talk to dad about it.

The same form dad assumes four different grammatical functions: in [1] dad is a kin term, in [2] dad is a natural kind term, in [3] dad is a vocative form of dad in [1] and as such it is like a pronoun, and in [4] dad is a personal name or a private, family-only nickname.

When Errol Morris tries to define what is special about names he writes,

“It is interesting that the word “of” is used in portrait photography and proper names. We speak of a photographic portrait and a proper name being of someone.”

This of-ness of names is, for him, the indicator of rigid designation. But, as any specialist in kinship studies could tell Morris, kin terms are rarely, if ever used in abstract, so that a father is always the father of someone. (Just like boss is always someone’s boss.) Just as names, according to Kripke, originate in the act of baptism, kin terms originate in the event of birth (or re-birth, as some cultures holding reincarnation beliefs would have it), marriage, death or adoption. A refrigerator magnet in my mother-in-law’s house reads “When a child is born, so is the grandmother.” Of course, there’s a pool of kin terms from which individual kin terms are drawn, just like there’s a pool of personal names from which parents select a name for their child, but they are activated as a result of a life-cycle event.

By virtue of their relational property, kin terms are called “relational nouns.” If they are used in abstract, they switch their grammatical function to become a natural kind term, as in example [2] above. Around the time when the great British philosopher John Stuart Mill quoted by Morris discussed names in “A System of Logic: Ratiocinative and Inductive” (1843), the British logician Augustus De Morgan came to realize that kin terms are relational nouns (see “The Genius of Kinship,” p. 23ff). One of the founders of semiotics, the American philosopher Charles Peirce, developed De Morgan’s insight into the well-known classification of signs into icons, indexes and symbols. About a kin term he writes (Peirce, Charles S. “The logic of relatives,” in Collected Papers of Charles Sanders Peirce, 1933, Vol. 3, 459),

“A relative is just that, an icon, or image, without attachments to experience, without ‘a local habitation and a name,’ but with indication of the need of such attachments.”

It’s important to note that Peirce likened a kin term to an icon in the same way as Errol Morris likened a name to a photograph. The difference between a photograph and an icon (in the original sense of the icon) is that a photograph captures the appearance, while a (religious) icon captures the essence. (This may not be a true difference but rather an artifact of the specific term usage by Peirce, though.) Kin terms define one person by referring to another person: X is the father of Y. One person is defined in multiple ways depending on his relationship to other people: X is the son of Y, the father of Z, the uncle of A, etc. Every kin term has at least one other kin term as its direct reciprocal: X is the son of Y, Y is the father of Z. Kin terms are used by everyone in society but it’s only a small group of people that know which specific individuals are addressed by those terms by other individuals. Personal names are usually bestowed on a person by his close relatives (including adoptive parents), often they are recycled names of more distant/deceased relatives, but they can become very widely known outside their bearers’ immediate circle. People may be known all over the world by their names, but the people whom they call “father” and “mother” most of the time remain obscure.

It’s likely that kin terms and names are so closely related semiotically that one cannot tease them apart by invoking rigid designation, similarity, photographic imagery or the act of baptism. They break the terminological conventions and disciplinary silos established in analytical philosophy, logic and linguistics and require a conceptual apparatus of their own.

Gender Neutralization in Kinship Terms: Putting the Swedish Experiment in Perspective

Friday, April 13th, 2012

Sweden is going through another “sexual revolution.” Traditional gender roles and stereotypes have been so thoroughly challenged in all spheres of life (politics, labor laws, advertising, clothing, sports – you name it) that time has come for change the lexicon and the grammar of the Swedish language.








The linguistic devices that are affected by a societal shift to gender-neutrality are personal names, pronouns and, potentially, kinship terms. There are currently 170 legally recognized unisex names in Sweden. Swedish parents are increasingly likely to give a girl a traditional boy’s name and a boy a traditional girl’s name. The online version of the country’s National Encyclopedia now contains a new pronoun hen defined as a gender-neutral personal pronoun instead of han ‘he’ and hon ‘she’. It’s kinship terms’ turn now and a male writer, Jasper Lundqvist, is publishing a gender neutral book for children (see above) called “Kivi och monsterhund” (Kivi and monster dog). This book is full of neologisms such as mappor, pammor (instead of gender-specific mammor ‘mothers’ and pappor ‘fathers’), morbroster (from morbror ‘uncle, mother’s brother’ and moster ‘aunt, mother’s sister’) and bröstrar (from bröder ‘brothers’, and systrar ‘sisters’).

Time will show how sticky these new kinship terms, personals names and the pronoun hen are, but, in at least two ways, these changes are systematic.

First, there’s a tendency for kinship terms and personal names to develop similar gender marking. In Spanish, gender-specific affixes affect both kinship terms and personal names in a patterned fashion (comp. Mario/Maria and hermano ‘brother’/hermana ‘sister’).

Second, many Swedish kinship terms are already highly compounded, descriptive and artificial, being juxtapositions of simple kinship terms (e.g., farbror ‘father’s brother’, morbror ‘mother’s brother’). The portmanteau formations such as mappor don’t seem to come out of the blue in a pure response to sociopolitical forces related to gender; rather they could qualify as changes originating in the internal structure of the kinship terminology or at least as changes heavily constrained by it.

Third, kinship terminologies show a global trend toward progressive expulsion of Ego Gender and Speaker Gender (see The Genius of Kinship). This can clearly be seen in sibling terminologies where the progressive removal of Ego Gender and Relative Age results in the dramatic simplification of sibling sets (Swedish systar and broder, just like their cognates in other Indo-European languages, are already heavily simplified sets). If we look at the opposite pole of terminological elaboration of sibling sets (and concomitantly at the geographically opposite side of the globe from Sweden), we find that societies differentiating their siblings in linguistically complex ways also exhibit a similar concern with gender segregation in siblings’ roles, outward appearance and behaviors. As reported by Garcilaso de la Vega (Royal Commentaries of the Incas. Austin and London, 1966, 4, 2, 211) about the colonial Incas, boys and girls were supposed to strictly adhere to the proper usage of kinship terms marked by Ego Gender, otherwise boys would become girls and girls would become boys.

While a Swedish toy catalog (see above, left) is experimenting with showing a boy pushing a pink stroller to appeal to gender-neutral parents, the Incas of colonial times were busy enforcing the opposite – the strict segregation of gender outfits and behaviors through language from an early age. The two behaviors are apparently associated or consistent with two radically different ways of manifesting gender in kinship terms because saying is being.

It’s possible that the reason why gender dynamics in society correlate so intricately with gender indexing in kinship terms, pronouns and personal names is because it’s precisely in these classes of language items that ontology impinges upon language.

Via Slate and Transparent Language.

Kinship Terms and Naming Taboos: Oroch

Friday, April 13th, 2012

Among the Altaic-speaking Oroch (Orochi) in eastern Siberia, personal names are subject to various pragmatic restrictions. Personal names can’t be duplicated or pronounced loudly, and they shouldn’t be used when cursing somebody or an evil spirit would harm the name-bearer. Kinship terms are not subject to any taboos and can be used freely. Parents can use personal names to address their children but even in this case there is a special set of birth-order terms to refer to children that’s used alongside names.

Source: Startsev A. F. “Nomenklatura rodstva,” in Istoriia i kultura orochei. St. Petersburg, 2001, p. 28 (in Russian).

It appears that cross-linguistically kinship terms, unlike personal names, are not subject to speech taboos. It’s also noteworthy that, by being taboo-free, kinship terms function as substitutes for personal names and, as such, approach the function of pronouns (Russian mestoimenie, lit. ‘name substitute’). The difference between kinship terms as name substitutes and pronouns is that kinship terms index deictically not the Speaker (vocative terms do), but the Name-Bearer.

Generational Reversal in Names and Kinship Terms

Tuesday, April 10th, 2012

Lucy Mair in “Marriage and Family in the Dedza District of Nyasaland,” Journal of the Royal Anthropological Institute 81 (1-2), 1951, (p. 103) reports on the pragmatic functioning of descent names among Bantu-speaking Ngoni and Cewa.

Not only are they used vocatively, as one would expect, in the respect register but also, conversely, in playful address to small children. Typically, cross-linguistically, older and younger relatives are treated differently: older relatives are addressed using respectful forms, while younger children are addressed using familiar forms. For instance, children can often be addressed by their personal names, while they in turn must use kinship terms when addressing adults. But there’s a known property to kinship terms whereby in some cultures (Arabic, Georgian, Hopi, see The Genius of Kinship, pp. 156-157, for a more complete roster) adults and children within the same family reverse their normative kinship terms in direct address to each other, so that a child becomes “father,” “mother,” “grandfather,” “grandmother,” “uncle,” or “aunt” to a corresponding adult, and the other way around. It seems that among the Ngoni the use of a descent name in playful interactions with children belongs to the same category of pragmatic phenomena. It may function as an educational device, so that children learn descent names from an early age, or as a restricted relaxation of verbal prescription and taboos in specialized playful contexts.

Kinship and Naming: A Note about Warao

Wednesday, April 4th, 2012

In his doctoral dissertation (The Warao: A Study in Microevolution.  University of California – Los Angeles, available online from HRAF), Mark Fleischman (pp. 65-66) reports on the centrality of kinship terms among the Warao:


“In small village groupings ego is unlikely to have much contact with very distant relatives. The social interdigitation of people living closely with one another would lead naturally to a greater complexity of terminology for people within the group. This specificity of kinship terminology within a small village group would negate any need for proper names for individuals, since, in most instances, the kinship term used by ego would show which individual he is addressing or talking about. Knowledge on the part of the others of all of ego’s relatives is necessary under these circumstances. Such information would be difficult to maintain in large living groups, or in groups where people are distantly, if at all, related. The Warao conform to the above expectations, and prior to missionary contact placed little importance on proper names.”


One of the key features of personal names is their referential precision. They are capable of referring to one individual and one individual only. The important point in the Warao quote above is that in small populations kinship terms possess greater specificity over personal names. It’s also noteworthy that Fleischman identifies the reason for why small populations often exhibit unusual complexity of their kinship categories and why complexity is not always of recent origin. This is the point I made in “The Genius of Kinship” when I postulated that Murdock’s “Complexly Differentiated Sibling Type” represents an evolutionary archaism and not a recent development. In a situation when other linguistic domains, such as personal names, numbers or pronouns, may be suppressed and underdifferentiated because of the lack of a strong functional need for these classifications, kinship terminological systems acquire a greater prominence as they take over some of the functions of those other linguistic domains.

Indo-European KInship Terms: A Discussion with P.A.Kerkhof

Friday, March 30th, 2012

More than a year ago I had a productive discussion regarding Indo-European kinship terms with P. A. Kerkhof (Katholieke Universiteit Leuven) at Wanana sculon Francon.