Indo-European phonology and morphology

A. Based on the example of simple compounds in American languages, Edward Sapir proposed that PIE *dak^ru-'tear' was really *wdr-H2ak^ru-'bitter water' or similar (published posthumously in 1939, see Manaster Ramer 2024). Though his *wd-> d-would not fit known sound changes (or *w-u > *0-u dsm.?), the semantics seem good. Other words for 'tear(s)' in East Asia are also clearly compounds for 'bitter water', 'salt water', 'eye water', etc. Kortlandt (1985) gave similar possibilities, including *drk^-H2ak^ru-, with *drk^-in G. drákos 'eye', H2ak^ru-'sharp / bitter / salty / saltwater', the u-stem equivalent of *drk^-H2ak^ri-> W. deigr; *H2ak^ri-> G. ákris 'peak', S. áśri-'edge/corner', *aH2k^ri-> L. ācri-'sharp' etc. (many IE nouns have cognates with i vs. u). I feel 'saltwater (from) the eye' makes sense in context, a better feel than his 'eye bitter'. With 2 r's and 2 k^'s, dissimilation could turn it to standard *dH2ak^ru-. However, many IE cognates show irregularities.

In this draft I adduce a number of examples showing forms that look like zero grades, but are rather full grades with vocalism changed by the interference of external factors: accent, sandhi, assimilation, dissimilation.
I wrote it after having read the paper of B. Vine: Myc. tu-wo-, Hom. θύος and the vocalism of s-stems in Proto-Indo-European.
https://www.academia.edu/84250689/

La protoforma céltica *bedo-'canal, mina', un sustantivo de resultado derivado de la raíz verbal PIE *bhedh(h2)-'(ex)cavar', gozó de relativa productividad en los dialectos célticos tanto insulares como continentales y en el terreno de los nombres comunes y propios por igual. En el presente artículo tenemos como objetivo, en primer lugar, confeccionar un panorama derivativo crítico de todas estas voces heredadas en céltico y, en segundo lugar, sumar a esta nutrida nómina otras formas que hasta la fecha no han sido analizadas como derivadas de *bedo-, ampliando de este modo nuestro conocimiento sobre las lenguas, la teonimia y la realidad material de los antiguos celtas.

Pisaniello & Giusfredi examined a large number of Anatolian glosses (mixed H. & Lw.) in medical context which seem to have been given by speakers of Anatolian languages, intended to help those less familiar with Ak. understand specialized usage or rare words. They say, "The fact that, as will be argued, a number of glosses were additions and commentaries rather than mere translations of the Akkadian original, furthermore, testifies to a local work of re-elaboration of the medical knowledge." However, The glosses all look like direct translations of the Ak. words to me, when discernible, with some of the original translated terms now lost. This is an idea they sometimes consider, but they say it is less likely due to the Anat. words not matching Ak. ones elsewhere. However, most of these cases also seem like direct glosses, unrecognized. Many of these have important implications for other IE. I will use their numbering for the convenience of those reading both at once. 3.1. Ak. i-te-eb : Lw. taršiyai 'vomits' (tarši(ya)-'vomit / belch') There might be a need to emend Ak. i-te-eb to Ak. i-te-eb<-bi> 'gets up' if it they are right. or maybe a causative. The meaning 'get up' or 'make get up / cause to rise up' is probably used in a technical sense for 'throw up', hence glossed by a term in order to let those less familiar with Ak. understand this usage. Based on *-eyo-> -ai-in *toubheyeti > Lw. dūbiti 3s, *toubheyonti > dūbainti 3p 'strike', this could be *trseyo(i) > taršiyai 3s. 'vomits', a mid. form. This is found in "His face keeps spinning tarpalli[…] of the night he is affected". Its insertion directly after 'spinning' shows that PIE *terp-'turn' (sometimes referring to weaving or plaiting), G. trépō 'turn to/around/back', is meant. If not, it would surely be quite a coincidence. The use of Lw. tarpalliš 'ritual substitute' is a technical sense in a H. lw., not its base meaning, clearly < 'turned around, the other/opposite' or 'in turn, in its place' or similar senses, return, (ex)change, etc. If it had this meaning in a medical text, what could it mean? Why insert it in the middle of an unrelated part? "His face keeps spinning" does not seem like the best translation. Indo-European derivatives of *weip-'turn / twist / bend / etc.' seem helpful. Since there are many Baltic verbs like Latvian vaîbît 'distort/adjust (one's face)', Lithuanian viepti 'make a face / gape' (also for 'rotate/twist / grimace', etc.), the simplest explanation is that 'twisted up / distorted (in pain/spasms)' is meant. Such a meaning might not be readily apparent to someone only moderately familiar with Ak. In fact, the same *-sr-> -zdr-(not normal -s(t)r-) also in *mHemsro-'flesh' > Sl. *memzdro-> OCS męzdrica 'membrane of egg', R. m'azdrá 'fleshy (inner) side of pelt', so it might have something to do with the presence of *H elsewhere (maybe H-r > H-R, a voiced uvular, and sR > zR), like *gWhaH2is-> Li. gaĩsas 'glow / gleam (of fire)', gaĩsras 'glow in the sky / (glow from a) fire / conflagration', gaĩzdras 'glow in the sky / glow from a fire'. The same in *g^(e)isro-'sand / gravel / pebble(s)' > Li. žie(g)zdrà 'gravel / grain', žìzdras 'gravel / rough sand', vs. *gis(ul)o-> OE cisil \ ceosel 'gravel / sand', MHG kis(el), NHG Kies 'gravel', Kiesel 'pebble', but it is not known if *H appeared (or is needed) here. If Kloekhorst is right about H. mah(u)rai-being an ai-stem, but still related to G. mērós, then maybe Lw. *mahhuršani-is from *mahhuršami-, with dsm. *m-m > m-n vs. *m-m > m-0 in H. With no other ex. of *-msr-in H., *-mfr-> *-mxWr-> is a reasonable outcome (for other *f / *xW in H., see Whalen 2025d), with dissimilation and metathesis explaining -a-u-vs. -u-a-, *mm/n/0, etc. With all this, maybe : *mHomsro-e vs. o like *mems-> Go. mimz 'meat', *momson-> mammō 'flesh' ? *mHomfro-*mHomxWro-*moHxWrom-*moHwrom-reg. or dsm. of H-H (if xW similar to H3) ? *moHurom-*moHuromi-few or no m-stems, shift > i-stem (as many C-> i-stems in other IE) or? (if mērós exact match, both < e-grade; unlikely, but maybe *-em-/ *-om-varied in some IE) *mHemsro-*mHemfro-*mHemxWro-*memxWro-reg. or H-H > 0-H (if xW similar to H3)? *momxWro-opt. rounding for PeP, like *penkWe > *kW-> Gl. pempe-, O. *pompe '5' *momxwro-*moxwrom-*moxurom-*moxuromi-few or no m-stems, shift > i-stem (as many C-> i-stems in other IE) Note A. *luH1-> *per+ > L. polluō 'soil, defile, pollute'; *mido-luH1o-> H. mitalha-š 'red mud' *luweH1-[or later affix?, ana. < v. in *-eH1 ?] > L. luēs f., luis g. 'plague' *leuH1tiHno-> Li. l(i)utýnas 'loam pit' [opt. H-H > 0-H ?] *leuH1no-> Li. liū́nas 'morass' [*eu > iu near *H?, irr. change said to be behind many BS *yu-]

The article offers a new solution to the unexpected occurrence of -θ-, rather than **-τ-, in οἶσθα 'you know' < *uoid-th2e. It is proposed that this form first became *woitsta, as expected, and then underwent the same development as seen in *pst > *phsth (> phth, cf. *seps-tos > ἑφθός 'boiled') and *kst > *khsth (> khth , cf. *h1ekś-tos > ἐχθός 'outside'), i. e. *tst > *thsth, before losing the first stop of the cluster. The isolation of the ending -θα meant that there was no source for restoration, unlike in all other cases of *tst (e. g. ἴσ-τε 'you (pl.) know' , ἄισ-τος 'unseen'). Incidentally, the proposed sound law by which all voiceless stops in TsT-clusters, including those which were later simplified to sT, were aspirated by *s also allows us to straightforwardly trace ἔσχατος 'uttermost' back to *h1ekś (*ἔξ-κατος > *ἔχσχατος > ἔσχατος), to consider the possibility that ἐσθίω 'to eat' was based on the 3sg.pres. *h1ed-ti 'eats' > *ἔσθι rather than on the imperative, and to connect the word group of αἰσχρός 'causing shame' < *h2eid-sk- with that of αἰδώς 'shame' < *h2eid- .

[Iranian] *namH 'to strike, beat' Oss. I. naemyn/nad, D. naemun/nad 'to hit, strike ', OKh. parnam-'to touch, feel', ? Sh. nimů, (Baj.) nimaw, Khf. nimaw, Rosh. nimōw, 'reproach, abuse; regret ?' The existence of an IIr. root *namH-'to strike, beat' was first postulated by Schmidt 1959: 113 ff., and accepted by Bielmeier 1979: 201; Abaev II: 169 f. The laryngeal presence for this root is most clearly indicated by the Ossetic past participle nad (< *nmHto-). The IE cognate forms that are quoted here, Gr. némesis, etc. can hardly contain the IE root *nem-'to take, assign, etc., as assumed by Pokorny (IEW: 763). IIr. *namH-would then derive from IE *nemH1-'to strike, beat', as reconstruced on the basis of the Gr. evidence. IE COGNATES: Gr. némesis 'divine retribution', nemétōr 'avenger', OIrish námae 'foe', (?) Alb. (Tosk) nëmë, (Gh.) namë 'curse' > It seems clear that Sanskrit nabh-'strike / break apart / tear' should be added as a cognate of Iranian *namH 'to strike, beat, abuse'. If from *nemH1-, dissimilation of *n-m > *n-b would create *nebH1-, with IE *CH > *ChH > Ch. It is also not likely that 2 roots *nemH1-existed in PIE with differing meanings. Here, 'reproach, abuse' seems to show that older *nemH1-'attack' fit all meanings above. If so, its connection to *nemH1-'seize / distribute' would be from 'seize (from others) / loot / raid / attack'. A similar shift in other IE roots covers a wide range of derived & metaphorical meanings : G. hairéō, Cr. ailéō 'take/grasp/seize/win/gain', Lt. sirt 'to loot', OIr serb 'theft', H. sāru 'booty' *slH1gW-\ *slH2gW-? > OE laeccan 'grab', G. lambánō 'grasp/seize / plunder / catch/discover/perceive/get', lêpsis 'seizing / receiving/accepting' G. láphūra 'spoils of war'. Li. lõbis 'big possession / treasure / riches / good(s)' and others that show 'decide/determine' vs. 'beat', possibly showing 'judgement' > 'punishment' or *nemH1-> Iranian *namH 'to strike, beat, abuse' *nebhH1-> Skt. nabh-'strike / break apart / tear' *samH2dho-> E. sand, G. ámathos *sabhH2dho-> L. sabulum, Arm. awaz *domH2no-> L. dominus 'master' *dobhH2no-> L. dubenus 'master' (related to *domH2(o)-'house') *kolH3mon-> L. columen > culmen 'top / ridge of house' *kolH3bhon-> G. kolophṓn 'summit' Skt. meṇḍha-'ram' Skt. *mheṇḍa-> bheṇḍa-'ram' *molHo-> Skt mala 'dirt / filth' *mHol-/ *bhHol->> G. molúnō / pholúnō 'soil/defile/debauch / stain/pollute / dye / (pass.) become vile/disgraced' *mHor-/ *bhHor->> phorū a nō 'defile/spoil', *phorúkh-yō > phorússō 'defile/knead/mix', *morúkh-yō > morússō 'soil/defile/stain', perf. memórugmai, Mórukhos '*participant in debauchery / *follower of Dionysus > Dionysus' (as in other words for 'follower of Dionysus / Dionysus') *Hmerwo-> W. merw 'weak / slack' *Hmarwo-> G. amaurós / maurós / maûros 'withered / shriveled / weak / feeble' *mHarwo-> *bhHarwo-> G. aphaurós 'weak / feeble', phaûlos / phlaûros 'petty / paltry / slight / low in rank / insignificant / easy', phaûros 'light' *mHegWno-> Av. maγna-'naked', Arm. merk, G. gumnós, Skt. nagná-*mRegWno-> *bhRegWno->> *b(r)agnaka-> MP brahnag, Os. bägnäg 'naked', Sog. ßγn'k *pumHe:s ? > Skt. púmān 'man', stem púmaṃs-/ puṃs-*puHbhe:s ? > L. pūbēs 'adult' ? > Skt. kiṭibha-m 'kind of exanthema' ? > Skt. kiṭima-m 'kind of leprosy' (see relation below; perhaps all IE words with *-(V)mo-and *-(V)bho-came from *-mHo-, etc.) *mraru-> Skt. mallu-'bear', *mrarw-on-> Greek Braurṓn (the site of an important sanctuary of Artemis where girls imitated bears) *mRaru-> *mhRaru-> *mharRu-> Skt. bhalluka-'bear' *wei(H)-'curved / bent / bend / wind / twist' >> *wimHon-> *wimon-'seaweed' > Middle Irish fem(m)ain, Welsh gwymon *wibhHon-> Latin vibō, gen. vibōnis, 'flower of Britannica' (the change of 'winding' to plants that wind around others things (and seaweed, known for this) is possible) *gWerHu-> L. verū 'spit/dart/javelin', *beru > Gaelic bior 'stake/spit' *gWerHu-masko-> Pamir *garimaška-> Shughni žīrmesk 'mullein', Yazghulami γurmešk *gWerH-mhasko-> *gWerH-bhasko-> L. verbascum 'common mullein' (it could be derived from 'stake/spit' based on the look of the large prominent stalk; this much similarity in unrelated words for the same thing would be too much for chance in IE, see Krzysztof Tomasz Witczak, verbascum ) Further notes on origins : 1. The evidence for *krstHmo-> kiṭibha-m / kiṭima-m comes from metathesis > *kHrstmo-> MP xurmā 'date' in:

Greek changed Kw > kWkW / pp: *H1ek^wos > L. equus, *yikwos > LB i-qo, G. híppos, Ion. íkkos 'horse'; *laku-L. lacus 'basin/tank/lake', *lakw-> G. lákkos 'pond/cistern/pit'; *pel(e)k^u-> G. pélekus '(double-edged) ax', *pel(e)k^wo-> pélekkon \ pélekkos 'ax-handle'. LB also had optional -pp-: G. hippophorbós, LB *yikwo-phorgWoi / *yippo-phorgWoi 'horse-feeders / ostlers' / i-qo-po-qoi / i-po-po-qo-i-qe ( < *kWe 'and') Now, why would a word like G. hippophorbós have 2 outcomes? Some say it was KW-KW dissim., but that does not apply to lákkos. Either way, it would not be regular. In fact, in *kWolpo-> OE hwealf 'vault/arch', G. kólpos 'bosom/lap / hollow space', not *pólpos, the lack of *p-is supposed KW-P dissim., so if anything had been regular, why not *yikko-phorgWoi? How could KW-P > K-P & KW-KW > P-KW apply in a word with KW-P-KW anyway? The double outcomes might be seen elsewhere, also iregular. Since these also do not have KW-P, etc., they come from *kp (based on kp elsewhere in the area, Paeonian Lúkpeios (from either 'wolf' after *kW > *kw or a derivative of *l(e)uku-'light / bright'). Knowing this, I looked for other ev. of *kw > *kp, *kp > p(p) / k(k), etc. : PIE *w(e)rp-> G. rháptō 'sew', Li. verpti 'spin', *wr̥ p-ko-> Aeo. brákos, rhákos 'garment/rags/ cloth / tattered garment / strip of cloth' L. stupēre 'be stiffened / be stunned / be struck senseless / stop', *stup-ko-'stiff fiber/hair' > G. stúp(p)ē \ stup(p)íon 'coarse hemp fiber', topeîon 'rope/cord', Skt. *stupka > stúkā-, *stukpa > *stuxpa > stūpa-'knot/tuft of hair / mound', Os. styg 'lock of hair' The optional nature of many of these resembles Tw / Tp > (t)p : *dhwn-dhwl-> G. pamphalúzō, tanthalúzō 'quiver / shake', *dwal-dwol-ye-> Arm. dołdoǰ 'quivering', yołdołdem 'shake/move / cause to totter/waver', dandałem 'be slow / delay / hesitate', dandał 'slow'. ? > *dhven-dhvreHn-> G. pemphrēdṓn, tenthrēdṓn 'a kind of wasp that makes its home in the earth' (likely 'cicada', thus connected to 'loud sound', see Skt. dhvánati 'roar / make a sound/ noise', dhvraṇati 'sound', dhvāntá-'a kind of wind') *stel(H)-ye-/ *stH2al-ye-? > Skt. sthal-'stand (firm)', OE stellan 'stand', OHG stellan 'set up', *stwélyō > G. stéllō 'make ready / equip / prepare', Les. spéllō *stolHo-> L. stolō 'shoot/branch/twig', *stwolo-> G. stólos 'equipment', Thes. spólos 'stake' Other loans show the same : 2 cities in south central Anatolia: H. Azatiwada-'ruler of Karatepe', Azatiwadaya-'Karatepe' G. Áspendos, Pamp. gen. Estwediius *walto-'hair' > OIr folt, Li. valtis 'yarn', G. *wlatiyo-> *wlatsiyo-> lásios 'hairy/shaggy/ wooded', Lasía, *latswiyo-> Lasíā, Lésbos >> H. Lāzpa This raises the possibility that some Greek *-kw-could become *-kW-,*kw-could become *kW-, but other dialects had *-kp-> -pp-/ -kk-, *kp-. LB would usually have q, sometimes pp (not knowing if q could stand for both kw / kW makes any more exact analysis hard). Knowing that G. pépāmai 'possess' is usually derived from *k^waH2-'swell' as in 'become big/strong' > 'have power/possession', with the assumption that *k^w-> *kWkW-> *pp-or similar (since *ppāstos 'possessed/acquired' seems to exist in the name Thió-ppāstos "Possessed by a god") makes it likely that pp-really came from *kp-. Since there are many pairs with p-vs. kt-of the same meaning ( pépāmai \ kéktēmai 'possess', égktēsis \ émpāsis 'estate/property', Zeús Ktḗsios \ Pāśios Zeús (protector of house and property), the names Ktḗsarkhos \ Pāśarkhos), I assume that in some dialects *kp-> pp-, in others *kp-> kt-. L. merda, TB melte 'dung', G. mínthos 'human ordure' (also with e > i, d > th, as in Crete) For *lss > lass, see alternation of lC / rC with lVC / rVC : *l-> G. Odusseús / Olutteus / Ōlixēs G. kalúptō vs. krúptō 'cover / hide / conceal' G. skórodon / skórdon, Alb. hurdhë, Arm. xstor 'garlic' G. kárphos 'dry stalk / stick / twigs', Mac. kárabos 'door' ? > G. thálamos 'inner room', oph-thalmós '*socket > eye' ? > L. ervum, G. órobos 'bitter vetch', orbopṓlēs 'vetch-seller' *derwo-> Li. dervà 'tar', G. términthos / terébinthos 'terebinth' *H1esH2r(K) > *ehar > G. éar 'blood', *eharǝ > *eara > poetic íara *H2albho-> L. albus, Greek alōphós 'white', alphoprósōpos 'white-faced' *bher-tro-m > L. ferculum 'bier / litter', G. phér(e)tron, Skt. bharítra-m 'arm' Skt. gárbha-, Av. garǝwa-, G. delphús 'womb', adelpheós, Lac. adeliphḗr 'brother' Skt. álpa-'small / slight flimsy', Li. alpùs 'weak', G. alapadnós 'easily exhausted / feeble' *mortyo-> OSw. merði, OIc merð 'fish-net', *-ts-> G. mórotton 'basket made of plaited bark' *gWrtro-'throat / pit' > Li. gurklỹs 'crop (of a bird)', *gWǝrǝtro-> G. bárathron, Ion. bérethron 'pit' *skalps 'digger > 'mole / snake' > *skal(a)ps >> skalapázein / skalpázein '*slither? / go by rolling/turning' *(s)parsa > Latin parra 'bird of ill omen', *parasos > Mac. paraós 'eagle', *sparsios > G. spalásios 'bird like the sparrow' *dhra(H2)g(^)(h)-> OE dreccan 'torture', G. tarássō \ thrássō 'stir/agitate/trouble/disturb', Skt. drāghate 'torment', R. draznít´ 'tease/pester' 2. Chiapello has done more with this. He compared pictures of vessels next to LA words he took as G. ( ), finding the same in kálpē : ka-ro-pa3. I should also mention Chiapello's idea ( ) that LA da-ro-pa next to an LA logogram *403VAS, which looks like a Minoan basket-shaped vessel, can be explained by da-ropa : G. dárpē 'large wicker basket'. These both have IE cognates : *kelp-> OIr cilurnn 'urn', W. celwrn, G. kelébē 'cup / jar' Linear A ka-ro-pa3, G. kálpē 'pitcher' >> L. calpar 'wine cask', Calpurnius > G. Kal(o)pórnios LA da-ro-pa, G. terpós \ tarpós \ tárpē \ tarpónē \ dárpē 'large wicker basket', Arm. t'arp' 'large wicker fishing-basket / creel', t'arb 'framework of wooden bars / wicker trellis-work' from *terp-'turn' (referring to weaving or plaiting) This probably shows that dárpē was a Cretan version of G. tárpē. This alternation of voicing and aspiration also seen in Cr., and many other G. (some of which could have come from a dialect like Cr.), for th / t / d : G. dáptēs 'eater / bloodsucker (of gnats)', Cretan thápta 'fly' G. tárpē \ dárpē 'large wicker basket' *dyeus > Zeús, acc. *dyeum > *dye:m > G. Zēn-, Dor. Zāń, Zāś, Cr. Tāń, Tēn-, Ttēn-Cr. óthrus 'mountain', Óthrus 'a mountain in Thessaly', *odrus / *odurs / *oduros LB o-du-ro, gen. u-du-ru-wo 'Zakros (in Cr.)' *andhroHkW(o)-'(hu)man' > G. ánthrōpos, Cr. antrōpos *dwi->> G. dí-sēmos 'of 2 times / with a double border, haplo-dísēmos/tísēmos *d(e)mbh-> Skt. dambh-'slay / destroy', Os. davyn 'steal', G. atémbō 'harm / rob' *bhled-? > G. phledṓn 'idle talk', pl. blétuges 'nonsense talk' *meld-'soft', *mld-ako-> G. malthakós tarássō \ thrássō and other C's from Cr. : G. kúmbē \ kúbē 'head', Cr. kuphḗ G. glágos, Cr. klágos 'milk' All this is very important in showing that LA contained IE words, all of which could be Greek. That some did not resemble standard Greek is due to the presence of sound changes found only in some dialects, as above. Without looking carefully and considering all evidence, no progress can be made. The spread of dárpē 'basket' from Crete might have to do with the basket-shaped vessels known from there. These might just be made for artistry, or could show that baskets were important in sacrifice and/or ritual (as he points out for the scene on the sarcophagus from Aghia Triada) and so kept the shape even when technology had made the use of plant-based objects used to carry bloody objects or liquids unneeded. As more support, another word for dárpē is sargánē. This s-could come from *tw-, which would allow dárpē to be directly related to tárpē (when *t-> d-would show that IE cognates like t'arb were certainly related). This could be due to another -pk-compound : tárpē 'large wicker basket' + kánna 'reed (mat)', kánathron 'cane or wicker carriage', LB *konōn < *kanōn << Ak. qanūm *tarp-kanā > *tpar-kanā > *twar-kanā > G. sargánē 'plait / braid / basket' Since G. terpós \ tarpós also shows e / a, the similar segánion 'wicker creel' could also be < *terp-kan-. This would further link all words. Loss of *r as due to r / R / x (as in kr > *kx > kk ): *proti > G. protí, Dor. potí, Skt. práti, Av. paiti-, etc. *mrkW-> G. márptō 'seize/grasp', mapéein 'seize' nebrós 'fawn', nebeúō 'serve Artemis (by imitating fawns)' *trVkso-'badger' > L. taxus, G. trókhos *drp-drp-> *dardráptō > dardáptō 'eat / devour' *dr(e)p-'tear (off / apart) > G. drépō 'break off', *dráptō > dáptō 'devour / rend / tear' G. daitrós 'person who carves and portions out meat at a table', Mac. daítas *smiH2-s > *smi:H2 'one', fem. nom. (like *-or-s > *-o:r, etc.) *smi:H2-ro-> G. (s)mīkrós 'small', Dor. mīkkós < *mīkxós *karsto-> Gy. karšt / kašt, G. káston 'wood', Arm. kask '(chest)nut' *bRuHk-> G. brūkháomai, Skt. bukkati 'roar', SC bukati *sprag-> ON spraka 'cackle/patter', Li. spragù, *spRag-ato-> G. spataggízein 'be in an uproar?', pátagos 'clatter / crash of falling/thunder / splash', platagḗ 'rattle' *dhwrenH1-? > Skt. dhvánati 'roar / make a sound/noise', dhvraṇati 'sound', dhvāntá-'a kind of wind' *dhwen-dhreH1n-> tenthrēdṓn 'a kind of wasp that makes its home in the earth', *tenthēdṓn > *tīthōn / *tinthōn 'cicada' >> Tīthōnós, Etruscan Tinthun That Dor. & Mac. might have retained this *R longer (making r-loss more common there) might be ev. of a close relation (a Dor. dia. is now known within the territory of Mac.). Other ev. shows the same in Cr.: PIE *ksustom > G. xustón 'spear/lance', Cretan rhustón 'spear' shows the odd change ks > *xr > rh, and since ks-/ rh-is also in Aeo. xímbā, (dia. not specified) rhímbā 'pomegranate', these both probably came about by the same type change. *ks > *kx > *kγ > *xR > *hR > rh would show a velar > uvular fricative (many languages have uvular r's of various types; xx-is not odd for G. with other CC-from various changes, like pp-, Cr. tt-). Also, in Crete *tw-> tr-(*twe 'thee' > Cr. tré), allowing *tw > *tγW > *txW > *tR > tr (which would show a change known...

The cause of most alternation is probably dissimilation or assimilation near a 2nd m / n or P / KW / w / u. Others are unexplained (some possibly caused by *H, if *H3 = xW, etc.). For some examples, often in Tocharian, see : m / n by m / n *mene 'mine' > OCS mene, Av. mana, Skt. máma *mems-> Go. mimz 'meat', *mensinks > G. mḗnigx 'membrane' Skt. mand-'rejoice / be glad/drunk / shine / praise', nand-'rejoice / be glad/pleased' *ni-dr̥ mH-> A. níidrum h-'fall asleep', met. > *nimdraH > Lahnda nindr, Skt. nidrā́ 'sleep (noun)', middha-m 'drowsiness', Kati mīnO, Sa. minī-'sleep' *-man > *-mam in OIr and Av. (Byrd 2006) likely a similar change in Ph. : Ph. iman 'memorial? / marker? / grave marker? / headstone?', G. ídmēn 'care / consideration' < *wid-men-'knowing' The Ph. name Iman would then be 'wise', with its equivalent in Armamaic zmam (appearing on coins) showing *w^iðman > *yizmam > *izmam *H1newn/m '9' (or caused by W below) 9 OE nigon, L. novem 9th > L. nōnus, Skt. navamá-, TB ñunte 90 > TB ñumka *mHegWno-> Skt. nagná-, Av. maγna-'naked ', Arm. merk, G. gumnós (or caused by W below) This also could explain the different Germanic outcomes of *mn as some *mn > *mm > _m (lengthening the V) : *Hnomn-ye-'name' >> G. onomaínō, Go. namnjan, *nammjan > *nōmjan > OF nómia *men-'think' > *men-mn > Skt. mánman-'thought/mind', OIr menme *men-mn-yo-s 'wise' > NPic. Mimnis, *memniyo-s > *mimmija-z > *mīmija-z > ON Mímir (or directly from perfect *me-mon-/ *me-mn-'have thought/known > remember / be wise/ knowledgeable' ) m / n by labial P / KW / u Li. nugarà 'back', Lt. mugura Skt. ámīva-'disease / distress', G. anīā, Aeo. onīā 'grief/sorrow / distress/trouble' *(H3?)nogWh-> TB mekwa 'nails', TA maku *n-Hed-we-'not eat' > TA nätsw-'starve', TB mätsts-*negWhró-'kidney' > *meghwró-> TA mukär Skt. viḍa-lavaṇa->> TB wiralom 'a kind of salt' (a medical ingredient) Skt. cūrṇa->> TA cūrṇ / curm '(medicinal) powder' IIr. *nastula-/ *mastula-'of nose(s) / nasal' > Kh. nastùḷi 'runny snot', Skt. nastakarman-, *nastulakarman-/ *masturakarman->> TB nastukārm 'nasal medicament', mastukārm 'medicine applied via the nose' *nebh-> G. néphos 'cloud', Skt. nábhas-'cloud/fog/mist', L. mefītis 'poisonous gas from swamp/volcano' L. pugnus 'fist', G. pugmḗ (maybe many others with -mo-vs. -no-with same meaning, hard to tell if all had same origin) *k^witro-> Skt. śvitrá-'white', *k^witi+ in compounds > śviti+, *k^wityano-> G. títanos / kíttanos 'chalk / lime / gypsum', Cr. Kíssamos, Kísamos *muH-'silent / mute / unable to speak / in a low voice / whisper', L. mūtus, G. nuthós 'dumb/ numb/dark', noûthos 'dull [of sound]' *nuH-'loud' > Skt. navatē \ nāuti 'sounds', Old Irish núall 'scream / din/fuss/noise / proclamation', G. móthos 'battle din', mûthos 'word/speech / saying / story' *-wVn > -wVm in G. *selwḗn > G. Seilēnós (the strange shape suggests a source in -ēn (common in G.), changed to ostem by analogy (like Tīthōnós from *tīthōn 'cicada') *selwḗn > *serwḗm > Linear B se-re-mo-ka-ra-o-re '(decorated with) siren heads', G. seirḗn 'siren' *H1newn/m '9' (or caused by m / n above) 9 OE nigon, L. novem 9th > L. nōnus, Skt. navamá-, TB ñunte 90 > TB ñumka *mHegWno-> Skt. nagná-, Av. maγna-'naked', Arm. merk, G. gumnós (or caused by m / n above) m / n by xW / H3 ? *dr̥ mH-> L. dormiō, *dr̥ -dr̥ mH-> G. darthánō 'sleep', Arm. tartam 'unsteady/wavering/ sluggish/idle' *n(o)H3-? > nóthos 'bastard/base-born/cross-bred/spurious', móthax / mó(th)ōn 'class of noncitizens raised as foster-brothers of Spartans / impudent fellow / licentious dance' The shift base-born/non-citizen seems to fit, and both might be connected to (with other alternations of their own) : nōthrós 'sluggish', nōthés 'sluggish/dull/stupid' Since Greek has many words of unknown origin, and the IE ety. of some beginning with n-& mhas been questioned ( 35402227 ), it makes sense to look for this same m / n alternation here. Not all words beginning with n-invariably came from *n-, etc. I see : G. malakós 'soft/weak/gentle', mal(a)kíō '*become weak > become numb with cold', málkē / nárkē '*weakness > numbness' This also shows r / l, known in other dia. (like Cretan) and -a-/ -0-like cognates already known for malakós 'soft' : *melH2du-'soft' > W. meladd, *H2mldu-> G. amaldū́nō 'soften' This same -a-/ -0-also in : *mar(a)thuro-> G. márath(r)on 'fennel', LB ma-ra-tu-wo 'fennel?' *nárthrāks > G. nárthēx / náthrax 'giant fennel' Keep in mind that without LB, we would have no evidence that m-w-existed to dissim. > *n-w, etc. That both these words for types of 'fennel' had (r)-th(r) and Nar()thr-makes their common origin nearly certain, with IE m / n in mind. For more on the origin of these, see below. The reason for thinking *mar(a)thuro-> G. márath(r)on, LB ma-ra-tu-wo is that r-(r) already is known in G., so with -u-/ -0-seen in other G. (thal(u)p-'warm', etc.), this dissim. taking place twice, before & after -u-> 0, explains all data. From a previous paper : Greek has several adj. in -uro-not in other IE: halmurós 'salty', kapurós 'dried by the air', etc. There is no reason to think the affix itself is not IE, seen in *seg^hurHo-'holding' > G. ekhurós / okhurós 'durable / secure', Skt. sáhuri-'mighty / strong / victorious'. I see no reason for Beekes' "Pre-Greek" This includes relatively old loans : Aramaic neṭāpā / nāṭōpā 'drip / aromatic resin' >> G. métōpon 'galbanum', metṓpion 'Egyptian aromatic ointment containing galbanum or oil of bitter almonds', netṓpion / nétōpon / níōpon 'oil of bitter almonds' Other Hamito-Semitic words that would easily be seen as the source of G. words, if from *mnot n-, include : *knuk-s 'nut' > *xnuxs > Ak. nušhu 'almond?' >> Arm. nuš 'almond', nši 'almond tree' *xnuxs-alo-> *Hmuγzalo-> G. amúgdalos / -on / -ē / amusgélā / amusgúlā 'almond' Here, the source of Ak. nušhu is not known within Sem., so its similarity to *knuk-s 'nut' allows older *xnusx-with dissim. of x (like some IE did for *k-k here). This loan taking place before *H-was vocalized in G. allows *k-> *x-> a-. Optional sg / *dg > gd caused by the same change as *zd > dd in G. dia. (likely zd / ðd, zg / ðg, etc.). This might also explain Cr. ákhnula 'nuts' by a version w/o *nu > mu : *xnuxs-alo-> *anuxsalo-> *anukhslo-> *anukhlo-> ákhnulon* (with the *-V-> 0 also seen in G. oísupos / oispṓtē 'lanolin ', etc.). Maybe also (if n > m could cause nt > mp ) : *pntH2-'go' > Go. finþan 'find out', OE fýsan 'send forth / drive away', G. pémpō 'send/ dispatch / send forth / escort'

This Guidebook is a comprehensive language method for learning Modern Indo-European from scratch. So far two volumes of the Guidebook have been published, with a total of 50 lessons.  Seven of the lessons are devoted to revision and notes. This second volume contains lessons from 26 to 50.

This Guidebook is thought not only for beginners, but also for advanced students and language researchers. The Guidebook is based on current scholarly work on Indo-European studies and aims at presenting a reconstruction as similar as possible to the original Proto-Language which developed at a late period of community. This self-learning method has the advantages of being intituitive, funny and easy to follow, while focusing on the most important lexical and morphosyntactic elements in gradual order of complexity.

Its content is presented in the form of dialogues and narrations with addition of insightful grammar and style notes.  A vocabulary list appears at the end of each lesson. By following this method students will:

• have a better understanding of the main characteristics of Modern Indo-European, in parallel to the ones featured by the Proto-Language;
• be able to ‘fly with their own wings’, and hold basic and medium-level conversations;
• command an extended lexical corpus of Indo-European;
• have an accurate clue of where and how to tackle scholarly references;
• acquire solid skills whereby they will find learning Indo-European daughter languages much easier. This Guidebook is highly recommended to all those who want to experience progress in Modern Indo-European and apply this interesting language in real living situations.

There has been a long-standing debate regarding the featural representation of `voiced-aspirated&#39; stops. Traditional models, including Laryngeal Realism (Honeybone, 2005; Iverson & Salmons, 1995; Beckman et al., 2013), assume that `voiced-aspirates&#39; are represented with a combination of two separate features: [voice] and [SG]. Other proposals (Ladefoged, 1971; Schiefer, 1989), however, contradict the traditional ones, arguing that `voiced-aspirated&#39; stops involve a completely di erent phonation type, presumably represented with a [breathy] feature. While earlier studies relied primarily on articulatory data, this dissertation sought to investigate the question of featural representations of `voiced-aspirated&#39; stops with data of several types: speech production, perception, phonological patterning (alternation/neutralization), and learnability of features via arti cial language learning experiments. Data were obtained from native speakers of (Bangladeshi) Bangla excep...

Revisionist trilaryngealism, consisting of the hypotheses CC•C, *h1 h2 h3, and at least two different vowels PIE *e *o (and optionally PIE *a), has split the laryngeal theory into several mutually incompatible models. The models of EICHNER (1973, 1978, 1980, 1988) and MELCHERT (1987)/RIX (et al. 2001) are characterized by symmetrical, but opposed reconstructions for Hitt. a-(*h3e-vs. *h1o-) and Hitt. ḫa-(*h2e/o-vs. *h2/3e/o-). In KORTLANDT's (2003-4) model the preservation of *h2 and *h3 is conditioned by the distributions of *e/o. Both laryngeals are allegedly retained before *e, but lost before *o in Old Anatolian. In addition, it is theoretically possible to define a variant of KORTLANDT's model in which the distributions are reversed. The present paper focuses on KORTLANDT's model in both its original and reversed form and demonstrates the internal inconsistency of this model, due to which it has to be discarded as a serious option for PIE reconstruction. This leaves us only the models of EICHNER and MELCHERT/RIX to compete with SZEMERÉNYI's (1967, 1970, 1996) monolaryngealism for the solution of the PIE laryngeal/vowel problem.

Ancient Greek famously shows an alternation between suffixal e-grade in the simplex of certain ablauting nouns and o-grade in the second member of compounds, e g πατήρ ‘father’ : εὐ-πάτωρ ‘having a good father’. First identified by Meillet, Old Armenian anjn ‘person’, pl. anjinkʽ : mi-anjn ‘monk’, pl mi-anjownkʽ has been cited for over a century as a counterpart of Greek compositional o-grade suffix (COS). However, the parallel is imprecise for two reasons: the inflections of anjn and mianjn differ only in the nominative and accusative/locative plural; and COS in Greek is mainly found with r- stem kinship terms, whereas in Armenian it is restricted to n-stems.
It emerges that COS in Armenian is practically restricted to compounds
of anjn and azn ‘nation, people’, raising the possibility that semantic factors and/or lexical analogy have played a role. After examining the
prehistory of n-stem inflection in Armenian, it is proposed that these compounds have secondarily acquired plurals in -own- by lexical
diffusion from other nouns denoting individuals, the core of which goes back to derivatives in *-ōn- < PIE *-o-(H)on- with the substantivizing or individualizing suffix *-(o)n- and/or the possessive suffix *-H(o)n-. It follows that the Greek pattern should not be automatically reconstructed for the protolanguage.

This article deals with the identification and explanation of several Iranian words and names preserved in Greek sources (απομεναμα < OP *apām vauvīnām māha ‘Month of the Good Waters’; πισάγας < OIr. *pisanga- ‘leper, spotted’; Τέασπις < OP *Θiyāvāspi-; ψιττάκη < OIr. *xšviftakā̆- ‘sweet, nectarous (bird)’; *τίγρης < OIr. *tigra- ‘sharp, pointed’). For each of them, a thorough discussion is provided on the problems connected to the transmission within the Greek tradition and to the ultimate Iranian etymology.

The aim of this contribution is to demonstrate that the form μᾶλα ‘sheep’ in Theocritus is not (as is commonly assumed) a hyperdorism, but rather points to an inner-Greek substantivization of an adjective μᾱλός, attested in Theocritus (as an epithet of a goat) and as (Att.-Ion.) μηλός ‘tame’ in the Etymologicum Magnum. Inducing a discussion of the development of the cluster *dm- in Greek, I will propose to reconstruct this adjective as *dm̥h₂-ló- (from *√demh₂ ‘tame, subdue’).

Taking inspiration from the biological taxonomy of mammals, this paper explores the diversification of the Indo-European language family through a zoographical lens. It investigates shared innovations of phonology, morphology, and semantics in zoonyms across language branches. The aim is to uncover evidence for early splits within the family tree. The study primarily centers on Anatolian versus Core Indo-European and features an extensive discussion of *h₂ŕ̥tk̑o- ‘bear’ (Hittite ḫartakka ) vs. *h₂ŕ̥k̑þo- (Vedic ŕ̥kṣa , Greek ἄρκτος, Latin ursus, etc.), *h₁ék̑u- ‘horse’ (Hittite */ekkus/, Luwian /azzu-/) vs. *h₁ék̑u̯o- (Vedic áśva-, Latin equus, Tocharian B yakwe, etc.), and *u̯ĺ̥kʷo- ‘lion’ (Luwian walwa/i-) vs. *u̯ĺ̥kʷo- ‘wolf’ (Vedic vŕ̥ka-, Lithuanian vil̃kas, Tocharian B walkwe, etc.). Thorough analysis of these examples will determine their relevance within the proposed scenarios.

https://www.stockholmuniversitypress.se/site/books/e/10.16993/bcn/

Traces the history of accented morphemes from Saussure back to Sanskrit sources (primarily Pāṇini) and casts a glance at the Greek grammarians. This paper weighs in on current debates about Proto-Indo-European morphophonology. The narrative may be hard to recover from the handout but feel free to write me for a follow up (a version of this handout will be advanced for publication)

The Hittite ḫi-conjugation is described as basically incompatible with the Gaeco-Aryan verbal system and consequently as reconstructionally insolvable. There might, however, be no ḫi-conjugation at all. There might be just an upgrading vowel -ó- in the „strong“ alternant which is always tightly combined with the stative endings („combining theory“; here below: "TEN").

Only one sixteenth of the Hittite verbal area (6,25 %) is concerned: not the stative section, nor the mi-verbs, nor the middle of the ḫi-verbs, nor the plural of the active of the ḫi-verbs (which keeps the original endings, especially the far-beyond-doubt-attested 3pl ending énti, untouched).

Ergo, the conclusive refutation of the nine traditional ḫi-conjugation theories may run as follows:

One: énti is not a 3-pl-perfect ending („perfect theory“; e.g. Eichner).

Two: énti is not a 3-pl-middle ending („middle theory“; e.g. Rosenkranz).

Three: énti is not a 3-pl-perfect-middle ending („perfect-middle theory“; e.g. Meid).

Four: énti is not a 3-pl-middle-déclassé ending („middle-déclassé-theory“; e.g. Neu).

Five: énti is not a nominal ending („nominal theory“; e.g. Cowgill).

Six: énti is not a 3-pl-thematic ending („thematic theory“; e.g. Couvreur, Pedersen, Watkins).

Seven: énti is not a 3-pl-h2e-conjugation ending („h2e-conjugation theroy"; Jasanoff).

Eight: énti is not a 3-pl-inactive ending („inactive theory“, Gamkrelidze-Ivanov).

Nine: énti is not a 3-pl-voice-opposition ending („ voice-opposition theory“, Sarah Rose, Memorial University St. John’s, Newfoundland).

TEN: Does not exist, because there is no ḫi-conjugation. In other words, énti solely is an athematic 3-plural-indicative-present-active ending. Or, in short: there are, to be sure, ḫi-verbs in Anatolian, but there is no ḫi-conjugation.

Reconstructing sound changes through lexicographers and grammarians: the case of a post-sibilant aspiration in Greek Data coming from lexica, grammarians, and scholia traditionally played a cardinal role in the reconstruction of Greek (including Mycenaean) outcomes of Indo-European roots. Which caveats should be considered when reconstructing sound changes based on these sources? In this work, I discuss a potential Greek sound change, i.e. the aspiratory effect triggered by a sibilant on a following stop, which was posited during the 19 th century almost exclusively on the basis of forms attested by lexicographers of the Imperial and Byzantine Ages.

The Saussure Effect is stated in various ways (see below), but in its simplest form it describes loss of *H near *o in Greek, and seeks to find regularity in its cause(s) : *oCHC > *oCC *bremH1-> bremetḗs 'roar', *bromH1taH2-> brontḗ 'thunder' (1) *terH1-> téretron 'borer / gimlet', *torH1mo-> tórmos 'hole / socket' *peraH2-> peráō 'pass/go through', *porH2tmo-> porthmós 'ferry / strait' *gWorH3mo-> bórmos \ brómos '(wild) oat' *HCo-> *Co-*H3lig-> olígos 'small / few', *H3loigo-> loigós '*diminishing > decimation' *H2nēr 'brave / strong / hero?' > anḗr 'man', *+H2nōr >> *H2nōreH1eti > nōreî 'is active' *H1lektro-? > ḗlektron 'amber / electrum', ēléktōr 'shining', loggoúrion / luggoúrion 'amber', log(k)oúrion 'glass' Though the source of ḗlektron, etc., is not clear, I include them in case ē-e-vs. 0-o-is part of the same change. It also did not affect *-oH-in : *dheH1món-> G. thēmṓn 'heap', Bu. dúuman 'pile / heap' *dhoH1mó-s > G. thōmós 'heap', Ph. dumas 'barrow?', Go. dóms, E. doom which would simply be a feature of the scope, if regular, but see below for other examples of *o(h)m-. Others are not clear; blōmós 'morsel of bread' could be from *gWlH3mo-or *gWloH3mo-, but its resemblance to *gWorH3mo-suggests o-grade. Other words appear to violate one or more of these principles for *-o()CHC-, often of unclear origin : *HmeigW-> ameíbō '(ex)change', amoibḗ (not *moibḗ) *HleipH-? > aleíphō 'anoint', aloi(m)mós '*oiling > polishing/plastering of wall-decoration' *pelH1-> ON felmta 'be frightened / tremble', G. pelemízō 'shake / cause to tremble', ptólemos / pólemos 'war' *klH3mo-> OSax. holm 'hill', *kolH3mon-> L. columen > culmen 'top / ridge of house', G. kolophṓn 'summit' *k^orH2k-> R. soróka, Li. šárka 'magpie', G. kórax 'raven', Skt. śārikā-'mynah / preacher crow' *kelH2-> Li. kálti 'strike/hammer', L. per-cellere 'strike down', clāva 'club', G. kólaphos 'buffet/blow/box on the ears' *(H?)loup-eH1k^o-'fox' > Skt. lopāśá-\ lopāka-, etc., G. alṓpēx \ alōpós, Arm. ałuēs *melH-'grind / dust' > G. mólophthos 'loaf baked in the ashes', Arm. młeł 'dust / chaff / ash' *morHtyo-? > OSw. merði, OIc merð 'fish-net', *-ts-> G. mórotton 'basket made of plaited bark' ? > G. skórodon / skórdon, Alb. hurdhë, Arm. xstor 'garlic' ? > L. ervum, G. órobos 'bitter vetch', orbo-pṓlēs 'vetch-seller' (which seems to show loss of *H in compounds, like G. thálamos 'inner room', oph-thalmós '*socket > eye') and if G. adj. in -amos, -anos & -aros came from *H2 (also see plókamos below), it could explain 'sphodrós \ sphedanós 'vehement/violent/impetuous' (sphendónē 'sling', sphéndamnos \ spéndanos 'Montpellier maple'), but their retention in : wólH2mo-> G. oulamós 'throng of warriors', [w]ólamos 'chase' *H3orbh-> L. orbus 'bereft', G. orpho-, orphanós ( >> E. orphan) *wrd-? > G. rhadinós 'slender/tender/pliable', rhodanós '?' *(H)wors-? > G. Ouranós '(god of) sky/etc.' etc., would also argue against regularity. For *-(a)ros & *-(a)mnos with loss of -a-, but not caused by *o, also see *new(a)rós > G. nearós 'young/new/recent', nebrós 'fawn', Arm. nor 'new' G. rhádamnos \ rhádāmos \ rhódamnos \ oródamnos \ *w\oradmnos > óramnos 'branch', *wradmnos > rhámnos 'box-thorn' Also some unclear cases, *k^elH2-? > kélados 'noise/din/clamor / sound/cry/shout / twitter/ chirp', kolo-surtós 'noisy tumult/uproar', kolōiós 'racket/brawling', koloiós 'jackdaw', koloíphrux 'Tanagrian cock'. Pairs like G. thalámē 'cave/den', ophthalmós '*socket > eye', with apparent *H > 0 in compounds, would support -amos < *-H2mos, or maybe PIE *-mHo-with optional metathesis, to explain -mos & -Vmos (as if from *-mHos, Italic *-amos, Gmc. *-umaz, etc.). Also against regularity, there seem to be doublets, or near-doublets, like : *plek^-> plékō 'plait', *plok-Hmo-> *plok-[H/h]mo-> plókamos / plokhmós 'braid' ( ) *petH2-'extend', pítnēmi 'spread (out/open)', potamós 'river', and the nearly(?) identical : *petH3-'fly / fall', ptôma 'a fall', pótmos 'what befalls one / fate / lot' (3) *nemH1-'give/take', *nomH1o-> G. nómos, Dor. noûmmos 'usage / custom / law' (2) *smogH1-> G. (s)mogerós 'suffering hardship', *smog[h]-> mógos / mókhthos 'work/toil/ hardship', *smogh-> Li. smagùs 'heavy' *terH1-tro-'gnawing / scraping / boring / cutting' > téretron 'borer / gimlet', térthron '*point > summit / tip' (4) *temH1/2-? >> Tómaros \ Tmáros 'a mtn. with a flattened top' (others say << *temH2-'dark', but the point remains) and others with unclear loss of *H in dialects (*o not certain) : *g(o)lH3-kiH2 ? > *gloH3khya > G. glôssa, Ion. glássa 'tongue', *gluxa > Alb. gjuhë If plókamos / plokhmós shows that o could cause *H > h / 0, it would be evidence of more irregularity. Maybe the same in mógos / mókhthos, depending on the timing of how some *KH and *HK changed to k/g/gh of all types (Whalen 2024e). Some might show analogy, but other words seem old, like G. alṓpēx matching Arm. ałuēs, *potH2-mo-> potamós (matching ON faðmr, OHG fadam, OE faeðm 'outstretched arms', E. fathom), or if aloi(m)mós were recent and analogical it would not show -m-vs. -mm-& would be likely to stll mean '*oiling / *anointing'. I don't include all proposed examples, but I think there are plenty of good cases in favor & against. It seems impossible to find total regularity here to me, but others disagree and pick and choose which examples are "real" to suit their purposes, based on no set method. I disagree with attempts by opponents of either to try to sweep contradictory examples under the rug, often by proposing very unlikely PIE roots without *H (with bad semantics) instead of with *H, loans, unlikely analogy, etc. Both sides sometimes give evidence that includes bad etymologies or impossible statements: Pronk, "balsas 'voice' is certainly not of Indo-European origin either". This is clearly ridiculous. For the group *bhalso-? > Skt. bhaṣá-s 'barking/baying', bhāṣa-'speech', Li. balsas 'voice', what possible reason would make one think, let alone pretend it was proven, that these were not IE? That the same happened in *golHso-> *golso-> Li. galsas 'sound/echo', R. gólos 'voice', (likely also >> Gmc. *kalz-'call / shout' > ON kalla, OE ceallian, E. call ) makes these graspings unlikely to succeed. It seems like *H > 0 / *h near o must be real due to its prevalence, but it does not seem regular. Similar cases in other IE exist, but with *H > 0 near o much more rare, and other *H > 0 with unknown causes. That this was different in each branch can be seen in it also affecting *h from *s in G. Evidence of *s > *h > 0 near *o in : *H1ois-m(n)-> G. oîma 'rush / stormy attack', Av. aēšma-'anger/rage' (vs. retained *sm > sm near *e, *tweismo-> G. seismós 'shaking') *kosmo-> OCS kosmŭ 'hair', OPo. kosm 'wisp of hair', G. kómē 'hair of the head' *k^ons-mo-? > G. kósmos 'order / government / mode / ornament / honor / world', kommóō 'embellish / adorn' (the relevance here depends on whether *-sm-or *-nsm-; since it is disputed, I include it to be safe) *sokWo-? > G. opós 'juice of plants', Alb. gjak 'blood', R. sok 'juice/sap', Lt. svakas *sorp-> OHG sarf 'sharp/rough', Lt. sirpis, G. hárpē 'sickle', (h)órpēx 'sapling/lance/goad' These might show the same alternation seen in other G. words from *H merging with h : *H2aps-> G. hápsos 'joint', TA āpsā '(minor) limbs', Skt. ápsas-'front side', H. happeššar-'limb / part of body' *H2aps-? > G. haphḗ '(sense of) touch / grip', Arm. *hap' \ ap' 'palm of hand / handful' (h-in *haph-haph-> hap'ap'em 'kidnap') *H2ar-mo-> G. harmós 'joint / bolt / door fastening', Arm. armanam '*be fixed in place > be stricken with amazement' *H1ek^wos > G. híppos, Ion. íkkos 'horse', L. equus With partial merger of *H / *h, it would also be hard to say if *gWorH3mo-? > bórmos \ brómos was due to a stage *gWroH3mo-(with kómē the same) or not. Disputes about the Saussure Effect's nature include timing (PIE or only Greek), exact environment, or whether it even existed in the first place. The Leiden School basically seeks to disprove it, and others to prove it. I won't get into every detail or dispute, but Yamazaki, for the opposite side, says that *polH1wya-> G. pollḗ 'many' exists. There is no evidence for *-o-in this root, especially for an adj., outside G-Arm., and most clearly require *-l-. Instead, I would say that the 2nd -l-resembles mega(lo)-, and is likely analogical, and that in *plH1u-> G. polús, Arm. yolov 'many (people)' syllabic *l often appears as al / ol in G. anyway, and syllabic *l > ol between *p_w resembles *wlkWo-s 'wolf' > L. lupus, G. lúkos, Alb. ulk (maybe exactly for *wolkWo-s, since there was also optional o > u by P / KW / w (*morm-'ant' > G. bórmāx / búrmāx / múrmāx; *wrombo-> rhómbos / rhúmbos 'spinning-wheel', *wodo:r 'water' > G. húdōr, *megWno-'naked' > Arm. merk, *mogWno-> *mugno-> G. gumnós). Other examples by Yamazaki seem to show *H > 0 in Baltic, but, again, not regularly. These are all due to tone, but this is not necessarily regular, even by *H. Most *HV do not show tone shifts, but *Ha also caused á in : *H2awso-m > L. aurum 'gold', Li. áuksas *H2awgs-to-> Li. áukštas 'high', L. augustus *kH2arwo-> OCS krava 'cow', R. koróva, Li. kárvė, OPr kurva-'ox', Alb. kau-'cow' (met. of *k^(e)rH2wo-'horned animal / stag' > L. cervus, W. carw, OPr sirwis) If one environment is not regular, why would *-oCH-need to be? The presence of Slavic -x-in *pelH-'grey', *plH-ko-'grey (thing) > mouse' *pĭlxkŭ > R. poloxók, Po. pilch 'dormouse' suggests that laryngeals were similar to x (maybe uvular). Maybe when *x remained, there was no tone shift, when *x became *h ( > glottal stop?), there was. Preservation of *Cxk would make sense, with KK acting different than other KC (if *HC = xC, etc.). If of PIE date, it would explain some cases of *o-H > *o across IE, but leave many more preserved *o-H unexplained. For ex., in apparent *domH2-s > *dom-s 'house / home' > Av. ušidå, *H2 is also gone in gen. *dem-s (*dems+pH2ter-, etc.). This could be analogy, but then we'd need stages of...

Pedersen’s law describes the development of the IE
*s after *i, u, r, k, causing the shift of IE *s into *š. The
process is well known from Indo-Iranian and Slavic.
The process also affects Lithuanian but not Latvian or
Old Prussian. The paper demonstrates that the Baltic
area was fully affected by the ruki-rule; the “defective”
results in the Baltic languages are results of secondary
depalatalizations. The paper also provides examples
of similar processes affecting “Pedersen’s” sibilant in
other satəm-languages.

Standing on the shoulders of many giants and some normal-sized scholars too, including but not limited to de Saussure, Noreen, Anttila, Ringe, etc., I argue that the Axe word is (pace Ringe) a perfectly formed IE compound, in fact, a bahuvrihi ending in a perfect active participle.  I also propose that the root *h2eg'- originally meant not 'drive' but 'poke/hurt with a sharp object' (which of course is how you drove both people and beasts, tho of course as we know only some animals lend themselves this).  Much as you cannot make a chimp or a racoon obey you this way, you cannot make me do it either with your slings and arrows or your walls of silence.  In next I hope to discuss my etymology of the HEX words as also such a partiicple, tho here is it less clear whether it is a compound or not (because the ha- syllable is hard to pin down), this etymology being one of de Saussure's best ideas--and of course generally ignored.  And yes I do think it hilarious (and sad too) that when I point this out some Indo-Europeanists who would be lucky to clean de Saussure's shoes tell me (and apparently they mean it) that well of course he was a great theoretician but not such a good etymologist.  Is there no limit to the academic mafia's hubris?

This paper surveys the previous interpretations of NPhryg. (-)τετικμενος and argues in favor of a translation 'condemned' and derivation from PIE *dei̯ḱ- (Ved. diś-, Gk. δείκνυμι, Lat. dīcere), against the equation with Old Irish tongid, ·toing ‘swears’. The proposed root connection finds support in the new etymological solution offered for Hitt. tekri- 'derogation, condemnation' < *do/ei̯ḱ-ri-. The third piece of evidence for the root *dei̯ḱ- used to express the idea of censure comes from PGmc. *teihan 'to accuse'. The source of the semantics is sought in Indo-European poetic language; it is argued that the PIE root *dei̯ḱ- could be employed in verbal contexts related not just to praise (Nikolaev 2012), but also to blame.

The shared features of Greek and Albanian include *r > ar\ra \ ri\ri (G. *k^rnaH2- > G. kírnēmi ‘mix (liquids)’, *plnaH2- ‘approach’ > pílnamai), KW palatalizing before front V’s but not plain K, etc.  However, these are not distributed in a way that fits a branching structure.  G. *r > ir is very rare, *r > ar \ or is very common; since Phrygian has most *r > ar but also Ph. mrotiē ‘to the dead?’, it seems Greek being directly between them is needed.  Other IE, like Celtic, more closely match Albanian *r > ar\ra \ ri\ri but for other reasons would not be near at all if tree theory were true.  A branch with Arm. and Ph. in one group, Alb. and Greek in the other, would not account for KW vs. K in the other, nor why it seems to be so irregular:

*gWhen- > G. eu-thenḗs ‘in abundance / wealth’, Arm. yo-gn / yo-k’n ‘numerous / abundant’

*gWhen-ye- ‘drive away, beat (off), hit’ > Greek theínō, Alb. gjanj ‘hunt’, Arm. ǰnǰem ‘destroy’

*ek^s-gWhen-yo-meno- > Ph. e[]gentoumenos ‘cast out’

*penkWe > Ph. pinke, Arm. hing ‘5’

*kWetwores > Arm. č’ork’ ‘4’

*gem- ‘hold (together) / seize’ > Arm. čim \ čem ‘bridle’


Seeing palatalization as an areal change, as if all others are not, is not useful as a classification scheme when so many changes in this group are the same (some only in Ph. and Arm., others only in Ph. and G., etc.).  Also irregular is K^ > K in Phrygian.  Fights over whether K^ was palatalized in Ph. miss the point; both outcomes exist, showing irregularity (like most changes described here).  Both of these are equally clear, and even found in the same roots.  PIE *dhg^hmVlo- > Greek khthamalós ‘on the ground / low’, Phrygian *ðγ^ǝmǝlo- > *zγ^ǝmǝlo- > *zj^ǝmǝlo- > szomolo- \ zomolo-  \ zemelo- ‘man (mortal) / *lowly > slave’, *dhg^homyo- > Greek khthónios ‘under the earth’, likely *upo- > Phrygian pokgonio- ‘the dead?’.  This has many implications for IE and classifying Ph. (Arm. has similar alternation of K^ and K before front (even *d \ *dz ), so the lack of regularity in close relatives can’t be waved away).  Since Arm. does NOT show any retained K^, with a larger number of ex. than Ph., a branching structure would require Arm. to split first.  Since the related changes to K(W) before front V are the opposite, more similar for Ph. and Arm., a number of changes with separate areas of complete and partial adoptation in a dialect continuum seems best for all.

The paper discusses the New Phrygian aorist form <ε>δικες (or αδικες < *adedikest) and argues against its identification with PIE *deiḱ-, proposed by V. Orel. Since Phrygian δ requires PIE *dʰ and Phrygian κ appears to go back to PIE *g / *ǵ , a comparison with Greek θιγγάνω, aor. θιγέ/ό- (along with the Delphic hapax θιγανα on the Cippus of Labyadai), MHG tîchen, and perhaps OIr. ·ding is proposed instead (PIE *dʰeiǵ- 'drive, handle, push'). The Greek verb has sometimes been compared to Ved. dih-, Lat. fingere, Toch. AB tsikā-, Gk. τεῖχος and derived from PIE *dʰeiǵʰ- 'build up, work clay, fashion', but apart from the semantic difference, "Miller's Law" (deaspiration after a nasal) will not account for the -γ- in the Greek root, as the paper argues. At the same time, the possibility of deriving NPhr. <ε>δικες from PIE *dʰiǵʰ-s- via deaspiration in *-Dʰs- cluster (argued to have been a PIE rule predating Bartholomae's Law) is tentatively raised.

Given the overall number of lines in the Homeric poems, the amount of verses that are considered unmetrical is very low. In most cases, the unmetricality is easily resolvable by considering a metrical license or by undoing the workings of language change. The verse that contains ἀνδρειφόντης, however, is arguably more unmetrical than any other Iliadic verse, and disentangling its hopeless scansion has challenged generations of scholars. It appears four times in almost exactly the same shape in the Iliad, which makes it a formulaic verse κατ’ ἐξοχήν, rendering its metrical obscenity only more problematic.

Greek adjectives in -ης (-ᾱς) such as ὑβριστής ‘violent, wanton’ are generally considered a secondary type, originating in an adjectivization of masculine substantives that became predominantly used in apposition. While this is certainly the preferred analysis for a former agent noun such as ὑβριστής (: ὑβρίζω ‘wax wanton, run riot’), there is a second type of adjectives in -ης that behave (in meaning and function) just like the the- matic adjectives they are seemingly derived from. Compare αἰχμητής ‘having a spear, spearlike, warlike’ (: *αἰχμητός ‘id.’), ἐτησίαι ἄνεμοι ‘the Etesian winds’ (: ἐτήσιος ‘yearly, annual’), ἀργεστής ‘bright’ (: *ἀργεστός ‘id.’), all of which are traditionally interpreted as substantivizations of the underlying adjectives. After introducing nine features to help determine whether a given noun can indeed be considered adjectival, this paper discusses the second type of adjectives in -ης (-ᾱς) and proposes an analysis as former “weak adjectives.”

When we think of the feminine gender in Indo-European and about how it manifests in the languages, we are faced with a paradox. On the one hand, feminine adjectival agreement with the ending *-eh₂- for thematic adjectives can be reconstructed quite safely for (at least) post-Anatolian Indo-European based on the ubiquitous pattern Gk. καλὴ κόρη, Ved. priyā́ kanyā̀ , Lat. puella pulchra ‘pretty girl’, etc. On the other hand, however, Ancient Greek exhibits thematic adjectives “of two endings” that have the masculine set of endings throughout the paradigm when agreeing with both masculine and feminine substantives (e.g. nom. sg. ἡ νῆττα ἥμερος ‘the tame duck’), which cannot easily be explained as secondary. This suggests that the agreement pattern with *-eh₂- for the feminine adjective is a rather recent development that had not been fully generalized and grammaticalized when the languages split up. This paper sets out to investigate the origin of these agreement forms in *-eh₂-, arguing for a revised version of the recently endorsed “weak adjective hypothesis”.

The meaning of Lith. mãžas ‘small, little’ and Latv. mazs ‘id.’ (compare also OPruss. massais ‘less’) has long puzzled scholars of Baltic and Indo-European linguistics. Etymologically speaking, there can be no doubt that Lith. mãžas and Latv. mazs belong to the root PIE *√meǵ ‘big, large’. But how are we supposed to understand the diametrically opposed meaning?

Das zu rezensierende Buch stellt den zweiten Teil einer bislang noch azephal gebliebenen Urindogermanischen Grammatik dar, die sich laut ihrem Verfasser Thomas Lindner (fortan L.) als "deduktives Begleitwerk" (S. 1 VI) zu der von L. herausgegebenen "induktiven Indogermanischen Grammatik" (ibid.) versteht. Bei dem vorliegenden Band handelt es sich abgesehen von einem einseitigen Vorwort ("Zu m G e l ei t "; S. V), einem ebenso einseitigen Inhaltsverzeichnis ("INHALT"; S. VI), sowie 22 Seiten Anmerkungen ("ANMERKUNGEN"; S. 349-370) und einer Drittelseite Literaturhinweise ("Siglen und Literaturhinweise"; S. 370) 2 ausschließlich (i.e. 348 Seiten lang) um rekonstruierte Flexionsparadigmen, wobei der Abschnitt zur Deklination (nominale Endungssätze, Paradigmen von Nomen, Numerale, Pronomen) mit 162 Seiten (S. 1-162) geringfügig kürzer ist als derjenige zur Konjugation (verbale Stammbildung, verbale Endungssätze, Paradigmen des Verbums, Verbum infinitum) mit 186 Seiten (S. 163-348). Die Aufgabe des Rezensenten stellt sich daher vergleichsweise einfach, in anderer Hinsicht aber auch einigermaßen schwierig dar. Die abertausenden Wortformen auf den 348 Seiten Flexionsparadigmen sind ja weitgehend rezensierungsresistent, und zwar nicht etwa, weil sie unzweifelhaft die tatsächliche urindogermanische Flexionsmorphologie abbilden würden, sondern einfach aus demjenigen Grund, dass sich über sie nicht trefflich streiten lässt. Rekonstruktion ist immer auch Interpretation. L. hat jede Form mit Gewissenhaftigkeit und gestützt auf eine valide Material-und Hypothesenbasis rekonstruiert. Doch auch wenn er damit nahe an einer zumindest im deutschsprachigen Raum verbreiteten communis opinio bleibt, sind die Flexionsparadigmen schlussendlich von L. selbst ersonnen, die Beispielwörter von L. selbst ausgewählt. Programmatisch sind in dieser Hinsicht also bereits die ersten drei Worte des Vorwortes-"Meine Urindogermanische Grammatik" (S. V); sie sind durchaus wörtlich zu verstehen. Dieses Umstandes ist sich L. freilich bewusst; im durchaus lesenswerten Zum Geleit (S. V) rechtfertigt er sein Buch als "Pionierwerk", in dem "auch Idiosynkrasien oder gar Irrtümer vorkommen [dürfen]". Auch anderer, an einem derartigen Werk berechtigter Kritik kommt L. im Zum Geleit zuvor. Man wird ja durchaus die Beharrlichkeit beneiden, mit der L. über hunderte von Seiten hinweg Formen wie den Vokativ Dual des 'Leber'-Wortes "**hxi̯ ḗk u̯-r-(i)h1 **hxi̯ ók u̯-r-(i)h1 [N[eben]F[ormen]. **hxi̯ ék u̯-n-ih1, *hxi̯ ek u̯-1 Die Seitenangabe (S. ...) bezieht sich in weiterer Folge auf die jeweilige(n) Seitenzahl(en) im zu rezensierenden Buch, wobei auf den von L. verwendeten nachgestellten Asterisk verzichtet wird. 2 Zitiert werden nur L.'s eigene Werke; siehe auch weiter unten.

The huge mythical snake defeated by the Storm God in IIr. was called *Vṛtrá- > Sanskrit Vṛtrá-, Avestan Vǝrǝθra-.  This word has been compared to Skt. vṛtrá- ‘defense / resistance / enemy?’, Av. vǝrǝθra- ‘attack? / victory?’ from *(H)wer- ‘defend / cover’ (possibly 2 separate roots).  For this, either ‘covering / concealing (the world or waters)’ or ‘obstruction’.  Cognates include a wide range:

Skt. vṛti- ‘enclosure’, OE worþ ‘piece of land / farm’

Skt. varaṇa-m ‘surrounding / enclosing / keeping off / prohibiting’, *varaṇā > Kh. waraáng ‘protective goatskin cloak’

G. érumai ‘defend / protect’

*w(e)r(e)tro- > Os. vart ‘shield’, Skt. vártra-m ‘dike / dam’, varatrā́- ‘strap’, várdhra-s ‘strap / girdle / belt’


I see no evidence that any of these are necessarily related.  Instead, there is evidence that it came from a word for ‘snake’ from a root meaning ‘turn / twist / wind’, etc. (PIE could be *wl-tlo-, *wr(t)-tro-, etc.).  There are several pieces of evidence that show oddities not compatible with origin from *(H)wer- ‘defend / cover’.


1.  Skt. AV vṛ́nta-s ‘caterpillar?’ < *wrt-no-.  This shows that such roots formed words in IIr. for ‘worm / snake’, with typical range of meanings in IE.  It is also similar to *wrton- > Arm. ordn ‘worm’ (most IE show *wrmi- > OE wyrm, E. worm, L. vermis or *kw-? > *kWrmi- > Skt. kṛmi-, Av. kǝrǝmi-).


2.  The names/words Skt. vṛtra-hán- ‘killing (a) snake(s) / Indra’, Av. Vǝrǝθraγna- ‘name of a god (probably the same as Tishtrya, who defeated Apaosha)’ seem to be related to Av. vārǝγna- ‘(representation of royal glory as) falcon/eagle’ (Vǝrǝθraγna- took the form of a bird of prey, among other animals).  IE names of raptors as ‘_-killer’ seen in Skt. śaśa-ghnī- ‘hawk-eagle’, G. kasandḗrion ‘kite’ (both < *killing/hunting rabbits), lagō-phónos, nētto-tkónos, passo-phóntēs, perdio-thḗrās, etc. (Witczak).  If vārǝγna- had a variant *vārǝǰan- like vṛtra-hán-, the loans Ks. váraš, Kh. yúruž / yùrǰ ‘falcon’ would confirm it (with optional *va / *vü; also similar *wi > yu in Skt. vīdhrá-, A. bíidri ‘clear sky’, Kh. yùdur ).  Since the mythical aspect of IIr. hawks includes killing snakes, and there is no other animal *va:r(a)- to be the victim, its origin seems clear.  There is no known way for *varta- to become *va:ra-, but the optional changes caused by IIr. *l vs. *r are well known, so *valtla-ghna- might work.  Also see below for other *l causing this change and others.


3.  Vṛtrá- is also called Valá-.  For the word Skt. Valá- ‘stone cave split by Indra to free Dawn and cows / Vṛtrá-’, there is no reason for these to need to have the same source.  Attempts to see Valá- as the brother of Vṛtrá- (with his exact characteristics and fate, as far as is known) seem pointless.  They would only make sense if it was impossible for one figure to have 2 names, which is obviously not true.  Both Valás being the same, one a personification (of a cave) is also unneeded.  Since both began with V- and were smashed by Indra it would be possible for confusion to apply one name to both, but if the optional changes caused by *l in *valtla-ghna- > vārǝγna- also applied to *vḷtla-, then dissimilation might create *Vṛtlá- > Vṛtrá- vs. *Vṛtlá- > *Vṛlá- > Valá-.  For the same type, see:

*kurt- > G. kurtía ‘wickerwork shield’, kúrtē ‘fish-basket’, Skt. kuti(:)- ‘hut’; *kuta > *kota > F. kota, Sm. goatte ‘tent / Lapp hut’

*kurtlo- > Skt. kúla-m ‘crowd/family’; *külä > F. kylä ‘village/dwelling’, Mi. kül ‘house’

*kurtlo-o:wyo- > *kula-āwya- ‘home of egg’ > Skt. kulā́ya- ‘nest’, Iran. *kulāwa- > Kurdish kulāw

*sweltlo- > *xvarθ(r)a- > Av. xVarǝθa- ‘food’, *swillo- > *sillë / *sullë ‘food’ > Alb. sillë ‘food brought out in the morning to a farmer working in the field / late breakfast’

*gWaHu-swl-tlo > *gaušt(r)a: ‘meat/flesh’ > NP gōšt, Ps. ǧvax̌a

and there does not seem to be total regularity in other clusters with l :

Skt. huḍa- ‘ram’, Dk. hʌldin ‘male goat’, ON *galtuz > göltr ‘boar’

*meldno-? ‘slow’ > Skt. manda-, Kh. malála ‘late’, Ku. mǝlaŋ ‘slowly’, R. medlennyj

*melno- > MIr. mell, Gae. meal ‘hill’, Irish meal ‘sphere/lump/knob/knoll/heap’, Skt. máṇḍa- ‘circle/circular/round’


4.  Though these words might have disputed origin, and thus not show various outcomes of *(a)ltl, the same can not be said for Iran. ‘snake’ (Manaster Ramer), which are supposedly *martra- > *marθra- > *mara / *ma:ra / *maθra / *ma:θra > mar / mār / mahr / māhr (NP mâr, Mz. mahr, Yushij māhr, etc.).  It seems very unlikely that both *martra-, that definitely meant ‘snake’, and *v(a)rtra-, which COULD have meant ‘snake’ would show the same alternations without being related.  If *w(e)ltlo- ‘snake’ existed, and IIr. alternation of w / m (Whalen) applied, since it is fairly common (Skt. -mant- / -vant-), both words would have the same origin and same reason for alternating.  This would allow *w(e)l-tlo- to be the source of all.  Though no prediction of what *ltl would become is possible in this theory, that both groups showed the same outcomes, with no other possible regular cause, makes it very likely.


5.  The alternative, that *martra- meant ‘killing’ or ‘deadly’ < PIE *mer- has no specific evidence.  An odd parallel would be G. máragna ‘lash / scourge’, Syriac māragnā supposedly being loans from an Old Persian compound formed from ‘snake-killing’.  If true, it would exactly match the changes in Vǝrǝθraγna- ~ vārǝγna- as ‘snake-killing’.  However, this seems like a meaningless description to apply to a lash, and it is much more likely that it is related to:

Skt. mṛgáyati ‘hunt/chase’, mṛgayás- / mṛgayā- ‘wild animal’, mṛgá-s ‘game / stag/gazelle / large bird / beast/demon’

*mṛga-ghna- ‘driving beasts’ (a whip/goad used to make domestic animals move) > Iran. *mǝrǝγa-γna- > *mǝrǝa-γna- (dissim.)

The exact sequence caused by g-g dissim. would be uncertain within Iran., especially since it’s only known from loans that can’t show all the specifics.  PIE *gWhen- meant both ‘drive’ and ‘hit / kill’, so there is no reason to think that a tool not used for killing (usually) had only this meaning.

Proto-Indo-European *paH2uson- is seen by Prósper as a compound of *paH2- ‘feed / protect’ (often used with animal names to form jobs in -herd), from his functions protecting and strengthening cattle.  Thus, originally Cowherd and/or Sheperd God.  In my mind, this makes the most sense if *paH2uson- came from *paH2-ukson- ‘protecting cattle’ (*ukso:n > E. ox, Av. uxšan- ‘bull’,  PT *wïksö:n > *(w)okso: > TB okso, TA opäs).  This could be due to dissimilation of back fricatives, if H2 = χ and k became x before fricatives:

*paχ-ukson- = *paH2-ukson-
*paχ-uxson-
*paχ-uson- dissimilation of back fricatives

This name had variants *paH2uson- / *pH2uson- / *puH2son- (possibly ablaut) and is the source of several IE gods:


*paH2uson- > *pauho:n > Greek Pā́n / Pā́ōn ‘Pan’

Pan is the son of Hermes, but essentially identical to him.  Both were likely originally names for the same god (wikipedia).  His particular aspects are likely old, from PIE.  Since he is part goat, with goat horns on his head, he bears a similarity to aspects of Thor and Pushan (Whalen 2023a).  Hermes’ characteristics are often shared by Indo-European counterparts (including Loki and the Divine Twins), but not his name (suggesting that it is a title based on his connection to the Herms and fertility in general, which would be nearly certain if first Pan under another name).  Many gods are connected to fertility, but almost every feature of his is also that of another god (cloak of invisibility and leading the dead suggest Hades, youth and cunning like Apollo, flying/swift like the Dioscuri, patron of travelers like Zeus is of guests showing respect for rules of society/strangers, etc.).


*puH2son- > Skt. Pūṣáṇ-, Scythian Pountas

In the Rigveda the god Pūṣáṇ- (Pushan) has a chariot pulled by goats and his name is also from a Proto-Indo-European word associated with gods and goats (Greek Pā́ōn / Pā́n).  Pushan’s chariot carries the Sun in its daily journey across the sky.  He also leads souls to Heaven (wikipedia).  These features are similar to those of the Aśvins (chariot which carries the Sun, carrying people away to save them) and Thor (chariot pulled by goats).  Since Pushan has a golden axe, which might correspond to Thor’s hammer (both representing (bright/golden) lightning, Mjöllnir is cognate with PSlavic *mulnijo- ‘lightning’).

This is as much certainty as anything that can be known considering how little is directly said of Pushan in the oldest sources.  He has a braided hair and beard, and his connection to gold and the sun might suggest, like the Sun and Agni (fire), he had a golden beard.  In many myths, the person pulling the object that was the sun was the Sun-God himself (Helios Hyperion).  Of course, there is also evidence that one of the Aśvins was the Sun, the other the Moon (Whalen 2023b).

Though Thor is now depicted with a horned helmet at times, it is uncertain if older versions had this.  Thor might have had horns, as shown by the use of ceremonial horned helmets in Norse rituals (compare maks of gods used in rites, or the Hand of Sabazius), but no direct evidence exists.  One myth of Greek Dionysus (conceived by Persephone and Zeus in the form of serpents) was identified with the god Zagreus as “horned child”.

The Scythian god Pountas must be related to Pūṣáṇ- (if its source is not wholly unknown) and together these might show *pusan- / *pusn- > *puθn- > *punt- .  It seems likely that *ns > nt is seen in other Iranian (Massagetic *manah-, *mans- >> dat. Khorsománōi vs. nom. Khorsámantis).  The opposite likely in Scythian Argímpasa (a goddess equated with Aphrodite):  *arti-patni: > *arḍi-paθna: > *aRgi-pasna > *argi-pasa > Argímpasa (like the Bactrian Ardokϸo, a goddess equivalent to Fortuna < *arti-xši-, cognate with Av. arti- \ aṣ̌i- ‘reward?’ and *xšay- / *xša:- ‘rule’ (Middle Persian pādi-xšāy ‘rule(r)’), making her ‘lady of fortune’.  Since some Iran. changed some s > θ (Skt. sraktí- ‘prong/spike/point’, Av. sraxti- \ θraxti-) an intermediate *s > *ts > *θcan explain all data by sound changes (maybe regular outside of Av., where sraxti- \ θraxti- makes optionallity necessary).


*paH2uson- > *paH2usno- > *pausno- > *favsno- > Latin Faunus

The changes to the word are not fully regular.  No explanation for such alternation as p / f, au / ū, exists, but many Italic words seem to show oddities in words with P-P (Whalen 2024a).  This provides some evidence of *w > *v being old, most easily seen in Greek.  Not being able to predict such changes is less than linguists desire, but it seems like sorting and analyzing such changes is sometimes all that is possible.

Since most knowledge of Roman gods comes from a period after heavy contact with Greece led to a mixture of their mythology, not all later descriptions of him seem to be wholly native.  Descriptions of Faunus as the “god of the forest, plains and fields… herdsmen” are likely true, but without perfect evidence, so be careful in analyzing him.


*paH2uson- > Pauso ?

This Messapic name resembles *Paus- in others, but claims like:

“Mallory and Adams saw a possible connection with… Pan… and Pauso (Messapic)” (wikipedia)

“The Vedic and Greek names are a match of, respectively, the Venetic [or Gaulish] PN Puso and Messapic Pauso (name of a mythical [maybe] king)” (Prósper 2016:  45)

have no evidence behind them.  To me, Messapic clearly shows *-s- > *-h- > 0 (among many other examples of *s > h, like *dH1so- > *theho- > G. theóphoros ‘possessed by a god / (divinely) inspired’, Ms. t(h)abara- ‘priest’).  Other h : h (G. Hippikós, Híppakos, Ms. Hipaka / Ipaka / Hopaka) make it clear that Ms. names (and likely all words) are derived from Greek  (Whalen 2024b).  In this case, many G. names in Paus- exist, making any other origin unlikely; when every name has a Greek match, looking for a non-Greek match as the first attempt makes no sense:

G. paúō ‘make cease / check / stop (from _ing) / hinder’, Pausanías ‘allayer of sorrow’, Paúsōn : Ms. Pauso