• Welcome to the new Internet Infidels Discussion Board, formerly Talk Freethought.

Language as a Clue to Prehistory

Another example of the "thing" part surviving of "nothing": French rien "nothing", from Latin rem, accusative of rês "thing". Romance nouns and adjectives are usually derived from the Latin accusative forms and not the nominative ones, despite the nominative case being the usual case for reference forms.


Looking at Slavic aspects, most perfectives are formed with preposition prefixes, but there is one perfective suffix that I've found: -nu-. Imperfectives are all suffixes, like -ova-/-ava-.


Proto-Indo-European had numerous derivational affixes, and not just prepositions as verb prefixes. The Slavic languages are far from alone alone in having prefixed verbs of motion -- Latin and Greek have plenty of such verbs, and some of this prefixing ended up as borrowed words in English, like "exit" from ex "out of, from" + îre "to go". English has enough of Latinate words with ex- for its speakers to use that prefix more generally, like ex-(something) being a former something, like ex-wife.

English adjective-from-noun derivational suffix -ish has Germanic cognates -isk, -isch, from Proto-Germanic *-iskaz, in turn from PIE *-iskos

It appears as Greek -iskos - a diminutive suffix - and in Balto-Slavic as Lithuanian -ishkas and Slavic -ski/-sky, a common suffix for forming adjectives from nouns. Like Sovetsky Soyuz "Soviet Union".

Another one is -en, "made of", from PGmc *-înaz and cognate with Latin -înus, Greek -inos, and Slavic -in,-ny, from PIE *-iHnos

Another one is -y, with Germanic cognates -ig, -ich, from PGmc *-gaz, with cognates Latin -cus, -icus, Greek -kos, -ikos, Sanskrit -ka, from PIE *-kos

From Latin -ius and Greek -ios we find PIE *-yós with various other descendants.

Suffixes like English -ness and Latin -tâs had a rather complicated history, with such ancestors as PIE *-tus action noun and -*-
*-teh2 state of being, from *-tós a verbal-adjective suffix

English suffix -ize is from Latin -iz-, and in turn from Greek -iz-.

Suffix -ist is from Latin -ista, in turn from Greek -istês in turn from -iz-tês, an agent-noun suffix, likely from PIE *-teh2.

Suffix -ism is from Latin -ismus, in turn from Greek -ismos in turn from -iz-mos, from an action or result noun suffix, form PIE *-mos.

I think I'll leave off.

It's remarkable how much detail one can find out about long-lost languages like Proto-Indo-European, even if most of what one finds is not very culturally informative.
 
The cycle goes:
  • Isolating to agglutinative: running words together
  • Agglutinative to fusional: running word parts together
  • Fusional to isolating: erosion of words, like final sounds dropping out
Is it much much too simplistic to say that each of these three tendencies is a tendency toward efficiency?
* Providing affixes to mark plurality, past-tense, etc. streamlines a language and reduces ambiguity.
* Combining several affixes together reduces the syllable count.
* Replacing inflected forms with a single root word makes the language easier to learn.
But each move toward efficiency introduces a new problem, solved by the next step in the cycle.
That's a good point, and that explains Otto Jespersen's negation cycle. Bertrand Russell's metaphysician is an extreme rarity. Everybody wants to express negation and do so often. But as people do so, they wear down the negation markers, and they eventually invent extensions. But those also eventually get worn down. no + thing -> nothing -> no. Sometimes it's the extension that becomes the new negation marker, the thing part.
I think it takes at least THREE modes to get this sort of oscillation; TWO modes isn't enough.
Actually, two unstable states are all that is necessary - a system will then go back and forth between those states. Think of a swing as it swings back and forth. Both the forward and the rearward positions are unstable, and when the swing stops at one of those, it reverses direction and goes to the other one.
 
LINGUIST List 13.588: Borrowing of Verbs Versus Nouns

That entry's formatting is not very good; its text is run together. Someone asked
A standard textbook on historical linguistics, Hock's 1991 Principles of Historical Linguistics, 2nd ed., p. 386, says: "...it has been noted that verbs are crosslinguistically less easily borrowed than nouns..." Is this a generally accepted claim?
It apparently got a lot of responses saying that yes, indeed, nouns are more often borrowed than verbs in a variety of languages.

[PDF] Borrowability and the Notion of Basic Vocabulary
This paper reports on a collaborative quantitative study of loanwords in 41 languages, aimed at identifying meanings and groups of meanings that are borrowing-resistant. We find that nouns are more borrowable than adjectives or verbs, that content words are more borrowable than function words, and that different semantic fields also show different proportions of loanwords. Several issues arise when one tries to establish a list of the most borrowing-resistant meanings: Our data include degrees of likelihood of borrowing, not all meanings have counterparts in all languages, many words are compounds or derivatives and hence almost by definition non-loanwords. We also have data on the age of words. There are thus multiple factors that play a role, and we propose a way of combining the factors to yield a new 100-item list of basic vocabulary, called the Leipzig-Jakarta list.
Of content words, sense perception, spatial relations, the body, and kinship have the least likelihood of being borrowed.

The list:  Leipzig–Jakarta list
100 concepts that were found in most languages and were most resistant to borrowing formed the Leipzig–Jakarta list. Only 62 items on the Leipzig–Jakarta list and on the 100-word Swadesh list overlap, hence a 38% difference between the two lists.

25% of the words in the Leipzig–Jakarta list are body parts: mouth, eye, leg/foot, navel, liver, knee, etc. Six animal words appear on the list: fish, bird, dog, louse, ant and fly – animals found everywhere humans can be found.

The items house, name, rope and to tie are products of human culture, but are probably found in all present-day human societies. Haspelmath and Tadmor drew the conclusion that "rope is the most basic of human tools and tying is the most basic technology".
 
I will look at a very common - and very irregular - conjugation: the present tense of "to be". I'd done that for the Western Romance languages and Latin, now I will compare across IE.

Language1s2s3s1p2p3p
English (to be)amare, (art)isareareare
Old English (wesan, bêon)eom, bêoeart, bistis, bithsind, bêothsind, bêothsind, bêoth
Proto-Germanic (wesanan, beunan)immi, biumiisi, biusiizi, biuthiizum, beumizud, beudsindi, biunthi
Croatian (biti)jesam, samjesi, sijeste, jejesmo, smojeste, stejesu, su
Czech (byt)jsemjsi, -sje, jestjsmejstejsou
Russian (byt')jesm'jesijest'jesmyjestesut'
Old Church Slavonic (byti)esmiesiestuesmuestesontu
Proto-Slavicesmiesiestiesmuestesonti
Lithuanianesuesiyraesameesateyra
Latin (esse)sumesestsumusestissunt
Greek (einai)eimieiestiesmenesteeisi
Sanskritasmiasiastismahsthasanti
Proto-Indo-Europeanh1ésmih1ésih1éstih1smósh1stéhsénti

How was the h1 pronounced? It's a part of  Laryngeal theory a rather contentious aspect of Proto-Indo-European reconstructions.
  • h1 - neutral - likely a glottal stop alternating with a @ (schwa) when a syllable
  • h2 - a-coloring - a velar / uvular / pharyngeal fricative - like "h" but with the back of the mouth or the throat constricted
  • h3 - o-coloring - like h2 but with rounded lips: h2w
So *h1ésti was likely pronounced ésti.

While the present tense points to root *h1es-, such past tenses as Latin fui and Lithuanian buvau "I was / I have been" point to a different root, *bheuH- This is also the source of English "be, been" and its Germanic cognates. English past tense "was, were" and its Germanic cognates have an additional source: PIE *h2wes- "to stay, reside".

So we reconstruct for Proto-Indo-European "to be": imperfective *h1es- and perfective *bheuH-

Germanic adds *h2wes- and Romance *steh2- through Latin stâre "to stand"

The Latin form has a suffix that is from ()-éhye- a thematic version of an originally athematic suffix for making stative verbs: ()-eh1- with (é)-ye- The first one makes stative verbs from perfective ones or from nominals (nouns+adjectives), the second one makes transitive imperfective verbs.

The () is for the ablaut of the root vowel.
 
Slavic aspects:

Russian Verb Aspect Made Simple – StoryLearning

Says that the most common Russian perfective prefix is po- a prefix that makes verbs of motion perfective. The prefix za- makes a verb not only perfective but also indicating the start of something (inchoative).

Also mentions some suppletive aspects, like govorit' (impf), skazat' (pf) "to speak, say, tell" and brat' (impf), vzjat' (pf) "to take"

Microsoft Word - ASPECT PAIRS- 31.7.11.doc - aspektpaare.pdf for Czech: mentions brát (impf), vzit (pf) "to take"

Basic Croatian: 17 Aspect of Verbs - za- is also inchoative in Croatian.


Turning to PIE, I've found DAWN OF VERBAL SUPPLETION
IN INDO-EUROPEAN LANGUAGESf
- listing several suppletive-aspect verbs in the dialects. "To be" is only one of them.
 
Language Dispersal Beyond Farming - an open-access book as a PDF file

Farming/Language Dispersal - Food for thought - Martine Robbeets
Idiomatic English from a non-native speaker :D

Farming making language dispersal - a completely reasonable hypothesis, since farmers can easily outbreed foragers, something that there is some genetic evidence for in some places.

Mentions the "diversity hotspot" principle - "Assuming that the deepest splits within a family reflect the greatest age, the location of these splits on the map is thought to point to the area where the proto-language began to diversify. The principle is thus based on the assumption that the homeland is closest to where one finds the greatest diversity with regard to the deepest subgroups of the language family." - like the Austronesian family, which spread outward from Taiwan. However, "A second limitation of this principle is that the contemporary hotspot of lin- guistic diversity may diverge from the earlier one." (the first one is how reliably one can determine a language family's subgrouping structure)

"The prehistoric population movements out of Taiwan and through Island South-East Asia into the Pacific discussed by Gray et al. (2009), for instance, display pulses and pauses that closely match the stages of splits and spreads in the phylogenetic tree of Austronesian languages." - so there is a good match between population and language dispersal there.
 
Martine Robbeets on "The language of the Transeurasian farmers":
  1. Proto-Transeurasian, the language ancestral to the Turkic, Mongolic, Tungusic, Koreanic and Japonic languages, reflects a broad-spectrum subsistence strategy probably including some plant cultivation and yielding food surpluses.
  2. The assumed location and time depth of proto-Transeurasian associate the ancestral language with the Xinglongwa culture, the first farming society in Northeast China in the 7th and 6th millennium BC.
  3. The spread of the Transeurasian languages to their present-day locations is con- sistent with the spread of agriculture in Northeast Asia. However, agriculture did not necessarily cause language spread by boosting the farmer’s demography and pushing them to search for new land. It also followed ecological stress caused by climate change, disrupting traditional resource bases and replaced previous subsistence strategies.
Cultural reconstruction indicates that the speakers of proto-Transeurasian targeted a millet-like crop for its seeds, sowed seeds and maintained fields for cultivation. Their food surpluses were sufficient to permit labor-intensive and technologically complex activities such as weaving. They were familiar with a process of oxida- tion, probably in connection with iron-rich clay in hematite pottery production. In contrast to the communities in the Yellow River Basin, the speakers of proto-Tran- seurasian relied intensively on grinding for their food-production. The starches involved in this process were not limited to millets, but were provided by various nuts such as walnut, chestnut, acorn and pine as well as roots. The reconstructed vocabulary therefore suggests a broad-spectrum subsistence strategy with some economic dependence on the cultivation of plants such as millets.

The lexical evidence is in line with the diversity hot-spot principle, locating the homeland of Transeurasian in the West Liao River region and Bayesian inference, estimating the time-depth of the family at ca. 5700 BC. The location and time depth indicate that proto-Transeurasian may be connected with the Xinglongwa culture (6200–5400 BC) in Southern Manchuria. This culture depended on a broad-spec- trum subsistence strategy including millet cultivation.
 
"Farming-related terms in Proto-Turkic and Proto-Altaic" by Alexander Savelyev
Historical sources from different times describe Turkic, Mongolic and Tungusic traditional economies as based on pastoralism, with agriculture playing only a minor role among their subsistence strategies.

I show that the majority of the Turkic pastoralist lexicon has a secondary nature, being formed due to contact, derivation or lexical recycling. At the same time, farming-related terms in Turkic are mostly unborrowed and underived and a few of them have reliable Altaic connections. The very limited number of agricultural terms reconstructible to Proto-Altaic as compared to the preceding Proto-Transeurasian period can be attributed to a loss of farming-related lexicon over time after the break-up of Altaic.
Altaic = Turkic, Mongolic, Tungusic, as opposed to these families with Korean and Japanese (Transeurasian).

"Archaeological and historical sources from different times describe Turkic, Mongolic and Tungusic traditional economies as based on pastoralism, with agriculture playing only a minor role among their subsistence strategies."

Transeurasian - 5700 BCE - Altaic, J-K
Altaic - 4600 BCE - T-M, Tungusic
TM - 2800 BCE - Turkic, Mongolian


Farming and the Trans-New Guinea family - proposes ancestral words for sugarcane and banana. Finds that words for taro, a common root crop, are much more variable, implying that taro growing is much more recent.
 
Macrofamilies and agricultural lexicon - Problems and perspectives - George Starostin

It's generally agreed that the first Eurasian agriculture originated in the Middle East some 12 - 10 thousand years ago, meaning that these first farmers must have had words for their agriculture.
Unfortunately, if we consider all known linguistic families whose protolanguages satisfy the following three conditions:
  1. they have been reconstructed to more or less general satisfaction, so that their historical reality is not a point of contention between the majority of specialists;
  2. they are generally agreed to have contained at least a certain amount of semantically unambiguous agricultural terms;
  3. they are generally agreed to have been spoken either in the Levant area or in regions not too far removed from it;
– then none of these families, including such linguistic taxa as Indo-European, Semitic, Dravidian, Kartvelian, North Caucasian (Nakh-Daghestanian and Abkhaz- Adyghe), can be reasonably claimed to have had their ancestral languages spoken as early as the required date.
Unfortunately, all these protolanguages split up at most by 5 - 6 thousand years ago, far too recent for agriculture.

The languages of the first farmers might have disappeared long ago, with their only surviving evidence being contributed agricultural vocabulary.

So we must turn to less well-established language families, families that go much farther back in time, like Nostratic and Afro-Asiatic.
 
He first considers Nostratic, and he notes that there is not much agreement on what is a member, but he notes an agreed-on core membership of Indo-European, Uralic, and Altaic. Possibly members are Kartvelian and Dravidian, while Afro-Asiatic is likely separate.

Of these, IE, Kartvelian, and Dravidian have some agricultural vocabulary, Uralic has none, and Altaic has a little bit. Proto-Nostratic had no clearly reconstructible words for anything agricultural.

Turning to smaller-scale hypotheses, Indo-Uralic is supported by grammatical features and basic vocabulary, things that are likely to be inherited. Vocabulary like "water", "name", "to lead", "to hear", ... and IE agricultural terminology does not have any good cognates in Proto-Uralic, implying that the IE agricultural vocabulary is from semantic shifts and borrowings.

Altaic agricultural vocabulary seems very limited and restricted.
 
Then Sino-Caucasian: North Caucasian, Yeniseian, and Sino-Tibetan. Sometimes extended with Basque, Burushaski, and Na-Dene, giving Dene-Caucasian. Of these, the original homelands of Yeniseian, Central Siberia, and Na-Dene, Beringia, are very unsuitable for agriculture. But Proto-Sino-Tibetan has some agricultural vocabulary and Proto-North-Caucasian a lot of it. But the agricultural vocabulary of Sino-Caucasian is very weak and doubtful.

The most success is with Basque and North Caucasian (Euskaro-Caucasian). It's hard to find much, but some correspondences are phonetically and semantically very good. There is also a lot of non-Indo-European substrate vocabulary in the Indo-European languages of Europe, and some of it seems to be related to Basque and North Caucasian. This suggests that the expansion of farmers into Europe in the early Holocene was of speakers of some Euskaro-Caucasian language, with Indo-European a later overlay.

Then discussing Afro-Asiatic (or Afroasiatic).

The deepest split in it is between (Semitic, Egyptian, Berber, Chadic) and (Cushitic, Omotic). The first part is North African and Middle Eastern, and the second part East African; Somali is a Cushitic language.

There isn't much evidence of agriculture-related cognates across Afro-Asiatic, meaning that the Proto-Afro-Asiatic speakers were unfamiliar with it.  Afroasiatic homeland - mentions a proposed homeland in eastern Ethiopia about 10,000 years ago. That's far from the Middle East. An offshoot moved northward to Egypt, then an offshoot of that one to the Levant and became the speakers of Proto-Semitic. So what happened to Europe also happened to the Middle East - an original population of farmers became conquered by people from their peripheries.
As an interesting curio, it could be instructive to mention a recent study (Agmon & Bloch 2013) that used statistical methods to ascertain that various terms reflecting hunting and foraging activities in Semitic tend to be shorter, i.e. are more frequently represented by archaic biconsonantal roots than agricultural terms, which, conversely, tend to be almost always represented by longer, triconsonantal roots. If this study checks out through detailed etymological research, this could be a serious argument in favor of a relatively late origin of agricultural terminology for ancestors of Proto-Semitic. For now, we simply have to accept the fact that a lot of research on various subgroups of Afroasiatic is still necessary in order to properly resolve the issue – and that, for the moment, strong evidence for agriculture in Proto-Afroasiatic is non-existent.
So we conclude that early Semitic speakers started out with lots of two-consonant roots and then extended many of them to form three-consonant ones. Something like Proto-Indo-European and Proto-Slavic speakers inventing verb-aspect systems.
 
Unfortunately, all these protolanguages split up at most by 5 - 6 thousand years ago, far too recent for agriculture.

The languages of the first farmers might have disappeared long ago, with their only surviving evidence being contributed agricultural vocabulary.

So we must turn to less well-established language families, families that go much farther back in time, like Nostratic and Afro-Asiatic.

Indo-European is the only well-established language family as old as 5000-6000 years old. It has been pushed back so far only due to evidence far greater than for any other language family: a plethora of descendant languages, and written texts that are much more than 3000 years old.

In addition to some of the ancient languages lpetrich mentions, Basque, North Caucasian and Burushaski may have had a common ancestor. Here is an excerpt from
This excerpt shows words associated with farming. Other word lists in this or other Bengtson papers show cognates for domesticated animals.

• Basque *gari / *gal- ‘wheat’80 = Cauc: Tindi q’:eru, Lezgi q:ül ‘wheat’, etc.81
• Basque *(gara-)ga ‘barley’82 = Cauc: Rutul q’ir ‘winter wheat’, Agul q’ir ~ q’ur ‘grain’83 = Burushaski *gur ‘wheat’.
• Basque: *bihi ‘grain, seed, kernel’84 = Cauc: Godoberi beč’in ‘rye’, Tindi beč’in ‘barley’, etc.85
• Basque *sikirio ‘rye’86 = Cauc: Rutul sɨk’ɨl ‘rye’, Khinalug sɨlg-li ‘rye’, etc.87
• Basque *olho ‘oats’, *alho ‘wild oats’88 = Cauc: Kabardian xₙə ‘millet’ < PWC *λₙə id.89
• Basque *arto ‘maize’ (earlier ‘millet’) = Cauc: Avar ro: ‘wheat’, Agul jerg ‘oats’, etc.90
• Basque *ilha- ‘vetch, peas, beans’91 = Cauc: Tsez hil ‘pea(s)’, Avar holó ‘bean(s)’, etc.92

Most impressive, in my opinion, is a whole suite of Basque agricultural terms, involving soil tilling and preparation, harvesting, threshing, sifting, and grinding, that have close Caucasian and Burushaski counterparts:
• Basque *laia ‘two-pronged fork (used for loosening and turning soil)’ 93 = Cauc: Bezhta ƛaχ-dami ‘rake’, etc.94
• Basque *haincu ‘hoe, spade’95 = Cauc: Chechen ästa ‘hoe, mattock’, Akhwakh ʕerc:e ‘wooden plow’, etc.96 = Burushaski *harṣ ‘plow’
• Basque *arhe ‘harrow’97 = Cauc: Avar ʁár-ize ‘to harrow’, Lezgi ʁar ‘harrow’, etc.98
• Basque *laain ‘threshing floor’99 = Cauc: Archi ƛorom ‘threshing board’, Andi loli ‘threshing, threshing floor’, etc. 100 = Burushaski *daltán- ‘to thresh’ < *rVŁV-n-.
• Basque *bahe ‘sieve’101 = Cauc: Tsakhur wex:ʷa ‘sieve’, Lak =ihi- ‘to filter’, etc.102
• Basque *eiho ‘to grind’ / *eihera ‘mill’103 = Cauc: Chechen aħ- ‘grind’ / ħer ‘mill’, Ingush ħajra ‘mill’, Lak ha=a- ‘grind’ / hara-qalu ‘mill’, etc. 104 = Burushaski *-hor- ‘to grind’.

The linguistic evidence presented here indicates that the western Dene-Caucasian speakers of ca. 7500 years ago (linguistic ancestors of the present-day Basques, North Caucasians, and Burushos) had a well-developed Neolithic pastoral-agricultural culture, including the husbandry of large and small cattle and the cultivation and milling of cereal grains and some other crops such as pulses.
How do we know that the Basques did not simply adopt these Dene-Caucasian Neolithic terms as loanwords, while retaining the rest of their original language intact? In fact the Neolithic terms have the same phonology and morphology as the most basic parts of the Basque lexicon.
For example, in Basque *olho ‘oats’ = PNC *λwʔwV ‘millet’ we see the same correspondence of Basque aspirated lateral (*lh) to PNC lateral fricative (*λ) as in Basque *e-lhu- ‘snow’105 = PEC *jĭwλV / *λĭwV ‘snow’, and ‘snow’ can hardly be considered a cultural word that is easily borrowed.106 Likewise, the phonological relationship between Basque *behi ‘cow’ and Andi buc’:ir ‘cattle’ is parallel to that of Basque *minhi ‘tongue’ = Andi mic’:i ‘tongue’,107 one of the most basic words in any language. Morphologically, the relationship between Basque *eiho ‘to grind’ (verb) and *eihera ‘mill’ (noun) is the same as that between Ingush aħ- ‘to grind’ and ħajra ‘mill’. The Basque allomorphs seen in *ahari / *ahal- ‘ram’ and *gari / *gal- ‘wheat’ are entirely parallel to those of the basic *heugari / *heugal- ‘abundant,
copious’ / to increase, multiply’ (cf. Tsez =eχora ‘long’, Akushi χala-l ‘big’, etc. < PNC *HāχułV / *HālχV ‘long, big’), and so on. In other words, there is no linguistic reason to suppose that Basque words for domestic animals, cultivated plants, and food-processing belong to a different or later layer than the most basic words (e.g., words for ‘blood, bone, tongue, tooth, horn’, etc.)
discussed above (page 161).

It is, IMO, a misconception that Basque is the residue of an old hunter-gatherer language of West Europe. Farmers arrived in Spain shortly after 6000 BC and the pre-Neolithic languages were long gone by the time of Julius Caesar. Basque arrived from the Eastern Mediterranean, probably via the Impressed Ware/Cardial Ware expansion. This expansion was driven by sea-faring adventurers, rather than mainly cultural diffusion. One can see remnants of Y-haplogroup G-PF3177 in isolated places like Sardinia. (Note that this is a DIFFERENT clade from the large majority of G's in Western Europe today (G-PF3345), who arrived with the Alan tribes during the decline of the Roman Empire.)

This links the early Neolithic language of Western Europe — proto-Basque — to an early Neolithic language in the Fertile Crescent. (Later arrivals like Hittite, Iranian, Semitic drove the early farming language sibling to proto-Basque into the Caucusus mountains.)

Burushaski — an isolate found only near the mountainous Hindu Kush — is also linked to this Basque-Caucasian family! The language of the Harappan civilization is unknown, but Burushaski has been proposed as a possibility; migration from the Tigris River area seems plausible.
 
Yes indeed - Notes on some Pre-Greek words in relation to Euskaro-Caucasian (North Caucasian + Basque)

Working from Beekes - The Pre-Greek loans in Greek
Noting lots of phonological features that are atypical of Indo-European.
 Pre-Greek substrate - good place to start. Here's a fun one: words for noisemakers that end with -nx, -ngg-

salpinx: trumpet / syrinx: panpipes, flute / phorrminx: lyre / larynx: voice box, throat / pharynx: throat

Category:Greek terms derived from Pre-Greek - Wiktionary - Wiktionary is a remarkably comprehensive resource

-

 Germanic substrate hypothesis
Non-Indo-European root nouns in Germanic: evidence in support of the Agricultural Substrate Hypothesis - sust266_kroonen.pdf
Category: Proto-Germanic terms derived from substrate languages - Wiktionary
Germanic words of non-Indo-European origin - Linguistics - Eupedia

 Goidelic substrate hypothesis - for the Gaelic languages
The substratum in Insular Celtic


Category:Terms derived from substrate languages - Wiktionary - look under the Proto- ones
 
Last edited:
The pre-Indo-European substrate languages of Europe had a feature that shows up in a lot of substrate words: variations in initials, like a- being present only some of the time.

John Bengton and Camilla Leschber conclude that they are fossilized noun-class prefixes, and that Basque also has them.

The North Caucasian languages have various systems of noun-class prefixes, and the language of the first Anatolian and European farmers thus likely also had such prefixes.

 Northeast Caucasian languages and N-D Noun Class lectures - N-D Noun Class lectures.pdf
Noun classes have agreement not only with adjectives and verbs, but also with adverbs and various function words.

The number of noun classes varies from 2 to 8.
  • Tabasaran (northern): 2 (human vs. non-human)
  • Avar, Dargwa, and most Andic languages: 3 (male rational, female rational, non-human)
  • Lak, Tsez, Hinukh, the Lezgic languages with noun classes: 4
    • Lak: male rational; (mature) female rational; animate; inanimate
    • Archi: male rational; female rational; complex division for remaining nouns
    • Tsez: male rational; female rational + inanimates; animates & inanimates; inanimates
  • Chamalal, Hunzib, Khwarshi: 5
    • Hunzib: male rational; female rational; animates and inanimates spread across other three classes (Forker 2014)
  • Chechen, Ingush: (traditionally) 6
  • Batsbi (Tsova-Tush): (traditionally) 8
But three of the Batsbi classes, VI to VIII, have only 20 words among them. Table of Batsbi (Tsova-Tush) classes:
ClassSingularPlural
Iv-b-Mostly male: stak' 'man', dad 'father', mar 'husband', ...
IIy-d-Mostly female: nan 'mother,' pst'u 'wife', johh 'daughter', ...
IIId-d-Largest class: bader 'child', dok' 'heart', ...
IVy-y-2nd largest class: tsark' 'tooth', q'ar 'rain', ...
Vb-d-phhu 'dog', cha 'bear', matkh 'Sun', ...
VIb-y-bak 'fish', b\ark 'eye', kok' 'leg', ...
VIId-y-bat'r 'lip', lark' 'ear', t'ot' 'hand', ...
VIIIb-b-borag 'knit slipper', kakam 'wool cut in fall', ...
 
Last edited:
Usually fewer plural noun classes than singular ones. For instance:
  • Godobery: (male, female), (nonhuman)
  • Hinukh: (male, female), (animals, inanimate I, inanimate II)
  • Bezhta (Tliadal dialect): (male, female), (amimals+inanimate, inanimate I), (inanimate II)
Proto-Nakh-Daghestanian (Northeast Caucasian) (Johanna Nichols):
  • v/() - male human
  • j/r - female human
  • b - many animates
  • d/r - inanimates (chiefly)
  • j - various nonhuman

Tsez:
ClassSingularPluralWhat's in it
I()-b-Male: human, divine
IIy-r-Female: human, divine; some inanimates
IIIb-r-Animals, some inanimates
IVr-r-Other inanimates
So Tsez has only two plural classes: sentient male and everything else.

The paper then considered how nouns for inanimates are assigned to classes II, III, and IV. For instance, vehicles are often in class III, so that class may mean "mobile entities". But abstract nouns are in all three, with those ending in -tli and -ni in IV.
 
To add to my previous post, inanimate nouns beginning with y-, b-, and r- are in noun classes II, III, and IV.

Gender Affixation on Caucasian verbs - An_Overview_of_Gender_Agreement_Affixes_in_the_Caucasus.pdf
Genders in the Caucasus are usually marked on verbs and adjectives by affixation

Almost all the languages differentiate masculine and feminine (except e.g. Tabasaran (Babaliyeva 2013)), and they also make a disctinction between human and non-human. The core of both the NEC and NWC systems is therefore:

– Masculine (Class I, in the noun class tradition)
– Feminine (Class II)
– Non-human (Class III)
– + Various inanimate genders

The largest NEC systems have up to six genders, i.e. Chechen, Ingush and Andi (Nichols 1994; Nichols 2011; Salimov 2010).
Two NEC languages' gender systems:

Dargwa (Daghestanian)
ClassSingularPluralWhat's in it
Iw-b-Male human
IIr-b-Female human
IIIb-d-Nonhuman
Chechen (Nakh):
ClassSingularPluralWhat's in it
Iv-b-Male Human
IIy-b-Female Human
IIIb-b-Nonhuman I
IVb-d-Nonhuman II
Vd-d-Nonhuman III
VIy-y-Nonhuman IV
 
Turning to Northwest Caucasian languages, they have very similar gender systems. Here, the verb prefixes vary by person.
Class2nd3rd abs3rd ergWhat's in it
Iw-d-y-Male human singular
IIb-d-l-Female human singular
IIIw-y-a/n-Nonhuman singular
I, II, IIIshw-y-r/d-Plural
abs = absolutive case (subjects of intransitive verbs), erg = ergative case (subjects of transitive verbs)

The rest of the article was mainly about trying to find patterns, like b- being very common, y- and r- being common feminine markers, w/v- being common masculine markers, and d- being a common inanimate marker.
 
Shevoroshkin (ed.) - Proto-Languages and Proto-Cultures (1990)_text.pdf

ETRUSCAN AS AN EAST CAUCASIAN LANGUAGE - Vladimír Orel and Sergei Starostin

Building on such work as
I.M. Diakonoff and S.A. Starostin. Hurro-Urartian as an Eastern Caucasian Language. Miinchen, 1985

East Caucasian ~ Northeast Caucasian

Also mentions Hattic ~ Northwest Caucasian (Abkhaz–Adyghe)

Etr. esera moon (?)' - PEC 'šVIHV- moonlight, light (cf. Urart. sél-ard moon, the Moon-goddess).

Here's a plausible comparison:

σέλας - Wiktionary - selas "light, brightness, the bright flame or blaze of fire; shine (n.)"

σελήνη - Wiktionary - selênê "Moon" (Doric selânâ, Aeolic selânnâ).
 
Euskaro- Caucasian Hypothesis: Current model (2017): | John D Bengtson - Academia.edu - looks like a set of slides

John D. Bengtson & Corinna Leschber 2019. Notes on Euskaro-Caucasian (Vasconic) Substratum in western Indo-European Languages. WEKWOS 5, Revue d' études Indo-européennes. R. Garnier, X. Delamarre, Les Cent Chemins, Arles: Éditions Errance, 11-50. | Corinna Leschber - Academia.edu

Proposes several that Latin has several Euskaro-Caucasian substrate words like baculum "stick, staff", bacillum "short stick" (Basque *mak-, Proto-North-Caucasian bhhänq.V "pole, post"), bucca "cheek" (Basque *beko "forehead", PEC *bek.wo "part of face, mouth"), câseus "cheese" (Basque *gastana "cheese", PEC *tsâk.wV "sour, raw"), cîmex "bedbug" (Basque *tSimitSa "bedbug", *tSimiri- "butterfly", PNC *ZimiZa "kind of stinging insect"), fovea "pit", fâvissae "vaults, underground chambers" (Basque *hobi "grave, tomb", PEC *fiwi / *fibi "grave"), pîla "squared pillar, column" (Basque *Hapal "shelf, benchtop, platform", PEC */apVtlV "board, cover, pole")
 
Then mentions some Germanic examples, like Proto-Germanic *hakô (masc. an-stem) "hook" (Basque kako, gako "hook", PEC *k.wäk.e "edge, point, corner") - Wiktionary lists some PIE etymologies for some of them.

Then some Balkan and Slavic examples, like Proto-Slavic *kuka "hook"

Then discussing how a- is sometimes present and sometimes absent in many substratum words and in Basque.

"In phonetics of substratal words, there seems to be a recurrent loan correspondence of IE velars to Euskaro-Caucasian sibilant affricates."

Like PGmc *sagjaz "sedge", Russian osoka "sedge", PEC *SertSV "sedge, ..."

Then a listing of such cognates as

Bean: PGmc *baunô, Latin faba, PSlav *bobu, Basque *baba

Chickpea, pea: Old Saxon erwit, Latin eruum, Greek erebinthos, orobos, Basque *garbancu (>garbanzo), (PNC *qorhhâ)
 
Back
Top Bottom