• Welcome to the Internet Infidels Discussion Board.

Language as a Clue to Prehistory

I still wonder when more progress will be made on the Harappan culture -- its language, its Y-chromosome, and its written script. Dravidian, para-Munda, proto-Burushaski are all proposed for its language, along with Sumerian/Elamite and Indo-European. (That last one is impossible but seems to still be believed by some India-born scholars). The early Hindus are noted for their mathematics, e.g. the decimal system and a rudimentary calculus. But I wonder if they may have built on the methods of the Harappans, one of the most advanced early civilizations.

Sequencing DNA from ancient skeletons is now relatively easy, but I see only an old result that some Harappans had Y-haplogroup L, prevalent also among Dravidians(?) The Burusho Y-haplogroup is mostly R2.

When the Indo-Aryans arrived and imposed Hinduism and a caste system (with the Aryans Brahmin) one might have expected the indigenous elites to get a high caste also! Since caste is passed father-to-son, the present-day Y-haplogroups may be informative. I think R2 is common among Kshatriya, L common among Vaishya, but I have NOT seen a thorough study of what strikes me as an interesting matter.

I'm posting all this because the Harappan script is in the news! "The Indus script is perhaps the most important system of writing that is undeciphered," says Asko Parpola, a leading Indologist. ... And a $1 million prize is offered for its solution.

The Mayan script was deciphered some decades ago with great difficulty. IIRC it was solved mostly by the teen-age son of two archaeologists. He had two advantages: (1) there were still people who spoke Mayan, though they couldn't read the script; (2) some of the symbols were "obviously" base-20 digits; the big numbers were regnal dates and birth-years for Mayan Kings; this gave a handle to the text.

But neither of these advantages is available for deciphering Harappan script.
 
In Defense Of The Comparative Method, Or The End Of The Vovin Controversy - Anna Dybo and George Starostin respond to Alexander Vovin's anti-Altaic "The End of the Altaic Controversy". AV has also strongly criticized the Dene-Caucasian hypothesis. AV used to be a support of Altaic, but for some reason, he turned against it.
I. Basic Vocabulary Or Morphology?

In both of his articles, AV places a particular emphasis on morphology as constituting the best basis for proving genetic relationship, and blames the Moscow school for neglecting it in favour of much less reliable lexical comparisons.
But inflections can be reworked, and there is not much Proto-Indo-European morphology that survives in Modern English, for example. But basic vocabulary survives much better, with most of English's Swadesh-list vocabulary being directly descended from PIE, often with the same meanings as the originals.

Another example is Latin vs. Romance -- only some of Latin's verb inflections survive, and not much much of its noun inflections.

"II. Phonetic Correspondences: Rigorous Or Realistic?"

How good must sound correspondences be to be acceptable? AD&GS give the example of which English vowels correspond to which German vowels: very complicated. English and German consonants correspond much better, however.

"III. Semantics: same meanings or typological reliability?"

AV objects to etymologies that he considers semantically implausible, like "horm" ~ "gums (of mouth)", though AD&GS claim that he extends that too broadly, to almost any semantic shift, even very common ones.

"IV. Convincing evidence: short lists or core etymologies?"

Short lists??? I prefer to avoid small-number statistics. AD&GS noted that many critics prefer to poke at peripheral issues.

"V. Internal vs. external reconstruction."

Internal reconstruction, while valuable, does have its limits. One risks getting into folk-etymology territory.

"VI. Philology or fantasy?"

AV implies that one ought to consult texts in some language, rather than use a dictionary. But it's hard to get such an understanding of several languages, even if it is only a reading knowledge. Dictionaries are often a good shortcut, especially if they go into detail.

"VII. Ignorance or irrelevance?"

AV claimed that the authors of the Etymological Dictionary of the Altaic Languages (EDAL) mostly worked in isolation. But considering what a big book that is, it has lots of previous research in it. Was AV expecting a history of the debate over Altaic?

"VIII. Mistakes: crucial or insignificant?"

That depends on how many mistakes one finds in what one looks at, I think.

I skipped over a LOT in that long document.
 
AD&GS finally get to

"IX. Religious belief or scientific research?"

They quote AV as saying
The only tangible explanation... is that the Altaic hypothesis at least in its Moscow version became a set of beliefs highly reminiscent of a religion. However, religion and science cannot coexist, because the first is based on faith, while the second seeks the explanation of the facts that are in need of explanation. EDAL does not explain these facts, it simply creates the ‘evidence’ for the pre-existing belief that is, of course, non-evidence.

...
The critical scholarship seems to be completely replaced by a religious belief in macro-families. No wonder that this attitude led to the situation when ‘sino-Caucasian’ is supported by virtually no one outside of Moscow Nostratic school and several Proto-Worlders, who compare anything with anything with a complete disregard to regularity of correspondences.
AD&GS continue with
AV’s penchant for branding his opponents as «religious believers» is well-known, although several years ago it used to be more frequently applied to anti- rather than pro-Altaicists (e. g., on Karl Krippes’ condescending review of [STAROSTIN 99]: «Krippes, like the majority of his fellow anti-Altaicists, gives his reader religious sermons, not supported by any evidence or documentation» [VOVIN 995]). In the light of this it is hard to understand whether, over the last decade, AV has really made the major progression from a «religious belief» in Altaic to «science» or has merely switched from one church to another. However, loud words aside, let us evaluate AV’s conclusion independent of its pragmatic context.
Then the debate over quantity vs. quality.
While there definitely has to be a certain «core» of «quality etymologies» (see above), neither the Altaic nor the Sino-Caucasian nor, in fact, any other hypothesis of relationship, long- or short-range, can be considered satisfactory if not backed up by a large amount of cognate forms.
That's what I like about the Kassian-Starostin and Ostapirat statistical work: testing the hypothesis of common descent by checking on the putative survival of core vocabulary.
As we have amply demonstrated above, the majority of AV’s rejections and accusations in [VOVIN 2005] lack absolute proof: at best, a few of them represent alternate explanations with widely varying proportions of «convincing force»; at worst (when no explanations are presented, only criticism), as forced arguments that make no influence whatsoever on the original probability of the hypothesis.
 
Likewise, there is much more potential evidence for Sino-Caucasian than there is for Sino-Austronesian, and much more potential evidence for Altaic than there is for, e. g., Japanese-Austronesian, Turkic-Sumerian or any other theory of relationship dealing with one or more Altaic branches. In addition, «Proto-Altaic» is a reasonably economic and realistic way to explain multiple similarities between Turkic, Mongolic, Tungusic, Korean, and Japanese languages without contradicting any of our knowledge (as opposed to implications) about these languages and the peoples that speak or spoke them. The same goes for «Proto-Sino-Caucasian».
Then grumbling that AV mainly discusses theoretical or methodological issues. Nothing like what one would do to assess the question of whether English is a Germanic or a Romance language.

As to Sino-Austronesian and Japanese-Austronesian, those are likely Austric substratum effects, from early Austric speakers inventing agriculture and then spreading outward.
It does not help matters much that AV’s writing style over the past few years has turned from moderately sarcastic to downright offensive. At times it seems as if the major goal of [VOVIN 2005] is not so much to present a critical overview of the Altaic dictionary as it is to convince the reader that its authors are little more than frauds. Time and time again, over the entire fifty-page length of the «review», the authors of EDAL are accused of ignorance («the authors do not know the Japanese language and are doing wordlist linguistics», «repeatedly manifested ignorance of the most basic facts about the Korean language history»), intentional carelessness («have no respect for and no need of cultural history of languages they compare»), and cheating («have attempted a willful misrepresentation of data on more than one occasion»).

...
All of this leads us to a sad, but, in our humble opinion, inevitable conclusion: [VOVIN 2005] should not be qualified as a true review of EDAL. Instead, it is a meticulously crafted propaganda piece, bent on misrepresenting EDAL as a work of unscientific fantasy and its authors as unprofessional frauds — coming as no surprise, since its author has already engaged in a similar, slightly shorter and milder, but even less argumented enterprise concerning Sino-Caucasian [VOVIN 2002].

...
In addition to these general problems, [VOVIN 2005] also sets a new low standard for academic debate in the field of Altaistics, which is wellknown, of course, for its amount of vitriolics on both sides; however, not even the unquestionable anti-Altaicist champion of the sarcastic approach, the late Gerhard Doerfer, has ever stooped to directly accusing his opponents of forgery and basic ignorance of «the most elementary facts» of the languages under study so many times over the course of one article. Even some of his most scathing criticisms, e. g. [DOERFER 995], read like the acme of politeness in comparison with much of AV’s writing — most probably, motivated by overtly populistic purposes, since generalized accusations frequently produce more effect on the unprepared reader than systematic discussions based on wholesale analysis of linguistic data.
 
The main anti-Altaic position is that Altaic is a  Sprachbund (German: language federation) the rest of the speakers of its branches exchanging linguistic features:  Areal feature

Some notable sprachbünde are the  Balkan sprachbund and  Standard Average European Both of them emerged in the Middle Ages, some 1,000 years ago.

The Balkan one has Albanian, Romanian, Bulgarian, Macedonian, and Greek.

sprachbund - Subtypes of Standard Average European - Linguistics Stack Exchange
One the one hand we have group A, which includes English, Norwegian, German, Spanish, French, etc. And then we have group B, including Latin, Russian, Czech, Latvian.
Number of features:
  • 9: French, German
  • 8: Dutch, most Romance, Albanian
  • 7: English, Romanian, Greek
  • 6: North Germanic, Czech
  • 5: Baltic, most Slavic, Hungarian
  • 2: Breton, Basque, Semitic (Maltese)
  • 1: Welsh, Kartvelian (Georgian), Armenian
  • 0: Irish, Indo-Iranian (?), most Uralic, Turkic (Turkish, Tatar), Northeast Caucasian (Lezgian)
Latin would likely rank low in this list, because it lacks several of the features. It has neither indefinite or definite articles, experiencers are often not subjects (mihi placet "it is pleasing to me": "I like"), passives are formed with "to be" + passive participles only for the perfective-aspect passives, with imperfective-aspect ones use alternative personal verb endings, ... though it does have some, like comparative particles (quam + abl. "than").

Proto-Indo-European would likely rank even lower -- it has alternative personal verb endings for all its passives.
 
Last edited:
Relevant to the Altaic debate, some opponents of Altaic common ancestry claim that these languages have converged in recent centuries. But recent convergence does not rule out an earlier divergence, as is evident in both the Balkan SB and SAE. Both of them have convergence that followed earlier divergence, since most of their members are Indo-European langs.

Balkans: Romance, Balto-Slavic, Greek, Albanian

SAE:
  • High: West Germanic, Romance, Albanian
  • Middle: North Germanic, Balto-Slavic, (non-IE) Hungarian
  • Low: Latin (?), Celtic, Armenian, Indo-Iranian, (non-IE) most Uralic, Turkic, Kartvelian, Semitic, NE Caucasian, Basque
The two highest scorers, French and German, have a common ancestor that was an early Indo-European dialect, some 5,000 years ago.
 
Finally,
As for AV himself, our response will probably be of little use to him, given how actively he now distances himself from ignorant frauds doing «wordlist linguistics». His prediction that «journalists and Proto-Worlders» who «will hail EDAL as a ‘great achievement’» will «all drown in the River of Time, since they do not know the Ford» [VOVIN 2005: 23] is easily applicable to the authors of EDAL as well (and is probably meant to be applied to them). The only thing that remains for us is to offer him our heartiest condolences for having drowned fifteen years of his own scholarly career in the same River, and hope that the rest of it will be spent in a suitably more timeless manner.

Seems like an allusion to this:

Vladislav Markovich Illich-Svitych | Nostratic Language
Language is a ford through the river of time.
It leads us to the dwelling of those gone ahead.
But he does not arrive there
Who is afraid of deep water.
 
I was reluctant to mention that link, because it didn't list its sources.

From elsewhere, the original:

His symbols:
K̥elHä wet̥ei ʕaK̥un kähla
k̥aλai palhʌ-k̥ʌ na wetä
śa da ʔa-k̥ʌ ʔeja ʔälä
ja-k̥o pele t̥uba wete

IPA symbols:
KʼelHæ wetʼei ʕaKʼun kæhla
kʼat͡ɬai palhVkʼV na wetæ
ɕa da ʔakʼV ʔeja ʔælæ
jakʼo pele tʼuba wete

My simplification:
kkelhä wettei hakkun kähla.
kkatlai palhakka na wetä.
sa da akka eja älä
jakko pele ttuba wete.
(j is /y/, ä is like English-spelling a /ae/)

From Vladislav Markovich Illich-Svitych | Nostratic Language

Word-for-word:
Tongue time-of water-of path/ford
gone-of dwelling-to us lead(s)
he but there-to come(s) no(t)
which-who fear(s) deep water

So it's even more simplified than the version that I was using:

Language is a path through the river of time.
It leads us to the home of the dead.
But he/one cannot arrive there,
Who fears deep water(s).

dead ~ ancestors ~ gone ones
home ~ residence
 
BTW, I spent the last few days finding out where all of it is mentioned in Aharon Dolgopolsky's Nostratic Dictionary: Dolgopolsky - Nostratic Dictionary (3rd edition, 2012) : Allan R. Bomhard : Free Download, Borrow, and Streaming : Internet Archive and other places. I used it because it follows in the tradition of Vladislav Illich-Svitych's work, likely incorporating much that work.

I succeeded, and I have some conclusions about AD's work, at least the parts of it in VIS's poem. Most of AD's etymologies have protoforms in at least one of Indo-European, Uralic, Turkic, Mongolic, and Tungusic, and usually more than one of them. There weren't as many from Korean, Japonic, Kartvelian, or Dravidian, however.

Afroasiatic is a difficult case, with protoforms for Semitic, Egyptian, and Berber, and not so much for Chadic, Cushitic, or Omotic. So it's rather easy to reach into the latter three for possible cognates.

Let's see how this works for the first word, K̥elHä "tongue (body part)" > "language" (a common semantic shift), entry 1035 in AD's big book, which he transcribed K̥EHlê

Indo-European *dnghweh2s > *dnghwâ > Tocharian, Armenian, Italic, Celtic, Germanic, Balto-Slavic, and Indo-Iranian forms, though these forms are somewhat irregular: Latin lingua, for instance. Greek and Albanian have different forms, of obscure origin, and I can't find anything on Anatolian, at least in Wiktionary.

Uralic *käle > Finnic, Samic, Mordvinic, Permic, Khanty, and Samoyedic. It coexisted with *nälmä > Samic, Mari, Hungarian, Khanty, and Mansi, from *näle- "to swallow".

Turkic has *til and it's widely represented across the members of that family.

Mongolic has *kelen, also well-represented, and Tungusic *xilngü, also well-represented, as far as I can tell.

AD also mentioned Korean hyeo, which seems rather distant, and East Chadic in Afroasiatic, kela. But he mentioned no other AA forms.

Reconstruction: Proto-Afroasiatic/lis- - Wiktionary, the free dictionary - a plausible case that AA had *lis- for "tongue": Semitic *lisän-, Egyptian ns (*lis), Berber *lis, and Chadic *lis, mainly represented in West Chadic, like in Hausa harshe.

So that EChad form may be a coincidence.

In summary, it seems to me that there is a relatively solid subset of Nostratic: IE, Uralic, and Core Altaic (Turkic, Mongolic, and Tungusic).
 
I earlier discussed Quest for the mother tongue: the story behind the search for "proto-World," a primeval language that most linguists believe will never be found, that many believe never existed, but that some say they're already piecing together. - Document - Gale Academic OneFile
and about Joseph Greenberg's "Language in the Americas", I quoted
Many simple mistakes have been found in Language in the Americas: words in the wrong language, words with the wrong meaning. Greenberg's detractors often preface their indictments by denouncing his methodology, but then, rather than elaborate, they almost invariably veer toward their favorite subject: his sloppiness. There seems to be an unofficial contest among them to characterize most powerfully the magnitude of the error. (Campbell may win with his pithy "The bulk of Greenberg's data is non-data.")
Seems to me that JG's critics have picked enough nits to supply a louse nursery.
 
Back to the old Lumper vs Splitter debates! It was about a decade ago that I looked at this, and just now I was motivated to hunt down some old links. The first two did not work, but showed up in the Wayback Machine. After that I just went directly to Wayback without checking whether the base URL worked or not.

Here's Greenberg defending himself from his vocal detractor Lyle Campbell:

This Bengtson-Ruhlen paper on Global etymologies has probably been mentioned already, but in case:

Here's the Greenberg-Ruhlen paper on Amerindian in the 1992 Scientific American

Here is a criticism of Bengtson-Ruhlen's work

What strikes me as peculiarly laughable about this last paper is the synopsis of Picard's complaints about halfway down. Realize that these were presumably the most egregious complaints they could come up with. Here they are:
* Incorrect Language: For example, the form woxos "shin" is identified as Blackfoot but is in fact Arapaho.
* Incorrect Gloss: For example, Natick mukketchouks is glossed as "boy", but is actually "son, man child".
* Incorrect Transcription: For example, the Shawnee word for "girl" is given as kwan-iswa but is actually kwaaniswa.
Is it not obvious these complaints are laughably picayune? For example, Blackfoot and Arapaho are both Algonquin languages, so how does that mistake factor into an argument about global etymologies? (Recall that Greenberg's method, and hence that of his disciples like Ruhlen, deliberately emphasized QUANTITY of data. These picayune "errors" showed up among THOUSANDS of data.)
 
Back to the old Lumper vs Splitter debates! ...

What strikes me as peculiarly laughable about this last paper is the synopsis of Picard's complaints about halfway down. Realize that these were presumably the most egregious complaints they could come up with. Here they are:
* Incorrect Language: For example, the form woxos "shin" is identified as Blackfoot but is in fact Arapaho.
* Incorrect Gloss: For example, Natick mukketchouks is glossed as "boy", but is actually "son, man child".
* Incorrect Transcription: For example, the Shawnee word for "girl" is given as kwan-iswa but is actually kwaaniswa.
Is it not obvious these complaints are laughably picayune? For example, Blackfoot and Arapaho are both Algonquin languages, so how does that mistake factor into an argument about global etymologies? (Recall that Greenberg's method, and hence that of his disciples like Ruhlen, deliberately emphasized QUANTITY of data. These picayune "errors" showed up among THOUSANDS of data.)
I agree that those are small issues. But do such critics point out bigger mistakes? Like such Indo-European false friends as English much ~ Spanish mucho, German haben ~ Latin habêre "to have", and Latin deus ~ Greek theos "god".
 
That's one reason why Greenberg's MASS comparison method starts with many thousands of words across hundreds of languages: large numbers drown out aberrations. And of course promoting a potential cognate (e.g. much/mucho) to be one of the twenty or so cognate pairs that best support a genetic connection would look at proto-Germanic and proto-Romance and quickly discard much/mucho.

A serious effort to evaluate Greenberg's method would start with his Amerindian Hypothesis. Lumpers regard Amerindian as especially clear-cut while Campbell et al view it as vile apostasy. @lpetrich -- what do you think of the Amerindian Hypothesis?
 
The Indo-Uralic sound correspondences
and an earlier work
IdentifyingTheIndoUralicLexicon.pdf by Onno Hovers

Something about his work makes me rather skeptical. Looking in its section on "Etymological Data" reveals 412 proposed IE-Uralic matches. One of them is "seven": IE *septm, U *ćäjćemä, but also Semitic *šabʕ-, Egyptian sfḫw (*ˈsafχaw), and Berber *saβ. That's often considered a wander word. I think that it was spread by Neolithic farmers that spread out from the Fertile Crescent and subsequently borrowed. Why 7 and not 6 or 8 or 9? 6 is 1 more than 5 and 2 3's, 8 is 2 4's, and 9 is 1 less than 10. So 7 may be a gap for people who don't have to count very much.

 Indo-Uralic languages - claiming not nearly as many, but some of them are in highly-stable vocabulary, like "water" and "name". Proto-Indo-European-Uralic comparison from the probabilistic point of view [JIES 43, 2015] concludes that these words are most consistent with common ancestry.

That aside, it will be difficult to resolve the mystery of the voicings of the PIE stop consonants, since PU had only one phonemic voicing: voiceless. One will have to look further, to Altaic, with more than one.
 
... One of them is "seven": IE *septm, U *ćäjćemä, but also Semitic *šabʕ-, Egyptian sfḫw (*ˈsafχaw), and Berber *saβ. That's often considered a wander word. I think that it was spread by Neolithic farmers that spread out from the Fertile Crescent and subsequently borrowed. Why 7 and not 6 or 8 or 9? 6 is 1 more than 5 and 2 3's, 8 is 2 4's, and 9 is 1 less than 10. So 7 may be a gap for people who don't have to count very much.

I'm a naive layman but I've wondered if the number SEVEN had special (mystical?) significance among some Semites, and that this provoked the word's wandering particularly among peoples previously able to count only to six or less.

*septḿ̥ ("seven") sure seems similar to /šabbāṯ/ ("Sabbath"). I know the two words are shown with different etymologies but some phonaesthesia might have been involved.
 
OH then gets to Proto-Uralic palatalized sounds t', s', and c'. Comparing them to PIE,

t' ~ ki, ik, ig, it, id
c' ~ sk-,-ik-,-Hs
s' ~ sC-, k(Vr), sti-, -Ts, -Ps, -ist, -Hs-
and some others

Seems like palatalization in Proto-Uralic.

Then velar fricative x. OH states that medial x defies sound correspondences, like PU *we/ixi~ PIE *wegh- "to transport" and *mexi "to sell, give" ~ *mey- "to exchange". Also in very basic vocabulary: *jewxi ~ *egwh- "to drink". There is also *teke- "to do, to put, place" ~ *dheh1- > *dhê- and *toxe- "to bring, give" ~ *deh3- > *dô- "to give"

Then PU *ng (velar nasal) ~ PIE *NH and PU *ng ~ PIE *nK, *nG

Then a big list of sound correspondences. Labial stops: PU *p ~ PIE *p, *bh ... dental stops: PU *t ~ PIE *t, *d, *dh ... velar stops: *k ~ PIE *k, *k', *kw, *g, *g', *gh, also laryngeals *h1, *h2, *h3 ... labial approximant: PU *w ~ PIE *w, *gw, *gwh ... palatal approximant: PU *j ~ *y, *g'h, (initially, before vowels *u, *e) *h1, *h2, (medially) PU *-jw- ~ *-h2u- ... velar fricative: *x ~ usually PIE laryngeals ... lateral: PU *l ~ PIE *l ... rhotic: (initial: rare), (medial) PU *-r- ~ PIE *-r-, sometimes PU *-Nr- ~ PIE *-Ns-, PU *-r- ~ PIE *-rs- ... labial nasal: PU *m ~ PIE *m ... alveolar nasal: PU *n ~ PIE *n ... palatal nasal: PU *n' ~ PIE *(consonant)-n ... velar nasal: (medial) PU *ng ~ PIE *NH, PU *nk ~ PIE *nK (K = velar stop) ... alveolar sibilant PU *s ~ PIE *s (not in a consonant cluster) ... palatal sibilant (initial) PU *s' ~ PIE *sC (C = consonant) ... palatal affricate (initial) PU *c' ~ PIE *sk ... palatal stop PU *t' (very rare) ... retroflex sibilant *s" (not common, lost of irregularities) ... retroflex affricate *c' (difficult) ... dental/alveolar spirant PU *dh ~ PIE *Tr ... palatal spirant PU *dh' ~ *KR ... zero consonant PU *() ~ PIE *H (laryngeals)

PU *a ~ PIE *h1, *h2 ... PU *ä ~ PIE *h2 ... PU *e ~ PIE *h1, *h2 ... PU *i ~ PIE *h1, *h2, *y ... PU *o ~ PIE *h1, *w ... PU *u ~ PIE *h2, *h3, *w (??)

Also some discussion of correspondences of vowels, though that is complicated by Indo-European ablaut.
 
(initial)
U *k ~ IE *K
U *kala "fish" > Finn kala "fish" ~ IE *(s)kwalos "large fish" > E "whale"
U *kulke- "to go, walk" > Finn kulkea "to go, travel, traverse" ~ IE *kwelh1- "to turn (end over end), revolve" > Latin -cola "inhabitant"
(initial)
U *k ~ IE *H
U *koje "dawn" > Finn koi "dawn" ~ IE *h2ey- "day, morning" > E "early"
(medial)
U *k ~ IE *H
U *teke- "to do, put, place" > Finn tehdä "to do" ~ IE *deh1- "to do, put, place" > E "to do"
U *x ~ IE *K
U *wixi- "to bring, take somewhere" > Finn viedä "to take away" ~ *wegh- "to bring, transport" > E "way"
U *jexi- "to drink" > Finn juoda "to drink" ~ *egwh- "to drink" > Latin êbrius "drunk"
U *x ~ IE *H
U *toxe- "to bring, give" > Finn tuoda "to bring" ~ *dheh3- "to give" > Latin dônum "gift"
U *j ~ IE *K
U *aja- "to drive, hunt, chase" > Finn ajaa "to drive ~ IE *h2eg- "to drive" > Latin agere "to act, behave, ..."

IE also has doublets like *h3ost- ~ *kost- "bone", and I think *Hant- "duck" and *ghans- "goose".

Not much to find any patterns in, I must concede. Did Proto-Indo-Uralic have both *K and *x?

 Indo-Uralic languages

Onno Hovers's work is
 
Oops. Onno Hovers's work is valuable, but I'm still skeptical about all the Indo-Uralic comparisons that he states.

I've found another Indo-Uralic paper: Preliminary evidence for a (non-exclusive) genetic relationship between Uralic and Indo-European by Gianfranco Forni

Proposing some 49 cognates, though I'm skeptical about some of them.

GF mainly worked on Indo-European and Uralic, but he also listed possible cognates in other language families, dividing them into "probable" and "tentative".

There were familier ones like "I/me" (1sg), "you (sg.)" (2sg), "that", "who?", "what?", "water", "name", and "fish": PU *kala "fish", PIE *(s)kwalos "large fish"> Latin squalus "large sea fish", English whale (we nowadays don't call them fish, despite their fishlike overall appearance and behavior) I'm using "that" as a generic demonstrative here.

There were some that I find rather unconvincing, like PU *tuli "fire" and PIE *tep- "(to be) warm, hot" > Latin tepidus "somewhat warm". How does one do l ~ p? Another one of those was PU *kakte "two" ~ PIE *dwoH > E "two"

GF also compared IE and U to several other language families, and I counted up what he considered good possible cognates. The others with the most cognates were Yukaghir, Turkic, Mongolic, and Tungusic, and many of those cognates were pronouns: 1sg, 2sg, "that", "who?", "what?" The others include "tongue".
 
Mother Tongue Journal splash page - a new issue dropped, #25 for 2024. Archive at Mother Tongue Archive – Mother Tongue and an earlier one, Long Ranger Archive – Mother Tongue

Two obituaries in it, for long-ranger John Bengtson and for Indo-Europeanist Raimo Anttila.

The JB one was done by Václav Blažek (Vatslav Blazhek), a fellow long ranger. JB was long interested in languages and linguistics, and he got a Master of Arts at a university. But he never had any official position, and he had to divide his time between linguistics and his day jobs, notably one which VB describes as "property description technician". After his retirement in 2013, he devoted all his time to linguistics, mainly long-range linguistics. But his first published article, from 1978, was about 10, 100, and 1000 in Indo-European.
As is apparent, the largest number of these publications are devoted to the Basque language and its genetic affiliations. In my judgment, John’s best work belongs here too, namely the monograph "Basque and its Closest Relatives: A New Paradigm."
One can find that book online as one of his papers at John D Bengtson | Santa Fe Institute - Academia.edu
After 37 years of collaboration, I am convinced that John was a man endowded with a strong but realistic imagination. He was not afraid to risk crossing the borders between language families that had, since the 19th century, been deemed unrelated and unrelatable. It is important to note that, in his work, he was always careful to employ the classical Comparative Method, as that method had been developed by the Neogrammarians in the early years of scientific linguistic study.
 
Then a paper on related word roots in Vedic Sanskrit: CaC ~ C(C)â, like ay (i) "to go" ~ yâ "to drive"

Then a paper of Proto-Uralic *k becoming k and h in Hungarian. It becomes k before front vowels and h before back ones.

Hungarian ~ Finnish < Proto-Uralic or Proto-Finno-Ugric
hal ~ kala < *kala "fish" ~ PIE *(s)kwalos "large fish" > E "whale", Lat squalus
három ~ kolme < *kolme "3"
hat ~ kuusi < *kutte "6"
hall ~ kuulla < *kule- "to hear"
kettő ~ kaksi < *kakta "2"
kő ~ kivi < *kiwe "stone"
kéz ~ käsi < *käte "hand"

Finnish also has a sound shift: t > s before i.

Back in the Middle Ages, Hungarian had ch, the kh fricative, sometimes written x, instead of h. Thus, k > x > h before back vowels. But what might have made it shift? Before a back vowel, k might become a uvular stop, q. Making the sequence k > q > x > h.

That paper also mentions various sources of exceptions, like later phonetic changes and borrowings.
 
Back
Top Bottom