Long-range shooting with recoil

S.A. Burlak, M.A. Zhivlov and I.B. Itkin have reviewed Jens Elmegård Rasmussen & Thomas Olander (eds.): Internal Reconstruction in Indo-European (2008) in the latest issue of Voprosy jazykoznanija (5, 2010:130-135). I find it positive that the reviewers take the time to thoroughly discuss most of the individual articles, and even more so that they devote almost one and a half columns to my contribution, “Internal Reconstruction vs. External Comparison: The Case of the Indo-Uralic Laryngeals”. 41HenwlmkxL__SL500_AA300_

However, that gladness quickly turns to indignation as it becomes clear that the reader is presented with a distorted impression of the content. My etymologies are rejected without motivation, and the reader is not provided with any information about why they were proposed in the first place. The four or five main points of the article, although deducible from the abstract and conclusions, are practically left out of consideration.

The reviewers start out by characterizing me as an “adherent of the Nostratic hypothesis”, although this cannot be explicitly inferred from the article. I have an open mind on these matters, but as it happens, one of my main points is that Nostratic etymology, because it is too focused on a search for cognates between Indo-European (IE) and Afro-Asiatic, can lead to wrong conclusions about Indo-Uralic (IU), and that it can prove fruitful to carry out bilateral or trilateral comparison before turning to the multilateral one. I am focusing on relations between IE, Uralic, and Yukaghir, but this does not necessarily imply that I have a firm opinion on more distant relationships (primarily because I do not yet have sufficient knowledge of the other languages involved). IU does not equal Nostratic. We must distinguish between these different levels of macro-comparison (or, perhaps rather, different degrees of khalepo-comparison), even though long-rangers on one side as well as their critics on the other hardly ever do so.

Saami (long-range?) hunters depicted in Olaus Magnus: Historia gentibus septentrionalibus (1555)

Saami (long-range?) hunters depicted in Olaus Magnus: Historia gentibus septentrionalibus (1555)

The reviewers claim that my comparison of the IE word for ‘hand’ and Uralic ‘limb’ fails since the PIE reconstruction is in fact *ĝʰes-r/*ĝʰos-to- rather than *ĝʰes-nt-, thus not corresponding to PU *jäsVnV. It is true that the IE daughter languages point to *ĝʰes-r, *ĝʰos-to- (or ĝʰes-to-), while my protoform is the result of internal reconstruction. The stem surely does not exhibit a typical heteroclitic pattern, and it is conceivable that ‘hand’ (Hitt. kessar, Skt. hásta-) was earlier inflected like Hitt. gipessar, gipesn- ‘cubit, ell’, cf. Skt. gabhasti-. Most heteroclitics have stems alternating between nominative -r and oblique -n(t)-, and these are probably ultimately identical to each other as shown by Birgit Anette Olsen in several papers. I admit that this should have been mentioned explicitly, and that the IE reconstruction should have been more precise. However, this does not at all affect the validity of my comparison; no matter how you reconstruct the stem-element under discussion, it does not form part of the root in either group. I show this by a hyphen which the reviewers have unfortunately omitted in their rendering (*jäsVnV instead of *jäsV-nV). They do not seem to take the root comparison proper into account, nor do they evaluate this or any other sound correspondences presented.

As for *kumte ‘wide; approximate number’, the reviewers claim that the latter meaning does not exist. In the article, however, I refer explicitly to Björn Collinder (in Die Sprache 13:182) to whom we owe the IU etymology, and the reconstructed meaning is based on Collinder’s examples: “Fi. kymmenkunta ‘etwa zehn’, satakunta ‘ungefähr hundert’, kyläkunta ‘Dorfgemeinde’. Wogulisch hånt bedeutet ‘Heer’, ungarisch had ‘Heer, Familie, Schar’”. The reviewers seem to care only about forms mentioned in Károly Rédei’s Uralisches Etymologisches Wörterbuch.collind1

They subsequently claim that I should reconstruct *kaswa, not *kawk-sa- for ‘high’, but this is not my idea either. I refer loud and clear to Jorma Koivulehto 1991 (Uralische Evidenz für die Laryngaltheorie); again, my references have been ignored by the reviewers.

And so it continues: Simply stating that PIE *h3meyĝʰ- ‘urinate’ cannot be compared to Proto-Yukaghir (PY) *onćə- ‘water’ does not count as proper scholarly criticism. One might just as well claim that Arm. get ‘river’ cannot be compared to Alb. ujë ‘water’. The claim obviously requires an accompanying explanation, and I give that to the reader, but the reviewers not only forget to mention the third relative, PU *kunće ‘urine; urinate’, semantically reminiscent of the PIE form while at the same time superficially similar to the PY one; they also withhold from the reader the sound-laws that fully account for every single part of the lexemes in question, including the loss of the h3 correspondent in Yukaghir. As I present it, the phonological correspondence is entirely regular. Ironically, while I regard this as one of the most striking cognate sets (it would never be taken for a loanword), the reviewers rather use it as a prototypical example of the invalidity of my comparisons.

Finally, I am criticized for equating the IE item *ĝʰalgʰ- ‘pole, stake’ with as many as three Uralic ones. Now, first of all, I write “and/or” between the Uralic items, obviously indicating that they are normally considered distinct, and that, although they can formally correspond to the IE item, they need not all be related to it nor to each other. Secondly, it is quite normal in Uralic linguistics to separate roots if no known derivational relationship can be applied to them, even in cases where the semantics, the physical similarities and the distribution among the daughter-branches tell us that they are probably ultimately related, albeit in ways that we have not yet uncovered. It is important to note in this respect that the three reconstructed forms are from different chronological stages; *jälŋV ‘tree trunk’ is Proto-Uralic, *jalka ‘leg, foot’ is Fenno-Ugric, while *jalaka ‘elm’ is Balto-Fennic.

Of course none of the etymologies would be convincing if I presented them in the way that the reviewers claim that I do. I conclude that none of their objections appears to be valid.

I find it a bit odd that as many as three reviewers can agree on such ill-founded criticism, especially when their main point is that in Copenhagen, like everywhere else, we use the wrong methodology. It is very easy to refute critique of that kind, but it is still a pity that you have to meet it even from reviewers otherwise known as serious and excellent linguists.

Historical linguistics in a broader context

In recent years, a tendency has arisen for historical linguists, e.g. scholars specialised in Indo-European historical and comparative linguistics, to join forces with scholars from related fields of research, including general linguistics, archaeology, historical philology, religion, history and natural history (e.g. historically oriented botany, geology and genetics). The Roots of Europe, having as its stated goal and as its raison d’être to examine and uncover Europe’s earliest languages, cultures and migration routes through interdisciplinary collaboration, is, I believe, a perfect model example of this recent tendency.

A randomly chosen example of what has become more elucidated thanks to the cooperation between Roots of Europe scholars is that a clear connection can be observed between linguistic factors and genetic migrations in Western Europe.

Molecular lineage
Molecular lineage

Historical genetics claims that human migration routes and history can be traced back to a sort of “genetic Adam and Eve” through the study of spontaneous changes – so-called mutations – in the DNA representing the male sex-determining chromosome (Y chromosome) and the so-called mitochondrial DNA. (mtDNA). Mitochondria are, in popularising terms, our cellular power plants: the energy needed by the cells is released in the mitochondria through respiration. Those two types of genetic material – the Y chromosome for men and the mtDNA for women – share as a common feature that they are inherited unchanged and in a direct line from father to son and from mother to daughter respectively… until a random mutation happens spontaneously only to be maintained and given forth in a straight line to the following generations on the paternal and the maternal side respectively.

That way, a typical Danish man or woman, for instance, is characterised by him or her carrying on an entire row or chain of Y-chromosomal and mtDNA mutations respectively that arose in earlier generations, whereas a newcomer from e.g. Africa will be characterised by a completely different mutation chain seeing that he or she has other ancestors in whom entirely different mutations have arisen. For the sake of convenience, a well-defined mutation chain is usually expressed in so-called haplogroups. Take for instance haplogroup R1b mentioned below: It covers the male Y chromosomal mutation of the chain “genetic Adam” -> M168 -> M89 -> M9 -> M45 -> M207 -> M173 -> M343 where each M followed by a number indicates a specific Y chromosomal mutation.

And now for what is relevant for this case: On the one hand, the geneticist Peter KA Jensen, one of Roots of Europe’s external researchers and collaborative partners, has informed the comparative and historical linguists employed at the centre about some of the results of the Genographic Project, viz. that Y chromosomal haplogroup R1b (see above) and the mutations characteristic for this group are particularly strongly presented in Western Europe, however strongest on the Iberian Peninsula.

Percentage of haplogroup R1a (pink colour) andf R1b (red colour) in Europe

Percentage of haplogroups R1a (pink colour) andf R1b (red colour) in Europe

On the other hand, the Iberian Peninsula is also the centre of what can only be seen as a unique language. I am, of course, thinking of the Basque language that seems to have no known, close relatives, maybe apart  from the now extinct Aquitanian language. Basque is, in other words, a linguistic isolate. All other languages of modern Europe fit nicely into either the Indo-European or the Finno-Ugric (Uralic) language family – but not Basque.

Now, the above considerations borne in mind, I find it extremely interesting that, according to another of Roots of Europe’s external researchers and close collaborators, the linguist Theo Vennemann, Basque(ish) elements can be found in a considerable number of languages of Northern and Western Europe – above all in the Germanic and the Romance languages. Most important and abundant are toponyms (e.g. river and mountain names) and metallurgical terms, but we also find many other words. Vennemann has, for instance, suggested that Latin grandis “large” has been borrowed from a pre- or proto-form of Basque handi “big”, and he has even managed to reconstruct phonological patterns for the loan word relationships between this pre- or ptoto-Basque language, named Vasconic, and the Northern and Western European languages. According to Vennemann, such circumstances are likely to indicate that the distribution of the Basque/Vasconic people has previously reached much further into the north-western part of Europe – at least compared to today’s situation with this once so widespread people having been forced back to a retreat in its core area: the Basque Country in the Pyrenees.

Proposed area of Vasconic languages
Proposed area of Vasconic languages

The distribution and prevalence of the above-mentioned haplogroup R1b, which the majority of scholars tends to associate with the Iberian Peninsula and the Basque people, goes perfectly well in hand with the alleged, former distribution and prevalence of the Basque/Vasconic language into the rest of northern and western Europe.

However, this case study has further perspectives: Through this very kind of collaboration, geneticists and linguists dealing with macro or long range comparison of languages and language families even further back in time might succeed in mapping the linguistic and genetic cohesion and migration routes of man all the way back to the exodus out of Africa.

The essence of this is, on the one hand, that all humans outside of Africa (and in a narrower sense even all from south of the Sahara) can unproblematically be traced genetically back to the mtDNA haplogroups M and N on the maternal side and to the Y chromosomal haplogroup CF, carrying the mutation M168, on the paternal side; and on the other hand that virtually all languages north of the Sahara may also be tentatively reconstructed back to a single language: Borean, including, inter alia, the Indo-European (e.g. Danish, Latin, Russian or Hindi), the Uralic (e.g. Finnish), the Sino-Tibetan (e.g. Chinese), the Eskimo-Aleut (e.g. Greenlandic/Kalaallisut), the Altaic (e.g. Turkish), the Afro-Asiatic (e.g. Arabic), the Dravidian (e.g. Tamil), and the Austronesian (e.g. Hawaiian or Indonesian) languages. In other words: Practically all language families outside of Sub-Saharan Africa. The linguistic part of this is, however, still highly controversial and uncertain, and only through future scholarship can it be shown whether this theory holds water.

Uncertain or not: What I believe to be the sticking point here is that we do find collaboration between comparative and historical linguists on the one hand and scholars from related fields, e.g. genetics, on the other hand.


This blog post is based on the following article:

Hansen, Bjarne Simmelkjær Sandgaard (2009): “Sproghistorie i en større kontekst”. In: Sprogforum, vol. 47, pp. 49-51.