Review Apua5

[Arnaud Fournet]

Dear All,

I've written a review of APUA5 for LinguistList.
Unfortunately they seem to be very busy these last weeks and nearly two months later it's not yet published.
I think it's interesting to discuss these issues here slightly in advance.
The review is not neutral as you'll see and I tend to have my own point of view on some of the issues.

Best 

A.F

***

AUTHOR: Kari, James and Ben A. Potter (Eds.)

TITLE: The Dene-Yeniseian Connection

SUBTITLE: 

PUBLISHER: (APUA) Anthropological Papers of the University of Alaska

YEAR: 2010

ISBN: 0041-935400000 

 

Arnaud Fournet, La Garenne Colombes (France)

 

SUMMARY

 

The book investigates two sets of languages: (1) Yeniseian (YE), spoken along the Yenisei River in Siberia and severely endangered, and (2) Na-Dene (ND), spoken in Northwest America. The central claim proposed in the book is that potential genetic connections exist between YE and ND, therefore leading to a new linguistic family of higher rank: (Na-)Dene-Yeniseian represented on two different continents. This hypothesis is primarily developed by Ed. Vajda, a specialist of Yeniseian languages. It is presented as the “lead article” of the book (p.1) but the book contains many other articles deserving much attention as well.  

 

The VI+363-page book comprises 17 papers, written by different contributors belonging to several fields. It includes (1) an introduction by the editors, (2) Part1, presenting linguistic evidence for the claim, (3) Part2, examining consequences and relationships of the linguistic claim for other fields such as archeology, physical anthropology, kinship terms, mythological themes, etc., (4) Part 3, peer-reviews of the claim. Part3 is a conspicuous feature as a book seldom includes contradictory sections. The aim of the editors is apparently to stimulate a world-wide discussion of the issues dealt with in the book. Judging from cross-references within the book most contributors had the opportunity to read other contributions in a spirit of mutual fertilization and emulation. Appendices explain how to read these not so well-known languages, which are sometimes written with conflicting conventions.  

 

Editors Introduction: The Dene-Yeniseian Connection: Bridging Asia and North America (James Kari and Ben A. Potter) (p.1-24). This part is an introduction to the wealth of issues discussed in the book. It also contains an account of the recent developments of the Na-Dene-Yeniseian issue. It is particularly valuable for the four (geo-)linguistic maps (p.6-9). A kind of summary of the whole book is provided (p.5). 

 

Part 1. The [Linguistic] Evidence for Dene-Yeniseian 

[Paper1] The Dene-Yeniseian Hypothesis: An Introduction (Bernard Comrie) (p.25-32). This paper is a kind of long abstract of Paper2 and can be reviewed with it.  

 

[Paper2] A Siberian Link with Na-Dene Languages (Edward J. Vajda) (p.33-99). This chapter is the longest of the book with 67 pages and can be divided into four subsections: (1) a short presentation of YE mentioning previous hypotheses about YE genetic links and including a long footnote expressing gratefulness to a large number of people, (2) EJV's approach of comparative linguistics (3) a comparison of some morphological elements of YE and ND, (4) a comparison of lexical items. 

 

[Paper3] Yeniseian, Na-Dene, and Historical Linguistics (Edward J. Vajda) (p.100-118). This section is mainly autobiographical with some theoretical considerations. 

 

Part 2. The Interdisciplinary Context for Dene-Yeniseian

 

[Paper4] Genes across Beringia: A Physical Anthropological Perspective on the Dene-Yeniseian Hypothesis (G. Richard Scott and Dennis O’Rourke) (p.119-137). This section is a bibliographic survey of currently available genetic data with a special focus on the issue of possible links between YE and ND populations. 

 

[Paper5] Archaeological Patterning in Northeast Asia and Northwest North America: An Examination of the Dene-Yeniseian Hypothesis (Ben A. Potter) (p.138-167). This section addresses the issue of the archeological (dis-)continuities and their potential correlations with human migrations or cultural transformations. It focuses on Siberia and Northwest America and proposes 6 scenarios which may account for a YE-ND link. 

 

[Paper6] The Palatal Series in Athabascan-Eyak-Tlingit, with an Overview of the Basic Sound Correspondences (Jeff Leer) (p.168-193). This section describes the sound correspondences of Proto-ND on the basis of concrete and detailed lexical data. One of the latest developments in ND phonology is the introduction of palatal(ized) stops.

 

[Paper7] The Concept of Geolinguistic Conservatism in Na-Dene Prehistory (James Kari) (p.194-222). This section deals with an unusual feature of Athabaskan: in spite of covering the largest area of all North Amerindian families, Athabaskan is extraordinarily homogeneous. 

 

[Paper8] Dene-Yeniseian and Processes of Deep Change in Kin Terminologies (John W. Ives, Sally Rice, and Edward J. Vajda) (p.223-256). This section examines the anthropological and social features pertaining to kinship in YE and ND languages. 

 

[Paper9] Selecting Separate Episodes of the Peopling of the New World: Beringian–Subarctic–Eastern North American Folklore Links (Yuri E. Berezkin) (p.257-278). This section deals with comparative mythology. 

 

[Paper10] Comparison of a Pair of Ket and Diné (Navajo) Myth Motifs (Alexandra Kim-Maloney) (p.279-284). This section deals with a mythological theme which seems common to YE and ND people. The name of the (dragon-)fly in association with supernatural powers and various cultural artefacts seem to be shared. This point supports the DE-YE connection. 

 

Part 3. Commentaries on the Dene-Yeniseian Hypothesis

 

[Paper11] On the First Substantial Trans-Bering Language Comparison (Eric P. Hamp) (p.285-298). This section is an enthusiastic eulogy of historical linguistics and of the supposedly proven ND-YE connection. The style is more personal than academic. As regards linguistic genetic studies as a field EPH suggests to call it “linguistic modern cladistics” (p.285) whereas “genealogical” is preferred to “genetic” in other papers (p.25).

 

[Paper12] Proving Dene-Yeniseian Genealogical Relatedness (Johanna Nichols) (p.299-309). The paper describes a very idiosyncratic alternative to the standard methods of historical linguistics. It is little likely to gain any acceptance. JN tends to accept the (most probably) false YE-ND connection, even though she is aware that the geographic distance between these two groups is a problem that would need a plausible answer. 

 

[Paper13] Yeniseian: Siberian Intruder or Remnant? (Michael Fortescue) (p.310-315). In this paper MF explains how he has moved from sceptical to moderately supportive of the ND-YE connection. He examines four scenarios of split and dispersal of this “family” with a focus on the idea that YE would be closer to Tlingit than to the rest of ND. One of the scenarios include a return from North America to Siberia. 

 

[Paper14] Transitivity Indicators, Historical Scenarios, and Sundry Dene-Yeniseian Notes (Andrej A. Kibrik) (p.316-319). AAK first describes himself as an experienced Athabaskanist and typologist. This peer-review is brief and strives to “count as constructive”. But it definitely reads like politely worded poison for the claim that Na-Dene and Yeniseian might be genetically close. The gist of AAK's objections is that Na-Dene Transitivity Indicators (TIs or classifiers) are largely coherent throughout the family, which suggests they were “established morphologically at the Proto-Na-Dene-stage” (p.317). In addition to this coherence, they are prefixed right to the verb stem, which confirms that they “must constitute the earliest acquisition of the proto-ND inflected verb” (p.317). In contrast with this situation “what bothers [AAK] most of all is that the ND transitivity indicators do not find a clear counterpart in Yeniseian” (p.317). AAK concludes: “I am afraid that, as long as the status of the immediately pre-root TIs is not clarified, morphological argument for the [Dene-Yeniseian] relationship largely fails [sic].” (p.318) 

 

[Paper15] Dene-Yeniseian, Phonological Substrata and Substratic Place Names (Willem J. de Reuse) (p.320-323). This peer-review is rather neutral about EJV's theory and is more about raising issues of principles and methods. WJdR first reemphasizes Fortescue's observation that “Yeniseian looks more like an intruder than a remnant.” (p.320) Yeniseian phonology significantly differs from those of its present-day neighbors which tends to show that their mutual areal interaction appears to be virtually nil and cannot be old. It could be added that the same is true for morphology. Next WJdR exemplifies the difficulty of dealing with toponymic "substrates" with the cases of Vasconic and Old European theories. One of the issues to be addressed according to WJdR is the distance between Yeniseian and Na-Dene without any apparent toponymic connecting dots which could account for the dispersal of a unique original population. WJdR prudently concludes “that Yeniseian placenames are just one layer in that [Siberian] picture, thus not necessarily a substratic one.” (p.322) 

 

[Paper16] Dene-Yeniseian, Migration and Prehistory (John W. Ives) (p.324-334). This section begins with the “apparent impasse” of the Dene-Yeniseian relationship that “the genetic data” are so “seemingly at odds” with the “linguistic evidence” and wonders how this contradiction could be solved. (p.325) This section is not exactly a review but more a kind of article about the relationships between archeology and linguistics as applied to the case of Apachean. JWI describes how small groups of Apachean speakers have managed to grow and adopt new cultural features while remaining fairly stable linguistically. 

 

[Paper17] The Dene Arrival in Alaska (Don Dumond) (p.335-346). This section discusses the different time windows for waves of migration into Northwest American and the potential connections between present-day groups, like Eskimo-Aleut and ND, and archeological cultures. DD somehow rhetorically and theoretically wonders which of Eskimo-Aleut and ND should be considered earlier. 

 

Appendix A Orthographic Conventions for Yeniseian and Na-Dene (compiled by James Kari); Appendix B Symbols and Abbreviations. Very nice and useful reference tables. NB: Phonemes f and v are inverted in table 12A, typos: vowles, trompe l'oeille [oeil] (p.353).

 

EVALUATION

 

Even though this only has an indirect bearing on quality, the book is not free from typos or erratic spellings in several papers: Tur[c]ologist (p.9), configu[r]ation (p.12), Yensieian (p.17), Yenisieian (p.314), Athapaskan, Athabaskan, Athabascan (p.20), Yuork [Yurok] (p.21), pedagocial [pedagogical] (p.22), toutes [routes] (p.24), Berling [Berlin] (p.24), geneticly (p.106). Transliteration from Russian is often terrible: iikh [i jih] ‘and their’ (p.22). The same work: Starostin (1982) is transcribed in several ways, sometimes erroneously: enisejskix (p.32), Enisejskikh (p.117), enis[e]jskikh (p.98, p.359). 

 

The evaluation does not cover the internal peer-reviews which are described in the summary.

 

[Paper2] A Siberian Link with Na-Dene Languages 

 

To put it simple and short the scientific value of Subsection3 dealing with morphology is very low. There are considerable problems. 

 

1. A general problem with the YE+ND connection is that EJV leaves previous proposals undiscussed. According to EJV “the position of Ket in Inner Eurasia has remained as enigmatic as that of Basque in Europe, Zuni in the American Southwest, or Burushaski in South Asia.” (p.36) Actually the position of YE is not “enigmatic”. YE has been compared with Caucasic, and Caucasic with Salish, so if a new link is added between YE and ND, then Salish and ND, which are geographically contiguous, should be closely related. Apparently nobody proposed a Salish+ND group, not even the boldest “macro-lumpers”. 

 

2. Another problem is the huge and increasingly exasperating gap between a number of claims and statements made in the section and the failure to deliver anything concrete and real. Formal equations can be found between Indo-European languages: Latin crE-dEre = Old Irish cre-tim = Sanscrit s'rad-dadhAti  = Avestic zrazdA ‘to put the heart in > to believe’. Formal equations between *attested* languages have the virtue of being highly suggestive and self-explanatory. One equation between Eyak and Athabaskan is presented in another paper (p.212). It is not clear (and maybe doubtful) whether such formal equations are possible between YE and ND languages. In all cases none is presented. Instead of concrete comparisons one has to read an abstruse meta-linguistic discourse, where it is most of time impossible to disentangle description, comparison, reconstruction, hypothesis and sheer speculation. 

 

3. The argumentation is mostly abstract and based on *reconstructed* patterns. One would like to see the real data which are accounted for by the reconstructions. It is often unclear if the forms are EJV's own creations and interpretations or taken from ND specialists' internal reconstructions based on ND data alone. Most forms have no references or are allegedly “based on” other people's works. 

 

4. Another methodological problem is this sentence: “The Yeniseian perfective/stative suffix is productive in both Ket/Yugh and Kott, showing that it belongs to the oldest [sic] layer of the verb morphology.” (p.42) The reasoning is egregiously false. As taught by the comparative method, only synchronically non productive forms belong to the potentially oldest layer of morphology. It is most of time unclear what substantiates the claim(s) that YE forms should be considered old or fossilized. Neither the synchronic descriptions nor the diachronic reconstructions of the YE+ND languages can be easily grasped or assessed. A similar and relevant critique is expressed by one of the internal reviewers (p.318). The claim that YE shares “a system of morphological homologies with the oldest [sic] layer of” ND is  pointedly questioned by one reviewer. See Paper14. 

 

5. The severest problem is the confusion of synchronic / typological with diachronic / inherited features. 

 

The backbone of the theory is that YE and ND supposedly share the feature of having a prefixal verb pattern. The reasoning can be described as follows: [Step1] EJV compares Ket and Proto-YE with other patterns: Sumerian, Caucasic, Burushaski, Bantu. EJV concludes that these patterns are different from YE and therefore dismisses any “southern Eurasian” links. [Step2] ND is also different from other Native American prefixal systems displayed by Algonkian, Caddoan, etc. [Step3] YE and ND patterns are similar. Therefore it teaches us something genetic about them. [Step4] Some of the elements which fill the slots are similar. Hence: proven. 

 

As reminded by BC (p.30) and EJV himself (p.34) this slot pattern is a typological feature which exists in a number of languages, including French which is not mentioned: for example je lui en ai donné, etc. Even though the elements are separated by orthographic blanks, the rigid slot pattern is there for all tenses and moods, including non finite forms. French is interesting because it shows that this feature can easily and rapidly spring out of a language, like Latin, where it did not exist. My objection to EJV's reasoning is that any of the cited languages, from Caucasic to Caddoan, can have tinkered and divergently created its own pattern, just like French did out of Latin material. In addition it is hardly believable that this typological feature could be preserved for milleniums by Na-Dene and Yeniseian from their proto-languages. 

 

To put it bluntly the whole reasoning is hopelessly flawed. It transforms a typological feature into an inherited feature. French shows that the (dis)similarities that EJV uses to extract a YE+ND perimeter out of a huge set of Eurasian and Amerindian languages presenting that feature prove utterly nothing. It is not even clear why the closest relatives of YE or ND should be looked for preferably among languages with a prefixal verb morphology, because this typological feature is irrelevant for genetic studies as is exemplified by French. 

 

6. The premice that Proto-YE had a prefixal pattern does not seem coherent with the examples given (p.49). Kott has verbal suffixes and does not seem to support the “reconstruction” presented (p.37). The table (p.50) cites a form i- ‘P2Sg’ for Ket but the only attested form is k(u)-.

 

7. §2.2.2 about pronouns does not show any transparent isolated or systemic look-alikes between YE and NC. As noted by EJV “Dene-Yeniseian differs from established families (as well some more speculative ones) in the relative inscrutability of its pronominal morphology.” (p.53) Even though the resort to pronominal forms may be excessive in comparative linguistics, this is one more (potentially serious) hitch. 

 

8. In addition the morphemic comparanda are conspicuous for being often made up of only one consonant such as l, n, x, s, y, which can be the worn-out residue of about any kind of complex morphemes and more marked phonemes. Their testimony is therefore exceedingly weak. 

 

In other words, and as a conclusion about grammatical features, it is impossible to adhere to the reasoning and the conclusions proposed in the section. There are massive obvious flaws and the kind of data, comparanda, reasonings and requirements that a comparative linguist would like to read is missing. The wording is abstruse and there is almost no real substance. 

 

The last part of the section deals with lexical items and potential sound correspondences. It is much more concrete, even though it raises several issues as well. [Issue1] The Proto-ND consonant inventory is two or three times richer than the limited Yeniseian inventory. That situation contains a serious potential for bogus matches by chance coincidence. [Issue2] There is no table of sound correspondences. [Issue3] The comparisons are atomistic. What are the systems of vowels and consonants of Proto-YE and Proto-ND and how do they match? It is possible that a different reconstruction of ND could rearrange phonological features so as to have fewer consonants and two more vowels accounting for palatalized or labialized features of consonants. Such a reanalysis might make ND more easily manageable from a comparative point of view, be it with YE or another group. [Issue4] As a rough indication of lexical relationships one would like to see a wordlist of  basic vocabulary, for example that of Swadesh-100 with primary data from all languages involved in the study. The claim (p.53) that “evidence of the genetic connection comes more obviously from [...] basic vocabulary” remains unsupported. See Paper3. [Issue5] Another problem is the system of correspondences and its internal coherence and plausibility. Apparently ND has a three-way contrast: voiceless ~ aspirate ~ glottalized. Some of the “cognates” proposed by EJV are: ND *t'ik'y ‘ice’ = Ket tik and Kott thik ‘snow, ice on the ground’, that is to say ND glottalized is YE voiceless. This can be compared with PA*ts'əq ‘finger’ = Ket tə'q, where YE glottal stop appears near a ND voiceless consonant. The glottalized and voiceless features seem to be inverted. This problem of inverted glottalization is mentioned by Comrie (p.30) [Issue6] Several items are in my opinion borrowed from Uralic into Yeniseian and certainly cannot be counted as cognates between YE and ND: PY *ses ‘river’ < PU *sos ‘wet’, PY *se's ‘larch’ < PU *sokse ‘cedar, pine, conifer’, etc. In addition these two borrowings show that the theory developed (p.70) about Proto-YE glottal stop is wrong. Glottal stop is and was a segment (PU [ks] > PY ['s]). The attempt to get rid of that phoneme as “optional” to facilitate comparison with ND is unacceptable. The phonemic status of glottal stop in (Proto-)YE ruins the system of correspondences proposed in the paper. 

 

As a conclusion, the section proves nothing about YE and ND's relative genetic positions. It is not even suggestive of a possibly close relationship. There are counter-intuitive features such as pronouns being completely dissimilar, tentative sound correspondences being inverted and some “cognates” being Uralic loanwords. All these features are disturbing to say the least. The premice that Proto-YE used to have a prefixal verb morphology similar to that of ND is not even proved. In addition the theory proposed is not a credible alternative to previous suggestions that YE may be related to Caucasic, which has also been suggested to be related to Salish, and these connections make sense in my opinion. (See Paper4) The claim that YE is a close (or the closest) genetic relative of ND is near doubtless false in my opinion. People who might be attracted by the DE-YE connection must be aware that pronouns, kinship terms and basic vocabulary have nearly nothing in common and that there is hardly any transparent formal equation between these languages. As a matter of fact the rest of the book shows that no real support for this claim can to be found in other fields like genetics, anthropology, etc. It is fascinating (and cruel...) to see how scholars try to handle a (near doubtless) false hypothesis in the rest of the book. 

 

[Paper3] Yeniseian, Na-Dene, and Historical Linguistics

 

The section confirms that EJV is not familiar with the methods and issues of historical linguistics and with the historiography of the field. It is now well-known that Jones' “famous” conference in Calcutta in 1786 is certainly not a starting point for the study of the Indo-European family (Cf. Boxhorn, Leibniz, Ten Kate, etc.). It is also amazing to read that “Yeniseian-Na-Dene cognates are probably no more than 10 percent of the basic vocabulary”. (p.115) That level is notoriously that of random and it tears apart the claim that the YE-ND connection is “proved” by basic vocabulary. It can also be noted that “the most striking proposed cognate [...] Ket qy'j and Proto-Athapaskan *q'ëj” (p.106) for ‘birchbark’ is no less strikingly similar to Uralic *koj-wa, koj-ku, koj-ma ‘birch, birch-bark (vessel)’, which undermines its status as a potential isogloss between YE and  ND. Quite obviously EJV underestimates the issue of (Uralic) borrowings in Yeniseian (p.). According to my own preliminary surveys Uralic (especially Ugric) loanwords into YE are certainly not “an inconsequential percent of the vocabulary.” (p.101)

 

[Paper4] Genes across Beringia: A Physical Anthropological Perspective on the Dene-Yeniseian Hypothesis 

 

The conclusion is coherent with the premice that genetics and linguistics are most often correlated: “there is no specific gene, haplogroup, or dental trait that provides a direct link between the Kets and any Na-Dene speaking population.” Unsurprisingly genetics and physical anthropology confirm that the ND-YE connection is (most probably) false. Haida would also be confirmed to be unrelated to ND. The section also brings support to the hypothesis that Salish and Northwest coast “Indians” have European features (a variant of mainly European gene X2).

 

[Paper5] Archaeological Patterning in Northeast Asia and Northwest North America: An Examination of the Dene-Yeniseian Hypothesis 

 

Among potential problems, some areas are nearly terra incognita and the differences in Russian and American academic traditions and the lack of descriptive standardization are also hindrances to synthesizing available archeological records. The section is modestly not presented as a “lead article” by BAP who is also editor of the book but it certainly deserves to be considered one. A key point is that there is no indication of human migrations between ca. -14000 and ca. -5000 calBP between Siberia and North America (p.154). This conclusion has a clear bearing on any ethno-linguistic scenario linking Amerindian languages with Eurasian families. On the whole BAP tends to accept a strong correlation between archeology and ethno-linguistics and is mildly supportive of the ND-YE connection.

 

[Paper6] The Palatal Series in Athabascan-Eyak-Tlingit, with an Overview of the Basic Sound Correspondences 

 

This section could also be considered a “lead article”. The addition of palatal(ized) stops increases the already high number of ND proto-phonemes and tends to make ND yet harder to handle in the Nostratic framework, to which Amerindian languages will inevitably be integrated in a way or another. It would therefore be interesting to tackle the distribution of all the phonemes, the root structure of ND and try to determine potential reductions of the exceedingly high number of proto-phonemes. 

 

[Paper7] The Concept of Geolinguistic Conservatism in Na-Dene Prehistory

 

Athabaskan is extraordinarily homogeneous. JK explains this situation with the typological peculiarities of Athapaskan verb and noun morphologies, which are also embedded in toponyms. Athapaskan toponyms are nearly always meaningful and they used to play a major role in (inter-tribal) economy and wanderings. JK favors a high chronology for ND (much earlier than 8000 BP), especially in the perspective of external links with Siberian languages like YE.  

 

[Paper8] Dene-Yeniseian and Processes of Deep Change in Kin Terminologies 

 

Plain terms like husband, wife, mother, etc. are replaced by symbols, which makes the whole section unnecessarily hard to read. It is unclear what the authors really want to say. The section is interesting but somewhat lacks a linear thread. 

 

[Paper9] Selecting Separate Episodes of the Peopling of the New World: Beringian–Subarctic–Eastern North American Folklore Links 

 

This section is extremely interesting and tends to show that comparative mythology shares with historical linguistics a large number of issues and reasonings among which the dichotomy: diffusion or heritage. It appears that the usual syntheses involving mainly linguistics, genetics and archeology overlook the potential input provided by comparative mythology. Unsurprisingly there is no mythological theme that YE and ND speakers would specifically share. 

 

CONCLUSION

 

On the whole the book is extremely interesting to read and contains a wealth of information. The Dene-Yeniseian connection is in my opinion false but this should not deter potential readers from looking at the numerous papers which generally have rich and interesting contents more or less independent from this connection.  

 

 

ABOUT THE REVIEWER

 

Arnaud Fournet, unaffiliated scholar, La Garenne Colombes (France)

His research interests include (historical) phonology, descriptive linguistics and macro-comparison, especially in the Nostratic perimeter. 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/athapbasckan-l/attachments/20110111/2d2cfc32/attachment.htm>