Linguistic shadow-boxing

(International Bad English version of a book review, published in German in Finnisch-Ugrische Forschungen, Bd. 58 / 2004)

Angela Marcantonio, The Uralic language family. Facts, myths and statistics. Publications of the Philological Society, 35. Oxford – Boston: Blackwell, 2002. 335 S.

In her new book, Angela Marcantonio (henceforth AM) intends to reveal the unfounded and mythological nature of the Uralic language relatedness, showing that there is no scientific evidence to support Proto-Uralic. The small and vociferous group of the so-called “rebels” or “revolutionaries”1 has already hastened to bring forth its all too expectable praise (Julku 2002 – cf. Kallio & de Smit 2002; Künnap 2003). Despite this kind of dithyrambs (that could rather be understood as negative merit for any book), I have done my best to shed my prejudices while reading AM’s book. I was happy to see that the most eye-catching characteristics of “revolutionary” Uralistics are missing here: there are no quotation patchworks, no bombastic rhetorics and no aggressive self-praise. At first glance, the book looks like a serious linguistic study: the style is pleasant and precise, the bibliography is of an impressive length,2 and incorrect translations or transcription errors very seldom hit the reader’s eye.3 However, AM’s book shows several fatal misunderstandings that finally leave its central claims completely unfounded.

Re-evaluate – but how?

The book consists of an introduction, a concluding chapter and 9 chapters in between, dealing with individual questions of Uralic language relatedness. Regrettably, these analyses are often left unfinished. Handbooks of Uralistics are quoted, mostly correctly, but the conclusive descriptions of traditional Uralistic standard or mainstream theory are often deficient or wrong. In many fatal cases AM departs from an assumed but unproved (no source references!) “received wisdom” that does not really exist. The actual criticism of this incorrectly described paradigm is, in comparison, strangely sparse: some irregularities in the standard theory are presented, as well as some “correlations” outside the Uralic language family, but no convincing picture arises from this criticism.

In the introduction, AM summarises the fundamental tenets of the Uralistic standard theory. According to her, all of these need not be wrong; some of them cannot be falsified, but, nevertheless, they are not sufficiently founded. As a theory, they are “not even wrong”, i.e. too vague and flexible in their formulation. Even if AM admits that some Uralic etymologies may represent “a true linguistic correlation” (whatever this means), these cannot be distinguished from chance resemblances. On the other hand, AM claims that there is a lot of evidence to the contrary. These “cons”, overlooked or consciously forgotten so far, will now be systematically dealt with, together with the “pros”. Among other things, the possible Uralo-Altaic correlations will be given a more precise scrutiny.

The introduction includes a short critical analysis of the “history of the Uralic theory”. One of AM’s basic tenets is that the present-day theory of Uralic language relatedness is an outdated axiom that modern research has somehow forgotten to re-evaluate, even though its basis, the research by József Budenz, is clearly not valid any more (for example, of a sample of Budenz’s etymologies, 81% are not accepted in modern literature, cf. also Marcantonio, Nummenaho & Salvagni 2001). AM’s criticism of Budenz is certainly justified, like that of other linguists who have already critically analysed the first language comparisons. However, it is still impossible to understand what this criticism has to do with the credibility of modern Uralistics. The culmination of this section is hard to understand as well: after blaming Uralistics for blindly following outdated ideas, AM criticises Uralists for rejecting Budenz’s original ideas of Uralo-Altaic relatedness. Regrettably, this kind of illogical turns are very typical of AM’s argumentation.

Rehabilitating the Uralo-Altaic hypothesis?

Similar research-historical conclusions are elaborated further in the 2nd chapter, “The historical foundation of the Uralic paradigm”. Typically, AM begins with an assumption without source references: it is “usually taken for granted” (!) that the Uralic language relatedness was scientifically proven in the last decades of the 19th century, on the basis of historical evidence on the past of the Uralic peoples (!) and the work of two linguists, József Budenz and Otto Donner. This is followed by a detailed presentation of the oldest historical sources where possibly Finno-Ugric ethnonyms appear.

In principle, AM may be right in questioning the traditional “Hungarian” interpretations for certain ethnonyms in Byzantine and Arabic sources (such as, for instance, majgeríja in a text by the Arabic geographer Ibn Rusta, generally supposed to be the first appearance of the ethnonym magyar – according to AM, this is rather a variant of the ethnonym “Bashkir”!). However, I fail to understand how this criticism could constitute evidence against the Uralic language relatedness. The fact that Arabic and Byzantine authors always mention the alleged ancestors of Hungarians as “Turks” is, of course, symptomatic of what has been assumed by other researchers as well, that is, an ethno-cultural symbiosis between the Hungarians and certain Turkic peoples. Language relatedness is irrelevant here – compare, just to mention a few Finnic examples, the fact that some Ingrians have called themselves “Russians” and the Latvian ethnonym for the Krevins means “little Russian”.

This somewhat baffling analysis of the oldest historical sources is followed by a more detailed version of the critique of Budenz, and then by a criticism of the arguments brought forth by Otto Donner in defense of the Finno-Ugric-Samoyedic language relatedness. Unlike Budenz, who paid a primary attention to phonological relationships and etymological cognates, Donner based his argumentation on certain suffixes, numerals, pronouns, postpositions and morphosyntactic typological characteristics. AM wipes this argumentation off the table with one simple explanation: as these elements do not always appear in all Uralic languages but are often attested in Altaic and other Eurasiatic languages4 (!), they cannot constitute compelling evidence!

AM concludes this chapter with a version (true, a very carefully and tactfully formulated one) of the Turanist conspiracy theory: there are political motivations behind the Uralic relatedness hypothesis, such as the West-European-Anti-Osmanic political tendencies in the Habsburg monarchy, the German ethnic background of the founding fathers Budenz and Hunfalvy,5 and later the connections between Turanism and Fascism in Hungary, which discredited all ideas of a Hungaro-Turkic relationship in the eyes of the new Communist power. In Finland, as well, the relatedness with the “Altaic” peoples was rejected – AM claims – for ideological reasons above all!

Instead of these ideas, which AM claims to have read “between the lines” in many studies on Hungarian prehistory (p. 53),6 one would expect a more solid and linguistically founded apology of the Ural-Altaic relatedness. Regrettably, AM does not go beyond short references to a handful of papers where, for instance, the Hungaro-Bashkirian toponyms (?) are presented as “some of the strongest pieces of evidence contradicting the U[ralic] origin of the Hungarians” (!). Such sparse remarks as the endnote 38 (p. 294) where sound correspondencies such as Hung. sereg ‘army’ ~ Chuv. sarâ ~ Turk. cärig / Hung. sátor ‘tent’ ~ Chuv. catâr ~ Tat. catïr are mentioned – according to Lajos Ligeti – as examples of a “convergent Turkic/Hungarian sound development” (?) do not help the reader any further.

Criticism of the comparative method?

In Chapter 3, “Modern interpretations of the Uralic paradigm”, AM concisely presents some alternatives for the standard theory: the modified family trees or bushes by Kaisa Häkkinen and Tapani Salminen, the isogloss model by Péter Hajdú, macro-family hypotheses (e.g. the Nostratic or Uralo-Dravidic relatedness) and the contact or convergence theories by Kalevi Wiik, Ago Künnap and János Pusztay. Her conclusion is baffling: if numerous scholars working on the same material, on the basis of the same comparative method, have come to present so different results, this must mean that the method itself is either unsuitable for these languages or simply “not even wrong”, that is: too flexible and vague.

However, what is completely ignored here is the fact that these “modern interpretations” are not based on identical materials or methods. The alternative models often have completely different goals and aims (Hajdú’s isogloss diagram) or they have explicitly rejected the comparative method (Wiik, Künnap, Pusztay). Besides, AM does not give the slightest reference to the devastating criticism that the views of the latter three “dendrophobiacs” have provoked, thereby giving rise to the illusion that the works of these three represent an unquestionable and generally acknowledged linguistic approach.

In Chapter 4, “Reconstructing the sound structure and lexicon of the Uralic family tree”, the proto-language reconstruction is critically analysed. The point of departure, however, is not the allegedly canonised work of Budenz and Donner but Janhunen (1981), now suddenly recognised as the “starting point for most modern reconstructions”. With a statistical analysis AM claims to show that Janhunen’s corpus does not comply to the “significance criterion” of the comparative method, that is, there are more sound rules than regular etymological correspondences and thus no “cumulative effect”.

Regrettably, these calculations are hampered by a fundamental misunderstanding. AM has counted the correspondences in Janhunen’s paper as sound laws, although Janhunen’s “vowel rules” do not apply for individual sounds but the reflexes of certain sound combinations. A similar lack of understanding is evident in her criticism of Uralic etymologies. Minor phonological or semantic irregularities are presented as serious evidence against the credibility of the comparative method, not as a natural result of the nature of sound changes, which often are only superficially regular and often obscured by various background factors. In effect, this means that the historicity of language is not understood.

The fact that AM does not understand the historical dimension of language actually becomes evident everywhere in her book as an unability to distinguish present-day superficial similarities and differences from historical linguistic processes and taxonomies. Thus the acknowledged fact that Hungarian clearly differs from the Ob-Ugric languages is transformed to “Hungarian recognised as an isolate”. AM discards the hypothesis of a common Ugric proto-language on the basis of well-known phonological difficulties with the reconstruction, as well as an historical argument: she claims that the belief in Hungaro-Ob-Ugric relatedness is merely based on the erroneous etymology hungarus ~ Yugria. Hereby she almost completely ignores the linguistic evidence that speaks for the Ugric relatedness; the detailed work by László Honti (e.g. 1979, 1998) is only mentioned in passing, later in the book.

The following detailed presentation of the Uralic sound correspondences ends, again, with baffling conclusions. The reconstructed corpus contains too many irregularities (as already seen, this conclusion is based on fatal misunderstandings) to be credible. These irregularities could, of course, be explained with the great time depth of the reconstruction (see below) or with the affective character and high frequency of the words in question. AM, however, does not accept these explanations.

From a handbook of historical linguistics (Bynon 1977), AM quotes a statement on the “stability and resistance to change” of words like tooth, foot, mouse, man, woman, to be, to go, […] will, can, do, which is explicitly connected with the high frequency and early acquisition of these words. However, this stability obviously refers to something else: the stability of a lexeme, not the absolute stability of its form. It is well known that particularly the most frequent words often display “irregularities”, erosion and allegro forms – recall word forms such as won’t, can’t, Finnish (dialectal, colloquial) oon, meen < olen, menen ‘I am, I go’, the well-known analogical sound substitutions in certain numerals (e.g. Slav. devȩtĭ ‘9’ by analogy with desȩtĭ ’10’, instead of the expected **nVvȩ-), or the likewise well-known irregularities in (tabooised?) basic vocabulary, for instance, the IE words for ‘tongue, language’ showing different onset consonants (*l-, *d-, *dh-, *t-, *Ø-, *s-, *g’-; cf. Hock & Joseph 1996: 233-234). True, words like “affective” or “descriptive” have sometimes been used too sloppily in Uralistic tradition (cf. the detailed criticism by Kulonen 1996: 50-61). However, this problem is of a technical-terminological nature and has nothing to do with the credibility of the comparative method.

Statistical evidence?

The critical analysis culminates in Chapter 5, “False matches or genuine linguistic correlations?”, where AM deals with the Uralic etymological correspondencies with a statistical method, following the example of Ringe (1995). She claims to show that the words reconstructed as Proto-Uralic may, in principle, be related but cannot be distinguished from chance resemblances. Could not the same have been done in a much more simple way, by showing that the traditional comparative method shows false (or at least as many) correspondences between, say, Hungarian and Italian (or between two corpora created with a random generator, or between Hungarian and Bashkir) as between Hungarian and Finnish? And if this proves impossible (as I very much suspect it will), can this negative result be falsified with Ringe’s method?

I must confess that I cannot understand the logic of AM’s statistical method. It seems to be completely atomistic, that is, does not take the possible influence of the sound context nor the processual character of sound change into account. As the possibility of the complete loss of a sound is, very unrealistically, excluded for statistical reasons (since Ø in Hungarian can correspond to so many reconstructed consonants, such as *s, *w, *k, *m, *j, *l, *p, correspondences between Ø and a consonant – such as Finn. suoni ‘vein’ ~ Hung. ín ‘sinew’ – must be counted to the “false matches”!), too many traditional etymologies lose much of their credibility.

With a “control case which is equivalent to random words” – and which is not documented anywhere in the book! – AM claims to have shown that the Uralic etymological correspondencies could as well be accidental similarities. On the other hand, however, she takes all Uralo-Altaic, Uralo-Eurasiatic etc. similarities mentioned in etymological literature at face value: words such as Finn. muna ‘egg’ ~ Hung. mony with an alleged cognate in Dravidic are simply “present elsewhere”. Does she mean that the possibility of chance resemblance is, for these cases, automatically excluded?

Borrowed or inherited?

Chapter 6, “Borrowed or inherited?”, begins again with a baffling statement, again without any source references. According to the Neogrammarian approach – AM claims – the inherited vocabulary complies to regular sound laws, while loanwords are irregular and identifiable as loanwords on the basis of this very irregularity. For this reason, comparative Uralists have arbitrarily interpreted the Uralo-Altaic shared lexical items as “irregular” in order to write them off as loan words! However, there has never been such a principle of “irregularity of loanwords” in historical linguistics, and the complete adaptation of loanwords mentioned by AM on p. 155 is nothing new in Uralistics. On the contrary: adapted loanwords and their sound substitutions belong to the most central materials in the more than century-old tradition of Uralic etymological research.

It remains unclear to me, what this alleged “irregularity principle” has to do with the subsequent analysis, where the phonological problems of Turkic loanwords in Hungarian, in particular, are dealt with. AM may be right in questioning the simplified textbook view of the indubitably (Pre-)Bolgar-Chuvash character of many loanwords. There are, as many experts admit, many cases where the donor language cannot be certainly identified. However, I cannot understand how the complexity of the phonological correspondences in the Turkic loanwords of Hungarian could constitute evidence against the Uralic language relatedness.

The traditional explanation model where a Hungarian word is brought back to an unknown, extinct or unidentifiable Turkic language is nevertheless often supported by Turkic or, perhaps, “Altaic” cognates. AM has nothing concrete to substitute this model with – should one take the same irregularities that were used as arguments against the Uralic relatedness and now use them as arguments in support of a Hungaro-Turkic relationship? The concluding remark (p. 179) that an isogloss model “is better suited […] to account for the intricate correlations” between the languages of Eurasia is symptomatic of a fatal misunderstanding. An isogloss model – unlike the family tree which includes an in-built genetic model of explanation (historicity and transmission) – can illustrate synchronic relationships very effectively but is unable to account for anything by itself, particularly for anything diachronical.

Younger and more simple than believed so far?

In Chapter 7, “The Antiquity of Proto-Uralic”, AM presents arguments against the traditional chronologies. As there is no archaeological or ethnographic evidence (!) nor historical sources (!) to testify to the great antiquity of Proto-Uralic, the evidence available, according to AM, consists of palaeolinguistic and palaeobotanistic research on words for certain trees, as well as IE loanword etymologies. Here, again, I cannot follow the logic of AM’s conclusions. There are not, nor can there be, any direct archaeological pieces of evidence for the traditional chronologies, but pattern explanations do exist, where the results of archaeological research together with comparative linguistics contribute to a convincing description (even if there still are considerable differences between the opinions of diverse researchers, cf. Carpelan & al. (eds.) 2001). And as there is no evidence for more recent great migrations, the wide geographical distribution and the cultural and anthropological differences between present-day speakers of Uralic languages just cannot be explained in any other way than by assuming a great time depth. This idea is, contrary to what AM claims, in principle independent of any concrete localisation of the primeval home.

According to AM, there is a “received wisdom” (without source references!) that Proto-Uralic acquired loanwords from Proto-Indo-European, Proto-Finno-Ugric in turn from Proto-Aryan, and so forth – the debated character of these hypotheses, above all, the great debate brought about by Jorma Koivulehto’s etymologies, is here passed by in silence. In the second part of this chapter, “Indo-European loanwords and the chronology of Proto-Uralic”, AM presents an incoherent mixture of conflicting hypotheses and chronologies. Finally, Kalevi Wiik’s substratum model is dealt with in brief (p. 199-201). Here, finally, I find AM’s criticism well-founded indeed. Even if there are numerous traces of a substratum language in Germanic, this substratum is phonologically clearly un-Uralic, and many of the alleged substrate phenomena could rather reflect much more recent contacts with Uralic language forms, which means that they do not lend support to Wiik’s hypothesis.7

In Chapter 8, “Morphology”, AM argues against the traditional reconstructions of Uralic morphology. Unlike Proto-Indo-European, PU presents no complete reconstructible paradigms nor ancient suppletions or syncretisms. Most Uralic case endings, verbal endings and plurality markers, according to AM, have arisen in historical times (!) in individual Uralic languages; the Uralic morphology (like the vocabulary) is not analysable using the comparative method, and statistical analyses similar to those used in the criticism of etymologies rather point towards larger Eurasian contexts. Here, as well, AM lists several Ural-Altaic correspondencies and similarities.

Regrettably, AM’s arguments cannot stand a more careful scrutiny. She reduces the case repertoire rather arbitrarily by trusting, among other things, Ago Künnap’s (in my opinion) too weakly founded arguments against the Uralic *m-accusative (which does exist, for instance, in South Sámi and Mari!), so that finally only two primary case endings remain: a locative/lative *-n and a locative/separative *-t (!). The so-called secondary case endings are written off with the simple argument that they have come into being only very recently: the agglutination process in fieri can still be seen in the oldest documents of Hungarian. This may be the case in the case of Hungarian, but it does not say anything about the pre-agglutination state of the language. Beside the classical typological cycle hypothesis (isolating > agglutinating > fusional > isolating), which actually represents a coarse simplification of the complicated, polysystemic linguistic reality,8 there are many good examples of retrogradic developments, typological cross-currents (cf. Rätsep 1981) and structural stability. There are classical examples of how affixes or categories characteristic of certain languages repeatedly disappear and regenerate themselves (Latin cantabo > cantare habeo > French chanterai >? je vais chanter). Why could not the local cases, which actually constitute the “hard core” of Uralic noun morphology (which, as a system, is clearly more stable than the verbal morphology, cf. Kangasmaa-Minn 1993), belong to such regenerated categories?

The conclusion of this chapter leaves the reader rather perplexed: as the morphology of the Uralic languages is young (!), the most simple assumption is that these languages, or their ancestors, are young as well! (P. 251: “I personally believe that the ancestor(s), or at least the ‘immediate’ ancestor(s) of the U languages, whatever their nature, were relatively young and had simple morphological structure.”) This, not exactly professionally formulated9 statement can only be understood by assuming that AM brings the present-day Uralic languages back to language varieties of an unknown origin. However, “young” is something else than “of an unknown age”, and what AM really means remains a mystery to me. What is evident is merely that she is unable, or reluctant, to distinguish typological preconditions from historical-genetic processes.

In place of a conclusion

Before the conclusive chapter, Chapter 9 (“Completing the picture”) presents some further details that bring nothing new to a critical reader. The Turkic origins of Hungarian proper names, AM claims, have been systematically ignored in research history (recall what was said above about the close Pre-Hungarian-Turkic ethnocultural relationship: names are important cultural items as well!). Similarly, possible Altaic cognates for Finnish names have been passed by in silence, so that the similarity, noted by Castrén, between Suomi ‘Finland’ and Sumi (a toponym in the Sayan Mountains) has fallen into unmerited oblivion, in favour of various competing, more or less arbitrary Indo-European etymologies for Suomi. Subsequently, AM presents her opinions on genetics in the length of one page (the reviewer withholds her own ones), and, finally, the picture is “completed” with a page-long survey on archaeology, where the criticism against Siberian primeval home hypotheses is misinterpreted as applying for Urheimat hypotheses in general. At this stage, even an unprejudiced and well-meaning reader, provided s/he has any knowledge of diachronical linguistics, will lose her/his patience.

If AM, despite the fatal flaws in her argumentation, has managed to prove something, it is that the traditional model of linguistic relatedness cannot be completely and exactly algorithmised; rather, it is a pattern explanation consisting of many interlinked parts, complex and yet tolerating gaps in its construction. In many details it seems to be based on intuition and fingertip feeling, but, actually, it is dependent of various external and internal background factors. These, however, are well-known facts. A method which only accepts exact phonological correspondences, without taking the various, indubitably real influences of semantic, pragmatic, intrasystemic and other factors into account, factors which cause “exceptions” and “irregularities” in any natural language, is meaningless.

As concerns AM’s statement that dialectological and diffusionist methods should be more widely used in historical Uralistics (p. 134-135), any serious Uralist will certainly agree with her (cf. Laakso 2001). The problem is just that speaker-related approaches to reconstructed or corpus languages must face numerous practically insurmountable obstacles. AM herself fails to present any examples of what such dialectological and diffusionist investigations in historical Uralistics could look like. Above all, the importance of historical dialectology in itself is no argument against the traditional model of language relatedness.

Trying hard to find the positive sides, AM’s book could be regarded as a positive challenge to Uralistics, giving impulses for further discussion. Could it be that the tradition of Uralistic rhetorics has been too arrogant in trusting the acknowledged nature of linguistic relatedness as a fait accompli? Have Uralists fought for the right cause with wrong weapons? Should something be done to counter AM’s claims that the counter-evidence has been systematically minimised and ignored?

Above all, this book is a sad memento for all Uralists about the looming marginalisation of historical linguistics, Uralistics included. It seems that the principles of diachronic linguistics are known to all too few linguists today. Otherwise it remains a mystery why an obviously honest and industrious, academically trained linguist can misunderstand so many fundamental facts. And otherwise it is impossible to understand what kind of policies and expert opinions can have made a renowned publishing house issue this book.


Bynon, Theodora 1977: Historical linguistics. Cambridge: University Press.
Carpelan, Christian & Asko Parpola & Petteri Koskikallio (eds.) 2001: Early Contacts between Uralic and Indo-European: Linguistic and Archaeological Considerations. MSFOu 242.
Hock, Hans Henrich & Brian D. Joseph 1996: Language history, language change and language relationship. An introduction to historical and comparative linguistics. Berlin – New York: Mouton de Gruyter.
Honti, László 1979: Characteristic features of the Ugric languages (observations on the question of the Ugric unity). – Acta Linguistica Academiae Scientiarum Hungaricae 29: 1-26.
—- 1998: Ugrilainen kantakieli: erheellinen vai reaalinen hypoteesi? – In: Oekeeta asijoo. Commentationes Fenno-Ugricae in honorem Seppo Suhonen sexagenarii. MSFOu 228. 176-187.
Janhunen, Juha 1981: Uralilaisen kantakielen sanastosta. – JSFOu 77: 219-274.
Julku, Kyösti 2002: Maanjäristys. – Kanava 30 (7/2002): 489-492.
Kallio, Petri & de Smit, Merlijn 2002: Missä ovat richterit? – Kanava.
Kangasmaa-Minn, Eeva 1993: Aspektista ja sen sukulaisilmiöistä suomalais-ugrilaisissa kielissä. – In: Valma Yli-Vakkuri (ed.): Studia comparativa linguarum orbis Maris Baltici 1. Publications of the Department of Finnish and General linguistics of the University of Turku 43. Turku. 13-23.
Kulonen, Ulla-Maija 1996: Sanojen alkuperä ja sen selittäminen. Etymologista leksikografiaa. Suomi 181. Helsinki: SKS.
Künnap, Ago 2002: Divergent or convergent linguistic development? – Linguistica Uralica 38: 4: 271-275.
—- 2003: [Review of the book reviewed here, in English, untitled.] – Eurasian Studies Yearbook 75: 171-173. [Same in German: Linguistica Uralica 39 [2003]: 1: 55-57.
Laakso, Johanna 1990: Reflections on the problem of Uralic N/V word stems. – László Jakab & László Keresztes & Antal Kiss & Sándor Maticsák [ed.]: Congressus septimus internationalis fenno-ugristarum 3A. Debrecen. 153-157.
—- 1997: On verbalizing nouns in Uralic. – FUF 54/3: 267-304.
—- 2001: Jenseits von Sprachgeschichte und Dialektologie: Sprecherorientierte Aspekte in der Erforschung des Ostseefinnischen. – Finnisch-Ugrische Mitteilungen 23: 105-111.
Marcantonio, Angela & Pirjo Nummenaho & Michela Salvagni 2001: The Ugric-Turkic Battle: A Critical Review. – Linguistica Uralica [2001]: 91-102.
Rätsep, Huno 1981: Some tendencies in the development of Estonian. – Sovetskoe Finno-ugrovedenie 17: 202-211.
Ringe, Don 1995: “Nostratic” and the factor of chance. – Diachronica 12/1: 55-74.

[1] For criticism of the so-called new paradigm see, in particular, the detailed online bibliography by Merlijn de Smit.

[2] However, it remains unclear how correctly the sources have been read, quoted and understood. Mikko Korhonen’s opinion that the search for the primeval home of a people is a hopeless enterprise has been misinterpreted by AM (p. 183) as a criticism of linguistic Urheimat hypotheses. I also find it strange to see my paper (Laakso 1997) as the only source reference for the supposed lack of word classes in Uralic (p. 243). However, I have – even more clearly in Laakso (1990), with more source references – explicitly argumented against this out-dated hypothesis of “Uralic lacking the parts-of-speech distinction”.

[3] Yet, there is a considerable amount of minor technical errors. For instance, Viru as the alternative name for the Estonian language does not exist in any language I know of (p. 2; AM systematically uses obsolete linguo- and ethnonyms such as the exonyms – “Ostyak” instead of Khanty, “Votyak” instead of Udmurt etc. – or “Finnic” in the meaning “Finno-Permic, non-Ugric”). While criticising the alleged irregularities in the consonantism of Indo-Iranic loanwords (p. 187-188) AM forgets, in her triumphant comparison of Mordvin purcos ‘pig’ with Latin porcus, that -c- in the FU transcription system denotes an affricate. A more detailed investigation would certainly reveal more such minor errors that I now have noticed by chance, for instance, the erroneous statement that the Finnish postpositions takana and luota are “just recently formed from former nouns” (p. 215).

[4] Here AM incorrectly claims (p. 45) that the plural suffix -t does not appear in Sámi.

[5] The German ancestry of the Donner family in Finland was obviously unknown to AM…

[6] For AM (ibid.), the frequent use of words like “problem” or Hung. kérdés ‘question’ in the titles of Uralistic studies is psychological evidence of the fact that the authors have not been able to fully suppress their honest doubts concerning the language relationship. A better knowledge of German and Russian sources would have convinced her at once that these expressions simply belong to scholarly rhetorics in more eastern parts of Europe.

[7] Astoundingly enough, Künnap (2002, 2003) simultaneously adheres to AM’s and Kalevi Wiik’s ideas. This is just as astounding as his ability to accept both AM’s criticism of the comparative method and Janhunen’s reconstruction of Proto-Uralic (as a reconstruction of a deficient pidgin).

[8]AM¹s typological conclusions are extremely weakly founded. In connection with the cycle hypothesis, she merely refers to Raimo Anttila¹s textbook of historical linguistics and Dixon¹s Ergativity (!).

[9] Any linguist should know that today¹s languages are neither “young” nor “old”; they simply represent the most recent, often not clearly delimitable, part of a lineage, the other end of which disappears in the mists of prehistorical times. The only possible exception, beside artificial languages like Esperanto or Klingon, could be found in pidgin and creole languages; however, AM does not explicitly mention this possibility.

[10] Although with certain reservations, AM defends this etymology with the existence of a “historically attested”, “original” form sum in old Russian sources. Hereby she ignores the fact that Fi. -uo- goes back to an original long -oo-, that the -u- in Old Russian texts is the regular reflex of Finnic or Scandinavian long -o-, and that the supposed development -u- > -uo- remains completely unfounded. Above all, it remains a mystery to me, here as well as in the whole book, why a listing of obscure, superficial “Uralo-Altaic” similarities should be methodologically superior to traditional etymologies, which are normally argumented for with detailed, sometimes even historically attested phonological processes and the phonology of the languages in question.

Updated April 23, 2003

Comments are closed.