Dr. Angela Marcantonio

                        Historical linguistics and the origin of the Finns:

                        The debate between ‘traditionalists’ and ‘revolutionaries’

1. The traditional Uralic theory and modern research

In the last couple of decades there have been in Finland as well as abroad a number of publications, in several fields of science, such as linguistics, anthropology, archaeology, genetics, which, one way or the other, all question what one can call ‘the standard, traditional Uralic (U) theory’ about the origin of the Finns and their language. Compare for example the following works in fields outside linguistics: Dolukhanov (2000a & b); Julku and Äärelä (eds,1997), Julku (2000); Niskanen ( 2000) and Nuñez (1987); see also the article ‘Palaeontology: science or fiction?’ in this volume for further references.

  According to the standard U theory, the Finns and their language are relatives, that is, genetically related, with other peoples /languages (such as the Saami, Mari, Mordvin, Khanty, Mansi, Samoyeds etc.), located between North-eastern Europe and Western Siberia, and with the Hungarians, isolated in the middle of Indo-European languages and peoples.  

   All these languages are claimed to form a ‘language family’ (‘Finno-Ugric / Uralic’ family), which means that they all are assumed to be derived from one single, genetic parent language, or ‘proto-language’, called ‘Proto-Uralic (P-U)’. P-U is believed to have been spoken in an area extending between the Ural Mountains and the Volga bends (the traditional home land), at least 6,000 - 8,000 years ago. From this area, the originally small, close proto- community, in the course of the years, divided itself, like the branches of a tree, into several branches, which migrated toward the areas where we find the various Uralic populations nowadays. In particular, the Finns and the other Finnic peoples, such as the Estonians, migrated toward the costs of the Baltic Sea, which are supposed to have been empty before the arrival of the migrating peoples. This theory is still widely taught in school and universities, not only in Finland, or Hungary, but also elsewhere, in Italy, for example. However, as mentioned, a group of scholars, from various field of research, have been recently claiming that all this is not correct, is not true. This academic debate between scholars of opposite views, between ‘revolutionaries’ and ‘traditionalists’, one could say, has intensified during the last couple of years as a consequence of the publication of two works on behalf of two ‘revolutionary’ linguists, the book by Professor Emeritus Kalevi Wiik (2002, Eurooppalaisten juuret), and by myself (2002a, The Uralic Language Family). 

    In spite of this on-going, at first clear-cut debate, I am often asked by Finnish and not Finnish friends, or by linguists who are not experts of U studies: ‘what is all the debate about?’. After all, usually, both traditionalists and revolutionaries claim that there has been a Finno-Ugric/ Uralic family and that the Finns do belong to this very family. Thus, what precisely do the two opposite schools argue about? It is exactly this issue that I would like to address in this paper. Being myself an expert of historical linguistic and Uralistics, and being an Italian who lives and works in Italy, therefore well away from the emotions that often go hand in hand with this kind of debate, I shall try to illustrate as objectively and as clearly as I possibly can, the basic tenets of the two opposite views, together with the ‘evidence’ on which these views are (supposedly) based as well as the ‘counter-evidence’ they are faced with. I shall also try to separate out the ‘evidence’, that is, the actual linguistic or extra-linguistic facts, from the many assumptions, interpretations and speculations that are an intrinsic part of both views (whether explicitly stated or implicitly embedded in them).

2. The methods of historical linguistics

At this point it is necessary to step backward for a moment, in order to illustrate what are the methods of analysis usually adopted by linguists to individuate and classify language families, and how do they work. In other words, why, on the basis of which kind of analysis and which kind of data, Finish and Hungarian are considered related to each other and belonging to the U family, whilst, say, Ancient Greek, Sanskrit, English and Italian are equally considered related to each other, but are classified as belonging to the Indo-European family? And, are these methods reliable, that is, have the known language families been really scientifically proven also in the sense of the term ‘scientific’ as we understand it nowadays?

   Let us start with the first question, which consists of two ‘sub-questions’. Question (1a): which method /methods of linguistic analysis are adopted? Question (1b): what kind of data are considered as ‘evidence’ in favour and what kind of data are considered as ‘evidence’ against an assumed language family?

   The answer to question (1a) is: the methods adopted are (mainly) the so-called methods of ‘historical, comparative linguistics’, which work as follows. Usually linguists, for the very fact that they deal with languages all the time, observe similarities among them, as any speaker / reader may also do (and as we all do all the time). Then they aim to make sure that the observed similarities are genuine, real similarities, indicative of some sort of connection among the languages under investigation, and not just the result of pure chance. In technical terms, linguists attempt to make sure that they are dealing with so-called ‘correspondences’ (among the sounds, the words and the grammar of the languages in question), and not with so-called ‘false matches/ chance resemblances’. Indeed, chance resemblances among similar / identical sounds, words or grammatical endings belonging to totally different languages are much more common than one would imagine, as shown by the following pairs: Italian villa ‘country-house vs Finnish villa ‘wool’; Italian (Rocca-)Ravindola, the name of a little village in Central Italy, and Finnish  ravintola ‘restaurant’. In this case at least, the very different meaning of the two sets of words alerts us that we are most probably dealing with chance resemblances. However, one can also find strikingly similar / identical words, both in sound and meaning, as shown by the following pairs (as proposed by Alinei 2003: 23, 304): Etruscan  maru  vs Hungarian  mérő, both meaning ‘surveyor’; Etruscan  parliu ‘to cook’ vs  Hungarian  párol ‘to boil, cook’. Another quite striking similarity, this time in the area of grammar, can be observed between Hungarian and Sanskrit: compare the Sanskrit ending -van, used to create gerundive forms of the verb, with the Hungarian ending -vá(n), having the same function, as shown in the following examples: Sanskrit  dā’-van ‘giving’ vs  Hungarian  vár-ván ‘waiting’.

   Thus, how do linguists discriminate between the two cases, that is between correspondences and chance resemblances? Usually, they first observe whether the languages under investigation share extensive similarities of grammar, in particular elements such as case endings, verbal endings, possessive endings, etc., because similarities of grammar are widely considered to be the safest hint of genuine correlations. If the grammatical correlations hold, then linguists assume that the other observed similarities in the sound structure and lexicon are real too. Then they further assume that all these similarities are due to the fact that the languages in question derive from the same parent language. Thus, having ‘postulated’ a given ‘language family’, although on a rather intuitive basis still, linguists can now proceed to analyse the observed similarities in a more rigorous, systematic manner, to ascertain that their original assumption of relatedness can be upheld.

   At this point we are already answering the second part of the question, part (1b): what kind of data are considered as ‘evidence’ in favour and what kind of data are considered as ‘evidence’ against an assumed family? In brief, how do linguists test their hypotheses scientifically? Usually, by applying the so-called ‘comparative method’ (of historical linguistics), linguists compare sets of words that are similar (or identical) in sound and possible meaning, to verify whether these (intuitively) observed similarities are also ‘regular and systematic’, and not just random. If this turns out to be the case, then the similarities are considered to be genuine, that is, they are considered to be ‘correspondences’. If in turn it can be shown that the assumed related languages share a ‘good number’ of  ‘regular and systematic correspondences’, then the original hypothesis of relatedness, and therefore the existence of the assumed proto-language, is considered to have been proven ‘scientifically’ (at least within the traditional methods of historical linguistics).

   As to the second question, whether the established language families have been proven scientifically also in the sense of the term ‘scientific’ as we understand it nowadays, a possible answer will be proposed in section 4.2 below. 

Coming back now to the issue of the U language family, let us explain what is meant by regular and systematic correspondences through some examples drawn from Hungarian and Finnish (of course, the comparative method has to be applied systematically to all the languages assumed to belong to the family for its validity to be proven):

As one can see, here there are two sets of words that are ‘obviously’ similar in sound, and also have the same meaning. Now, if we want to verify whether these words are not only ‘similar’, but also ‘correspond’ to each other ‘systematically and regularly’, (that is, sound by sound, and for all the sounds constitutive of the word), we apply the comparative method. By doing so we realize, for example, what follows:

n      the initial sound /p-/ in Finnish corresponds to the initial sound /f-/ in Hungarian

n      the first syllable vowel /e/ in Finnish corresponds to the first syllable vowel /é/ in Hungarian (examples (1) and (3))

n      the first syllable long vowel /ä/ in Finnish corresponds to the first syllable vowel /e/ in Hungarian (example (4))

n      non initial /l/ in Finnish corresponds to non-initial /l/ in Hungarian (example (3))

And so on. These words are, therefore, ‘correspondences’. If, by repeated application of the comparative method to many sets of similar words, linguists can demonstrate that, for example, each initial /p-/ in Finnish corresponds ‘only and always’, to an initial /f-/ in Hungarian, or more generally, that each initial sound /x/ in Finnish corresponds only and always to an initial sound /y/ in Hungarian, or to an initial sound /z/ in, say, Samoyed, and so on, whilst, for example, each final sound /x/ in Finnish corresponds to a final sound /w/ in Hungarian and to a final sound /u/ in Samoyed, then we have established a rich number of correspondences that are not only regular, but also systematic. As a further step, we can now attempt to get an idea of what the sounds, the words, the grammar of the original proto-language might have been like, that is, in technical terms, we can attempt to ‘reconstruct’ the U proto-language at any level of language. For example, drawing from the correspondences reported in the table, we can reconstruct the sounds constitutive of the four words, and therefore, ultimately, we can reconstruct the words themselves. As an example, considering once again the initial sounds /p/ and  /f/, we assume that the /p/ sound is original, because we know (from other fields of linguistics) that usually it is a sound /p/ that changes into /f/, and not viceversa. Therefore, in this case we safely reconstruct an original, proto-Uralic sound /*p/. By establishing correspondences and reconstructions of the type shown in the table, we also establish the ‘sound changes’ which might have occurred from the assumed parent language to the daughter languages: for example: initial  /p-/ ‘has changed’ into initial /f-/ in Hungarian (whilst remaining unchanged in Finnish), and this sound change too, like the related correspondence, is ‘regular and systematic’. At this point, having established many regular and systematic ‘sound changes’, so-called ‘sound-laws’, we have proven ‘scientifically’ the validity of the U family, and of the U proto-language, according to the requirements of historical linguistics.

However, this is a rather simplified picture of the full story. In fact, languages are complex entities, they develop and change in complex ways, under the influence of many factors, so that in reality one finds ‘irregular’ sound-changes that need accounting for in various ways in all language families. Similarly, a full reconstruction of the proto-language is hardly ever possible to achieve. However, if one can demonstrate that the great majority of sound-changes/sound correspondences are regular and systematic, and that major areas of the proto-language can be reconstructed, then the proto-language in question can be indeed considered as ‘scientifically proven’, at least within the framework of the adopted method.

3. What traditionalists claim and what revolutionaries claim

Let us come back to the issue under discussion, that is: what precisely the two opposite schools, the traditionalists and the revolutionaries, argue about? The answer is simple now, and can be formulated into two parts, the first part dealing with the linguistic status of the U family, the second part dealing with the archaeological, anthropological and genetic implications embedded in the U theory.

part 1 answer): the two opposite schools argue about ‘how good’, how ‘well-behaved’ the U family is. In other words, does it display a good number of good sound-correspondences / sound-changes? And, have major areas of the U proto-language been reconstructed, in particular, major areas of grammar?

    The traditionalists claim that the U family is a ‘well-behaved’ family, being full of good correspondences / sound changes. They also claim that the proto-language has mostly been reconstructed according to the requirements of historical linguistics, as shown by the traditionally reconstructed family tree (as reported in all text books).

     The revolutionaries, on the other hand, generally claim that this is not at all the case. On the contrary, according to them, the similarities shared by the U languages are neither good enough nor numerous enough to justify the thesis of a genetic relationship, at least certainly not in the terms purported by the traditional theory and family tree.

part 2 answer): the two opposite schools debate the issue as to whether the predictions made by the traditional U theory do actually respond to reality. For example, the U theory claims that the modern Finno-Ugric people come from the east, where their original homeland is assumed to have been located, and that they moved westward, always splitting like the branches of a tree. Now, is there evidence outside linguistics that this has indeed been the case? In other words, do archaeology, history and anthropology support the thesis of the existence of an ancient, very close community, its U homeland and its westward migration?

   Furthermore, the concept of proto-language (at least when it was first conceived about two hundred years ago) has always been equated with the population that speaks it. In other words, if somebody speaks a U language, he must also belong necessarily to an originally U ethnic group; if somebody speaks an Indo-European language, he must necessarily also belong to an originally Indo-European ethnic group. However, as is now well known, the equation ‘language = ethnic group’ does not always hold true; indeed instances of ‘death’ or of ‘switch’ of languages are quite frequent and might have happened also in prehistoric times, in any area of the world. Besides, in recent years, several scholars have started to use genetics to trace back the origin and the movements of peoples in prehistoric times, giving linguists the opportunity to test the implications of their theories also on this front. Thus, do the findings of genetics support the existence of a genetically unified, possibly a genetically ‘Uralic’ population?

      As already mentioned, these fields of research have all, and quite independently of each other, come to conclusions that clearly contradict the factual, historical predictions and implications embedded in the U theory (see section 4.1 below for details). Now, the revolutionaries demand that these results are taken into account, because only a multi-disciplinary approach can allow us to trace back the origin of languages and people with a good degree of certainty, particularly if we are dealing with prehistoric times. Therefore, the traditional U theory has to be modified so as to encompass these new results, in particular the findings from archaeology and genetics. However, traditionalists tend to ignore this multi-disciplinary approach, claiming, implicitly or explicitly, that linguistics is the queen of the humanistic disciplines, having its own scientific methods of analysis, and therefore does not need much help from any other disciplines. Occasionally, one may find traditional scholars who recognise that one or the other of the tenets of the conventional model are not quite correct, and propose a modification for it (these scholars are called ‘revisionists’ by Janhunen (2001); see also Marcantonio (2002b)). For example, some scholars recognise that the traditional family tree diagram is inadequate to represent the complex historical processes that have brought forth the modern U languages (as repeatedly pointed out by Hajdú; see for example Hajdú (1987: 67)), and propose different models of development (see for example the comb-model family tree by Häkkinen (1983:384) or the family tree diagram proposed by Viitso (1995 &1977). Others recognize that there is no enough evidence to locate the homeland in the area of the Ural Mountains, and suggest alternative homeland areas (there have been in fact several different proposals at this regard, see for example Suhonen (1999) and Koivulehto (1999)). Nevertheless, these scholars still believe that the standard theory is, in the main, correct.

4. What is wrong with the traditional Uralic theory?

Having presented the points of view of the opposite schools (in broad lines of course; see Fogelberg (ed., 1999) and Grünthal (ed., 2002) for an overview of the debate), I can now proceed to express my own opinion on the matter.

   In line with the revolutionaries, I personally believe the traditional U theory to be fundamentally wrong, for the following reasons: a) as mentioned, the conventional model is inconsistent with the results of other disciplines, as summarised in the points (1)-(5) of section 4.1; b) the conventional linguistic model has never been in fact scientifically proven, neither in the traditional nor in the modern sense of the term ‘scientific’, as discussed in section 4.2 below.

4.1. The contradictions between the linguistic model and the results of other disciplines

1. According to the traditional model, the genetic transmission of language from parents to children is the dominant, if not the only factor, in the process of language transmission and diffusion. Thus, one would expect all the people speaking U languages to have, basically, the same genetic characters. However, the analysis of human genes is at variance with this assumption. In fact, for example, the Samoyed and Ob-Ugric people have largely ‘mongoloid’ genetic character, whilst the rest of the (traditionally classified) U populations, including the Finns and the Hungarians, are largely ‘europoid’. That the Finns are ‘europoid’ is also confirmed by palaeo-anthropology, whose recent research on human fossils has shown how all the North-European population share most features with the originally europoid Cro-magnon peoples (see the article by Niskanen in Kaltio 2003/5).
2. There is no evidence for a “Uralic gene” – as one would expect from the equation: language = peoples.
3. There are no archaeological traces of migrations from the Ural / Volga area toward the West. Indeed, populations and technology  (such as arrow-heads, ice picks and ceramic technology), appear to have spread in a direction generally from the Southwest to the Northeast, that is, in the opposite direction than the one predicted by the conventional model.
4. The supposed migration from the homeland into ‘empty’ European areas is contradicted also by evidence that North-eastern Europe has been inhabited, without interruption, by local populations during the last 10.000 years.
5. There are no anthropological or archaeological traces of an ancient, close community living either in the area of the traditional homeland or in the areas of the alternative homelands proposed so far.

Thus, the result of genetics, archaeology and palaeo-anthropology constitute independent, clear and consistent ‘evidence’ counter to the factual implications proper of the U theory

4.2. The linguistic model

Let us now turn to what I believe to be the shortcomings of the U theory itself, at the linguistic level. Again, I find myself in agreement with the revolutionaries’ view, that there is not enough linguistic evidence to postulate a U family, at least certainly not as traditionally conceived.

    At this point the reader could ask: the opposite school claims just the contrary, whom should we believe then? The answer, I think, is now quite simple. Contrary to what traditionalists and text books usually claim, the specialist literature points out to the existence of many irregular sound changes, that is, many exceptions to the stated sound laws. Furthermore, the specialist literature also shows how the vast majority of the grammatical and functional endings across the U area have not actually been inherited from P-U (as would be expected), having instead formed during the separate lives of the various languages (see Korhonen (1996), Marcantonio (2002a: 203-251) and Suihkonen (2002)). Thus, going back to the table above, all text books emphasize the regularity of the correspondences: Finnish  /p/ vs  Hungarian /f/ and Finnish /e/ vs Hungarian /é/, as I have done myself, for the sake of the argument. However, these same texts omit to mention the observed exceptions, or the fact that not all the sounds constitutive of the words do correspond to each other regularly and systematically. For example, regarding the supposedly regular sound change:

proto-Uralic /*p/    à      Hungarian /f/

the reality is that in several occasions P-U /*p/ becomes in Hungarian /b/, and not the expected /f/ (see for example bőr, derived from *per', according to UEW 374). Or, considering again (I should say now) the ‘presumed’ correspondence: Finnish  puu  vs Hungarian  fa  (example (2) in the table) one can observe that the vowels do not correspond. Similarly, if we consider the words meaning ‘tree’ in the other U languages, we observe that they hardly display other regular sound correspondences apart from the correspondence of the initial sound (see Marcantonio 2002a: 100-101 & 161-162). Thus, of the four sets of words listed in the table, only the pairs  pelkää vs  fél  and  pesä  vs  fészëk  appear to be regular in all their constitutive sounds (according to the reconstuction of P-U by Janhunen (1981)).

    Of course, exceptions to the stated laws, or the formation of new morphology in the separate lives of the daughter languages, are phenomena present in all language families, as mentioned. However, if it turns out that the irregular sound changes /correspondences outnumber the regular ones, then the assumed language family cannot be considered as scientifically proven within the adopted framework, for the simple fact that the established ‘sound laws’ do not have general validity, they are not ‘laws’, being instead simply ‘post-dictive’ descriptions of the sound relations. In this case, the observed, presumed ‘correspondences’ are in fact mere ‘similarities’  (these similarities being in turn either false matches, or borrowed words not yet integrated into the sound structure of the receiving languages, within the framework of the comparative method). Equally, it would be difficult to sustain the thesis of an inherited, common morphology, if the vast majority of the morphological material to be found across an assumed family is clearly language specific. Now, as far as I know, linguists in general, including specialists of Indo-European and U, as well as specialists of those language families or sub-families whose validity has always been higly disputed (such as Altaic, or the Oceanic sub-group of the Austronesian languages), have not paid enough attention to the fundamental issue of the balance between proper ‘correspondeces’ vs mere ‘similarities’. Only if this balance can be shown to be decisively positive in favour of the ‘correspondences’, that is, if it can be shown that the identified correspondences and related regular sound changes are ‘statistically significant’, then one can claim that the families in question have been scientifically proven also in the modern sense of the term ‘scientific’. In other words, in my opinion, linguists have being arguing on whether or not a given language family/sub-family has been scientifically proven without reaching any sort of resolution (just like in U),  because they have failed to verify the balance between regularity and irregularity by carrying out a proper, quantitative analysis.

    In my book on the U languages I have attemped to do just this. I have attempted to ‘quantify’,  ‘to mesure’ the rate of regularity vs the rate of irregularity of the U comparative material (‘comparative corpus’), on which the traditional ipothesis of relatedness is based. I have found out that the number of the irregular sound changes greatly outweigh the number of the regular ones. Similarly, I have found that, in the great majority of the cases, the stated sound laws turn out to be rather ‘ad hoc’, being obeyed by a very small number of items (often just one or two words). I cannot of course go into the details of this quantitative analysis here – the interested reader might want to take the trouble to read my book…. Here it suffices to draw the reader’s attention to the following ‘fact’ (a fact widely recognised in the specialist literature; see Marcantonio (2002a: 75-78)):

The key Ugric node of the traditional U family tree, consisting of Hungarian + the Ob-Ugric languages (Khanty and Mansi), has never been reconstructed, because Hungarian is radically different in morphology, lexicon and phonology from its supposed sibling languages (see for example Abondolo 1987:185 & 1998: 428).

As a consequence, the Finno-Ugric node has never been reconstructed, because its reconstruction presupposes the previous reconstruction of the Ugric node. The top U node itself has likewise never been reconstructed, because this in turn presupposes the previous reconstruction of the Finno-Ugric node. Indeed, what is normally referred to as the reconstruction of the U node, the reconstruction proposed by Janhunen (1981) – the only rigorous reconstruction of P-U actually ever attempted – despite its name, is not at all a reconstruction of the top, P-U node. It is instead a partial reconstruction, a reconstruction of only two of the branches (supposedly) constitutive of the conventional tree: Finnic and Samoyedic. The key Ugric node is missing from the systematic comparison, for the reasons already exposed. As the Author himself declares (Janhunen 1998: 461), this P-U reconstruction is “a simplified but very useful approximation of proto-Uralic [that] can be obtained in the meantime by comparing proto-Samoyedic with proto-Baltic-Fennic, with additional information drawn, as required, from Saami and Mordva”. Thus, even if one does not share my views on the need to use statistical analyses within historical linguistics, the fact remains that a proper, rigorous reconstruction of all the branches of P-U, and therefore, ultimately, of P-U itself, has so far proven an impossible task.

   Last, but not least, it is worth mentioning that many of the similarities observed among the U languages are also shared with the nearby Altaic languages and also with  Palaeo-Siberian language.

5. Conclusion

To summarise, the validity of the U language family has not yet been scientifically proven thus far at the very linguistic level, contrary to what is generally claimed. Furthermore, the factual predictions of the theory are consistently contradicted by the results of other disciplines.

   At this point it must also be emphasized that the family tree model itself does not reflect adequately the way in which languages actually spread and develop, for two reasons: first, as mentioned, it disregard processes of language transmission such as borrowing, language acquisition, language switch, language mixing, language death, etc. Second, the tree model presupposes that the languages / peoples, once separated, are no longer in contact with each other, which is clearly false. Indeed, many linguists, including myself, believe that this unrealistic, out-of date model should be abandoned, or at least it should be used together with more realistic models of language spread and development. Furthermore, through linguistic research only we cannot trace the origin of languages /peoples much far back in time. In fact, the reconstructed proto-language in itself can tell us little about the antiquity of the supposed language family/community, if there are old written documents (like in Indo-European), and practically nothing, if there are no old written documents (like in the U linguistic area). In fact, reconstructions (of sounds, words, grammar), even the very good ones, are nothing but  ‘abstract formulas’ that help us to represent the similarities we observe among languages in a sort of systematic manner. Therefore, being only formula created by linguists, these reconstructions hardly have any real, temporal (and spatial) dimension.

  Taking all this into consideration, I believe that the revolutionaries are quite right in rejecting, or at least significantly modifying, the conventional model. In particular, they are quite right in adopting a multi-disciplinary approach, trying to fit together the results of linguistics, anthropology, archaeology, and genetics, these last three disciplines allowing us to trace the origin of peoples, although not of languages, quite far back in time.

   Coming back to the specific issue of the origin of the Finns and their language, if we dismiss the believe that they are ‘Uralic’ people coming from the east, we can start exploring new hypothesis. For example, we know from archaeology that there have been no migrations from the east, and that the costs of the Baltic Sea have been inhabited without interruption for the last 10.000 years. We also know from genetics and palaeo-anthropology that the Finns are basically ‘europoid’ peoples, coming from central Europe, possibly from the ‘Ukraina  refugium’,  together with the other European populations, according to Wiik (2002). Putting this ‘facts’ together, it is reasonable to assume that the local people who inhabited the costs of the Baltic Sea were the ancestors of the modern Finns and the other Finnic populations. It is also reasonable to assume that they moved from the South toward North in concomitance with the receding of the ice-sheet about 15,000 /10,000 years ago. This is indeed the new, basic idea of the revolutionaries. The traditional scholars have objected, among other things, that there is no way to know whether these old, local Baltic populations spoke a form of P-U, therefore we are dealing here only with assumptions, and not with facts. This is certainly true, but is does not matter, given that a P-U language probably never existed, or, if it did exist, it certainly did not originate and develop as proposed by the traditional model. It is therefore more than appropriate to explore other, different, assumptions regarding the proto-language, or at least regarding the direct proto-language of the Finns, if one insists on this concept. For example, P-U could have been a lingua franca, as Wiik (2002), Künnap (2000) and other scholars claim, or a ‘chain’ of languages / dialects extending from Scandinavia up to the Palaeo-Siberian linguistic area, as other scholars claim (for example Pusztay 1995 & 1997). Whatever the case, whether or not we accept these new assumptions, whether or not these new theses will turn out to be correct, one thing is crystal-clear to me as a linguist: the traditional U theory has never been so far ‘scientifically proven’, not even according to the traditional requirements of the methods of historical linguistics!

Bibliography

Abondolo, D. 1987. Hungarian. In B. Comrie (ed.), The Major Languages of Eastern Europe. London: Routledge.185-200.

Abondolo, D. 1998. Hungarian. In D. Abondolo (ed.), The Uralic Languages. Routledge Language Family Descriptons. London: Routledge. 428-456.

Alinei, M. 2003. Etrusco: una forma arcaica di ungherese. Bologna: Il Mulino Ricerca.

Dolukhanov, P. M. 2000a. Archaeology and language in prehistoric Europe. In A. Künnap (ed.). 11-22.

Dolukhanov, P. M. 2000b. “Prehistoric Revolutions” and languages in Europe. In A. Künnap (ed.). 71-84.

Fogelberg, P. (ed.) 1999. Pohjan poluilla. Suomalaisten juuret nykytutkimuksen mukaan. Bidrag till kännedom av Finlands natur och folk 153. Helsinki: Finnish Society of Science and Letters.

Grünthal, R. (ed.) 2002. Ennen, muinoin. Miten menneisyyttämme tutkitaan. Helsinki: Suomalaisen Kirjallisuuden Seura. Tietolipas 180.

Hajdú, P.1987. Die uralischen Sprachen. In P. Hajdú and P. Domokos. Die uralischen Sprachen und Literaturen. Hamburg: Buske. 21-450 [Budapest: Akadémiai Kiadó].

Häkkinen, K. H. 1983. Suomen kielen vanhimmasta sanastosta ja sen tutkimisesta: Suomalais-ugrilaisten kielten etymologisen tutkimuksen perusteita ja metodiikkaa. Publications of the Department of Finnish and General Linguistics 17. University of Turku.

Janhunen, J. 1981. Uralilaisen kantakielen sanastosta. Journal de la Société Finno-Ougrienne 77: 219-274.

Janhunen, J. 1998. Samoyedic. In D. Abondolo (ed.), The Uralic Languages. London: Routledge. 457-479.

Janhunen, J. 2001. On the paradigms of Comparative Uralic studies. Finnisch-Ugrische Forschungen 56: 29-41.

Julku, K. and Äärelä, M. (eds).1997. Itämerensuomi-eurooppalainen maa. Studia Historica Fenno-Ugrica II. Jyväskylä: Atena.

Julku, K. 2000. Die ältesten Wurzeln der finno-ugrischen Völker im Lichte der heutigen Forschung. In  A. Künnap (ed.),125-130.

Koivulehto, J. 1999. Varhaiset indoeurooppalaiskontaktit: aika ja paikka lainasanojen valossa. In P. Fogelberg (ed.). 207-236.

Korhonen, M. 1996. Typological and Historical Studies in Language. A memorial Volume Published on the 60^th Anniversary of his Birth. Edited by T. Salminen. Mémoires de la Société Finno-Ougrienne 223. Helsinki.

Künnap, A. (ed.) 2000. The Roots of Peoples and Languages of Northern Eurasia II and III. Fenno-Ugristica 23.

Künnap, A. (2000). Contact-induced perspectives in Uralic linguistics. München: LINCOM EUROPA.

Marcantonio, A. 2002a. The Uralic Language Family. Facts, Myths ans Statistics. Transactions of the Philological Society  35. Oxford/Boston: Blackwell.

Marcantonio, A. 2002b. Comment on the article by J. Janhunen:“On the paradigms of Comparative Uralic studies”. FUF 2001, 56: 29-41. Finnisch-Ugrische Forschungen 57: 466-470.

Niskanen, M. 2000. The origins of Europeans: population movements, genetic relationships and linguistic distribution. In A. Künnap (ed.). 33-59.

Nuñez, M. 1987. A model for the early settlement of Finland. Fennoscandia Archaeologica  4:3-18.

Pusztay, J. 1995. Diskussionsbeiträge zur Grundsprachenforschung (Beispiel: das Protouralische). Veröffentlichungen der Societas Uralo-Altaica  43. Wiesbaden: Harrassowitz.

Pusztay, J. 1997. Ajatus uralilaisten kansojen ketjumaisesta alkukodista. In K. Julku and M. Äärelä (eds). 9-19.

Suhonen, S. 1999. Uralilainen alkukoti. In P. Fogelberg (ed.). 245-248.

Suihkonen, P. 2002. The Uralic languages. Fennia 180: 165-176.

UEW = Uralisches Etymologisches Wörterbuch, I-VIII (ed.: K. Rédei; Budapest: Akadémiai Kiadó 1986-1991. German edition by  Harrassowitz).

Wiik, K. 2002. Eurooppalaisten juuret. Jyväskylä: Atena.

Viitso, T.-R. 1995. On classifying the Finno-Ugric languages. Congressus Octavus Internationalis Fenno-Ugristarum (eds.: H. Leskinen et al.; Jyväskylä). IV: 261-266.

Viitso, T.-R. 1997. Keelesugulus ja soome-ugri keelepuu. Akadeemia  9: 899-929.