Speculation in Palaeolinguistics: file B - further information

Speculation and error correction in Palaeolinguistics; file B - further information

See the essay of this title on this website for the main arguments. Numbers relate to points in the main essay.

3: Northumbrian and Celtic
The extended Northumbrian manuscripts post-date the arrival of the Vikings by a few generations, so the 'code breakdown' they show could connect to Scandinavian influence rather than Celtic. Without a cluster of Celtic loanwords to back it up the theory of Celtic substrate influence completely reshaping Northern English looks like a loser.

In Nh., wesa wosa is the word for 'to be'. This shows the loss of final n.
The link with Celtic looks rather unlikely to me as the loss of -n is also found in Frisian and North Germanic, so that the evolution of the morphology looks simply like an isostatic adaptation to the primary event of losing the n. There was presumably no Celtic substrate in Sweden or Frisia. In manuscripts, the form biothun alternates with biothu because the n comes and goes.

A study of Fueckel's monograph (Anglia, 1901) on one Northumbrian text suggests to me that there is an amount of 'morphological collapse' in Northumbrian and that the question of a distinctive linguistic arrangement north of the Humber is quite possible. The point of departure may be the "loss of -n in unstressed syllables" which obviously destroys a large swathe of the inherited Germanic endings both nominal and verbal. I would appreciate an evaluation of this in the round, given that Fueckel’s treatment is monographic and makes no observations at all on system or even on date. He is very thorough and we do notice:
a words with no endings at all, i.e. cases no longer being shown by sounds
b loss and simplification of endings (which had been based on the vanished ns)
c alternations of forms e.g. with -n and with the -n missing (loss of interest in case endings?)
d loss of distinction between different vowels (once the -n has disappeared) which F interprets as them being 'Murmellaute' so that the ending is only an indistinct vowel and again there is levelling of quite different endings

But what does this mean? a bad scribe? a stable spoken language in which realisations of words varied from minute to minute? a system in decay? degradation by copying of an older text which was much more consistent? It is regrettable not to know.

It may well be that the Vikings destroyed all the manuscripts in the North as part of their cultural programme, but in any case the lack of Northumbrian writings predating the Vikings makes the whole ‘Celtic substrate’ thing shaky and speculative.

The Finns add English dialect material, names of places and rivers etc., to increase the amount of Celtic loanwords. They do not seem to realise that if you double the sample you have to double the amount of positive results, or else you have not increased the percentage share of loanwords found. Much of their effort seems futile. The revisits to words of very shaky and questionable etymology points in the same way. Five very bad etymologies have as much weight as one good one. The conclusion is that they are biased, they prefer one conclusion to all others and are selecting material which fits in with it. I find this problematic. Meanwhile the presentation of evidence is plodding and very often frustrating and unpersuasive.


3a The legacy position is that there is no trace in English dialects of speech variations in the pre-migration homeland. However, we find that the loss of -n in the infinitive is found in Scandinavia, in Frisia, and also in the northern dialect in England. Obviously Scandinavia is, equally, to the north of the core Germanic speech area. So the idea that Northumbrian has no connection with Scandinavia and that its Anglian settlers did not come from regions further north than the havens from which the migrants to southern England shipped out has some puzzling aspects and may repay a further probe. Indeed, knowledge of early Scandinavian may give the insights needed to unravel the mysteries of Northumbrian language history.

3b Tolkien suggested that Northumbrian bioðun/ biðun (3 Ps Pl of the present of the verb 'to be') was derived from Welsh byddant (the same, although it is a future/habitual tense). This is very interesting. but meaningless in isolation. This is one of the earlier 'breakthroughs' in which Celtic material was detected in Old English. The resemblance is closest in Northumbrian but the Nh is very close to the forms in all other OE dialects. While the paradigm of this verb is chaotic (and so vulnerable to substrate influence?), the b- stem is present in other Gmc languages and is no doubt part of the inherited material (matches f- in Latin fui etc.)


3c There are a number of features shared by Welsh and Gaelic which are puzzling from a strictly Indo-European point of view. These included "the cluster of features comprised by prepositional pronouns, relative verbal forms, polypersonal verbs and initial mutations". An Internet article by Ratka Matasovic (Zagreb) at http://opus.kobv.de/ubp/volltexte/2007/1568/pdf/celtic_languages_in_contact.pdf describes an Irish Sea Alliance which gave rise to essential aspects of mediaeval Welsh (and Gaelic). Matasovic sees a phase of many bilinguals speaking both languages, making the languages converge. This phase produced features which we see in the modern languages but which were not represent in the ancestral Celtic. They typify ‘Insular Celtic’ (IC). Thus "(1) We know that between ca. 350 A.D. and ca. 550 A.D. there was intensive language contact on the British Isles. British and Goidelic, as already separate languages, as well as Vulgar Latin, and (at least since around 400 A.D. in Eastern Britain) Anglo-Saxon, were all spoken in the British Isles during that period in sociolinguistic conditions favourable to language contact.
2. Common phonological developments show us that those languages influenced
each other, and there is ample evidence for widespread bilingualism, perhaps
even plurilingualism during that period.
3. Several features shared by the IC languages, but absent in other forms of
Celtic and European languages, cannot, for reasons of relative chronology, be
attributed to Proto-Insular Celtic." (Ranko Matasovic, Insular Celtic as a Language Area, 2007)

Thus, Welsh and Gaelic may have been closer together in 700 AD than they were in 0 AD. Also, the Celtic of the Continent may have had none of the features which we are familiar with as linking Gaelic and Welsh. (By Alliance linguists mean a group of different languages which are geographically close and which converge by borrowing due to bilingualism. It translates the German word Sprachbund and the classic case is the Balkan languages. It is the opposite of 'descent' or 'genealogy'.) Matasovic's brilliant paper does not give every detail, but we can take it that the geographical site where this mixing took place would have been along the Atlantic seaboard of Britain: Western Scotland, Cumbria, Lancashire, Wales, and the shores of the Bristol Channel. Gaelic immigration into those areas is very well attested. He says "In Early Britain and Ireland, after the withdrawal of the Roman legions in 410 A.D., the dominant type of bilingualism seems to have been one in which at least Goidelic and British were idioms of roughly equal status. Code-switching must have been frequent, as well as exogamy, with children growing up in mixed marriages speaking early forms of British and Goidelic, and in some cases also Vulgar Latin, equally fluently. This type of situation facilitated the spread of structural features, but not necessarily of lexical material.
There is a vast amount of evidence for the presence of Goidelic-speaking communities in Britain in the period ca. 400-600 A.D. Those communities thrived chiefly in Wales, and, to a lesser extent, in Cornwall, i.e. precisely in those areas where the British languages survived the expansion of Anglo-Saxon. The evidence in question consists of historical records pointing to the immigration of Déisi, an Irish tribe, to Wales, which was facilitated by the weakening of the Roman military presence there in the late 4th century. There are also historical records confirming the existence of Irish kingdoms in Dyfed and Gwynedd in the Early Middle Ages, and the presence of Goidelic-speaking population in Wales is confirmed by the Ogam stones." He also finds it possible that the mixing took place within Ireland (and that P-Celtic groups preceded Q-Celtic in Ireland). (The Goidel did not thrive 'principally' in Wales if the whole of Northern Scotland was taken over by Gaelic-speaking incomers who are still there.) The postulate of an Irish Sea Alliance points logically to a grand-scale distinction between the Celtic speech of the western seaboard and that of what became England, the centre and east. This area is however the group of most interest to the Finns’ theory. It would follow that the only areas which left any written record of their speech, or had heir languages capable of testifying to their nature, are of diminished relevance to the question of how the Celts submerged by the Anglo-Saxons spoke.


6. Vennemann
His version of late prehistory resembles Renfrew's. He takes Hans Krahe's data but demolishing his IE etymologies, he finds in the oldest stratum of river names an 'agglutinating source with initial accent and predominance of A', which is not compatible with IE phonology. He connects these and old place names to Basque.

The Vanir are a 'matrilineal' mythical memento of the Semitic element in W Europe, and seaborne.

The omnibasque theory is incompatible with Palaeolithic Continuity Theory - it casts the IEs as late migrants evidently from a Mediterranean-Aegean homeland.

In the 19th C, linguists noticed that the word 'Basque' corresponded, not only to 'vascones' (now 'Gascon') but also to 'Vosges', the name of a range of hills in Alsace. So they detected a pre-IE language in which something like 'vask/vosk' meant 'mountain' and which called the Basques 'mountain people'. This language was not Basque as the Basques call themselves by a different word, Euskara. So traditional linguistics has no problem with an ancient language being spoken all the way from the Pyrenees to the Rhine, leaving behind a scatter of unusual words. No one really has a problem with this.
'Feasgar' is not an old Gaelic word but a loan from Latin 'vesper'. 'Euskara' does not mean either 'west' or evening', but 'Basque', a people of north-east Spain, so the semantic match with words for 'evening' is bad. I find at least two "west" roots in the accepted IE languages, giving respectively (English) 'west' and 'evening'. I think 'vecher' belongs with 'west, vesper, [W]esperides' and not ucher etc. On the whole, the phonological match is also bad. I am interested by a possible connection between 'Abend, avond, evening' and 'ucher', which would show a velar-labial alternation.
http://www.cls.psu.edu/pubs/pubs/LINGUA1158.pdf



5. Renfrew

(Polythematic) "These terms describe the lack of stem variation in the verb system to produce optative and subjunctive moods. Other features in the monothematic variant are the lack of separate feminine stems of nouns, the lack of a comparative form of adjectives. In fact, the whole idea of temporal strata of the root language."

Other scholars have discussed the resemblances between Hittite and Germanic. The overall shape of the Gmc tense system is similar to that in Hittite and so consistent with being archaic-faithful and so not creolised. This simplicity though represents a shorter features list and so is also compatible with creolisation. Shorter lists may not be structural
resemblances - they are also lists of missing resemblances, I would have thought. The patterns which Feist identified as resembling creolisation cross all features of the language and are not easy to dismiss. Clearly the ‘Hittite=German’ theory is one which gets Germanic off the hook of being creolised. If Germanic represents an older and more pristine version of Indo-European than Latin or Sanskrit, that may also serve a political purpose - in salving the egos of German scholars.

I think both the proposed solutions to this problem - ‘Germanic is Creole’ and ‘Indo-European went through a phase of complexification after Hittite and Germanic had split off from the core’ - are unprovable and suffer from obvious and unanswerable criticisms.

Szemerenyi notes that the consequences of the laryngeal theory have yet to be fully worked out although 100 years had passed since it was first broached by de Saussure. This sheds light on the Renfrew thesis, which after 25 years has not even been fully worked out, let alone evaluated and accepted. Maybe after 90 years it will be finally there with an answer for every language and province, and by that time everyone under 70 will be a follower of Renfrew. I can't categorise this, but the complexity of the evidence seems to be out of all proportion to the number of people qualified to work on it.

8. NWB
I am surprised at the lack of any scholarly follow-up to the major work of Kuhn, Hachmann, and Kossack.
I have the impression that Kuhn did not wish to draw attention to the links between his work and the earlier work of Gysseling, which reached similar conclusions, so that here we also have a lack of continuity: Gysseling does not use the word Nordwestblock and the efforts of Netherlandish scholars do not connect up with parallel work by German scholars. On a more emotional note, Kuhn was simply a great scholar and it is very emotional to handle the four volumes of his collected papers and ponder his originality and scope.
We are bound to ask the question whether the NWB peoples also throve in England - so close for most of history to the Netherlands. I am not aware of any work on this except Pederson, an amateur publishing on the Internet. I found his work interesting but his grasp of Celtic etymology is weak. One would expect the invasion of England by the Belgae, as described by Caesar, to bring linguistic material related to the Nordwestblock, especially if the invasions were in the 4th and 3rd Cs BC and not just in the 1st. I find it very surprising that English scholars have ignored Gysseling’s very credible and careful results on the Belgic language, when the presence of numerous Belgae in England was described already by Caesar and has been accepted by so many archaeologists. It almost looks as if English scholars do not consult work written in Flemish. Yet, the prehistory of the Netherlands must be of great importance to the study of prehistoric England.

Another major line of work on non-Germanic words in NW Europe has been pursued by Eric Hamp. He wrote in particular on the 'pig plough apple' group, referring to three words which are not inherited words in Germanic and so must have been borrowed into Germanic from another language. Hamp's work does not upset Kuhn's but also does not seem to connect to it. Hamp if I am right does not try to tie his word history to a
particular geographical space or to archaeology.

additional questions


11. How in the 4th and 5th Cs do we differentiate the Franks from the Saxons?

Some of the evidence I looked at for point 2 indicated that the Anglo-Saxons did not come from 'towards Denmark' but from the Low Countries, especially what is now Belgium. Okken's powerful summary of the language process in the prehistory of the Netherlands opens, not too explicitly, an exciting question. Were the Franks and the Saxons just two successive collective and functional names for an uninterrupted flow of marauders drawn from the whole Germanic hinterland and funnelling into the vulnerable frontier of the wealthy Roman province? If we could go back to 400 AD would we find any difference - of speech or customs - between Franks and Saxons? Or is it more that they acquired different names based on where they settled and aggregated to territorially bound authorities (compounded from arriviste marauders)?

Old English is either closest to Frisian, or to Old Saxon, or to Old Dutch. Each opinion has some linguists who support it. The evidence is difficult. We normally expect languages to have descent trees in which splits are both reversed. That is, once Italic has split from Old European (let’s say) it does not flow back and fuse with another Old European language. Any language descended from Italic is closer to any Italic language than it is to Greek or German. So Old English too should belong unambiguously with one of these North Sea languages and not the others. However, the classification idea does not work with dialects within a language: you do not say that Yorkshire dialect is closer to Somerset than to Kentish, that is meaningless. The classification depends on clear splits, where an object A is firmly not-B. The problem with the North Sea languages is that they did not split 'properly' - there was intensive exchange around the North Sea, perhaps as late as the 11th century. Cnut's Danish-English kingdom is just a political expression of the links. (Hans Kuhn wrote very persuasively on these sea-spanning links.) Looking for the affinities is a productive task, or it seems so to me. The disagreements are of great interest. Different tests give different results.

Gotthard Lerchner's exceptionally detailed analysis of cognates in Germanic languages around the North Sea starts with about 1300 selected words and then analyses their distribution in these languages in order to give a very close definition of the affinities between those languages. His prized asset is unique pairs, the idea being that where dialects A and B share unique pairs which dialect C does not, then A and B are more closely related to each other than either to C. A word which A and B and C all share is of no geographical value. Lerchner seems to have spent about three years studying a set of 1300 words, a depressing count which also suggests that better results may be reached by pursuing this method, unexhausted because it takes so much time. The dramatic result of Lerchner's work is that Anglo-Saxon is lexically closest to South Netherlands (Flemish, to use another name). That is, the direct ancestors of the English speech community (or, a dominant share of them) were living in Belgium before crossing the Channel. This is unsurprising in geographical terms. The problem (is problem the word?) is that in terms other than lexical English is, notoriously, closer to Frisian, spoken in the northernmost part of Holland (and adjacent parts of Germany and Denmark).

This can be supported by good written evidence for Saxon activity along the French Channel coast, known indeed as the 'Saxon shore'. The trouble is that this is where the Franks are supposed to have been living, after their successful invasion of what is now France, at the same time. Either there were good traffic cops in the Dark Ages, and the two streams of people were kept perfectly separate, or else perhaps there was one big flow of barbarians from the interior of Germany, beyond the frontier, and its constituents became either Franks or Saxons on arrival, becoming subjects to states which were known by those names. We can readily separate Saxon and Frankish dialects a few centuries later, when the territory of the Netherlands had a dividing line between Saxon dialects and Franconian ones (this line runs through the eastern part of the Netherlands still today), but that is not clinching proof that such separation was already there in the 4th century.

I couldn't find a study of the phonological nature of the hundreds of Frankish words in French (taudis ecurie guerre marechal berge trepigner fanfare etc.) so I do not know if they can be differentiated from early English words. Gamillscheg's classic work Romania germanica lists about 700 Frankish words in French, but Lerchner does not give an analysis of these in terms of 'shared innovations'. They are a bounded and specialised list.

O's paper [have forgotten name, sorry] analyses the place-names of the whole north-west Germanic realm and again comes to the conclusion that early English names are closest to the evidence for the Southern Netherlands and Pas de Calais. This matches the zone of Europe where the sea crossing to England is shortest and safest.


12. Frisians
The North Sea Germanic classification problem reveals a flaw in the nineteenth century heritage of Germanic linguistics, widely felt to be one of the most solid and thoroughly researched areas of the whole human sciences. This has to do with the relationship of Anglo-Saxon and Dutch. Thus, in Krahe’s standard university textbook of the Germanic languages, (originally 1942?), AS and Frisian are placed together in one branch of West Germanic, with Dutch in a separate branch. However, a quite different classification is defended in works by Klaas Heeroma and Thomas L Markey, where Dutch is shown as closer to English than Frisian. This affinity of English and Frisian is represented as the product of marginal archaic retention, as waves of innovation give an appearance of unity to spatially central dialects, and spatially marginal dialects miss out on the innovations and thus appear closer to each other only by what they omit. It is genuinely surprising that something so basic could have been missed. I suspect that the German scholars who made almost all the discoveries just weren’t very interested in Frisian or Anglo-Saxon. The Frisian/Saxon unity is known as ‘Ingvaeonic’ after a sentence in Tacitus. Heeroma’s classic paper is ‘Wat is Ingvaeoons?’.

Secondarily, the new research dissolves the Old Saxon/ Anglo-Saxon bond, so that again Old English is closer to Dutch than to Saxon as spoken in Saxony. The change owes much to a revolutionary 1955 paper by Hans Kuhn. The linguistic gap between Old Saxon and Anglo-Saxon suggests that the first English did not set out from Denmark and Saxony, but from the South Netherlands and the Pas de Calais - the ‘litus saxonicum’. They probably lived in Angeln and Saxony at an earlier date, but migrated to the Flanders area as a staging-post and sailed to England from there.

This result provides an explanation of the finding in topic 11 that the Anglo-Saxons came from the south Netherlands and not the North (or from the region between Frisia and Denmark). Heeroma gives powerful reasons for thinking that the whole Netherlands spoke a dialect resembling Frisian in early times, say the 5th and 6th centuries AD; that France, and the Frankish of France, were mighty sources of cultural authority already then (or soon after), and that waves of linguistic influence spread out from there and gradually southernised the Frankish speakers of the regions north of the Meuse; and that this was the origin of the Dutch:Frisian linguistic boundary. Heeroma gives excellent reasons for identifying quasi-Frisian (better: Ingvaeonic) relict features far to the south of what is now Frisia. England, for reasons not hard to identify, was relatively immune to these waves, and so acquired a peripheral and archaic character in relation to the Frankish speech area, which in time left it in the same category as Frisian - although at an earlier stage they were not the most closely related dialects.

The end point of the process was the reduction of bilingual Frankia, in the Seine valley (roughly) to a single language, so that the most prestigious Frankish dialect ceased to be spoken and the most influential part of the Frankish realm came to speak French only. Thus this Latin dialect acquired the name of ‘Frankish’.

If this whole account is right, then Krahe’s textbook is wrong. This suggests that adjustments to the inherited picture of the relation of English to other Germanic dialects of the North Sea area may yet be possible. Heeroma also thinks that Old Saxon is the closest language to Old English - a finding which contradicts Lerchner’s lexical study.

Searching for free material on the Internet, I came across an essay in Frisian which attacked Heeroma very thoroughly on the basis that his views of Ingvaeonic were a derogation from solid Frisian nationalist views.
(NB I rely on the summary in English as my Frisian is nugatory.) The writer was hostile to the idea that Frisia was part of the Netherlands. Thus “The linguist Klaas Hanzen Heeroma (1909-1972) wrote several articles in the thirties and fourties in which he attempted to disprove the existence of Frisian substrate outside the province of Fryslân itself, by means of a clever manipulation of the terms East-Ingweoanic and West-Ingweaonic. In this article, I relate the tenacity with which he held on to those views to his language-political background: as a staunch supporter of the “Groot-Nederlandse” (literally “Great Dutch”) movement, he viewed any attempt to establish a regional linguistic standard as a threat to the unity imposed by the Dutch standard language.“ I believe I am right in saying that Heeroma came from one of the eastern provinces around Groningen, where the language of daily intercourse is a kind of German, and that his first language was one of these Low German dialects (known as ‘Saxon’, sassisch). Much of his work has been on Low German and the history of ‘Saxon’ in the Netherlands, and he edited a magazine which was about synthesizing Low German and ‘oosters’ or ‘sassisch’. 'Istvaeonic' is a term for another alliance of Germanic peoples, in this case specifically the Franks and their language, which includes Flemish and Dutch in our terms.

I realise this is subjective, but speaking personally I thought Frisian would be easy for me, as a native English speaker who has spent a good deal of time reading Anglo-Saxon. Whereas the Frisian I have looked at has been quite puzzling. I found Dutch much easier to learn. Sample of Frisian:
>>Oars as de bêsten fan ‘e ginnerative tradysje hat Heeroma it konseptueel tinken foaral misbrûkt, mei it taalpolitike doel om troch de Fryske taal hinne de Fryske identiteit oan the taasten.<<
(Otherwise than the best of the generative tradition Heeroma mainly misused conceptual thought with the linguistic-political goal of affecting Frisian identity through the Frisian language.)

Phoenicians
Sammes has been taken up by the ‘British Israelite’ movement and there is an extensive body of fringe scholarship on ‘Hebrew-Israelite origins of Europe’, connected to fundamentalist Bible readers. There is a whole literature on supposed Hebrew names in Europe, which is essentially pre-20th century and which I have not explored. ‘Sammes, working on the supposed likeness of British place-names to
Phoenician words, proposed a wholesale colonisation complete with language and religious practices. He argued that Welsh was the surviving form of Phoenician spoken by ancient Britons who had been driven into the mountains by Saxon invaders. The book has illustrations of deities and rituals, many associated with Stonehenge, but absolutely no material evidence.’
(from http://static.royalacademy.org.uk/files/antiquaries-optimised-v6-192.pdf )

Cyrus H Gordon is a recent writer on the dissemination of Semitic languages into Europe (and North America). See Gordon, Cyrus H.: Before Columbus: Links Between the Old World and Ancient America; Crown Publishing, New York 1971.


One of the sites I found by idle internet searching was owned by the ‘Christian White Nationalist Alliance’. Tim McVeigh supposedly had Christian Identity beliefs. The sensibility is a radically white racist development of British Israelitism: >>Where else in the annals of history is there a record of nearly an entire nation suddenly converging on a wilderness? Only the migrations of the Anglo-Saxon-Gothic tribes into early Europe can fit the picture, and that occurred at the very time that Israel was dispersed and became lost to history. The Angles, Saxons, Celts, and Goths, who overspread Europe, are said to have originated in the region of Medo-Persia, about 700 BC, the very time and place in which the nation of Israel was lost to history. The early Christian church noted a remarkable fact: There was a distinct resemblance between ancient Israel’s religion and that of the early inhabitants of Europe. Early Christian writers used the Latin phrase, “Preparacio Evangelica,” meaning that European mythology constituted a good “preparation for the Gospel.” We now know why Norse mythology, Celtic Druidism, and Greek mythology all bear such striking similarities to the Old Testament -- it’s simply because these peoples were the physical descendants of ancient Israelites who migrated to Europe in ancient times, bringing deep-rooted traces of their religion with them when they came.<<
The political line here is straightforwardly to eliminate the Jews from the Bible and its message about history. This is a big issue for certain fundamentalists, who seem to divide into two groups, one which would give Israel any weapons system that exists and one of which wants to write Israel out of the Bible.
Other groups who are keen on Sammes still believe that the Earth was created in 4004 BC, in what is a piece of scholarship related to Sammes.


Substrates
There is no suggestion that the people who spoke the geminate language or any other substrate were psychologically different from the better-known Germanic. The literature does not refer to any substrate of virtuous, egalitarian, unpretentious, low status Geminate People who did not conquer anybody because they did not believe in violence (etc. etc.) This may follow, but actually we are looking at the history of particular words, not people.

A hot topic in European archaeology is the nature of the interaction between incoming Neolithic farmers and abiding Mesolithic fisher-hunter-gatherers. While this may have produced linguistic substrates, there is no evidence that the geminate language or any other was spoken by Mesolithic natives, rather than by another farming group. There is no reliable evidence that some of the Mesolithics did not speak an Indo-European language. Contact zones may quite well have included early Germanics coming south and occupying more advanced and richer areas, thus acquiring loanwords from people on a higher cultural plane than they. There are no bases for dating the substrate lendings, but the bilingual contact zone may have persisted for thousands of years. (A comparison might be the English:Welsh linguistic border, which is still there and may already have existed in the 5th century AD.)



10. Erich Roeth.
Was Illyrian spoken in Thuringia until the Middle Ages and is there a separate 'basic language' at village level obeying different rules from the German which appears in the written record?
This is the theory put forward in Roeth's book Sind wir Germanen? das Ende eines Rätsels (1967; 350 pp.)

comment. This is actual nonsense so I am only describing it for fun.
Q. Did peasants in Thuringia speak an Illyrian language instead of, or alongside, German, into the late Middle Ages?
Answer: no.

Roeth's theory claims:
-That all linguistic changes in early Europe emanated from the
Illyrians of Thuringia as the centre of events
- that there was a separate language spoken by the peasants and that you have to master this in order to understand language history; and that in Thuringia there was a double separate language, so that peasants spoke, up until a certain point (around AD 1250?) Illyrian as well as German
- that the Illyrian stratum of words bypassed the sound shifts of the past 3000 years or so and were preserved unchanged when recorded any time after 1800 AD
- the Illyrian language is characterised by a number of sound shifts identified by Roeth but apparently unknown to all other linguists
- many of the 'Illyrian' words have cognates in the Baltic languages (Lithuanian and Latvian)
-the ancestors of the Romans came from Thuringia and so what seem to be Latin words in the dialect are really ancient local words which have stayed put and not gone gallivanting round the world

I don't have a problem with the idea that a language was spoken in Central Europe before the arrival of Germanic or proto-Germanic. We could call it 'Illyrian'. Illyrian is a name attested only for an area in Yugoslavia and Albania, so I would prefer Venetic for Central Europe, but I concede that a number of books and articles in the 1930s and 1940s familiarised this name for describing cultures in central Europe which included Thuringia. (This was a current led by Julius Pokorny and Hans Krahe which was very powerful for a time but seems to have ebbed away to nothing after a while.)

No more am I against the idea that 1/3 of the German vocabulary is non-Indo-European, or that, consequently, the 'alien' words come from other languages -or that these languages were spoken by the people who made up the prehistoric cultures of Central Europe, reasonably including Lausitz, Anjetitz, and so on.

The problem is more with the engine room of language history, the processes of validating and checking individual etymologies and connections. Roth is ignoring necessary checks and as a result his findings are globally unconvincing. What he says is fanciful. For example, he finds a word in local field-names which apparently has no equivalent in standard German, traces it back to Illyrian, and claims to know what it means in Illyrian. But he has no source of Illyrian so how does he know what the word means? The field may be named after the person who owned it, the crop it carried, its shape, a bird that nested there, etc. Since he can make up the meaning and the source word, he has unlimited freedom - and his results are not worth serious consideration. We may well reflect that since he makes up the rules as well as playing the games for Illyrian words, he will win every time and other scholars cannot check his results.

Thuringia does not have great communications and is not highly urbanised. However, it is part of the North European Plain. Anything rolling through would also roll through Thuringia. It is not a likely place to find an archaic language clinging on. Also, it is in the heart of Germany, the home of the majority of the world's historical linguists since the discipline began. If there were a second language lurking in Thuringian villages then someone would have discovered it. Roth quotes quite a number of scholarly studies on the peasant culture and dialect of the region, the fruit of pervasive academic interest and pride. Yet this amazing discovery remained on the bough for him to pluck.

I think this is a work of literature. Roeth was a publisher and too much attuned to what an audience wanted to hear. Lulling you into a dream somewhere between Heimatkunde and proto Green sensibility. An agreeable saunter through a rural landscape where his ancestors lived. Abandoning any intellectual standards is part of the trip. He is animated by a local patriotism which wants history to have happened in Western Thuringia -where, so far as I can make out, nothing of significance ever happened at all. There is a certain piety about his lack of realism.

It is interesting that R's words do not coincide with the postulated 'geminates language' at all. This suggests to me that his approach is arbitrary.

Roeth was not an academic and this makes it likely that no one has bothered to follow up his results in the 40 years since they were published. This is sad in a way, but after all it could be a completely wasted effort. He wrote a second volume which remained unpublished for many years after his death, but has now come out.

13. the DO auxiliary, a supplement to point 5.
This material has been expelled because of being like a remark, 'Once you've read all the medical evidence about the Kennedy assassination you're going to wish you'd never heard of the medical evidence.' Once you've seen all the evidence on the DO auxiliary you’re going to wish you’d never heard of it.

Summary.
Older English said 'forsakest thou the devil', modern English says 'do you forsake the devil'. The latter construction is known as periphrastic. It wraps up a number of changes as a package. The ‘forsakest’ type is called ‘synthetic’ because the stem and the tense marker are ‘put together’ in the same word. In English, the word 'do' is very frequent and carries out an undifferentiated function:

don’t you do that. I didn't do it. Did you do it or not.

It is an 'auxiliary' verb, it supports other verbs. As 'it' the pronoun can stand for any inanimate noun, so 'do' is a 'pro-verb' that can stand for any verb (except be).

The thesis of the 3 Deadly Finns is that the spread of periphrastic DO is due to 'Celtic speech habits' among populations in England which had given up Celtic speech roughly half a millennium earlier.


comment
(a) The following features are organically linked:
-periphrasis
-auxiliaries
-reduction of morphology and concentration of it so that only auxiliary verbs are conjugated
-reduction of all other verbs to unchanging verb-nouns

There are quantitative, I mean within the mathematics which links information and verbal form, links between loss of word endings and the rise of auxiliaries as markers of tense and mood. It would be an error to deal with them separately. If the tense marker leaves the 'meaningful' verb it has to re-emerge in the auxiliary and not just exit from the sentence. Thus a cluster of innovations may be structurally linked - and this mandatory innovation may not be either borrowing or inheritance.

The rise of undifferentiated, unmarked, verb-nouns, implies a sweeping replacement of conjugation by periphrasis in which auxiliaries become the only marked words alongside neutral verb-nouns, and all tenses or moods are signalled by auxiliaries. This terminal phase has not been reached in English, although reports say that it has been reached in some dialects of Welsh or Gaelic (not in the contemporary written languages).

(b) The prevalence of 'do' as an auxiliary reached a peak in the 16th C and declined after that. 16th C texts use it in contexts where we would not use it.

(c) Continuous Welsh prose doesn’t turn up until about the 11th C so I am doubtful about getting at 5th C Welsh enough to determine how it influenced an arriving pre-English language. The early poetry is very stylised and does not record a wide range of linguistic situations.

The link between celticity and later English only holds if the DO auxiliary is not a feature of related continental Germanic languages. Unfortunately, the unbounded development of a DO auxiliary is a key feature of Dutch and Low German. Examples at http://listserv.linguistlist.org/cgi-bin/wa?A2=ind0506E&L=lowlands-l&P=R1114
- this is really important. It seems that to find the DO auxiliary you need to look at dialectal forms of NW Germanic languages, actually the ones closest to (putative) pre-boat-Anglo-Saxon. So in modern Low Saxon people say 'so wat dou ik nit maken' 'I don’t do a thing like that'. In a Dutch book on archaeology quoted above I found two sentences:
Men mag zelf niet graven, doch de officiele instanties doen het ook niet!
(One isn't allowed to dig, but the official bodies also do it not!)
and
Ik wil dat in het volgende proberen te doen (I will that in what follows try to do) (do stands for another verb meaning 'trail, pursue')
Both of these show the 'proverbal' use. If this construction is present in 20th C Dutch, it is hard to argue that it was completely missing from Invasion Period Anglo-Saxon. It would be equally convincing to present its development in English as a 'coastal Sprachbund' process,
with England adopting models from the near abroad, relatives of English on the coasts of the North Sea.

The idea of a 'linguistic unconscious' is to be fought off wherever possible, but it is difficult to explain why these languages could have been so wary of the DO word in their early stages, separated from each other via migration and politics, and then developed the DO word in parallel and in the same ways. Scholars have neglected plattdeutsch
because it has not produced a significant literature. But if you take Germanic as a central theme of linguistics then you have to pay attention to Platt because it may have unique evidence on some processes. We have to consider a counter thesis that the use of auxiliaries was imitated by Welsh speakers from English - the dominant partner in linguistic affairs on the island. If we accept a 'Sprachbund' in Britain then we have to consider the total process. It would follow that modern Welsh may not be purely Romano-British with zero per cent influence from English.

If we are seeing a 'North Atlantic language alliance' that includes English, Welsh, Gaelic, Dutch and Low German (at least), we would then ask where the 'leading sector' is, the source of innovations which after a lag are shared ones. This search would normally look for bilingual groups and then in particular for influential bilingual groups.

If we are looking for lurking and profound ‘speech habits’ then the simplification of declension & conjugation in 14th century England connects rather directly to the origins of Germanic in circa 1000 BC. The sets of changes bear strong resemblances to each other. The ‘buried’ speech habits would then be rather English than Celtic.

I am not persuaded that the whole process can be explained as a Celtic substrate influence spreading over France, Spain, and Britain, or even that the earliest records of Celtic show the use of auxiliaries in a significant way. In fact, the frustrating absence of DO auxiliary, present participle, etc., from early Saxon texts can be paralleled by the puzzling absence of periphrastic constructions, modal auxiliaries, etc., in early Irish and Welsh texts. The ever increasing advance of these features in the last 800 years or so has been a striking feature of Celtic history. That does not mean that the trait was already there in 600 AD - or that it can be demonstrated at all in Celtic remains in Gaul and Spain.

One question is whether the use of periphrastic verbs began to advance in Insular Celtic before it began to advance in Old English. In the earliest Welsh texts, it does not seem to be there. In the Mabinogion (late 12th C), where it is present, that seems to be a point of contrast between this prose text and the older poetry. Heinrich Wagner, the Dublin scholar and pupil of Ernst Lewy, was emphatic on the predominance of periphrasis in the late spoken Celtic languages which he studied in the field. He was actually indignant about the difference between spoken and written forms, odd as that may sound. Evidently complex morphology is a feature of the Gaelic literary language and evidently when we get records of spoken Gaelic the morphology is barely present and the use of periphrasis is highly advanced. This situation cannot be projected back to 400 AD. Yet the extensive corpus of written Celtic literature uses synthetic forms for the most part. If modern dialects have lost these complex verb and noun endings, that is a recent development in the history of the languages, and it would be very risky to project it back to 500 AD. We are looking at a linguistic revolution here, not at a literary:spoken opposition. There is a related problem that people assume that everything Celtic is ancient, so that the idea of linguistic revolutions grinds against intuition (or cultural prejudice) even when it looks like the most reasonable explanation of the evidence. Everyone goes to Ireland to find the European past, irrespective of whether it is there.

I have done a very modest search in early Celtic texts I have in the house, and the periphrastic line seems to be missing altogether. I have the impression that the condensed manner of ancient poetry excluded the use of periphrastic verbs and so used simple verbs more than prose (and than speech?). This would apply both to Old English and to early Welsh texts. So the DO auxiliary may be stylistically marked as colloquial. This reduces the value of our early texts, which are highly stylised and formal. Old English has laws and poems, Old Welsh has only poems.

Prising the story of the DO auxiliary out of this fraught historical context is dubious. My starting point was the claim of the 3 Finns that this development reflected the influence of Celtic speech habits (in monoglot English speakers of circa 800 AD to 1500 AD). To return to this, I don't find the claim proven.

(d) There is a whole book, of 1953, on the DO auxiliary in English, by the Swedish scholar Ellegård. He was aware of a supposed link with Celtic, proposed by Preusler in 1938. He investigates it and dismisses it. Incidentally, one can count 5 or 6 books about the DO word.

15. Krahe
In the 1950s, Hans Krahe studied a group of elements in river names and decided that their unity preceded the break-up of the Indo-European into separate languages. He called this phase Old European. He assumed that this group was Indo-European in nature, in line with his pre-existing belief that the oldest IE was spoken in Germany and radiated out from there. He moved into this field of study while coming out of the stage of 'Illyrian theory', which found Illyrians everywhere in Europe - a theory which now has no supporters.

Beekes points out that IE has no a, and most of the early river name-elements include an a. Also there are non-IE alternations in the oldest material. For this reason, the 'Old European' linguistic stratum cannot have been IE. The Spanish linguist, Antonio Tovar, confirmed this conclusion in his lecture reviewing Krahe. He confirmed Krahe's conclusions about distribution but denied the IE nature of the river names. A similar point was made much earlier by FBJ Kuiper. 'Old European' is a language of unknown affinity. River-names are accepted as being the oldest part of the verbal world in many regions, as words which migrants are always likely to have borrowed from the people already there. Thus many rivers in England have Celtic names, and thus Old European river-names could credibly be words which arriving Indo-Europeans learnt when they got here. This would also imply that there was a pre-Indo-European language of very wide geographical spread. (The 'no a' theory is part of the 'one vowel' theory and is not accepted by every Indo-europeanist.)

Krahe is too important to leave out. His work is some of the most stimulating in this field, although 60 years later some corrections are needed.



16. MacKenzie

WC MacKenzie in ‘Placenames of Scotland’ (1931) advances the idea that the Germanic component in Scottish was nothing to do with Angles or Saxons but was Frisian. (He excepts the Lothians, i.e. the region which includes Edinburgh and areas south and east of there.) Thus Dumfries means ‘fort of the Frisians’. As stated above, Frisian has no -n in the infinitive and Northumbrian shares this feature. A link between Frisian and the Anglian spoken in Southern Scotland thus exists. MacKenzie is unaware of this. It is plausible that the settlers who landed north of the Humber came from havens further north than those who landed in Kent, Essex, Hampshire, etc., and so that they could have included Frisians and Saxons among other groups. Frisians must have reached England, because there are villages called ‘Frisby’. Frisia is considerably closer to England than to Scotland.

He does not feel that he should check his argument by studying old languages or collecting facts of any kind. He expects to triumph simply by force of character. Why should studying language history qualify you to make pronouncements on language history? His whole position is motivated by nationalism, a wish to deny any connection between Scotland and England in the ethnic and linguistic spheres. The undeniable fact is that the Scots language is very similar to the English spoken just south of the border, which has massive resemblances to the English of the rest of England. MacKenzie is simply looking for a way to fly in the face of the facts. The Frisians (who are also Calvinists, which helps) are a convenient intellectual weapon.

This is the way things were a hundred years ago. Possibly the Internet is giving this amateur scholarship (or, amateurship) an unheard-of break, so that every bad idea will be published and available to the whole world. There was a whole world, in the 19th and 20th Cs, of books ‘of local interest’ in which amateur scholars, men of leisure without professional training in the disciplines which attracted them, wrote sentimentally on ‘place names’ or whatever and wrote down, while doing so, a wealth of charmingly crackpot ideas. Many of them collected useful information. Some of these ideas trickled on to the ‘New Age’ thing in the 1960s, with ley lines, neo-paganism, and so forth. These books, little read and little considered, were a haven for unconventional ideas. The new Internet world looks like swelling this sector to incredible dimensions. This may be beneficial for an area like the history of the Scots language, something noticeably under-worked and under-developed. However, crackpot ideas are a symptom of underdevelopment and a component of it.

In international textbooks on language Scots is simply presented as a variant of English. Haarmann’s standard work does not present Scots as a separate language or offer any count of its speakers. This is intolerable for nationalists to whom the five mile journey from just on the English side of the border to just on the Scottish side is the biggest distance in the whole universe. Scots has to have a separate history. Without this imperative (which as a genetic Scot I understand very well) MacKenzie’s theory would never have been broached. MacKenzie also presents an even more ludicrous theory that Pictish did not die out but simply evolved into Scots. Thus any resemblances between Scots and English would be due to chance coincidence.