Geographical Development Of European Languages | ||
Grover S. Krantz (1931 – 2002) Professor of Anthropology at the Washington State University Geographical Development Of European Languages American University Studies, Series XI, Anthropology and Sociology, Vol. 26 New York, Peter Lang, 1988 Chapters 4 and 5 |
||
Links |
||
http://onlinelibrary.wiley.com/doi/10.1111/j.1467-8306.1991.tb01708.x/references |
||
Posting Notes |
||
Posting introduction and Table of Contents are on the
Title Page. The posting's notes and explanations, added to the text of the author and not noted specially, are highlighted in blue font, shown in (blue italics) in parentheses and in blue boxes. Page numbers are shown at the end of the page in blue. |
CHAPTER FOUR — RACIAL PROPORTIONS
Population Mixing It is assumed here that the introduction of Neolithic farming and herding practices into Europe was accomplished primarily by an incoming wave of immigrants from the Near East. To assume otherwise would require postulating and supporting some unique circumstance for this area of the world. It is also assumed that a small proportion of the native peoples was genetically incorporated into this intruding population at a regular and cumulative rate. This section summarizes my efforts to quantify this incorporation on a geographical basis. There are two basic aspects to this population mixing. One is the introduction of Anatolian genes into the European pool by the farmers. (Primitively, the genetic barrier between the Balkan Peninsula and Anatolia would have led to a considerable contrast in the gene pools of the two areas.) The results of this intrusion can be perceived in modern populations in terms of a few classical traits like hair and eye color, and some measurements, but the contrasts are not great. Blood-type frequency distributions seem to offer a somewhat better record of this event.
The other aspect of mixing is the expansion of Caucasoid populations (whether European or Anatolian) over Mongoloid peoples in the northeastern part of Europe. The results of this mixture are much more evident in terms of visible anatomical traits, and can also be traced in the frequencies of blood types.
These two aspects overlapped in time, but they are discussed mostly in the order given above. The focus here is on the geographic and demographic factors that would have determined these mixtures, and not the results in terms of distribution of the characteristics themselves. Beyond a superficial check for consistency, I leave it to others to compare these results with detailed racial data. Anatolian-European Mixture Recently it has been suggested that agricultural practices spread into Europe by diffusion (imitation) as opposed to being carried in by a wave of immigrants (Dennel 1983). If true, this would be the only known instance of a hunting people who eagerly adopted the farming way of life. The documented history of hunter-farmer contacts around the world in recent centuries has always shown a great reluctance on the part of nomads to shift to a sedentary existence. In all known cases the spread of agriculture has been accomplished primarily by the spread of the actual farmers themselves (Smith 1976). It does not follow from this that all, or even most, of Europe's present population is descended from the Near Eastern immigrant farmers. A native admixture of only 2% for each 40 km of frontier advance would cumulatively reduce the Anatolian gene pool to only 20% of its original component at the far end of Europe. This decline would be gradual, but not linear. At each step the rate of retention of Anatolian genes would be .98, so one multiplies .98 times .98 times .98, and so on, for as many 40-km steps as are being measured. This leads to a rate of decline that is necessarily rapid at first, and becoming more gradual as it progresses.
At this rate the incoming genes would be reduced to half their frequency just to the west of Vienna,
and their remainder halved again at the border of Scotland and by lands' end in most other
directions. In all, the collective gene pool of Europe would end up about half native and half
intruder, with a strong cline from the entry point to the far west and north. The known distribution
of genetic traits in Europe conforms quite well to this description (Mennozi, Piazza, and Cavalli-Sforza
1978) .
A similar intrusion can be postulated for the first appearance of agriculture in northern India, and the distribution of genetic traits again conforms with this pattern. Agriculturalists also overran most of sub-Saharan Africa from a Negroid source just north of the Equator. In this case the percentage of native incorporation (of Pygmies, Bushmen, and Hottentots) was apparently much less, judging from their gene frequencies (Krantz 1980). Distributions of external traits also tend to follow the same trends in Europe, India, and southern Africa. Anatomical features characteristic of the source areas gradually decline in each subcontinent as one moves away from those sources. One objection to the immigration scenario is that there was not a great population reservoir in
western Anatolia at an early date to supply this influx (Dennel 1983). This objection apparently is
based on an analogy with modern migrations, where
the migrants are drawn from a vast area by sophisticated means of mass transportation. By contrast,
the earliest Neolithic expansion was strictly a frontier phenomenon. Each increment of population
expansion that needs land to farm will simply move to the nearest suitable place that is known and
accessible. At or near the frontier there are new lands available for occupation. A hundred km behind that frontier these new lands become inaccessible and probably unknown, so expansion moves into the less desirable local places immediately at hand. No great reservoir of population was needed for this kind of expansion — just an open frontier with some ordinary people who are overreproducing themselves. In the following discussion I draw the line between the European and Anatolian (Asian) gene pools at the Dardanelles and along the Caucasus Mountain divide. A similar line between North Africa and Europe is put at Gibraltar. This may seem to be a rather arbitrary geographical boundary, but it is probably the best available for present purposes because these would be the original major barriers to gene flow. The model of frontier expansion makes some reasonable, though indemonstrable, assumptions that have already been described. The farmers' initial occupation of some 500 sq km out of each block of 1600 should probably involve little or no mixing with the local inhabitants. It is their subsequent expansion into the other two-thirds of each block that leads to the absorption of some of the natives. As this would begin in earnest only after the initial colonizing activity had passed on by more than 40 km, the local foragers would find it increasingly difficult to survive or even to escape envelopment. As the farmers eventually fill out a 40-km square to a density of one person per sq km they reach a population of about 1600. If they incorporate 2% of their own numbers in each such block (i.e. hunters that would constitute numeric 2% of farmers, or 32 out of 1600), this means taking in 32 out of an original hunter population of 160. The other 128 hunters can be accounted for by a mixture of killing and dispersing some, and absorbing others without allowing them the opportunity to reproduce. Actually, the hunter population in this square would have been reduced to more like 110 when the farmers took the best one-third of the area; the missing 50 could have moved elsewhere as stragglers. Then taking in 32 people out of a trapped 110 constitutes a 30% absorption rate of the natives. This of factor 30% is used throughout this reconstruction.
Where adverse climatic conditions significantly delay the frontier advance, this model should still hold for each 40-km geographical unit, on the average. The added time allows for more casual interbreeding between invaders and natives. Extra native genes would enter the farmers' gene pool. Extra farmer genes would also enter the natives' gene pool. These displaced farmer genes will be recovered as the frontier advances, thus negating the extra incorporation of native genes. Presumably, those natives who most resemble the farmers are those most likely to be added to the frontier population. This incorporation model may be followed over most of Europe, from Greece all the way to Sweden, Ireland, and Spain. At a .98 retention rate for each 40-km movement, these results are mapped at 10% intervals in Fig. 14. This shows the percentage of western Anatolian genes that ought to have survived in each area. With a few changes discussed later, this should also be a close approximation of the present condition.
A major exception to the above model occurs in eastern Europe outside of the Balkans. Within this area it already has been proposed that a pastoralist economy had spread out from a source in the Hungarian Plain, and this was composed entirely of native Europeans. Their population would have been 0.25 per sq km, or two and a half times that of native foragers. Thus, when farmers overran this area (excluding the plain itself) they would outnumber these herdsmen by only four to one, instead of the usual ten to one as elsewhere. The same model for farmers overrunning hunters is followed here, with just one basic change being required. In the original model 30% of the natives who were trapped behind the frontier were incorporated into the farming population. The same factor of 30% can be used here, but now applied to a larger base number of natives. In a given 40-km square the herder population would ideally be 400, at .25 per sq km. Upon losing the prime one-third of the area to the farmers they are reduced to 275 people. When the farmers fill the rest of the square and absorb 30% of these, it means 82 people are taken in and the remainder are neutralized in other ways. These 82 people will then constitute 5% of the population of 1600 farmers. In other words, the genetic retention rate will shift from .98 down to .95 when farmers overrun each 40 km of herders. In those parts of eastern Europe where Uralic pastoralists preceded the Indo-European farmers, a .95 retention rate for Anatolian genes has been used in the map (Fig. 14). In the Hungarian Plain itself the farmers were mostly excluded, but they heavily penetrated some parts of it, so the retention figure for European genes was arbitrarily put at 60 to 70%, thus matching the adjacent peoples. The population of the lower Danube is similarly assigned the same gene frequency as its neighboring farmers. Beyond Lithuania and Belorussia the advance of the Indo-Europeans was halted for a long time, so the farmers beyond here would be 100% European. The Indo-European frontier also halted part way up the Scandinavian Peninsula at the 120-day growing-season line. Part of the Neolithic technology, primarily cattle pastoralism, would have been transferred to the native Nordics who were completely European in their genetic make-up.
Still one other halt in the Indo-European advance apparently occurred in the highlands of Scotland. Here farming is almost impossible, but sheep herding and fishing do well. The Sea Shepherds out of the Aegean are almost certainly the source of the pastoral population here (Again, not the E1b1b Aegeans, but R1b Celtics). Sea Shepherds would have been fully Anatolian in genetic make-up at their source point in Rhodes. There is no obvious logical structure to calculate their rate of incorporation of native peoples following the model of other population movements. The major difference is that these Boat People spread inland from the coasts, generally only short distances at first, so native hunters would not tend to become trapped by an engulfing frontier of new peoples. Also, their postulated advance rate of 200 km per generation is the fastest of any in this reconstruction. The normal rate of genetic retention for herders overrunning hunters is shown below to be .92 per generation, or .815 per 200- km advance. This low rate should not apply to the Sea Shepherds for the reasons noted in the last paragraph. The rate should be higher than the .92 per generation, and I have arbitrarily put it at .95 simply as an educated guess. Allowing two units of mixing in crossing the Aegean, and 54 more 200-km units to reach Scotland, the Anatolian component would have been reduced to just 6%. Thus they were 94% European in their genetic make-up at the end. In contrast, the farmers who reached that point somewhat later in time were reduced in 70 steps of 40 km, but with only a 2% admixture in each, so they should have arrived in Scotland with 23.8% Anatolian genes (The Celtic Scots are R1b 77.1%, R1a 6.6% and I 11.2%, other admixtures 5%. They are 77% non-European, agglutinative, R1b Asian Kurgan folks). If the native foragers had continued to survive in Scotland, a transfer of sheep to their care would leave the Highlands 100% European in genetics. But if the Boat People had prevailed and occupied the area, as is most probable, they might have been about 6% Anatolian. If they mixed significantly with natives the result would have been somewhere between zero and 6% European, but closer to the latter. In any case the cline of Anatolian genes would drop from 23.8% near the Scottish border, down to about one-fourth of that amount. Caucasoid-Mongoloid Mixture The spread of pastoral people to the north and east out of the Hungarian Plain, in a wave that was followed for a short distance by farmers, is the major one that had racial implications of the visible type. These herders were of pure Caucasoid race (not to be confused with Caucasian languages) in terms of such classical traits as skin, hair, and eye color; hair form and distribution; and nose form and body type. In their advance to the north-east these herders encountered and overran hunters of the Mongoloid racial type (Fig. 15). The mostly invisible genetic contrasts between Europeans and Anatolians, as discussed above, are here combined into the commonly understood racial type of the West.
The Caucasoid-Mongoloid racial line used here is based on the most likely areas of development of their specialized traits during the last glacial epoch, and on their expectable northward expansions as the ice melted. I have allocated all of subglacial Europe to the Caucasoids, also the area around the Black Sea, the Caucasus Mountains, and Iran. Mongoloids are assumed to have lived east of the Caspian Sea and throughout central Asia; they were the more cold-adapted people. With climatic amelioration both racial types would expand northward, but the Mongoloids had the
distinct adaptational advantage and would have spread well to the west of their original area. The
ultimate racial line can be drawn from the north shore of the Caspian Sea to Leningrad and extending
it through Scandinavia, but leaving all of Finland on the Mongoloid side because of the early Baltic
Sea barrier. This line in Scandinavia also fits well with the last melting away of the local ice
sheet, where early hunters would meet after entering the peninsula from opposite ends at the same
time. This dividing line might appear to be highly speculative, but any attempt to place it as much
as 200 km in either direction would raise serious geographical problems, if not impossibilities.
Obviously there would have been considerable racial mixture along this line, with a dilution of both
types for some distance to either side. But for purposes of counting gene frequencies at greater
distances, a sharp line or a mixed zone would both produce the same results. The same general model for genetic mixture can be followed here, just as in the previous cases, where farmers were the overrunning population. The herding intruders here move in 80-kra squares, first occupying the prime area of a 50-km circle that is arbitrarily placed in its center. The 6400 sq km of the square originally should be inhabited by some 640 native hunters. A tribal unit of 500 herders then moves in to occupy the 2000 sq km of ideal territory, displacing 200 of the hunters and leaving 440 of them behind. The herders then expand into the entire square with a final population of 1600. If they also incorporate the usual 30% of the 440 remaining hunters, these constitute 132 of the new herder population. These 132 are now 8% of the 1600 herders in this 80-km square. This means a genetic retention rate of .92 for each step of 80 km, and is the equivalent of a rate of .96 for a 40-km farmer-sized move. For purposes of long- range mapping, the three cases may be compared.
A new complication enters the picture when the pastoralists change their emphasis from cattle to reindeer at the line of pure coniferous forests (Figs. 12 and 15). This was postulated to have reduced their density of occupation, and increased the size of their tribal areas and made their rate of advance faster. It may be assumed that hunters in this environment were similarly reduced in density, so no correction need be made for this aspect. The only significant difference would be the increased distance these herders now advance in each generation. Previously this was expressed in the reduced basic time for them to advance each 200-km unit, from 75 down to 60 years, plus penalties. In this context the movement of reindeer herders should first be thought of as an increase from 80
to 100 km for their advance per basic generation. With the same genetic retention rate per
generation of .92, this is now their rate per 100 km.
Their rate per 200-km unit is then the square of this, or a .8464 retention. The reindeer boundary on the main Finnic (suggested Caucasoid) path is crossed 200 km before they enter Mongoloid territory. On the Samoyed (i.e. Nenets, suggested Caucasoid) path it is crossed 200 km after passing this racial line. The Vogul/Ostiaks (suggested Caucasoid) cross it 400 km into the Mongoloid area. The expected retentions of Caucasoid racial genes may now be mapped with careful attention paid to the rates according to which type of animal is presumed to be the major domesticate. The results of this are shown on the Eurasian map in Fig. 15.
The move of the Finnic branch through Finland and into the Scandinavian Peninsula enters Mongoloid territory already with the reindeer adaptation. They overrun and incorporate these natives with a retention rate for their Caucasoid traits of almost .85 per 200-km unit. Their component of Caucasoid genes begins with 100% at Leningrad, declines to 51% at the head of the Gulf of Bothnia, and reaches a low of 37% when they cross back over the racial border in mid-Scandinavia. After that point they are incorporating more Caucasoid genes from the early hunters who entered Scandinavia from the other direction. Their Caucasoid proportion then rises to about 53% when they finally meet the Indo-European farmers and Nordic herders. Within the total area of greater Lappland, including two-thirds of modern Finland above the farming line, the overall average of Caucasoid genetic background works out to 50%. This would have been the condition as of 3130 BC when the herders met in Scandinavia and closed out the last of the pure hunters. During the 5000 years since then, these Lapps have been in contact with northern Europeans in and
around much of their territory, but with almost no contact with Mongoloid populations. The
expectable gene exchange with their more numerous neighbors would have somewhat increased their
proportion of Caucasoid genes. It is also likely that social selection, especially in more recent
centuries, would have favored the European physical appearance beyond their total genetic
composition, thus further increasing the Lapps' resemblance to their fully Caucasoid neighbors. The Samoyed (i.e. Nenets) path to the north-east passed over Mongoloid territory for 200 km with presumed cattle
breeding. The retention rate of Caucasoid genes would be .815 for this segment. After this the
reindeer herding rate of .8464 applies to the rest of their expansion. At this rate their Caucasoid
component is reduced to 50% in 800 km, and halved again to 25% in just over
800 more km. At the eastern limit of Samoyed (i.e. Nenets) distribution in the Taymyr Peninsula it reaches a low
of only 7%. The recent Samoyed (i.e. Nenets) distribution runs from 1000 km beyond the old Mongoloid line all the way to the Taymyr Peninsula. The Caucasoid component for the area ranges from 42.5% down to 7.1% across this distance in the initial spread. This gives an average of just under one-fourth Caucasoid genes, and three- fourths Mongoloid. This conforms reasonably well with their observed physical traits. They would have reached their eastern limit at 880 BC according to the present schedule. The subsequent almost 3000 years of internal mixing should have evened out the east-to-west discrepancy to a considerable degree. Further selection for Arctic-adapted Mongoloid traits should also have occurred in place, thus increasing their Monogloid appearance. The Vogul/Ostiak history is similar to that of the Samoyeds (i.e. Nenets). The distance traveled and timing involved is close enough to theirs, so that similar Caucasoid-Mongoloid proportions should have been strung out over roughly the same west-to-east distribution. Both this group and the above Samoyeds (i.e. Nenets) are spread over such large areas that significant genetic contributions from outside were unlikely. Only with the recent Russian penetration and control of the entire region has there been much mixture, especially along the southern fringe. The racial make-up of Europe presented here remained constant until the ethnic movements began in
500 BC. The subsequent changes are mostly minor because population increase, in place, was the
general rule. Some estimates of these later changes are summarized in the later chapter on the Final
Adjustment, and are shown in Figs. 29 and 30. |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
CHAPTER FIVE — GERMANIC SHIFT
The first major break in the Indo-European dialect mesh, at least in Europe, resulted from the Germanic expansion out of Scandinavia. This, and a few other events, inaugurated a series of changes that totally broke up and redistributed the European language picture over the following two thousand years. Farming peoples have the ability to evacuate their homelands under adverse circumstances, while hunters cannot. This is possible because they can take in their last crop and move, taking this food-and-seed with them. They also may have better information about where a more suitable environment might be found. More important is their ability to concentrate in large numbers for a move through a narrow space. This concentration can give them immediate military superiority over almost any resistance. Of course, upon settling in a new area they must necessarily disperse themselves to about the same degree as the local residents, and then their temporary advantage is lost. Simple migrations of farmers, into already farmed areas, are usually absorbed by the previous
residents within just a few centuries. If the intruders and residents share similar technologies,
economies, and cultural prestige, then it is just a matter of time and mate exchange until the more
numerous residents linguistically absorb the newcomers. Ordinarily, such intrusions number no more
than a few thousand, or maybe tens of thousands, of people moving into a region that is many times
more populous. Migrations of much larger numbers would be decimated by problems of sanitation and
disease if they were similarly concentrated. A very large migration might avoid such problems by
subdividing into many separate units. But these would usually be distributed in various directions,
each of which would ultimately meet the same fate of all other such smaller migrations. There are some special circumstances under which migrating peoples can and clearly have maintained their ethnic identity indefinitely. If the group occupies an insular or peninsular location in which it becomes the sole or major population, then it might persist. The physical absence or small number of immediate neighbors might serve to insulate them against linguistic absorption. The other circumstance is where the separate elements of a very large migration maintain communication and close proximity to one another. In this manner they might divide the resident population into relatively small groups between their own immigrant elements. Under ideal circumstances it could be the divided groups of residents who become linguistically absorbed on a large scale. It was apparently a massive migration of just this kind out of Scandinavia that moved into and created what eventually became Germany. Exodus from Sweden The Sub-Atlantic phase of climatic deterioration in Europe was a relatively mild episode that
effected farming only along its northern fringe of greatest vulnerability. It probably was not even
noticed on most of the continent. Its major effect of cooler and damper climate
devastated the Scandinavian Peninsula; the East Baltic region apparently was not much bothered, but
this is disputed (Fodor 1976 vs Kivikoski 1967). The Sub-Atlantic spans the last five centuries BC.
During its most intense part, from 400 to 50 BC, Sweden almost disappears from the archeological
record (Pounds 1973:14, Scott 1977:9, and Stenberger 1962:109). During the late Bronze Age, just before this episode, about one-third of present Sweden had been occupied by farmers — with much of it very densely populated. The coast was occupied halfway up the Gulf of Bothnia. The Norwegian coast was then held by Europeans (i.e. farmers, linguistically presumed Germanic, genetic mixture unstated) as far north as Narvik. By 400 BC Swedish occupation was reduced to just a thin coastal strip in the south, and the offshore islands. Norway was almost emptied, and its coastal occupation drew back to near Bergen, at about the latitude of Stockholm. Denmark does not appear to have been affected. (See Fig. 16 for the emptied areas.)
The obvious conclusion is that nearly the entire population of Sweden, and most of Norway as well, simply went away somewhere. These areas began to be repopulated after about 50 BC, presumably by recovery from the fringe settlements and by new immigration, mainly from Denmark. Other explanations have been offered for this lack of an archeological record for nearly four centuries. The developing Iron Age on the continent at this time has been credited by some authorities with cutting off the affected area from trade and communication, hence the scarcity of datable remains (Stenberger 1962). I fail to understand why such a unique event as this cutoff should have occurred here. More seriously, some Swedish scholars I have talked to (but failed to note their names) think that
there are enough undated Bronze Age graves to fill the gap if they are mostly assigned to this time
period. This strikes me as dubious for two reasons.
The percentage of undatable graves ought to be roughly constant for all centuries, but we are to
suppose that they concentrate on just the afflicted time period. This is so odd as to require some
special explanation that is not forthcoming. Also, we do in
fact have a continuous record on the coast and islands where the material is clearly datable to the
otherwise empty period. We can safely proceed on the assumption that the major part of the Bronze Age population of Sweden simply moved out. Being farmers rather than hunters, they had this option. It is also evident from the record that they left rather abruptly — the entire exodus must have taken place in less than a century. The only logical direction for them to go was south. It is tempting to see this event as the origin of the German-speaking population on the continent. Before we can draw this conclusion it is necessary to show that such a move is consistent with the rigorous approach used in the rest of this reconstruction. The recent German area, and especially the rough mapping of the earliest Germanic occupation here, conforms closely with the combined drainage of the Rhine and Elbe Rivers, and the lesser rivers between them. I will go further and postulate that this was exactly their initial area of intrusion. The Oder drainage to the east, and the Vistula basin still farther east, may also have-been invaded. Judging from the following calculations, and also from the later pattern of linguistic events, it would appear that these eastern rivers received no more than minimal Germanic immigration, which did not long persist.
As Swedish Bronze Age agriculture was failing in the 5th Century BC, the affected people would
naturally be looking for more suitable lands to the south. They would have had enough geographical
knowledge at that time to know that warmer climates lay in this direction. They would also know that
the land in this direction was already occupied by settled farmers (and
horse pastoralist tribes of Sarmatian circle). There was enough time for
explorers to discover (if they didn't already know) that beyond the Danish area there were river
valleys that were not well populated. Most of the Celtics in the area of modern Germany farmed the
better-drained, lighter, upland soils away from the rivers (Pounds 1973), and some of the
Scandinavians would certainly know this. A major ecological opening was available at that time for a
population that was already oriented to exploit it. For transportation, Scandinavia was well supplied with large boats. These are well known, indirectly, from two sources. Rock engravings and other art work of the Bronze Age show unmistakable boats, often shown containing numerous men with oars. Of similar age are burial places or monuments — patches of ground outlined with many upright stones in the shape of a ship. It is not presumed that a ship was actually buried at such sites, but the size and shape of one is clearly represented in each case. Most of these range from 6 m to 20 m in length, with a few measuring much more (Stenberger 1962) (Typologically, the burials are clearly connected with the R1b Kurgans). There has been much discussion as to the construction of these ships, whether they were skin-covered frames, or were solidly planked. The rock engravings have been interpreted by many authorities to represent the wooden ribs of a ship showing through an ox-hide covering. A simple consideration of their absolute size rules this out. Ever-larger ship frames can be made from timbers of increasing sizes, but ox hides have a fixed thickness and strength. Beyond a few meters length, such a ship would be so heavy that even a gentle scrape against shore gravel would tear through any hide. Multiple layers of tanned animal skins only postpones the inevitable failure. While we do not yet have actual remains of ships from exactly this time, there is one dated to the early 3rd Century BC from Denmark. This specimen is 10 m long and is built of planks (Clark 1977:203). Planked craft, much like the later Viking ships, would be capable of transporting large numbers of people to new locations. Let us suppose a million such people are to be moved out of the southern one-third of Sweden into the middle of the western German area in half a century. The people would weigh an average of 60 kg (including infants and children) and may have had
baggage, equipment, and livestock of 240 kg each, for an estimated total of 300 kg per person.
(Modern refugees often move with only a small fraction of this weight.) A fair-sized Bronze Age ship
of 15 m could carry 15 to 20 metric tons of cargo, thus about 50 people with their possessions, and
a crew of perhaps ten. At just walking speed, and moving by day only, such a ship could travel from central Sweden to half-way up one of the German rivers and return in less than three months. (Some trips would be longer, most shorter.) Two trips each summer and, allowing for bad weather days, just one winter trip could easily be made by each ship. One fair-sized ship could then transport 150 people per year to their new homes at a leisurely pace. In 50 years a million people could be transported this distance by 133 such ships. If a thousand ships were available (a low estimate) this entire population could have been moved in less than seven years. Obviously, most of the existing ships would continue to be used for ordinary commercial purposes, fishing, and war during this time. But then, Danish sources would probably have had at least an equal number of ships that would have been available for this move, giving a total of maybe 2000 craft. In a 50-year period, the mass migration of a million heavily supplied people could have been accomplished with the diversion of only about seven percent of the available sea transport. The Swedish immigration into Germany need not be viewed as a military invasion. Rather, it would have been more like a massive infiltration of settlers into the lightly occupied river valleys. Of course if resistance were encountered, the intruders would have had overwhelming numbers of armed men available at each of their limited paths of entry. These rivers would have been familiar to many commercial shippers from the Denmark area, and no doubt to a few from Sweden and Norway as well. They may also have been regularly visited and fished, with occasional settlers being introduced even before this great migration. The geographic pattern of the move is both simple and regular. Emigrants from Sweden might naturally
assume that a warmer climate was the general necessity, although it was only warmer summers that
were actually needed. They would accordingly have felt compelled to concentrate their intrusion into
those rivers located to the west of Denmark — to the east winters were just as cold as in southern
Sweden. We can presume a limit to their adventurousness and the distances they were willing to travel, which would also be based on their area of prior familiarity. This limit could be put at 200 km to either side of the base of the Danish peninsula. To the east, in the least favored direction, this gives them access to the coast and lesser streams, but stops short of the Oder river. To the west they would enter the Elbe, the lesser rivers Weser and Ems, and reach into the Zuider Zee. Within this sea they would have access to the Isel River (also Ysel or Ejssel) which is one of the mouths of the Rhine. Progressing up the Isel they enter the main part of the Rhine upriver from its joining with the Maas (Meuse) River. If they were consistent in moving only upriver, then the drainage of the Maas would be omitted from their area of occupation. From this entry pattern we simply follow their advance up every branch of each of these rivers to the limits of their drainage basins and stop there. During the Roman conquest of Gaul, and over the next few centuries, note was made of the Celtic-Germanic boundary in much of this region. A few tribal assignments in Roman sources may be incorrect, but generally the Germanic area reported at that time corresponds closely with the line based on these drainage basins. My exclusion of the major mouths of the Rhine as well as the Maas tributary was in fact originally suggested by this early mapping of the Germanic tribes. Only later did I find this made perfect geographical sense. Occupation of Germany At first glance it might seem that an exodus from just one-third of a small country like Sweden could not possibly have overwhelmed the entire Celtic population of such a large area as modern Germany. However, more careful examination of the details shows it to be quite feasible, even inevitable.
The land area of Sweden that was seriously farmed in the late Bronze Age, and then evacuated, amounts to some 98,600 sq km by my measurements. In the 2nd Century AD it is estimated that the northern part of this, called "central" Sweden, supported a population of at least half a million people, at a density of something like ten per sq km. (Here and elsewhere I have used population estimates from Pounds 1973.) This would involve half of the farmed part of Sweden. For the remainder of settled Scandinavia, as well as for Germany, a density as low as five per sq km is indicated. Seven centuries earlier (5th c. BC), Bronze-Age Sweden would have supported at least as many people. Their settlement actually extended somewhat farther to the north at that time, as compared to the early Iron Age (2nd c. AD), and there were no significant improvements in agriculture during this interval. Assuming the same densities in 500 BC as for 700 years later (2nd c. AD), the other parts of Sweden and evacuated Norway would have held another 472,500 people. Adding these to the population of central Sweden, this gives us the potential for an exodus of 972,500 people. (I have not figured in the native Scandinavian cattle herders of the interior, nor the coastal people to the north. Some of them may have been involved, but this is not certain. Probably they spread into the evacuated areas.) At six persons to each sq km in the defined basins of the Rhine and Elbe, a high estimate of the target population in western Germany would be about 2,472,000 people in an area of 412,000 sq km. (Duby, 1973, estimates a population less than half this great, but he no doubt would also have reduced the Swedish numbers correspondingly if he had dealt with that area.) The residents should then outnumber the intruders by well over two to one; the invaders having just 28.2% of the combined populations. Casual examination of rivers and tributaries in the German area shows 6,275 km of total length of
significant streams that would be easily navigable for even the largest ships of the time. There is
also a northern coastline of 750 km that is of easy and immediate access. The likely scenario would be for these
infiltrating-invading Scandinavians first to occupy the coast and both banks of all these rivers to
an average depth of ten km in all directions. (The actual occupation zone would vary greatly from
place to place, but this may be used as an average for purposes of computation.) The area of these
shoreline strips adds up to a surprising 133,000 sq km, or 32.3% of the total.
(See Fig. 16 for a map of this initial area of occupation.) If the intruding population were spread
evenly over all this area, to the exclusion of the resident natives, their density would be 7.42 per
sq km. This is about the average for the entire source area, and is well under that estimated for
central Sweden. This riverine zone of the German area was originally the least populated part; as noted above, the higher lands were the most favored for their easy farming. If the density of native riverine residents is taken as two per sq km, then some 266,000 people would be living in these strips. This would be only 10.76% of the total Celtic population of the Rhine-Elbe basins, who were occupying 32.3% of the total land area. It may be assumed that the infiltrating Scandinavians would absorb these people into their own groups. Some residents might move away from the shores, while others would be attracted to the newcomers, and these should about balance. The intruders' 972,500 would be augmented by the assimilation of these 266,000, for an immediate total of 1,238,500. In this first action in a limited area the intruders outnumbered the affected residents by almost 4 to 1, so these could have been incorporated with no difficulty. This augmenting raises the population density along these river strips to 9.3 per sq km — almost the same as back in central Sweden, and a workable figure as will be seen. After the subtraction of these 266,000 people, the remaining Celtics between the riverine zones
would number 2,206,000. They would be occupying a badly subdivided area of 279,000 sq km, with a
density of 7.9. At this point the remaining Celtics will still outnumber the intruding Germanics,
but the ratio is reduced to less than 2 to 1. Now the intruders have 32.3% of the land and within the whole river drainage area. The number of rivers and their combined lengths that are used here is somewhat arbitrary. During the first century of intrusion the riverine occupation may have been somewhat less than the total; given time, the occupation could well exceed that calculated here. The rivers figured here are simply those that are shown most regularly on various European maps. The 10-km distance from each bank is based on an easy day's walk to the limit and back. The density of Celtic population along these rivers was arbitrarily set at just one-third of the average reported for the entire country. This riverine population density would then be one-fourth of the density in the remaining higher lands where most of the farming was concentrated. It just happened to work out that when the resident Celtics of the rivers are added to the Scandinavian intruders, their total would fill these strips to almost the ten per sq km, just as in the best Swedish lands. The ecological basis for this newly achieved, high density of riverine settlement must also be considered. Some explanation is needed for the Swedish population here being about four times greater than it was for the natives before they arrived. The two obvious means are fishing and short-term crops. Fishing skills were increasingly developed as Indo-European farmers expanded to the north.
Dietarily essential vitamin D for winter consumption was obtained from sun-dried fish. Water craft
and other appropriate fishing technology were devised and adopted from resident hunters, as well as
from the Aegean Boat People who were encountered there. Most of the ships from the affected Swedish
and Norwegian areas would ultimately end up in Germany, along with the skills to build more. The
German river system offered an underutilized ecological opportunity to exploit this large aquatic
food supply. This developmental gap eventually would have been closed by a slow filtering to the
south of the appropriate skills, and this was no doubt underway. But the process was fortuitously
accelerated by the climatic adjustment that drove large numbers of skilled fishermen into the area during one short time period. Scandinavian farmers had long been situated up against the limit of the 120-day growing season. This limit is somewhat arbitrary because it depends on the crop, and fluctuates from year to year. Still, it marks the approximate limit of farming that was reached by the first settlers around 3000 BC. Yearly variations in the length of the growing season would have adversely affected crop yields on many occasions. This problem would be most frequent near this upper limiting line, but at times it would have affected the greater part of inhabited Sweden and Norway. In bad seasons often only part of the grain crop would mature. This would cause a strong natural selection favoring those variations that matured slightly early; conscious selection by the farmers is not required. The automatic result of short-season varieties may also be at the cost of losing some other desirable qualities. Seed from these fast-growing strains would have been brought by the settlers into the German area and would be sown in the river valleys. The German rivers, especially the Rhine, have a spring flooding as they drain meltwater from the snowfall in higher altitude areas to the south. Accordingly, their bottomlands cannot be tilled and sown until much later in the season than the surrounding uplands. This results in a growing season that was too short for the Celtic farmers to use. However, selected northern seed was better able to mature in the available time. Scandinavians were able to take large yields from this unusually fertile soil. Again, the techniques and seed might eventually have filtered down to this area, but it was a climatic shift that brought them in abruptly at this time. Celtic and Germanic peoples would find their respective territories interfingered in a complex
pattern, with each group of one ethnic type partially surrounded by groups of the other. With total
population figures for the entire region still running almost 2 to 1 in favor of the Celtics, they
eventually would have assimilated the Germanics if all other factors were equal.
Several additional factors tipped the balance in favor of the Germanics. (Throughout this discussion I often use "Germanics" for Germanic-speaking people,
"Celtics" for Celtic speakers, and so on, for easy flow in the narrative. I am sorry if this
shorthand form bothers some readers.) The scenario of intrusion given here is somewhat artificial for the sake of simple presentation. I have described how the river zones were occupied as a first stage, which was followed by the taking of the intervening higher lands. Actually, many of these between-river spaces in the north would have been taken early in the movement, and long before the river zones were filled in their more distant southern parts. The actual process of filling in between the rivers was probably progressing as a "second front" a few hundred km behind the advance wave up the rivers themselves. The time taken to evacuate Sweden also imposes some constraints on this model. If the evacuation was accomplished in 50 years, there had to be enough land available at the end of this time to support all these people. If the riverine settlements had progressed just half of the 600 km straight-line distance through Germany in these 50 years, they would have acquired 66,500 sq km of river zone territory. The invaders’ full 972,500 people, plus 133,000 from half of the Celtic riverine population that would be incorporated at this point, add up to 1,105,500 Germanics (The genetic composition shows that transition was smooth, and that is reflected in the Classical writings, with the blurred difference between Germanics and Celtics is still being chewed over by the modern scholars). At the previously calculated density of 9.3 per sq km, only 609,150 of these could be settled in the partially occupied river zones. The other 496,350 must be settled in other places. At the assumed upland density of 7.9 per sq km (as with the Celtic farmers), another 62,800 sq km are needed to locate the rest of the Germanics. In order to take this much territory the full Germanic advance over the land area must have moved almost 150 km, or one-fourth of the total distance. Out of a total linear distance of 600 km, if the riverine advance had reached 300 km in 50 years, the total occupation must have gone half of that, or 150 km. A new aspect to the intrusion now appears in the above version of upland occupation following
shortly behind the
riverine expansion. No provision was made for the Celtic population that had been living on these
uplands. Many Germanics would have been lost in the inevitable wars, and others lost through
temporarily strained refugee living conditions. Their places would have been filled by many
thousands of Germanized Celts. I cannot imagine that any more than half of these upland Celtics
would have been incorporated into the Germanic ranks.
That still leaves the other half with nowhere to go but out.
The exodus of these Celtics is discussed in the next section. Something like this double pattern of advance is probably close to what actually happened. The advance wave of riverine occupation would eventually absorb and Germanize all of the 266.000 Celtics who lived there. The second wave of upland occupation that followed behind would Germanize about half of the Celtic inhabitants (ultimately 1,103,000 people) over a protracted period of time, and force the departure of the other half. The numbers involved can now be reviewed in terms of the originally described two-stage occupation of Germany. The initial riverine occupation by 972,500 Scandinavians would absorb 266,000 natives and become a population of 1,238,500. These would then assault the upland Celtics in a piecemeal fashion and eventually absorb 1,103,000 of them in modest increments. This leaves 2,341,500 Germanic people in a previously Celtic land, while another 1,103,000 of these Celtics have gone elsewhere under pressure. It still remains to be explained why the invaders should have prevailed over the residents so often that such an impressive Celtic withdrawal would have occurred. River shipping of large volumes of produce gives a significant economic advantage to those with the equipment and locations to accomplish it. They can also transport large populations both up and down the rivers as circumstances call for. Military reversals in local conflicts need not be total for those, who have ships. The simple expedient of crossing a river can often provide a safe retreat — an opportunity that is denied to land-based foes under comparable circumstances. It can safely be presumed that the recent mass migration out of Scandinavia and the continuing upriver settling would mean these people had a degree of cooperation and coordination far greater than that found on the Celtic side (Celtic side was on both sides). Finally, there is the simple fact that these northerners were bigger and stronger, on the average, than their Celtic adversaries. In combat with hand weapons this gives a substantial advantage to one side, all else being equal. The effects of a similar body-size discrepancy between Celtics and the still smaller Mediterraneans were often noted by classical writers among the latter (which allows to think that not much difference in body size was between two Celtic sides). Absolute body size in this part of the world varies with latitude in a regular geographic manner. It must be stressed that the details given here of exact land areas, populations, and densities are actually only approximate. The succession of events is also not necessarily exact. This picture is built up from the most probable parts known, and the result is not only inevitable, but it accords well with historical facts that were recorded by civilized people around the edges. Details might well have varied, but the general picture is probably true as given here. The most precise, and hopefully testable, prediction here is that German ultimately relates more closely to ancient Swedish than it does to Danish (Danes with Y-DNA Hg's I 38.7%, R1a 16.7%, R1b 41.7%, E1b1b 2.9% little differ from the Germanics and Swedes). The numbers and distributions of ethnic forces given here would have been just barely sufficient to
ensure ultimate Germanic dominance of the Elbe-Rhine drainage basins. Another factor that probably
contributed to their success was the sparse population of the region surrounding the Low Countries.
The presense of non-Germanic and non-Celtic place names here indicates that it must have been
ill-filled in the first wave of Neolithic settlement. Barraclough (1978:85) maps this area as
including the coastal 30.3% of the drainage basins under consideration here. If we assume this place
held just half the number of people as in other Celtic areas, then an across-the- board reduction of
15.2% can be made. Applying this correction factor throughout all the previous calculations would
mean that A large ethnic group of native Mesolithic people (reference to Y-DNA Hg's I) continued to exist as cattle herders for some time in the interior of the southern part of the Scandinavian peninsula. We do not know exactly when these (Y-DNA Hg's I) Nordics were assimilated into the Germanic agricultural population. If they still existed as a separate entity at the time of the evacuation of Swedish farmers, they could have expanded to fill some of that void. (Remaining Swedes, reverting to a Mesolithic economy with some pastoralism, could have filled it just as easily. Lappish reindeer herders may also have temporarily extended their area.) It is much more likely that these Nordic natives were absorbed into the Germanic speech community well before this time of exodus. These people are the only obvious source of the non-Indo-European vocabulary found in the Germanic languages. Since this includes German itself, the absorption must have occurred before the southward move began in 500 BC.
Celtic Diaspora According to many prehistoric reconstructions a great expansion of Celtic peoples began about 800 BC. This was supposedly because of their military and commercial power that was based on the ironworking in their central European homeland. Archeological evidence of the spread of Hallstatt iron work was probably just that — and no human movement need be supposed. Only after 500 BC is there clear indication of human migrations, and their explanation given here turns out to be quite different from the usual image of a conquering ethnic group. An ethnic expansion of the Celtics is no more logical than the supposed ethnic spread of the Kurgans out of the Pontic Steppes.
During the intrusion of Scandinavian immigrants into central Europe large numbers of the Celtic natives evidently left this area. The entire Rhine through Elbe drainage basins has an area of 412,000 sq km, and was then the home of about 2,472,000 people. The considerations developed above indicate that nearly half of this population, an estimated 1,103,000 people, departed under pressure within the period of 450 to 50 BC. Massive movements of Celtics were noted by Roman and other historians during this same time. These migrations clearly stemmed from a source generally in the German area, just as postulated here. While some authorities assume they are refugees fleeing from advancing Germanics, many others see this as part of an overall Celtic expansion on its own merits. There are no natural factors that would appear to support this latter interpretation. A ’’domino" effect is described here where Scandinavians make one shift, and this causes Celtics to make the second. The lasting effects of these two moves, however, are totally different. A presumed 972,500 Scandinavians made a concentrated move in one direction into a clearly deliminated (delimitated) territory. The circumstances were such that the ethnic composition of these intruders would be preserved and was automatically imposed on those of the fragmented resident populations that remained. The exodus of 1,103,000 Celtics was dispersed to the west, south, and east, in at least six identifiable groups. They were locally successful in conquering pieces of territory and maintaining their ethnic identity for some time. In the long run, however, their dispersion was too wide-spread for any of them to have preserved their languages longer than about four centuries.
If we assume these Celtics left in only the six groups identified here (an unlikely assumption),
they each could have numbered 184,000 people. More likely, these six groups were only about half
this large, while the remainder was distributed in many smaller movements. A wide dispersal of so
many groups, some of them very large and others small, could easily create the impression of an expanding ethnic type. This is especially true if they were militarily powerful,
and if their spread was accompanied by the diffusion of advanced ironworking techniques. The actual
cause of this movement has been obscured by the fact that historical sources were often so situated
as to observe the Celtics, but largely missed the driving force behind them. The Belgae of northernmost Gaul were noted by Caesar in the middle of the First Century BC, and he
was told that they had recently come from the region of Germany. These were Celtic people, in large
numbers, who had also entered Britain by that time. The area in question of northern Gaul should
have supported a minimum population of 450,000 at six per sq km. Obviously, even a full one-sixth part of the Celtic dispersal (184,000 people) would have constituted no more than a rather influential, and possibly a conquering, force. They apparently did not have the benefit of any major new technology, nor any unique geographical circumstances, that would have enabled them to make an ethnic replacement in settled agricultural regions. But is is possible that they were the first Indo-European farmers to occupy some parts of the Low Countries that up to then had remained in Mesolithic hands. In most of both Gaul and Britain they should have made little difference because the previous residents were also Celtics.
Similar intrusions are known to have occurred in Iberia in the 5th Century BC. These would therefore have been among the earliest Celtic groups leaving the German area. They are known from place names to have settled in various locations over half of the peninsula. The resultant ethnic mix is sometimes called Celto-Iberians or Celtiberians. Only in the northwestern corner did their language apparently persist until Roman times where it formed a distinct dialect after Latinization. This corner, still known as Galicia, was semi-protected from linguistic assimilation by the fact that there were no neighbors along half of its boundary. Assuming the Celtics entered Iberia around both ends of the Pyrenees, we could count these as two of the six major groups that exited from Germany.
Also in the 5th Century BC a massive Celtic intrusion occurred in the Po valley of northern Italy. This was previously a thinly settled area, and these "Cisalpine Gauls" were a powerful force. They were suppressed by the Romans some 400 years later, and ultimately assimilated.
Part of southern Poland (mostly in Ukraine) is still known as Galicia from the ancient Gauls, or Celtic settlers from the same diaspora. The Galatians of central Anatolia were well known to Christian missionaries of the First Century AD. They supposedly arrived early in the 3rd Century BC, and eventually disappeared as a linguistic entity. The movements into Italy, Poland, and Anatolia may be considered as three of the
six major Celtic migrations. (See Fig. 17 for a map of all six.) If this reconstruction of events is
correct, the absence of many more migrating Celtic groups should become increasingly conspicuous.
While this Celtic dispersal apparently involved a slightly larger population than the initial Scandinavian intrusion that ultimately forced them out, they left no more than indirect linguistic evidence of their fate. They do not appear on my language maps that summarize each time period. The predictable geographical and technological circumstances that assured Scandinavian success on the continent were equally definitive in leading to the ultimate disappearance of these migrating Celtics. Language Map Linguistic distributions in Europe at the end of this phase of Germanic migration overlap somewhat with the changes that are introduced by the next major event, the Roman Empire. These two events can be separated arbitrarily at the year One (Fig. 18). This assumes the Germanic move was complete, and that Sweden and Norway were repopulated — these were not quite true at that time. It also assumes that Latin had not yet supplanted any of the languages that it later did, though this was actually beginning to happen.
The map differs in just a few particulars from that of a few thousand years earlier. These can be mentioned in their logical, and somewhat temporal, order of priority. The separate ethnic group of cattle herders in Scandinavia is gone. These Nordics almost certainly
had merged with the northernmost Indo-Europeans before the Germanic move began in 500 BC. This
would represent the beginning of the Germanic separation which Lockwood (1969) argues must have
occurred in a small area in relative isolation from the rest of Indo-European. The Germanic expansion onto the continent changed their boundary from what may have been only a dialect transition into that of a fully bounded language. From this time on, its own dialect mesh drifts independently of the rest of the Indo-Europeans. This may have been the first break in the dialect mesh of the original Indo-European spread. When the Germanic expansion reached the Alps it cut off the western Celtics from all contact with the rest of the Indo- European mesh. This continuity had previously been maintained by a chain of dialects around the eastern Alps. The language line across southern France could no longer be circumvented, and Celtic also becomes free to drift on its own course. For some three centuries Greek had been the socially dominant language of the eastern Mediterranean and was being adopted by various peoples around the Aegean and up the Balkan Peninsula. This amounts to a folding back of one language over several dialects, though by adoption rather than by human replacement. It is clear that Greek had acquired linguistic independence from the rest of Indo-European by this time. Of course the Greek of Linear B was distinct from written Hittite a thousand years before, but the possibility of a chain of connecting dialects of nonliterate peoples between them at this time cannot easily be dismissed. Greek expansion also served to separate the original Indo-European mesh into its Asian and European halves and set them free to drift independently of one another. While little attention is called to this Greek phenomenon, it is actually a small version of the next major event that centers on Rome. Greek certainly had a much wider distribution in the cities of the eastern Mediterranean than is shown here, but it is primarily the rural speech that is estimated in this map. The category "Anatolian" continues to be used for all the Indo-European languages of Asia without distinguishing the Iranian area. Other language breaks might have been created in the Balkans by the repeated incursions of Kurgan
horsemen out of the Ukraine. These are known from archeological records going back to well
over 4000 BC, and are here assumed to have been made by Altaic speakers. Domestication of the
horse enhanced the military capacity of these people, but it would not have increased their
population, nor in any other way enable them to make ethnic conversions. I am now inclined to think
that the Indo-European dialect mesh survived their activities in good order |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||