In Russian (later)
Ogur and Oguz
Türkic and European Genetic distance
Classification of Türkic languages
Indo-European, Dravidian, and Rigveda
Türkic, Slavic and Iranian
Türkic in English
Türkic in Romance
Alans in Pyrenees
Türkic in Greek
Türkic in Slavic
|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
Introduction and Conclusion
We can only wonder how in such short time, a historical blink of an eye, our understanding of the migrations and diffusions grew so much, due to the development of the genetics and DNA genealogy. At the time of this book writing very little was positively known about pre-historic migrations: neither dates, nor routs, nor participants were known. In the following period, many pre-conceived Eurocentric notions were refuted. In this work some of them are taken for granted, others are justifiably rejected. Google gives to this work of Professor Grover S. Krantz a respectful 2000 book references. The weakness of G. Krantz's Eurocentric arguments does not overly compromise the basic concept, where the paradigm of the driving force for migration, in lieu of mobility gained by pastoral horse nomads sometime in the 6th mill. BC, is substituted with a spread of farming started sometime in the 8th mill. BC. For any educated person, it is a "must know" reading.
Insights: Logical and predictable expansions:
Dating and phylogeny for genetic Y-DNA marker of major groups in the text:
Tested and dated European Paleolithic remains in the table are highlighted in bold. Haplogroups C and D are Mongoloid, E thru T are Caucasoid. The phylogeny gives fairly clear identification of European Paleolithic with Hg I (Paleolithic languages), European Paleolithic with Hg G2a (Paleolithic languages, agglutinative daughter languages), farmer expansion to Europe with Hg J (Semitic languages), Uralic expansion to Europe with Hg N (Fennic agglutinative languages), Central Asian expansion to Europe with Hg R1 (Nostratic languages, > proto-Indo-European R1a + I + J, and proto-Türkic agglutinative R1a + R1b + N areal groups). The Caucasus Hg G2 (agglutinative languages) remained locked in the mountains, the presence of Hg G2 (G2a1а) in the Caucasus does not exceed 5000 years, likely the age of escape, together with R1a, from the overland Kurgan R1b's devastation in Europe.
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.
Grover S. Krantz (1931 – 2002)
|Table of Contents|
CHAPTER ONE — INTRODUCTION
Purpose and Method
This work presents a new approach for reconstructing the geography of prehistoric language differentiations and distributions. It is made explicit in the form of a complete reconstruction of the origin and distribution of all present European languages, and many extinct ones as well. Both the method itself and the results of its application to Europe are offered for critical evaluation by scholars in the appropriate fields.
History has usually concentrated on such things as battles, rulers, and political boundaries. It
can also be said that ethnic history, as measured by language groupings, is an equally important
aspect that is often ignored or given only perfunctory treatment. Prior to written records, the
usual subject material of history is missing, by definition, and informed speculations are then made
about ethnic distributions over time and space. I think that many of these speculations are at least
ill-founded, and are often incorrect. It has proven notoriously difficult to correlate archeology
with language history.
What is needed is a rigorous approach to reconstructing linguistic behavior in a geographical setting — one that recognizes the biology and economy of the human carriers of these languages. Such an approach should be founded on principles of human ecology, as well as linguistics, that are applicable to both the prehistoric and historic eras as well. This would take the form of a set of explicit rules of movement that are consistent with known human behavior, and which are universally applied. Such a set of rules would be set in motion at some remote time in the past and then consistently followed through to end up with the actual modern language distribution (Fig. 1).
This kind of reconstruction must also show reasonable agreement with the known facts of archeology, history, and linguistics, though it doesn't have to agree with the current speculations in these fields.
A reconstruction of this type is a model that may be tested against other kinds of reconstructions, most of which are not criticized here in any detail. Some archeological and historic data were used to set the times and locations of major events, but otherwise this model was developed to stand on its own merits and internal consistency. This approach is explicitly not a compilation of all available knowledge, from all fields, in order to create such a reconstruction. Instead, it is an essentially independent method whose results should be compared with reconstructions based on other methods.
The first of two principles of human movement to establish here is that people and their languages generally do not move.
All peoples will tend to occupy their own territory until some compelling force is brought to
bear to make them change their location (or their ethnic identity). When a group expands into new
territory, this is normally accomplished by a relocation of only the population surplus. There is no
tendency to evacuate most of the residents from their original homeland as long as they can still
live there. It is normally in the best short-term survival interest of all ethnic groups to maintain
their locations and languages as tenaciously as possible. Any suggestion of an instance of contrary
behavior should be accompanied by a sufficient explanation for it. I first argued this point in
print in 1976, it was also stated by Brues (1977:250-51), and it has been elaborated on by others
This means we cannot simply suppose that a group of barbarians was able to spread out and impose its language over large areas of both civilized and noncivilized peoples. Further, great migrations of nonliterate peasants cannot be postulated at will to account for more recent distributions. We cannot just move people all over the map solely to accommodate proposed linguistic connections. Anthropologically speaking, any migration proposal must include answers the following questions:
1. Why did the
move at all rather than simply remain where they were?
If any of these questions are unanswered, by statement or implication, the proposal is correspondingly weak. If none of them are answered, then it is hardly a proposal at all. To postulate a migration without giving answers to these questions is only to restate a linguistic distribution in other terms, as though such a restatement somehow explained something.
In 1977 I applied this principle of stability to a reconstruction of the immigration of the first
humans into western North America. In that article my typical answers to the above migration
1. They did not actually
move, but instead relocated their surplus population as an expansion of their territory.
It was thereby shown that at least 95% of the affiliations of the languages and their distributions in western North America could be exactly accounted for by using this, and one other, basic principle of human movements. Most particularly, the concept of a series of "waves" of immigration into the continent had to be ruled out as being inconsistent with the principle of stability. The few discrepancies between my reconstruction and some standard linguistic opinions were emphasized in four comments by Baumhoff, Elsasser, Fleisher, and Leonhardy — all 1978.
The second principle used in the North American reconstruction was that a technological innovation could so increase the numbers, power, and/or prestige of one group that they could overwhelm their neighbors. This provides specific answers to point 5, how they prevailed if the land was inhabited.
Individuals, families, and small groups will cross ethnic borders; this is a normal process at
all times. The usual result is that these migrants soon adopt the language of their new neighbors
and linguistically disappear. A great differential in population densities will mean there is a
correspondingly greater flow in one direction. If this one-way excess in flow is great enough and
fast enough for them to maintain regular contact with other immigrants, then they will keep their
language and can instead turn the natives into a minority that is eventually absorbed. As long as
the advantage of one group over the other is maintained, the linguistic frontier will continue to
move in the direction of lower population density.
With most innovations the neighboring peoples will eventually adopt the new item, increase their own influence, and then resist further encroachment by matching numbers, power, and/or prestige. These people will then press on their neighbors in turn, still farther from the source of the innovation. In this manner a series of linguistic boundary shifts, all in the same direction, were systematically described and explained for western North America.
In reconstructing the linguistic development of Europe one encounters a few innovations that moved ethnic boundaries in this manner. The most outstanding of these was clearly the initial appearance of agriculture, which swept over the continent with such force that the local natives were rarely able to adopt it. This Neolithic revolution would have caused so nearly total an ethnic replacement that this event is taken here as the effective base-line for all subsequent linguistic events. That this also represents the introduction of Indo-European languages into the continent is the most obvious hypothesis to consider first.
My North American reconstruction also included a major episode of depopulation and subsequent repopulation of large areas that resulted in major shifts of some language boundaries.
A comparable climatic event is identified in this European reconstruction, but its effect on farmers turns out to be quite different from its effect on hunters.
When a climatic change makes an area uninhabitable to hunters they simply decline in numbers in
that place, with no more than stragglers moving elsewhere. When it becomes obvious that the change
is not just a minor episode, they might want to shift their tribal territory somewhere else. But all
adjacent livable spaces are already occupied by other tribes who may also be suffering, yet who are
not so weak as those desiring to move in. Without large quantities of transportable food supplies,
the afflicted population has no recourse but to gradually fade out of existence.
With an agricultural population under similar pressure, there is at least the option of taking in their last crop, gathering together in large numbers, and moving against some small part of their neighbors. Lacking some advantageous technological innovation, such true migrations are normally doomed to eventual linguistic extinction. In the European reconstruction there is one major exception of this kind that survived and flourished because of special geographical circumstances. This is the one I like to call the Germanic Shift.
In many previous studies of European linguistic prehistory the authors have proposed substantial migrations of large bodies of people. These are sometimes pictured as nation-sized populations that shift their locations as units, making and/or breaking linguistic contacts. Individual tribes under stress, even groups of a few thousand people, are able to make this kind of move where the rear echelon is in contact with, and deliberately following, the leading edge. Some reflection shows that a peaceful, settled, farming population, numbering in the hundreds of thousands, cannot possibly move like this as a unit.
Suppose, for example, that a peasant group inhabits an area of 500-km diameter, and is going to
shift its location 50 km in one direction. One method would be that all of the included farmers
simultaneously load up their possessions and move their homes 50 km in the same direction. (We may
ignore how this would be timed to avoid interfering with their farming operations and with each
other.) Those at the lead of this move might welcome the new lands, but we must inquire why this
should be so, and how they prevailed. Those at the rear might welcome the relief from some natural
or human pressure, but we would like to know exactly what this was. For the great mass in the center
this would just be a 50-km move from one farming site to another equivalent, but unfamiliar one, and
to no apparent gain. Who or what could compel them all to make this move?
Alternatively, the great mass could be left in place, while those in the rear abandon their lands and move through the others to take up new positions in the front. An extraordinary set of circumstances could facilitate a move like this, but those circumstances must be clearly spelled out.
If one proposes that a large ethnic group shifts its collective location, this should be described in terms of two processes. There must be some force that is physically or ethnically eating them up from behind, while the other front is advancing successfully for some other reason. These factors may or may not operate at the same time, and the main body of people in the center is probably little affected by events at either frontier.
In recent years some detailed archeological investigations and reinterpretations have indicated that many previously supposed migrations might not actually represent ethnic replacements (Dennel 1983). While there may have been major changes in the civilized superstructures, the arts and crafts of the basic producers (workers and peasants) mostly continued unchanged. Even well-known periods of major destruction, as in Greece in the 2nd Millennium BC, do not necessarily represent ethnic changes (Xirotiris 1980).
The above considerations may appear to put insurmountable restrictions on any attempt to reconstruct linguistic prehistory, and especially to extend such a reconstruction up through historical times. But there are also three considerations that allow for some flexibility that might have been unexpected.
First, there are the time depths usually ascribed to the differentiation of the various European
languages. These are extrapolated from more recent, known rates of linguistic change. But there is
actually no compelling reason to assume that the languages of preliterate peoples changed at the
same rate. In fact, there is evidence that suggests a slower change in earlier times (Pulleyblank
1981). This at least opens the possibility of more time depth in which to encompass the necessary
movements, and a greater range of archeological events that might be included to explain these
movements. (I return to this subject in the concluding chapter.)
Second is the question of when these languages began their major divergence. Most current reconstructions of Indo-European languages begin with a dialect mesh in a limited area that began its spread and subdividing as much as 6400 year ago, at about 4400 BC (Gimbutas 1977). However, I find compelling reasons to think that Indo-European began to spread more than 3000 years earlier, at 8000 BC. On the other hand, it may have retained its integrity as a dialect mesh at least as recently as is commonly supposed, though over an enormous area. In fact, I find no clear argument for a break in that diverging mesh until much later — easily up to 2000 BC, maybe further. (It should be noted that a dialect mesh, as in recent India, can have fully distinct languages in its widely distant parts.)
Third is the proper identification of certain prehistoric languages. Scraps of linguistic evidence like a few monuments, some place names, and personal names of leaders, may in some cases be the products only of an elite group that has taken control of an area. It must be remembered throughout that the linguistic picture being reconstructed here is that of the real people, not just their leaders. A glance through any historical atlas of Europe will show how little relationship there usually has been between the political units and the known linguistic boundaries of the people at those times. Historically impressive movements and conquests have had major impacts on the lives of those who wrote the history, but often without having had any comparable effect on the basic population.
In a preliminary paper on reconstructing the linguistic picture of Europe (1981) I made the case
that the original spread of the Indo-European languages could have been accomplished only by the
first Neolithic farmers. Quite simply, no other human cultural force since the origin of modern man
could have had the power and persistence to cause a language spread of this magnitude. The origin
place of these Indo-Europeans accordingly would be in the area of southern Anatolia
The currently popular Kurgan hypothesis especially fails for lack of the ability of these marauders to change the languages of entrenched peasantry over so much of the world. There is no possible resource base in the preagricultural Ukraine that could have sustained a dense population, let alone fed them out in such numbers as to linguistically convert almost all the people from Ireland to Bangladesh. That such a conversion was based on conquests with the use of horsedrawn chariots is easily dismissed. Indic and Latin have common terminology for their use, yet they were invented (2000 BC) well after the separation of these two languages. Thus "...references to the horsedrawn chariot in IE ritual and mythology are probably later accretions and not derived from a common IE myth." (Mallory 1981: 216, 217.)
Moving on into the detailed distribution of Indo-European subdivisions one finds even further support for the Anatolian source of this family, as well as for the antiquity indicated here. Any attempt to draw Indo-European speech from the Ukraine at 4400 BC would make it tend to show natural divisions that in no way resemble the actual distributions, ancient or modern. On the other hand, the substance of this study shows that all of the Indo-European subdivisions, relationships, and locations follow automatically from a Near Eastern origin at the beginning of the Neolithic, some 10,000 years ago. Further, even the non-Indo-European languages of Europe are shown to be distributed in their correct ways as part of this same rigorous reconstruction.
The following picture of linguistic geography does not challenge any conventional classifications
of known languages, as my North American reconstruction did in a few instances. The unexpected
conclusions here are mainly in the area of increased antiquity ascribed to the original
Indo-European dispersion itself, and in the longer residence indicated for some of its subdivisions
in their present locations. This would include, for example, developing Greek in its present area
since 6500 BC, and Celtic in Ireland since 3500 BC. The antiquity of Magyar in Hungary may be
equally surprising; I find it to be a Mesolithic speech that predates the Neolithic entry.
The linguistic affiliations of many prehistoric populations are clearly predicted in this reconstruction, as indicated above, including the non-Indo-Europeans. The most interesting of these are the archeologically famous Kurgans of southern Russia, who turn out necessarily to be Altaic speakers, and specifically of the western, or Turkic branch. Perhaps less surprising is the proposed affiliation of the original Iberian farmers with the Afro-Asian language family.
In at least one major instance the commonly assumed direction of migration of a population is reversed here. It is usually stated that the Uralic Magyars moved into Hungary from an eastern source in the 9th Century AD. I find instead that all the other Uralic speakers expanded out of Hungary in the opposite direction, and at a much earlier date.
Particular details in this reconstruction, like those mentioned above, may not prove to be a major source of objection. These details are predictive, and many will be tested in the near future. Instead, I expect that the method itself may prove to be the most uncomfortable aspect to some readers. The problem is that so little room is left for "historical" explanations — great leaders, important decisions, ambitions, accidents, apathy, luck, and all the rest of what people believe in. Yet, the absence of this "human element" in the study of ethnic history should be no more disturbing than is its absence in such personal things as disease and heredity.
Notes on sources
The vast majority of the information given here is common knowledge within the various fields of
anthropology, linguistics, history, and geography. The juxtaposition of these well-known facts is
what is new. It may be considered improper, even insulting to the reader, to cite sources for what
can be found in basic text books. It is impossible to cite sources for what is original. For these
two reasons there are fewer specific references used here than might be expected.
Because many readers will be well-versed in only two or three of the basic fields involved here, mention should be made of my major sources of information. Some of these are drawn upon heavily throughout the text, along with other general works, but are referenced only for particularly unusual, unexpected, or critical points. Many of these are not cited beyond their mention here.
For most of the geographic data I have used the 2nd, 11th, and 15th editions of Goode's Atlas (Goode 1943, Espenshade 1960 and 1978), and the Oxford Regional Economic Atlas of the USSR. (Economic... 1956). Modern language distributions are drawn from Zaborski (1960), and certain boundaries for my base line of 1939 are drawn from Jastrow (1943). Basic data and comparisons among the Indo-European languages are largely from Lockwood (1969 and 1972) .
Conventional theories about the origin of the Indo-European peoples come from a diversity, of sources ranging from the works of V.G. Childe (1926) through M. Gimbutas (1973 and 1977). Interesting maps, as well as some speculations on prehistory, are found in Kinder and Hilgemann (1964) and in Barraclough (1978). For early technology and demography I found Forbes (1958) , White (1962 and 1967), and Pounds (1973) to be the most useful on a topic that is rarely touched on by others.
Archeological interpretations, especially as relating to migrations, are dealt with by Dennel
(1983), but he goes even further than I would in denying them. Human reproductive rates and early
population densities are discussed by Hassan (1981). Some ideas on Neolithic migrations and their
rates are found in Ammerman (1977). The basic idea that Indo-European languages were spread by food
producers was at least suggested by Smith (1976) in a brilliant work on the significance of the
Neolithic. Other sources are mentioned as pertinent.
Many of these works also give information with which I am in total disagreement. Such information is all in the form of interpretations and/or speculations that have often unwittingly become elevated to the level of "factual" knowledge. In a few instances I am only offering alternative possibilities to the commonly held presumptions. In all such cases I have attempted to indicate the differences of opinion, as well as the degree of speculation being indulged in here.
A dialect mesh is a network of tribal speech communities that are mutually intelligible adjacently, but which become increasingly divergent with distance. Hunting tribes of typically 500 persons, at a density of 0.1 per sq km, would occupy areas averaging 5000 sq km, with diameters of 80 km. At 400 km, five tribal boundaries are crossed and the cumulative effect of five dialects is a different language in a mature dialect mesh. At 800 km, and ten dialects removed, no discernable relationship is likely to be evident from direct comparison without considering the intermediaries.
This description applies to dialect meshes of great antiquity and stability, like the Australian
aboriginal languages.. Such a mesh will change gradually over time as a unit, with each of its parts
maintaining their degree of distinction from every other. A new dialect mesh develops when a
language spreads over a new territory of great size. Distinctions between tribal dialects will be
small at first, but after thousands of years they increase to the levels of contrast noted above. At
10,000 years ago I would assume that all of Europe and the Near East constituted a single dialect
mesh. This may have been continuous with the speech of most of the rest of the inhabited world as
With pastoralists, and even more so with farmers, the area occupied by a 500-person tribal unit necessarily decreases as population density increases. This would shorten the geographic distances between unintelligible and unrelatable dialects within a mesh. However, these Neolithic economies are so recent in appearance that there has not been enough time for their dialects to diverge over space even as much as those of hunters. In still more recent times, especially in Europe, ease of communication and transportation has stopped dialect divergence altogether.
Linguistic innovations may appear anywhere in the mesh, and will spread in varying directions at various speeds. At any one time the dialect mesh is thoroughly and almost randomly crisscrossed by hundreds of isoglosses, each of which separates two areas in terms of a single linguistic trait. An innovation that appears at one place is capable of eventually spreading to any and all other places in the mesh. This is analogous to a widespread biological species with genetic continuity throughout its range, but with gradually increasing morphological differences over distance.
When a dialect mesh is broken, its parts become free to evolve in their own independent
directions. Through the loss of intermediate and mutually intelligible dialects, two or more parts
of the original mesh will begin to diverge by what we call linguistic drift. Again, this is
analogous to a biological species that loses genetic continuity between two or more extremes that
subsequently cannot exchange genes, even indirectly, and thus become separate species. Bringing
together two divergent parts of a mesh (or a species) may produce between them a line of
unintelligibility (infertility), but as long as they remain connected through some route of
intermediate dialects (populations) they are not truly separate languages (species). A freely
drifting and separate language requires a full boundary of mutual unintelligibility from all others.
Just as in biology, there are indeterminate situations where innovation flow is just badly
restricted, and it is not clear whether the division should be labeled as dialect or language
(subspecies or species).
Bounded languages are created from a dialect mesh by moving one speech community out of its natural position and into contact with a contrasting linguistic environment, several dialects or more distant. If this is done without leaving a chain of connecting dialects, then a new language has been set free to diverge (drift) from its former mesh mates. It is not necessary that the actual human speakers must make the indicated move, although they often do. A change in the language spoken by one group may similarly put a barrier of unintelligibility between it and some of its neighbors.
It is argued here that human groups do not change their locations, or their languages, as a common practice. Such changes require compelling reasons. Any well-thought-out reconstruction of such events in the past must indicate these reasons.
The basic method of moving a speech community into another location on the dialect continuum is to expand its area at the expense of its neighbors. Those at the source area can eliminate, absorb, or linguistically convert their immediate neighbors, and thus have their speech come to occupy this new area in addition to their original territory. If this process is extended over a sufficient distance, two unlike speech communities will become juxtaposed, and a language barrier comes to exist between them. Such an expansion may be based on a population increase at the source that enables these people to infiltrate and/or conquer their neighbors and thus spread their own ethnic identity. It may also follow from sufficient cultural prestige at the source, that adjacent communities copy their language in ever-widening circles of imitation. These methods may be combined in various proportions in particular instances.
A roll-up of intermediate dialects may progress along a physical or linguistic peninsula, thus
assuring that no connections continue around the edges, and such a language then becomes fully
bounded. Alternatively, the roll-up may be radial — spreading out in all directions from its
source and thus serving to complete its boundary.
An advancing linguistic frontier can also separate additional languages by breaking their continuity with each other. A dialect chain along a coast line, for instance, might have its central part overrun from an inland source, thus splitting it into two independently drifting parts. Or a peninsula could have its base overrun and become separated from its previous contacts with dialects on the main land mass.
When an advancing language frontier encounters a physical barrier of some size, like a mountain chain or a body of water, it should divide and pass around each side of it. The resulting two arms of the frontier may eventually meet one another on the far side of this barrier. If the separation was of long duration an actual language boundary may exist between them at this meeting. As long as a dialect chain trails back from each arm to their point of initial separation neither side will become a fully bounded language. But if any part of this U-shaped dialect chain is removed, by any process, then the two arms become free-drifting languages.
A very different source of dialect contrasts that often leads to language separations occurs when
an intruding language overrides a large area that previously contained some linguistic diversity.
The original population might not be eliminated, and sometimes not even greatly outnumbered by the
overriding ethnic type. Many of the first speakers of the new language here are natives, and
actually grew up speaking one of the original languages of the area. They will continue to use
various aspects of their own pronunciation, vocabulary, and grammar — and their children will tend
to copy these. The dividing lines between these underlying languages often will automatically
surface as subdivisions in the new language of the area. Ideally, this phenomenon should be
demonstrated in each instance by documenting the particular linguistic phenomena of the preceding
language that appear in the same geographic area in the replacing language. Even without such
documentation the phenomenon can safely be inferred in many instances. This has been termed
"Mischsprachung," and has received less attention than I think it deserves. Several examples appear
in this reconstruction.
The history of European languages can be reconstructed in all its essentials by using a surprisingly small number of applications of the principles discussed above:
1. An initial wave of several different immigrant
Neolithic populations spread themselves and their languages over a sparse substratum of native
hunters throughout Europe and nearby areas. These major groups, in order of significance, are
Indo-European, Uralic, Altaic, Afro-Asian, and Caucasian.
Throughout these developing steps lesser distinctions appeared as a result of breaking dialect chains, from substratum language differences resurfacing, and from the shifting of technological innovations from one advancing group to another.
All of this linguistic history can be shown to be triggered by just a handful of technological
and cultural phenomena, and to follow from the uniform application of a set of simple rules of human
movement. Detailing this history, as the reader will see, involves lengthy descriptions - that
relate the above factors to the actual geography of Europe. But throughout, the basic simplicity of
the whole process should be evident.
Rules of Movement
The movements of human populations that serve to change the locations of language boundaries must be described in terms of consistently applied rules. Where exceptions are made, these too must follow describable rules of general applicability. These requirements parallel the rules of linguistic changes themselves, and so should come as no surprise. If something as personal as one's own language follows natural laws, then it ought to be expected that the location of one's home might equally follow from natural laws. Individuals are free to vary (within limits) in both speech and location, but human groups are much more predictable under given circumstances.
In order to describe the movements of early European populations I have devised a set of rules that are at least consistent with known human behavior. Within a limited range of possibilities, I have settled on those which generate a pattern of movements that is also consistent with most of the known archeological dates for the spread of Neolithic economies. Variations in these rules of movement have been made to suit different economies and climates. At later times, and within the scope of the historical record, other adjustments are made in these rules to conform with certain changing circumstances.
The basic rule of movement is that a population has the ability to double its numbers in each generation of 25 years. This will be true under the most favorable circumstances and where there is unlimited opportunity to expand its area of occupation. Further, this ability to expand will be applicable only to those small groups on a frontier of advancing occupation. Populations located well behind such a frontier will have neither the opportunity nor the means to expand in that direction, and often would not even be aware that such a frontier existed. Such populations can expand only by finding the means to increase their density of occupation in place.
Since any frontier expansion (in Europe) is always over a territory already occupied by some
other population, normally at a much lower density, an additional five years is added to compensate
for the losses of personnel in the inevitable minor conflicts. For all subsequent calculations I
simply use the figure of a 30-year generation for the doubling time to allow for this constant
The basic social unit used here is the 500-person tribe that is typical for hunting and gathering peoples, and is applied to all economies initially. Larger units come into play only later. The most probable population densities and their consequent tribal dimensions can be given here:
The typical pattern for an advancing farming frontier is for the surplus population to move out from filled areas into the most convenient new places to farm. On the frontier this will normally mean moving a short distance ahead into an ideal location rather than continuing to fill less desirable places closer at hand.
It is assumed that on the average only about one-third of the potentially usable land will be occupied, to any significant degree, in the first generation of occupation. Only after the farming frontier has passed on will the less desirable two-thirds of the land be occupied by the expanding population that no longer has access to truly new lands. At such a time the overall density will then reach the typical 1.0 per sq km. (See Fig. 2, upper.)
The gradual filling of the frontier zone can be visualized as a row of squares, 40 km to a side,
as shown in Fig. 2, lower. Each of these is occupied by 500 people who farm the best spots,
which can be represented by an arbitrary circle in the center with a diameter of 25 km. This circle
encloses 500 sq km, and is almost one-third of the area of the whole 40-km square. Over the next
generation each of these partially filled squares will generate 500 new people who are looking for
new places to farm.
These people will mostly move out into the next row of 40-km squares that have not yet been farmed. It would be more practical to move as far as 40 km to an ideal place rather than to move a lesser distance to a poorer place. This amounts to about 17 people per year, or three small families, making a move of this kind in each square.
At the end of a 30-year generation this new 25-km circle should be filled, and the source group will then continue to expand its population by filling in the less desirable lands immediately at hand. For them to continue moving away and into the best lands from this same source area, this would mean a move of 80 km, and this becomes impractical. New advances are now made from the newly filled 25-km circle discussed above. Actual frontier advancement would not have followed exactly this regular pattern of filling circles and squares, but this diagrams the process in the form of an easily visualized process. For purposes of mapping, this advancing rate is given in terms of 200-km units, thus of five generations, and requiring 150 years to accomplish.
Following this same reasoning, a frontier of herdsmen advancing over hunters will add successive tribes of 500 people in the circles of ideal territory, this time with a diameter of 40 km each. These will again constitute almost one-third of the potential area of a typical frontier square now measuring 80 km to a side. Given the same 30-year population doubling rate, these pastoralists should advance their frontier 80 km per generation. They will move the standard unit of 200 km in just 75 years, or twice as fast as the farmers. After their frontier has passed on, further population expansion will be into the less desirable areas immediately at hand, and the overall density then brought up to 0.25 per sq km.
A model for the initial advance of a hunting population would follow a similar pattern, but
faster for two reasons. They would cover 120 km in each generation because of their sparser
distribution. Also, each generation would be only 25 years because there would be no previous human
inhabitants to contend with. (Hunters cannot overrun territory of other hunters at the same density
and technological level.) This process is not detailed here because their occupation of the entire
continent is assumed from the outset.
Additional delay factors are introduced when the advancing farming frontier enters areas of low
winter temperatures and shortened daylight hours. It is assumed here that a January mean temperature
will have no appreciable effect until this reaches below the freezing level. At this point winter
problems will lead to enough unusual casualties to increase the time for doubling the population
from 30 up to 40 years. It must be noted that this ten-year penalty applies only to the earliest
people entering the colder winter zone, and that later adaptations would reduce or even eliminate
the problems. But the frontier advance rate applies only to those people initially moving in, and
later adaptations would have no effect on this.
In terms of the basic 200-km unit of five generations, this winter cold penalty increases the advancement time from 150 up to 200 years. Another 50-year penalty per 200 km is added when the frontier crosses the -2° (Centigrade) January mean temperature line. At -4° another 50 years are added, and again at -6° . It is assumed that by the time this temperature is reached, the frontier peoples will have devised sufficiently good winter adaptations to handle anything more extreme without adding even further penalties. The actual picture of increasing winter cold problems would not follow this neat pattern of steps, but this is at least an approximation (Fig. 3).
In northern latitudes the reduced length of winter daylight time can also be expected to result in added casualties as this begins to become significant. Since temperature and daylight do not decrease together in most places, their effects are calculated separately. When midwinter daylight reduces to less than nine hours the first penalty of ten extra years is added to the 30 years for population doubling. In terms of the basic 200- km unit of advance, this is again an extra 50 years added on to the minimal 150. At eight hours of daylight another 50-year penalty is added, and again at seven and at six hours. After this it is assumed that further daylight lessening need cause no additional penalties, only a continuation of those already assessed. Just as with temperature changes, the actual situation would be a gradually increasing problem, then a leveling off as it is solved. The stepped increments used here only approximate this curve. If one changes temperature to daylight losses, the graph in Fig. 3 should apply equally well to this situation.
With both temperature and daylight variations the penalties may be seen as ten-year increments
for each 40-km advance. Since the mapping is all done in 200-km units of five population doublings,
these are treated as 50-year penalties. With both factors, there are just four such increments used,
so the most extreme case is where 200 years are added to the 200-km advancement time from each
cause. Given the basic 150 years, plus these two delaying factors, the slowest possible advance rate
used here is 550 years per 200 km.
It was not exactly ideal to impose these penalties in such large increments at each climate line, but it made calculations of frontier progress much simpler. In following each route of advance I applied the penalty when more than half of the 200-km unit was in the new zone. Occasionally I would balance some nearly equal instances by choosing speed in one case and delay in the other along the same route. Clearly these factors, as applied, are not exact reflections of real conditions, but in the long run they should be at least a close approximation. (See Fig. 4 for the map of temperature and daylight clines used for imposing these delays.)
The basic rates of advance for peoples with other adaptations are calculated using the same kinds of penalties. The same geographic lines of temperature and sunlight are used, the same four increments are assessed and held constant thereafter, and the added increment is always one-third of the basic advance rate under ideal circumstances. Thus with (sheep) herders, whose basic advance rate is 75 years per 200-km unit, 25 more years are added for each step in each of these two factors, so their slowest advance rate is 275 years per 200 km.
In the following descriptive sections different advance rates for other circumstances are described as they are introduced. They mostly follow this same basic pattern.
No special penalties or delays were introduced for adjustments to new zones of flora and fauna as were used in my North American study. Here it was considered that Neolithic farmers were bringing with them so much of these important items (domesticated) that any changes in the local situation would have had only a trivial effect. Also, no time penalties were assessed for building boats at significant water gaps. As will be seen, there were always other people with shipping available at each such place.
Three more limiting factors of geography may be introduced at this point, as illustrated in Fig. 5. In many places the soil was (and is) too deep and/or heavy to be plowed with the early "scratch plow" that simply cut a furrow. Such soils are mostly randomly distributed, but some major and extensive areas are shown that effectively blocked the entry of large numbers of farmers until new techniques were developed.
A northern farming limit, for most practical purposes, exists along the line of the 120-day growing season. This line would move somewhat during different eras, but it largely marks the end of the Neolithic advance. Only much later did new species and strains of crops enable farmers to pass this line to some degree.
Mountain barriers would sometimes impede Neolithic progress and significantly retard
communications across them. Ethnic boundaries sometimes developed along these, and continued at the
meeting lines beyond.
CHAPTER NINE — CONCLUSIONS
The entire story of the development of European language geography, as just reviewed, turns out to be rather simple in its broad outline. It consists of just five major steps — the initial Neolithic occupation, the Germanic move, the Roman Empire, the Slavic expansion, and the Feudal adjustment. Three of these are based primarily on technological changes, one on a climatic fluctuation, and one on a spread of social factors. There may well have been other important events, but no more need to be added to account for the known linguistic distributions, past and present. The rates of movement of the human populations involved are brought down to just one basic pattern with a few regular quantitative variations. This is also certainly oversimplified, but again no significant differences from this are needed to account for the present language map. (The Slavic shift to the west after World War II is so well known that its inclusion seemed superfluous.)
All of the changes recounted here are rational, natural, and regular — in a word, predictable. To some
readers this may also appear to be mechanical, and thus neglecting the "human element". It is true
that a consistent theme of this reconstruction is that it does not take into account any conscious
decisions of the people involved, nor any chance events — no "historical" factors
Quite simply, it was not found necessary to introduce any such explanations to account for the big picture. In terms of finer details not covered here, these historical factors no doubt would come increasingly into play. That level of detail was not encountered in this study.
An interesting aspect of this investigation is the close parallel between its basic philosophy and
that used in purely linguistic reconstructions. In both cases a set of universally applicable rules
is devised to describe the known phenomena, and all variations on these are also formulated into
rules which admit of no exceptions. There is no inherent reason to connect these two very different
kinds of linguistic reconstructions; the rigorous application of rules is part of science in
A number of conclusions are reached in this study that contrast sharply with present commonly held
opinions about the prehistory of European languages.
There are ten such contrasts that are worth recounting here in some detail. The first four of these
should prove to be the more
important and surprising to most students of this subject.
1. That the Indo-European languages originated in the Near East and were spread by the first farmers at an early date. There is ample evidence that Neolithic farming spread over most of Europe since 7000 BC from a Near Eastern source by way of Anatolia. It is equally clear that this economy was brought by an incoming wave of settlers who greatly outnumbered the native foragers at all points of contact. It necessarily follows that the language of these immigrants prevailed over that of the previous residents in all cases. There was a limited entry area, at the Dardanelles, for those whose descendants eventually overran the majority of the continent. This means that most of early Neolithic Europe must have constituted a single linguistic unit. At no time in subsequent European history is there any evidence of another human wave of advance that had anywhere near such overwhelming numbers, or that spread itself over so great a territory. Even with no information beyond this, the most obvious conclusion would be to relate this Neolithic invasion to the major language family of Europe today. That all of the geographical details and subdivisions fit with this picture, as well as the distributions of other nearby families, only serves to strengthen this interpretation.
The timing of this event, as compared with the supposed time of origin of the Indo-European
languages, is the major point of contention. If the Indo-European dispersal could be put back
substantially in time, then the source and the mechanism for its spread would follow automatically.
The speed of linguistic change has been roughly estimated in some Indo-European languages where
documentation extends back to well over 2000 years ago. (Linear B to modern Greek has the longest,
though poorly known, history of 3400 years.) By assuming a constant rate of change, and
extrapolating back beyond these documents, a beginning time for Indo-European divergence has
generally been put at about 3500 BC. Many Indo-Europeanists now push this back another 900 years to make an archeological
connection with the Kurgans as the
presumed source. The fact that the Neolithic spread began much earlier, at 8000 BC, may pose no
problem for several reasons.
First, the dialect mesh was most likely intact up to about 2000 BC and maybe even much later, while full-speed drift would commence only after it broke up. Differentiation among its connected dialects until then may have been considerable, but was probably much less than what occurs in freely drifting languages.
Second, it is quite possible that language change was a much slower phenomenon in early times than it has been more recently. Most known cultural changes follow this pattern. I see no inherent reason why language diversification should also have followed an accelerating growth curve, but perhaps it has.
The third reason is that increased communication between different languages tends to provide innovations for further changes. Such contacts certainly increased in later times with the development of civilizations. This is mainly a supporting argument for the acceleration idea given above.
Finally there is my own observation that small speech communities appear to drift more slowly than large ones, all else being equal. This is easily shown with the languages at hand — Baltic is most like proto-Indo-European, Icelandic is the most conservative Scandinavian language, Rhaeto-Romanche is the most like Latin, and even Quebecan resembles an early form of French. Slavic is also close to the original Indo-European, while its population explosion is a very recent phenomenon. Similar examples are found in other language families, including some California Indians (Krantz 1977). The general rule is that among sister languages, these with the larger populations change the fastest. This is not generally recognized, I suspect, simply because most linguists have not arranged their data to look at this possibility.
A combination of some or all of these factors could easily put the Indo-European origin time back to
10,000 years ago, rather than the 5500 (or 6400 that has been commonly supposed.
Another major problem with this thesis is that proto-Indo-European vocabulary indicates a mainly pastoral economy (Goodenough 1970). Common words for "plow" and "to plow" as well as words for the major cereal grains are found in the European language branches, but cannot be matched with corresponding words in Anatolian or Indo-Iranian (Cowgill 1970). At first glance this would indicate that the original Indo-European speakers were pastoralists, but not farmers. A second look, however, suggests that this common European terminology may be significant. It shows that the original Indo-European speakers who spread over Europe must have been both farmers and herdsmen — exactly the basic Neolithic economy as postulated for them here. If they were herders only, then one is at a loss to account for their common farming terminology. On the other hand, my reconstruction is at a loss to explain why this same farming terminology is not found in the Asian branches. Clearly this is an impasse, with basic vocabulary capable of being used to argue both ways.
The regular occurrence of horse and cattle words in all Indo-European branches could be explained by their being broadly intruded into the pastoral terminology before the language break-up had progressed very far. On the other hand, the idea of a similarly late intrusion of farming terminology into the European half of the family would presuppose they were not farmers up to that time.
Benveniste (1969) shows that Indo-European societies all include a division between priests, warriors, and farmers, but that the etymology for these classes does not correspond in the various branches. This also suggests a broadly intrusive social structure at some later date. Likewise, words for kingship, slaves, sacred, sacrifice, and legal terminology do not go back to common roots; whereas terms for order, moderation, and control are comparable. Again, this all seems to point more to peasants than to warriors at the beginning (Utterly unconvincing and puzzling argument).
In addition, Goodenough (1970:262) pointed out that if one follows the Indo-European connection with
Caucasian (Lehman 1952:112) to its logical conclusion, this implies that its spread "...could go
back to the Neolithic colonization of Europe."
2. That Anatolia was mostly Indo-European speaking as far back as 7000 BC. This conclusion greatly simplifies all of our prehistoric reconstructions for that area by removing the need to look elsewhere for the origins of some known peoples of early history. The supposed migration of Hittites through Caucasian territory (however that might have been accomplished) is not needed in order to explain their remote similarities; they actually had been neighbors since 7500 BC in their present locations. Even the Mitanni, of presumed Indo-European affiliation, are first located exactly where this language family is shown to bulge down into the Afro-Asian area in the initial dispersal (Utterly unconvincing and puzzling argument. Simplification and analytic convenience is not a convincing argument for a controversial linguistic puzzle).
It is often assumed that southern Anatolia could not be the Indo-European homeland because its natural flora and fauna does not correspond well with the terminology of the protolanguage, as derived from its most widely distributed words. This does not necessarily follow if such words were distributed only as widely as are the natural items they refer to. Thus the original Indo- European would not be expected to include words for things that did not exist in its homeland, but such words would develop where and when the items were encountered and spread from there. While Indo-European natural vocabulary does not support an Anatolian source, I fail to see that it clearly excludes this source either (The key fact that Indo-European natural vocabulary does not support an Anatolian source is a fact, excludes clearly or dimly is a speculation).
3. That the Altatic language family spread from the Kuban rather than the Altai Mountain area. The only available impetus that would account for Altaic expansion would be their animal husbandry, initially based on sheep and goats. The most obvious source of this economy is the Near East. Putting the exact location of proto-Altaic just east of the Sea of Azov is not well substantiated, but it could not have been far from there. The timing of this origin at 7500 BC ought not to generate much controversy, especially if a long-persisting dialect mesh is allowed for in the early stages.
There will be disagreement with my identification of Kurgans, and especially Scythians, with the Altaic family of languages. Ecology and geography argue for a single ethnic group having spread pastoralism over the entire Eurasian Steppes, and there is no evident opening for even a partial replacement until the time of the Slavic heavy-plow farmers.
No mechanism has ever been suggested whereby Indo-European speech could have entered the steppe area at some early date, let alone spread out over most of its neighbors, and then be replaced by Altaic. Some Iranian names among the later Scythian rulers are no indication that this was the language of the population. A conquest from the outside, or even imitation of names from a higher civilization nearby, could easily account for this meager linguistic evidence. The fact that the Mongol conquest in the 13th century AD left only a few small language enclaves would certainly argue for ethnic stability of even this region.
4. That the Uralic language family stemmed from the Mesolithic inhabitants of the Hungarian Plain. According to most authorities Uralic originated in the Ural Mountain area and spread out from there, including a final intrusion into Hungary in the 9th century AD. This is most improbable because there is no accounting for how such a pastoral adaptation could have arisen in this starting location. Yet if the adaptation moved into the area from outside, it should have been brought by people whose linguistic affiliations lay elsewhere. On the other hand, it is certainly possible that some relatively primitive nomads from the Urals could have intruded themselves into the heart of Eastern Europe, but no good reason for such a move has been offered. It is especially difficult to explain how they could have ethnically replaced an entrenched peasantry that probably outnumbered them by more than ten to one.
The only resolution to these problems is to postulate Hungary as the origin place for the Uralics,
and having them spread to the north and east from there. As reconstructed here, they would
automatically have separated themselves into the three major sub-divisions that are now located in the
It also follows here that the easternmost of these subdivisions would automatically have the closest affinity with the Hungarian source.
Archeological evidence of a presumed early Uralic population in the Ural Mountain area fits just as well with the new interpretation offered here as it does with the traditional view. It is a simple consideration of who had the capacity to move in which direction, at what time, and against which population, that should decide the issue. The only problem that remains is the linguistic time depth that is apparently involved. Since this appears to be a common problem throughout this reconstruction, and the only one of note, the principle of Occam's Razor would argue that we should make a general correction in just this one aspect.
5. That the Basques are a remnant of an early Afro-Asian occupation of Iberia. This view should raise little objection except from those few who want to make Basque the oldest language in Europe. The only realistic alternative to the farmers out of northern Africa would be to make them descendants of the Aegean Sea Shepherds. In recent times the Basques have been mainly farmers. Their fame as sheep herders could arise from a farming base just as easily as their farming could have been added to pastoralism. It is their geographic situation, rather than any strong evidence of linguistic affiliation, that points to Afro- Asian connections. The Sea Shepherds ethnically disappeared from their coastal occupations, and there is nothing to indicate why they might have continued just here to become Basques. On the other hand, the Afro-Asian expansion should have stopped exactly along the present Basque boundary in France.
6. That the Germanic-speaking fringe of southwestern Finland dates back to the arrival of the first
Neolithic farmers out of Sweden. Traditionally, this intrusion is put sometime in the Middle Ages
during the period of Swedish military expansion. One problem with this view is that nowhere else
along the trans-Baltic coast did other Swedish speech communities survive where they should equally be expected. A
more positive argument is that this is the exact strip of coast, with the correct depth of
penetration, that would have been occupied from a Swedish source at 3050 BC according to the
present reconstruction. Its linguistic unity with Sweden could have been maintained since then by
regular shipping contacts.
7. That Breton has been the local Celtic language of Brittany since the earliest Neolithic. History would have this as an intrusion from western Britain during the 5th to 9th centuries. The improbability of such a linguistic replacement, without a known mechanism, should be evident. The difficult terrain of Brittany is physically similar to those areas of Celtic holdouts in western Britain. A similar continuity of speech is only to be expected. The especially close relation between Breton and Cornish is a problem that does not fit well with my reconstruction.
8. Deriving the German language directly from Sweden is not an original suggestion. That they initially occupied exactly the drainage basins of the two major western German rivers is highly specific here. It was not just the historical estimates of their early boundaries that suggested this mapping. The Germanic move out of Scandinavia was the only major population shift that had neither overwhelming numbers nor higher civilization to account for its success. It was therefore necessary to examine the exact geographical circumstances to see if there might be some kind of satisfactory explanation. These turned out to be unique, at least on this scale, and a river-basin pattern of occupation showed how this move was accomplished.
9. That the pattern and timing of the Roman Empire was inevitable, at least in rough outline. This
should raise little objection, but it will be insisted by many that it was a historical circumstance
that Rome was its center. I'm inclined
to think that if all the details were known, the triumph of that one Italic town might well prove to
have been inevitable as well. It was, after all, located right on the meeting line of Etruscan and
Greek civilizations, and it could profit maximally from both.
Lacking such knowledge, we can speculate on the results if a competing site had been successful. It is my contention that if Carthage had won the Punic wars it would have formed the same empire for all practical purposes. This would no doubt have had essentially the same linguistic result, except that it would have been Phoenician instead of Latin that dominated the western Mediterranean.
10. That the modern Turks entered Anatolia from Europe rather than from a central Asian source. The interjection of the Turks, which broke the geographic continuity of the Indo-European languages, may not be a proper part of this study because it occurred outside of Europe. Still, the people who apparently made this move, and the evident mechanism that was used, both originated in eastern Europe. Anatolia also has figured quite prominently in European linguistic affairs ever since the major entry of the Neolithic economy from there 9000 years ago, so its recent fate may be of legitimate interest.
Historical sources give no indication of how this Turkish interjection was accomplished — military conquest being no explanation. An ethnic replacement of this magnitude requires an ecological reason comparable to that which caused the great Slavic expansion. These Turks did not bring in a higher order of civilization, nor were they a massive exodus from some natural disaster. In fact, the mold-board plow did enter an area of Turkic speech on the lower Danube, as reconstructed here, giving these people sole and direct geographical access to Anatolia.
The ten innovative concepts summarized above are only a small fraction of those presented in this
study; the list could have been prolonged almost indefinitely. These are selected as being the
conclusions that are most at variance with the usual
historic and prehistoric reconstructions. It was never my intention to deliberately seek to disagree
with previously held views, but neither did I hesitate to do so when the evidence warranted this.
It might be supposed that the major common theme of these new conclusions is the assigning of greater antiquity to many events than has usually been assumed. I would rather describe the common theme as saying that most people have tended to inhabit their homelands since their differentiation. Only incidentally has this come to indicate longer residence in a number of instances.
Many other reconstructions of Indo-European prehistory have had the various peoples moving rapidly and widely around the map for some time, then somehow arriving at their first recorded locations just before their written history begins. After this they move very little except for a few who's languages did not persist. This contrast between high mobility before history begins, and then relative stability after this, seems much too extreme for people who were agriculturally based throughout. By my interpretation these migrations are substantially reduced in number, and those moves that are retained are spaced out somewhat in their timing. In each instance an ecological reason is given that would appear to make the move inevitable.
A notable aspect of this reconstruction is that the same procedures and rules of movement are applied throughout its development. Prehistoric peasants are assumed to have been just as tied to the land as their historic descendants. If early boundaries are to be moved from place to place, then these must be described in the same terms as for the more recent movements.
It will be noticed that many historically famous migrations are either specifically denied here, or are simply ignored. The waves of ethnic conquest that repeatedly swept over the Eurasian Steppes were, at most, replacements of the ruling groups. Any major change in the Altaic languages of this area would require some mechanism that has yet to be proposed. Likewise, the series of invasions of Hungary had no known method by which they might have changed the ethnic identity of the masses of people living there. A whole series of others, like Goths, Vandals, Vikings, etc., are left out entirely because they caused no permanent linguistic changes of significance. If the German army has not affected the French-German speech boundary with three crossings in one recent century, then it is not surprising that earlier and far smaller armies had no lasting linguistic effects either.
There are also a few aspects to this reconstruction that are not fully satisfactory to me. Some of these are worth recounting briefly.
1. The Caucasian language origin proposed here is very hypothetical and could use some confirmation or correction. I am particularly uncomfortable about postulating another center of origin for Neolithic farming in the Caucasus Mountain region, and passing over the known stand of natural emmer wheat in northwestern Iran for the source. Yet all considerations of timing and geography argue for this more northerly location.
2. The Aegean Sea Shepherds are treated here mainly in their frontier capacity as Boat People, but
they later would facilitate the spread of full Neolithic farming techniques and people along the
northern Mediterranean coast. Linguistically, this could mean that Aegean speech was a factor in
Ligurian, Etruscan, and various Italian languages. Racially, this means a greater Anatolian
component in Italy and southern France than my maps indicate. I have largely ignored this racial
factor simply because there was no obvious way to quantify its effects. A correction of some kind is
3. What caused the Magyar extension into Transylvania? There is a substantial population of modern Hungarians in the center of Romania for which I find no ecological explanation that is suitably rigorous.
4. If the mold-board plow originated inside Magyar territory, why is all its terminology Slavic in origin? Vocabulary argues for an origin in Slavic lands; ecology would place it on the Slavic-Magyar border; but subsequent geographical considerations fit best with the spread beginning just inside the Hungarian Plain. This reconstruction is based mainly on the geographical interpretation, although it would make relatively little difference which one had been used.
5. The historical impact and repercussions of the Slavic expansion to the west appear to have been somewhat earlier than they were in the south, although their timing ought to have been about the same. Some adjustment might be made to explain this, but it would also affect the solutions to problems 3 and 4 above. Every variation I've tried so far seems to affect these three problems in a slightly contradictory manner.
6. The proposed exodus of Finnic farmers out of Latvia and Estonia, and then across Russia, is an uncomfortably unique phenomenon. There is nothing logically wrong with this aspect of the reconstruction for its time and place, but it would be much better if there were other examples of this kind of move.
7. It is not clear why the Turks should have spread so far with the mold-board plow while the Magyars used it to advance their frontier less than half that distance. It is true that there were some major grasslands for the Turks to move through while the Magyars faced only forested lands to the west of Hungary. Other factors of ecology and demography may have been involved.
This list of remaining problems could also be prolonged, but not very far. In general, the procedures used here have handled all but some relatively trivial discrepancies. I would welcome any new attempts at this kind of reconstruction. Perhaps some of the problems noted above might be handled with new factors or new interpretations. More empirical data relating to some of my assumptions about human movements might strengthen or alter them. Perhaps the entire problem could be redone with some wholly new or greatly altered assumptions.
The overall picture of European linguistic geography is perfectly described and accounted for by the procedure used here. The conclusions reached are given in the form of clear laws, discrete events, and exact mapping — all of which may have been a bit overstated and oversimplified for the sake of clarity. In effect, the stage was set as of 10,000 years ago, the rules were given, and some reasonable innovations were tagged for their introduction at specified times and places. The whole thing was then put in motion and the results were observed through time. While much of the resulting story is not presently testable, some of it is. The fact that this process does fit well with most of known history and archeology, and that it ends up with the modern distributions and relationships — all this argues strongly that it represents the actual series of events.
It would seem unreasonable to discredit the method itself because of disagreement with some of its results. If this were done for the sake of a few details, then the explanation for the major part is lost as well. Either the method works or it does not. It would be peculiar indeed if 90% of European linguistic history followed natural laws, and the remainder were under no such constraint.
This is not to say that the reader must accept every detail of this reconstruction as given, or else reject it all. The method still has room for correction and change. Since the entire reconstruction follows automatically from the facts of geography and my basic assumptions, any error in its conclusions should be found to stem from an error among these assumptions.
In developing this picture I tried out many different sets of assumptions about human movement that were within the range of possible choices. All of the dates given at the 200-km intervals on the migration paths were calculated at least ten separate times. Many other paths were also figured out, several times over, that are not shown on the maps. There is also some leeway in the application of geographical barriers, climates, soils, and vegetation zones; the atlases do not always agree on these. I do not claim to have tried out every possible set of assumptions, but after finding some impossible situations at various extremes it was usually easy to close in on the most likely ones.
Hopefully, the spirit of this approach will be appreciated by critics and supporters alike. Ideally it will be used by those who would alter or redo this reconstruction, or who would attempt a similar reconstruction of another area. The spirit might be restated simply as: There is a reason for everything, it's just a matter of finding it.
In Russian (later)
Ogur and Oguz
Türkic and European Genetic distance
Classification of Türkic languages
Indo-European, Dravidian, and Rigveda
Türkic, Slavic and Iranian
Türkic in English
Türkic in Romance
Alans in Pyrenees
Türkic in Greek
Türkic in Slavic