Added
The dissertation
has a copyright. Please give proper credit when quoting its content.
A SPATIAL ANALYSIS OF MEGALITHIC TOMBS
18. TOMBS AND THE INDO-EUROPEAN HOMELAND
By
Maximilian O.
Baldia
1993, 1995,
1999-April 25, 2006©
All
rights reserved
18.1 Introduction
The correlation between languages or ethnic groups and archaeological cultures is wrought with technical problems and excites political and nationalistic feelings. This is unfortunate, because it prevents impartial scientific evaluation of the evidence, methods and theories. Further more, the concerns with the location of an Indo-European homeland and an archaeological culture may fall into the category of incorrect questions, based on false assumptions. Nevertheless, attempts should be made to evaluate the relationship of the TRB, megalithic tombs the homeland, since it has fired the imagination of professional archaeologists, linguists and the general public. In fact, since the beginning of the Classifying Period (1800-1900) the location of megalithic tombs has led to speculation about the relationship between their builders and the historically documented linguistic or ethnic groups of the same region. As early as 1815 Westendorp believed that the tombs were Celtic, raising the interest of W. Grimm and others who developed Historical Linguistics and hoped to find a correlation between the tombs and linguistic findings. Unfortunately, as in archaeology, there are shortcomings with Historical Linguistics:
As
one reaches further and further back into the past, very quickly one finds that
reconstructed languages are themselves based on other reconstructed languages.
... In reconstructing a language, linguists build dictionaries (albeit rather
limited) of word fragments, etymologies, sounds, and a notion of grammar. The
reconstruction necessarily is rather uniform and therefore not a language as it
would be spoken in the real world. ... Linguist Ernst Pulgram pointed out, “You
always have differentiation within a language - geographic differentiation
and social differentiation.” ... Most reconstructions are “atopic, achronic,
and aphonic.” ... One does not know where they were spoken, when they were
spoken, nor exactly what they sounded like. At best, reconstructions represent
abstractions of past languages.’ (Lewin
1988a:1229)
Thus the origin of Indo-European has been highly controversial (e.g. Stover and King 1978, Krupp 1981) so that one does not ask `Where is the Indo-European homeland?’ but rather `where do they put it now? (Mallory 1989: 143).
18.2
Models of the
Indo-European homeland.
Sir William Jones proposed in 1786 that resemblances between Latin, Ancient Greek, Gothic (the ancestor of German), Celtic, Iranian (Farsi) and Sanskrit were due to common descent (Lewin 1988a:1128). The Danish geographer Malte-Brun used the term Indo-Germanic, which was probably already used among intellectuals in Paris, in his 1910 publication for this widespread group of languages, while T. Young published the term Indo-European in 1913 (Häusler 1985:64-65 Footnote 4, 1988:1-2). In 1816 Bopp used comparative philology to establish relationships between Celtic, Greek, Latin, Germanic, Slavic, Iranian and Indian (Sanskrit) (Neumann 1975:674). Later Schleicher reconstructed what he believed to be the earliest Indo-European words and created his famous Stammbaum or family tree of the evolution of the modern representatives of Indo-European (Fig. 18.1, Mallory 1989:14).
Figure 18.1. The
Indo-European family tree as envisioned by Schleicher.
This undertaking presupposed a common root termed Proto-Indo-European (PIE) and led philologists to expect archaeologist to locate its homeland (Neumann 1975, Narr 1975:690-703). Ever since then archaeologists and linguists have attempted to determine not only the original homeland, but also the chronological position for Schleicher’s branches and the assumed concomitant emigration of the carriers of linguistic sub-groups from the homeland (Mallory 1989:16).
Figure 18.2. The wave model of Indo-European language change of J. Schmidt (after Mallory 1979:19 Fig. 8). The model has (relative) geographic implication. 1 = Shared e (Indo-Iranian has a), 2 = Shift from *k’ to s (centum vs. satem), 3 = word endings with m instead of bh.
Schleicher’s tree demanded an intensely close relationship between Germanic and Balto-Slavic ..., but the split into centum and satem languages, resulting from shifts in the Indo-European *k’ sound in Balto-Slavic, suggests a close relationship with Armenian, Iranian and Sanskrit (Mallory 1989:19). J. Schmidt resolved the impasse by proposing an undated synchronic model of overlapping circles (Fig. 18.2). He felt that changes occurring in one language must have spread in waves to some of the other languages, but not to all of them.
Figure 18.1 The isogloss “wave model” of R.
Antilla
Schmidt’s model had geographic implications. Germanic is placed in the north, Balto-Slavic is directly adjacent in the east. Related to the latter and further east is Indo-Iranian and Armenian. Celtic is located to the south and southwest of Germanic. It overlaps with Italic, which also shares traits with Greek.
A more elaborate model, based on twenty four isoglosses, was proposed in 1972 by R. Antilla (Fig. 18.3). It includes Tocharian and Hittite, which due to their peripheral geographic position appear to have undergone fewer changes. In this model, Germanic is divided into a North, a West and an East Group, which corresponds with the historic distribution of the languages. They are separated by the centum/satem line and three additional isoglosses in Baltic.[1] Baltic is in turn separated by up to two isoglosses from Slavic. Germanic is about as distant from Slavic as from Celtic, yet both Germanic and Celtic belong to the centum group.
In 1970, Friedrich employed linguistics together with paleo-botany in an attempt to localize the homeland by tracing the prehistoric location of trees whose names were held in common among several languages. He had noted the cohesion between the four western stocks of Indo-European (Italic, Germanic, Slavic and Baltic), which was reflected in the basic terms related to the names of trees and products of wood. He added that a determination of the location of such trees during the suspected period of cohesion in Proto-Indo-European (PIE) speech could have direct bearing on questions of early Indo-European dialects and migrations.
His analysis focused on the eighteen types of trees recognized in PIE as well as the thirty names common to Germanic, Baltic, Slavic and Italic, and to a lesser degree Celtic and Greek. The thirty names refer primarily to generic groups, although they occasionally apply to a single species.
His investigation focused on the arboreal history of
central and
In the process, he discovered that pollen analysis has its technical problems. Trees produce enormously varied amounts of pollen and the field of paleobotany often entails precarious and problematical inferences and depends heavily on personal judgment about data that is often inadequate and variables that are numerous (Friedrich 1970:15).
He found that the homeland was difficult to determine and
characterize. He observed that the natural boundary between the forest and the
(forest)steppe of the Russian Plain was established during mid-Holocene. The
northern European/German Plain was much warmer than today. Scattered but large groves of pine and
hardwoods were found along the southern
Bug and Dnepr, along the Doin and
Friedrich felt that the appearance and proliferation of this forest gave the Indo-Europeans their heavy focus on arboreal terminology. In summarizing his findings Friedrich (1970:18) stated that:
... from the foothills and steppes north of the western
side of the Caspian westward to what is now the Ukraine and on northwestward
into the north German plain there ran a
fairly continuous and fairly homogeneous ecological zone - by and large one of
temperate climate, open plains, and mixed hardwood forests. I assume that it
was precisely in this east European area during the Atlantic period that
speakers of Proto-Indo-European were distributed in a block of dialects about
three hundred miles wide and five hundred or more miles long: the area may have
been only a third as large, but in any case probably included the central and
eastern Ukraine. Subsequently, during
the last of the late
A considerably different view is expressed by Gamkrelidze and Ivanov (1990:113), who decided to reexamine the entire system of consonants posited for the protolanguage and proposed a new system of consonants for the language about the time Antilla published his wave model. In 1990 the authors point out that they improved the consonant system of the classical PIE system. Their reconstructed PIE consonants established a closer connection of Germanic, Armenian and the Hittite daughter languages with Sanskrit. Thus they reversed Grimm’s Law, which states that these languages underwent a systematic sound shift, while Sanskrit preserved the primeval system of sounds.
Illustrating their point of view with the “g” to “k” shift
in Germanic, they conclude that “the
Germanic languages are more archaic than Sanskrit and Greek” (ibid.). This
conclusion would seem to be at odds with the authors’ compulsion to seek the
original homeland in the archaeologically and linguistically least plausible
area just south of the
This is incorrect. As early as 1978 Schwabedissen produced
evidence of those grains and animal husbandry in the supposedly still
Mesolithic Ertebølle culture. The sites providing this evidence are the
Satrupholm Moor (Förster Moor) and Rosenhof, by Dahme in the area of
the Oldenburg Graben on the Baltic coast of
In addition, the fortified
Figure 18.4. Petroglyphs
of draft animals from Uzbek supposedly dated to the second and third millennium
B.C. (Gamkrelidze and Ivanov 1990:115).
Gamkrelidze and Ivanov further observe that oaks spread
into
To further support of the Transcuacasian homeland, they
state that: Petroglyphs ... found in an area from the Transcaucasus to upper
The strange thing is that of the eleven purported pairs of horses, at least six pairs depict
one or even two animals with long, cow-like horns. Two of these “horned horses”
actually look very much like the two oxen on the broken sections of the large
menhir that was reused in the building of megalithic chambers in
Figure 18.5. Various
proposals for the location of the Indo-European homeland since 1960 (Mallory
1989:144 Fig. 80).
Archaeologists also have attempted to locate the homeland.
In 1987 and 1989 Renfrew addressed the Indo-European homeland problem by
combining archeology and linguistics, placing the homeland in
A purely archaeological approach was taken by Gimbutas
who, similarly to V.G. Childe in 1926, placed the homeland in the area of
the lower
This ... ‘Kurgan Hypothesis’ ..., while critically received by archaeologists, has proved attractive to many philologists concerned with the Indo-European language group ... Radiocarbon dates ... show ... contemporaneity with the supposedly derived groups in Europe. ... Contacts existed between the late Pit Grave/Catacomb Grave phase and the Northern European Corded Ware cultures ... (But) where direct archaeological evidence is used to support the ‘Kurgan’ Thesis, it is too often treated in an uncritical, if not tendentious manner. ... It has a curious old-fashion air, reminiscent of earlier writings on the distribution of megalithic monuments ...
Häusler negates the purported hierarchical social
organization of the supposed “
The assembly and evaluation of data on the oldest finds
of wheel and wagon in the Pontine area shows, that these finds in most cases
belong to simple burials of the older Ochre-Grave culture, which contrast with
the opulent wagon graves of hither Asia. The wagons, or models of wagons and
wheels referred to do not allow a derivation from hither
Rebuking linguistic arguments based on Gimbutas’ model (1989), he points out that there is no evidence of an Indo-European invasion and argues for an autochthonous linguistic evolution, in which the TRB, KAK and EGK/Corded Were culture figure prominently (e.g. 1981a, b, 1985, 1992a, b, 1994). However, he and Kilian envision a PIE territory as long as 3000 km (Fig. 18.5), and place the initial linguistic development into the Mesolithic, if not the Paleolithic.
Mallory (1989:257) objects to the early dating, proposing instead that the original development of Indo-European occurred between ca. 4500-2000 B.C.,[4] when Northern and Central Europe would have been occupied by the TRB, followed by the KAK and Corded Ware cultures (ibid. p. 182). The TRB culture area is a likely homeland candidate, because of its geographic location, as well as early linguistic and archeological evidence for the wheel and plow agriculture (ibid. p. 162-164). The link between Indo-European mythology, which is preoccupied with domesticated cattle and horses also matches with the archaeological evidence from the TRB, KAK and Corded Ware (Fig. 1.3, 18.6-18.7).
Figure 18.7. Extent of Corded Ware culture (hatched lines) and Danubian complex (broken line) with superimposed Indo-European languages and Anatolian (Mallory 1970:109 Fig. 72).
In this connection it is vital to recall that Gamkrelidze
and Ivanov suggest, on the basis of sound shifts, that Germanic is older than
Greek. Furthermore Germanic and Baltic are linguistically and geographically
closely linked, while Germanic is also linguistically (and geographically?)
close to both Slavic and Celtic (Fig. 18.3,18,7-18.8). The earliest location known for all these
language groups falls at least partly into the geographic distribution of the
TRB (Fig. 1.3). Most importantly Old Prussian, extinct since
A.D. 1700, and now represented only by Latvian and Lithuanian, was once
located east of the
what is most striking is that Lithuanian shows roughly the same general retention of the Proto-Indo-European forms (naturally mitigated by minor sound sifts) as does Sanskrit, despite the fact that the latter language is attested nearly 3,000 years earlier than Lithuanian. This apparent archaism has mesmerized linguists for over a century now and has led some to the conclusion that the Indo-European Homeland must have lain in or near the Baltic. ... The ... region even retains the Proto-Indo-European names for rivers. ... Moreover, ... Lithuanian and a number of Slavic languages retain traces (of the Indo-European free accent) (ibid. p. 157). ... Because of this transparent conservatism, many linguists hold that the Baltic languages, like their Slavic neighbors, have probably moved but little since late Indo-European times. (ibid. p. 82).
Such observation suggest that there may be a connection between the distribution of megalithic tombs and the North/Central European Indo-European languages.
Figure 18.8. The oldest
accepted locations of German, Baltic, and Celtic (Mallory 1989:87, 82, 103 Fig.
58, 54, 69).
18.3
Homeland models and the
tombs
During the early part of the Culture History Period (1900-1950) Kossinna identified the TRB with the origin of Germanic tribes. But during the Data Reevaluation Period (1950-1970’s) Behrens objected strongly against a Nordic preeminence in the theories related to the origins of TRB megalithic tombs. However, in 1972 Dehnke published a paper in which he reviewed the state of megalithic tomb research and found that recent results seemed to favor older interpretations of the archaeological data.
Dehnke, following Schwantes and others, derived TRB
ceramics from Ertebølle pottery. He further suggests a likely cultural
continuity between that of the megalithic tomb builders and the subsequent
EGK/Corded Ware population, as well as the Bronze Age cultures of
If there is a connection between the tombs and the homeland
models, one should expect a rough correlation between projected linguistic
region and the spatial distribution of the tombs. Reexamination of a
preliminary spatial analysis (
Reevaluating the earlier conclusions by using the present
data base of trapezoidal mounds and their size suggests two lines of
communication. One line leads from the
Historically the
While the density of rectangular mounds forms an
intermediate position, confined largely to the same areas as the Iron Age
Jastorf and to a lesser extent the Harpstedt culture, the tumuli reinforce the
uniqueness of the Scandinavian part of the North Group through the consistent
use of round megalithic enclosures and the popularity of polygonal-dolmen. This
could imply linguistic differences, such as regional dialects, because the
tumuli concentrations near the Baltic coast indicates an intense, integrative
communication along the bays and between islands. Thus the high concentration
of tumuli and grand-dolmen on Rügen connotes a relationship with the
Scandinavian area, but the enclosures here and in many other parts of
The
Complementary conclusions are reached when analyzing the
chamber distribution. The Iron Age Jastorf culture coincides with the TRB
North Group in
The widespread, relatively cohesive grand-dolmen
distribution, centered in
Passage-graves exhibit one of the widest distributions
arising from intensive cultural interaction, but a regional differentiation
occurs in time, perhaps indicating linguistic differences as well. Thus complex
passage-graves occur almost exclusively in
The gallery-graves range from the Elbe-Saale region in the
east to the upper
Thus the comparison between reconstructed Proto-Indo-European linguistic patterns and the megalithic tomb distributions leads to the speculation that the western tongues of the Indo-European language family are already discernible in the TRB.
18.4
Cultural continuity and
geographic expansion
Although questions remain,[6]
some archaeologists see a continuity from the TRB to the KAK and the
EGK/Corded Ware culture. Together, these cultures exhibit most of the
cultural characteristics usually cited as necessary to the identification of
Indo-European. According to Häusler (e.g. 1994:45 ff. with references)
physical anthropology indicates a continuation of the TRB population in the KAK
and the archeological evidence reinforces this view. Indeed, there seems to be
a considerable congruity in some pottery from the Danish TRB MN V and
There may also be a similarly close relationship between TRB and the EGK/Corded Ware Culture (2900-2450 B.C.), which may overlap with the TRB/KAK by as little as 50 calibrated years (Bakker 1992:xiii). The EGK artifacts found in megalithic tombs are usually identifiable as part of later secondary interments (Schuldt 1972). Osteological evidence indicates genetic relationships (H. Grimm 1984:125-130). Tempel (1979a:124-26) also argues for a continued succession of interments from generation to generation in the passage-grave Ostenwalde 834. Inside was Tiefstichkeramik plus EGK pottery, followed by Bell Beaker pottery. He further seems to suggest a cultural succession into the Early Bronze Age, noting a continuation in the use of the same burial places and the kinship of Early Bronze Age Blocksteinkisten (cists) with megalithic tomb construction. This view may be supported by the cultural sequences at Wartin and Kierzkowo (Bakker 1992:73-79, 190-192 Fig 29-31).
There may be continuance of burial TRB burial customs, so
that the EGK continued interments of both
flexed and extended positions as secondary burials in megalithic
chambers and under tumuli with stone circles, and both cultures continually
deposited projectile points in megalithic tombs. Less well documented is the
use of archer’s wrist guards in the TRB, due to the lack of undisturbed
burials. Yet a long dolmen at
Frellesvig on the
At Rude the skeletal remains found within one of the sub-megalithic cists was radiocarbon dated to 2270 b.c. This may mean that the cists were inserted into the top of the mound at the end of the MN or beginning of the LN by the EGK. If 2270 b.c. is the actual construction date, the chambers are a part of the size reduction process of megalithic tombs arising from the same social forces that gave rise to the burial practice at the miniature passage-grave at Diever, Netherland and small TRB cists (Bakker 1979a, 1992). One could attribute this development to the changes in mortuary practices encountered especially in the Late TRB. Indeed the TRB flat graves of the West Group have a dimensions similar to those of the EGK and Bell Beaker graves according to Hogestijn (1991:94). In his evaluation of Neolithic flat-graves of Brandenburg Schoknecht (1976) argued for a continued chronological development from the TRB via the EGK (Obergrabzeit and Oderschnurkeramik) to the LN (Dolchzeit).
However, megalithic chambers did not disappear entirely. For example, in Beier’s (1991:179) research area the massive Corded Ware cists and occasional drywall constructions connected with Late Neolithic and Early Bronze Age burials are still reminiscent of the former megalithic tradition. All of which suggests a long and continued cultural evolution from the TRB to the Corded Ware to some archeologists (Krzak 1981:25-27) and a continuation into the Bronze Age is implied by cists and one long-mound with megalithic enclosure (Fig. 19.1).
However, the succession was not uniform. For example, the
KAK never became established in Netherland and although it existed in most
parts of
One may, therefore, argue that there is a continuation and
expansion of cultural influences starting with the TRB and continued during the
KAK and the far flung Corded Ware culture. Further more, the existing TRB
trade routes expanded during the Corded Ware, because a Scandinavian type
flint dagger has been found in a barrow at Orlivka near Berdyansk on the
Please send comments or questions to Max Baldia