Stammbaum

I have long felt that the Stammbaum theory of the Indo-European languages as rendered by August Schleicher in 1862 is an inaccurate representation of the genetic relationships between the Indo-European languages, because it was based solely on morphological and phonological data, and incorporated no archaeological or ethnographical evidence. I will be calling into question the methodology used to produce the Stammbaum, in an attempt to convince the reader that a new methodology is necessary.

It is very important to determine whether Schleicher's representation is accurate. If one takes the relationships as represented by Schleicher to be correct, it leads one to a certain series of conclusions regarding European prehistory, such as which cultures (archaeologically speaking) spoke which languages, and who moved where, when. If Schleicher is wrong, then it follows that any conclusions about history which use the Stammbaum for support must be treated as suspect at best. I do not propose that we abandon the Stammbaum altogether; on the contrary, it is a very useful representation of the commonalities between the Indo-European languages (see Appendix I). I believe, however, that it must be taken solely as such, and not as a family tree showing the actual descent of the languages discussed.

I would like to examine the Germanic branch of the Stammbaum theory in detail and the methodology involved in its genesis. I believe that this methodology may be faulty for the purpose of determining genetic relationships. It is necessary to combine the archaeological evidence with the linguistic and the historical information to clarify the actual genetic relationships between the languages. Alone, these fields do not tell us much. My approach will begin with the Proto-Germanic branch of the traditional Stammbaum and examine the process of the subsequent "branching out" or divergence which reflects the assumed development of the modern Germanic languages. This examination will incorporate the most current archaeological data available as well as the oldest historical accounts of Germanic speaking peoples, but I will likely place more emphasis on the archaeological evidence, the reasons for which will be discussed later.

The traditional Stammbaum accepts certain assumptions regarding the migrations of various Germanic tribes. Some of the relationships which Schleicher assigns would not be possible without the movement of certain groups of people from one area of Europe to another. Here I will look at the historical accounts of these migrations and for evidence of them in the archaeological record. Tacitus wrote of extensive migrations of the Germanii, but I do not think that this is grounds to say that these migrations actually happened exactly the way he described them, and it is also problematic to determine whom he was actually referring to as Germanii.

As a hypothesis, I believe that linguistics and archaeology are inextricably intertwined, and that it will one day be possible to synthesize the data from these areas and to construct a model which shows how each of the Indo-European languages came to be and to what other languages they might be directly related. Historical information must also be taken into account, as first hand testimony should be able to tell us exactly what our other two areas do. It is important, however, that we be very careful in using the data obtained from these different fields to prop up weak suppositions. To quote (translated from French) archaeologist J. Vansina: "The archaeologists use the conclusions of the linguists to support their inferences about the material culture, and the linguists use the conclusions of the archaeologists to attest to the existence of speakers of some sub-family or dialect of Bantu in a certain geographic sector" (Renfrew 283). Vansina was talking about Bantu, but the same sort of tail-chasing goes on in Indo-European circles.

If the archaeological evidence supports the traditional Stammbaum, then all is right in the world of Indo-European linguistics. If, however, it does not, then a serious reanalysis must be undertaken and a new method must be adopted. That method can only be an interdisciplinary one incorporating the fields of history, archaeology and linguistics.

I begin my discussion of this topic with a look at some of the oldest historical evidence regarding the Germans, most notably the work of Cornelius Tacitus. The first century AD, during which Tacitus was writing was one of great turmoil in the European world. The Roman Republic had only recently given way to the Roman Empire, and the Great Migration of the Germans was on the horizon. Oppida, fortified settlements with an average population of 2000 and an average life span of only 70 years, had been all the rage in temperate Europe up until about 100 years before Tacitus' time. The fact that they were fortified indicates that there was something to protect or something to protect against, and the settlements' very short life span indicates that people were on the move. The Celtic Iron Age had ended and the Roman Age had begun. Julius Caesar had subjugated Gaul, but the threat of the Germans remained. The hideous defeat of Varus in the Teutoberg forest in AD 9 temporarily ended Roman designs on German lands, but German victories over Roman armies were not easily forgotten; the Germans were still the enemy.

It was originally my intention to address the written histories of the Germanic people in conjunction with the archaeological evidence. However, after having looked into the ethnography of the German people written by Tacitus in the first century AD, I have decided that these are two separate topics; they simply do not mix. In order to really combine these two aspects of our problem, one or more of the ancient historians would have had to provide precise information on the cultural accoutrements of a specific tribe and placed that tribe in a specific location. Following that, modern archaeologists should be able to go to that specific location and find artifacts matching the descriptions. Unfortunately, this is not the case. Even the idea of taking a specific type of artifact found in a specific area and associating that object then with the tribe who was alleged to have lived there cannot be taken as methodologically sound because it cannot be proven.

During the first decade of this century, a German scholar named Gustav Kossinna worked under the assumption that, although cultures appear to be in flux, they remain static enough to show continuity and a definite identity. Adhering to this idea, one can establish rough geographic borders between archaeological cultures and assume that each of the areas thus created represents a tribe. All that is left to do is to decide then which tribe the area represents. Kossinna's work led him to the conclusion that the Germanic tribes of the Roman Iron Age were descended from a large Nordic Bronze Age culture (Todd 20). His work fit in well with the nationalist political climate under which it was done and shaped ideas about European prehistory for the first half of this century.

After the Second World War, people began to realize that this idea was fitting less and less with the archaeological evidence, and Kossinna's work was dismissed and deemed politically tainted. This leads me to the point that all archaeological theory, and dare I say all theory, must be taken in the historical context under which it was devised. Human beings are fallible, and we are easily susceptible to having our views colored by current political trends. Human beings have agendas, and whether we like to admit it or not, we are almost incapable of being impartial judges. That said, I will now embark on a discussion of the Germania of Tacitus.

The Germania of Cornelius Tacitus was completed in the year AD 98. It was originally titled De origine et situ Germanorum or "The Origins and Land of the German People." It was written more or less as an ethnography of the Germans so that Romans could learn about their northern neighbors (Goold 119). The Germania has long been one of the most widely used sources for information regarding the German situation in the first century. The editor of the edition I have chosen goes so far as to say that, "though there is much in Tacitus which we cannot cheek, everywhere we can hear the ring of truth. Tacitus' Germania is by far the fullest and most valuable treatise of its kind which has come down to us from ancient to modern times" (Goold 122). Despite this ringing endorsement, I would caution giving Tacitus too much credit. His knowledge of geography was apocryphal by modern standards and, like everyone else, he was not entirely unbiased.

Tacitus has the Germans located east of the Rhine, north of the Danube, and bounded elsewhere by mountains and oceans (see Appendix II). He gives no clear eastern border of their territory, but the Vistula is now generally accepted to have fulfilled this purpose (Todd 15). He claims that the Germans must be indigenous to the area because no one would want to migrate there due to the climate (Goold 131). He breaks the Germans down into three groups before addressing the individual tribes. These groups are the Ingaevones of Denmark and Scandinavia, including the Cimbri, Teutoni and Chauci, the Istaevones along the Rhine, and the Herminones of central Europe whose ranks included the Suebi, the Chatti and the Cherusci. These three groups, with minor spelling changes to their names, comprise the West Germanic branch of the Germanic languages. Tacitus says that the Germans break themselves down into these groups, and that the groups are named for the three grandsons of the god Tuisto, these being the fathers of their respectively named groups. This implies a cultural unity between the Germans as a whole, of which no evidence has been found.

Tacitus lists many individual tribes and their locations and occasionally makes reference to some of their artifacts. The locations are somewhat hazy, however, and it is possible that some of the characteristics which he ascribes to a particular tribe may actually be those of another tribe and that the characteristics which he attributes to the Germans as a whole may in reality be those of only a few. One thing which is fairly clear from Tacitus' account is that the Germans were in motion. The same Cimbri who were supposed to have lived in Denmark won a victory over a Roman army under Carbo at Noreia thousands of miles away in 113 BC. There were Germans who had moved to the west side of the Rhine around 200 BC. The Suebi moved into Baden-Württemberg, pushing out the Helvetii. The Usipi moved south from the lower Lippe to the Lahn, and so on and so on. These people are not smelling the flowers.

At some points, Tacitus does say some things which may be relevant. He describes the practice of drowning immorals in bogs (Goold 149), and indeed, many bodies of people who appear to have been victims of execution have been found in bogs. Mention is made of the preferred style of dress and coiffage (157) among Germans as well as the burial practice of cremation and inhumation of the ashes in a turf mound (171). All of these things he attributes to the Germans in general. He also comments on German housing, but although the description of the dwellings is good, he does not pinpoint exactly who lives in this sort of house; his statements are too general.

At two points he actually assigns specific artifacts to specific tribes (205). He says that the Rugii and Lemovii, both on the North German Plain near the coast, carry short swords and round shields. He also describes the boats of the Suiones who live "in the sea" (Goold 205). Here, at last, is something that we may be able to work with. As it turns out, however, wooden boats do not survive well except under very special circumstances, and the metal goods in the archaeological record of this area are inconclusive. Thus, even when Tacitus is specific, it is of little help.

One of the biggest problems with addressing the archaeological evidence of the Germans is that they did not have any large settlements. The oppida mentioned above were a Celtic phenomenon. A few of them were located north of the Danube and east of the Rhine, the southern and western borders (respectively) of the area we assume (according to Tacitus) to have been German, but these too were Celtic outposts. Most of the German settlements involved less than 100 people. A few of these sites have been excavated. Most of the work has focused on the northwest coastal region of the continent. Evidence of individual houses was found, and bodies of execution victims were found in bogs. Here the archaeology seems to hearken back to Tacitus. Whether this helps us or not is yet to be discussed. The sites excavated seem to have been fairly independent economically. The pottery and iron goods were manufactured on site, but there is evidence of trade with the Roman Empire in the form of wine vessels, coins, and various other trinkets. The small number of inland sites which have been carefully studied seem to show roughly the same pattern (Wells 190).

Based on the information at hand, the archaeological evidence in the area in question can be divided into seven main culture groups (Todd 55). Identifying the tribes associated with these groups is a risky business as, Todd points out:

This was a major goal of archaeologists before about 1945, but none of the identifications then confidently put forward are now widely accepted…. Normally archaeology will not assist us much with the mapping of tribal boundaries, or at the very best will be a halting guide. A tribal map of Germania will therefore not be established, and the positions of the major tribes in the Early Roman Iron Age can be indicated only by the literary sources with all their shortcomings. (55)

He too seems to recognize that Tacitus may not be the most accurate source of information.

Todd lays out the seven groups on the basis of their pottery types (see Appendix III). The groups are as follows: The Rhine Weser group occupies the east bank of the Rhine all the way to the upper Elbe with a few sites on the Main. The North Sea group extends along the coast from just east of the Ems, with a heavy concentration at the mouth of the Weser, through Schleswig-Holstein into southern Denmark. The North German group occupies the Jutland Peninsula and some of the Baltic Sea islands. The Elbe Germans occupy the entire Elbe basin, all of Schleswig Holstein and Lower Saxony to the Oder, and part of the Danube. The Lower Oder group is, of course, on the lower Oder with a smattering in Lower Saxony but not as far as the Elbe. The Przeworsk group occupies the upper Oder all the way to the upper Vistula, and the Oksywie group focuses on the lower Vistula, but extends along the coast almost to the Oder. As may be evident from the descriptions, many of these “groups” have very highly concentrated areas of overlap. If one assumes that languages, or at least dialects, go hand in hand with cultural groups, we are left with somewhat of a quandary. We would find two, or sometimes even three languages spoken in the same area at the same point in time. The real question is less one of how this is possible than it is one of how this is relevant. Making sense of all of this information and applying it to the task at hand is not easy.

The historical evidence, when taken in the light of modern knowledge, leaves much to be desired. Tacitus' broad cultural commentary does not provide us with enough information to determine who was who. Coupling this with his deficient geography only serves to make matters worse. The archaeological evidence is specific enough to give us some vague areas which were allegedly inhabited by different groups, some of whom, just for the sake of complicating things, occupied the same area at the same time. It does not, however, provide us with an answer as to which tribes were in which places at which times. This point cannot be neglected if one is to attempt to relate the different languages to one another. If the tribes inscribed their names in the cornerstones of every one of their dwellings, these have surely been eroded by the flowing sands of time, leaving us with nothing more than handfuls of dust and mute bricks. A sad fact is that even if we did by some miracle know exactly who was where at a given time, it may not be helpful because we still wouldn't know what language they spoke.

Addressing the linguistic evidence for the connections between the languages may prove just as problematic as our other forms of information. In 1860, after having read about Darwin's evolutionary theories, August Schleicher began to apply Darwin's ideas to language. In many respects he was right to do so. He says in his book Die Darwinsche Theorie und die Sprachwissenschaft:

Die Sprachen sind Naturorganismen, die, ohne vom Willen des Menschen bestimmbar zu sein, entstunden, nach bestimmten Gesetzen wuchsen und sich entwickelten und wiederum altern und absterben: auch ihnen ist jene Reihe von Erscheinungen eigen, die man unter dem Namen 'Leben' zu verstehen pflegt. (7)

He chose to see languages as living things, subject to the mercurial nature of life itself. Language is an organism, not a by-product like carbon dioxide or urine, the composition of which hasn't changed much over the last several millennia. It is used so frequently that innovations occur constantly and the language has a tendency to evolve faster than those who use it. The idea was that if language is an organism, it should be possible, as it is with other organisms, to trace the modern creatures back to their proto-parents by comparing various features. This comparison would also tell us which languages are siblings and which are merely distant cousins.

This was a good idea, but there are problems with it. One of the problems is that natural selection doesn't apply. The survival of a language and the continuance of its species are not a matter of whether the language is strong enough, but of whether the people who speak it are. If the people who speak a given language die, the language dies; it is not the fault of the language. The changes which occur in a language that does survive do not so much follow rules as they can be described by rules. The changes themselves can go in almost any direction and can be influenced by any number of factors. Perhaps we will be able to say more about the whimsical nature of language after we look at how the languages can be grouped.

The Germanic languages (since Schleicher's work) are traditionally grouped into three separate categories. North Germanic is made up of the descendants of Old Norse. East Germanic is now entirely extinct, and the only attested member of this forlorn family is Gothic. West Germanic comprises the largest group of independent modern languages in the Germanic branch of Indo-European and includes such celebrities as English and German. The traditional basis for the tripartite division of the Germanic languages is one of morphological and phonological features. Morphological features are those related to the derivation and inflection of a language. How a language forms its plurals or derives adjectives from verbs or verbs from adjectives, as well as differentiating between first, second and third person singular and plural in its verb forms are all aspects of a language's morphology. The phonological system of a language pertains to the sounds of the language. How those sounds change over time and how they react to one another can show us things which some languages have in common with one another. The syntax of the various languages has proven unreliable in helping to discover the connections between them, because many of our sources are translations into a Germanic language from another language, with the syntax of the translation heavily influenced by the syntax of the original. Robinson explains:

Many of the texts that do exist are undependable witnesses to the syntax of their languages. A majority of the translations from Greek and Latin originals are almost totally worthless as examples of native word order, government, use of participles and cases, and so on. Additionally, many indigenous poetic works are also somewhat suspect, given their propensity to violate grammatical rules in the service of a higher, esthetic standard. Descriptions of earlier language stages based upon an indiscriminate collection of examples from all available texts may thus significantly distort the true state of affairs in the languages involved. (163)

Whether or not the lexicons of languages can tell us anything about their inter relationships is a matter of debate which will be discussed later.

Right now we are looking at how the languages are traditionally grouped, so I will begin with a discussion of the morphological features (see Appendix IV). In order to get as close as I can to the time period from which our historical and archaeological information has come, I will be examining the oldest attested forms from each of the three groups.

The morphology of Old Norse contains many interesting features which can be used to set it aside from the rest of the Germanic languages. A Proto-Germanic ending for a-stem nouns which has been reconstructed as *-az rears its ugly head in Old Norse in the form of an r. This does not occur in any of the other languages. In fact, this ending has been lost altogether in all of the other languages except for Gothic, where it is represented as s. The definite article in Old Norse has an origin entirely separate from those of the other Germanic languages. The language also distinguishes between plural and dual in first and second person personal pronouns. Most of the other languages involved fail to make this distinction. The third person pronouns are different. In some areas Old Norse shows more inflection than the other languages, and it has a set of inchoative weak verbs which are formed in a manner similar to those of the same set in Gothic; the other languages are lacking in this respect. Old Norse also has a set of strong verbs which reflect what could formerly have been reduplication. Reduplication is a process by which the root onset of a strong verb is copied, a vowel sequence is added to it, and it is put to the left of the root as a prefix. This feature, along with possible ablaut, is used to inflect the past tense of some strong verbs. Gothic is the only language in our discussion which retains true reduplication.

Many of these features are seen as retentions from Proto-Germanic, but quite a few of them are innovations. Innovations which are shared between two or more languages are often assigned more value in determining how closely the languages are related to one another. If two languages which originally came from the same source retain a feature from that source, they could have simply split from the source and gone their separate ways with no further contact with one another. If however, two languages coming from the same source have the same innovative feature, they may have remained together (or at least in close contact with one another) until that innovation took place. It is important to note that just because two languages possess the same innovative feature, there is no guarantee that they were together when the innovation occurred; it is possible for the same feature to develop independently in different languages. Sharing innovations merely gives us better odds that the languages in question are closely related. If two languages do not share an innovation however, it is very likely that they split from one another before the innovation took place. So far, Old Norse seems to have a lot in common with Gothic, but many of these commonalities can be seen as retentions, and thus do not really tell us much.

A look at the East Germanic morphology should perhaps begin with reduplication as this feature seems to set Gothic apart from all of the other languages. In the Gothic preterite, the reduplication process takes the root onset and adds the vowel [ai], then prefixes it onto the root. Thus the preterite form of háitan surfaces as haíháit, and the preterite of fâhan becomes faífâh. Occasionally, the root vowel is changed to ô as in saian to saisô. Gothic also has the unusual class of weak inchoative verbs like Old Norse. The verbs are derived from adjectives by adding the ending -na or -nan. Many of the other morphological features of Gothic have already been mentioned in conjunction with those of Old Norse. Any other features which become important to the discussion will be mentioned at that time.

Now we turn to the morphology of the West Germanic languages. This is the only one of the three groups in which there is more than one attested old form. What sets these languages apart from the others? What keeps them together with each other? The morphology certainly helps. There are many aspects of the morphology which keep the West Germanic languages apart from the North and East branches, and, at the same time, unite the West Germanic languages with each other. The masculine nominative a-stem noun ending, present in both Gothic and Old Norse, is missing from all of the attested West Germanic languages. None of the West Germanic languages show reduplication, although Old English has a possible reflex of former reduplication. In the other languages the Gothic seventh class of strong verbs, the reduplicating class, has been leveled into the rest of the verb system. The West Germanic languages also lack the -nan class of weak inchoative verbs present in the other two groups. Our Western bloc also shows a few features which their northern and eastern brethren lack. All of the West Germanic languages have gerund forms of verbs, and they have short forms of the verbs <stand> and <go>. These features are nonexistent in the other two groups. Granted, there are some features which various languages in the West Germanic group have in common with the languages of the other two groups but not with other languages of their own. For example, Old High German has distinct case differentiation between dative and accusative in the first and second person personal pronouns. The other West Germanic languages have leveled this out, but the feature is present in both Old Norse and Gothic. There are other similar examples of specific features overlapping across the groups, but by far the weight of the features would lead us to believe that these groups are correctly delineated.

The morphology is used to group the languages by means of commonalities versus differences, thus the West Germanic languages have more things in common with each other than they do with the other two groups, and the other two groups, although they seem to have much in common with each other, have enough differences to be classified separately. The phonological features of the various languages are used in the same way, although sounding the deep sea trenches of differentiation is a little more difficult; the waters are murkier.

As far as the phonological features are concerned, there is a lot more overlap across group lines (see Appendix IV). In some cases, Old Norse will share a feature with Old Frisian, a West Germanic language, which all of the other languages lack (the loss of Proto-Germanic word final *-n). In other cases, Old Low Franconian and Old High German will share the same reflex of a Proto-Germanic vowel with Old Norse (the reflex ei from Pgmc. *ai). The overlapping phonological features are not consistent enough to require regrouping the languages, and as they do not actually contradict the current grouping (and some of the phonological features support the current grouping) this overlap can be dismissed.

As I mentioned before in regard to the morphology, innovations are regarded as more important in determining the relationships between the languages than retentions of older forms. When examining the phonological features of Gothic and Old Norse, one finds that they seem to have less in common phonologically than they do morphologically. There is, however, one phonological innovation which has helped drive a wedge separating these two languages from the West Germanic and pushing them uncomfortably close to one another. This feature is called sharpening.

Sharpening is a process by which two reconstructed Proto-Germanic glides *y and *w became geminated when they followed a short vowel and preceded a long vowel. In the next step of the process, Gothic and Old Norse took the geminated glides and turned them into sequences of two stops plus a glide. Thus the glide sequence *yy became ddj in Gothic and ggj in Old Norse, and the sequence *ww became ggw in both languages. In the West Germanic languages, the glide sequences yielded a diphthong plus a glide. For example, compare Gothic bliggwan (from Proto-Indo-European *bhlewo-) with Old High German bliuwan (Prokosch 93). Sharpening is a process which occurs in two stages. First, the gemination of the original glide takes place, then that gemination is affected by a second process which in two of our groups stops the sequence and in the other influences the vowel. It is the stopping of the glide sequence which interests us.

Suzuki looks at this problem from a phonotactic and syllabic perspective (172), whereby, in order to achieve ideal syllable structure, the glides are strengthened. Initially, because the onset to the second syllable is low in consonantal strength, thus producing an undesirable syllable structure, the second glide is strengthened. Then the first member of the glide sequence is strengthened, thus closing the first syllable and thereby maintaining the short vowel. Suzuki's explanation of the second strengthening is that it is a generalization of the identity of the segment. That is to say, the first glide was strengthened because the second glide was.

Many people have looked at this problem and attempted various solutions, but the most pertinent point for our topic is why Gothic and Old Norse would treat the sequence in such a similar manner, while the other Germanic languages all chose a different way. This fact, along with the morphological similarities between North and East Germanic, has led many to believe that they are more closely related to each other than they are to West Germanic. This is compounded by the belief that the Goths originally migrated from southern Sweden to the mouth of the Vistula and then to the Balkans. As I said before, just because two languages share the same innovation, in this case a very interesting treatment of a phonotactic dilemma, it does not mean that they might not have developed the innovation separately. Actually, given Suzuki's view, I would be inclined to say that the strengthening process is not as unnatural as it may seem, therefore making it that much more possible that the innovation occurred separately.

The analysis of linguistic data in this manner relies heavily on reconstruction, that is, on taking cognates from two or more different languages at the same point in time and deducing the older form of the words through the application of various phonological processes. Reconstruction can be a very dangerous means of acquiring information. One of the pitfalls of reconstruction is that there may be many intermediate steps from the oldest attested forms of each language used in the reconstruction to the language one is attempting to reconstruct. Linguists try to reconstruct these intermediate steps to explain certain phenomena. The gemination of a proto-glide, for example, would be an intermediate step in the process of sharpening, between the attested forms of Germanic and the supposed Proto Germanic. This process fails to take into account the fact that sometimes items are completely dropped from a language and other items appear almost out of nowhere. A good example of this would be if we tried to reconstruct an attested old language from its modem descendants without using any of the intermediate forms which in this case would also most likely be attested. This process was examined by linguist Ernst Pulgram, who found that if we try to reconstruct Latin on the basis of the modern Romance languages alone, the picture we get is somewhat skewed. We would produce words like caballum for horse and guerram for war, when the actual words attested in Latin are equus and bellum respectively (Renfrew 85). The worst part about reconstructing Proto-Germanic, is that we cannot go back and check our work as we can with Latin; we will never really know if we are correct. Renfrew also quotes linguist Leonard Bloomfield on this point:

The comparative method assumes that each branch or language bears independent witness of the forms of the parent language, and that identities or correspondences among the related languages reveal features of the parent speech. This is the same thing as assuming, firstly that the parent community was completely uniform as to language, and secondly that this parent community split suddenly and sharply into two or more daughter communities, which lost all contact with each other. (Renfrew 103)

If the comparative method (reconstruction) is a dubious methodology, then anything which is based on it must also be dubious. The method of grouping the languages and assigning their position on the Stammbaum based on that grouping is a process which is based on the assumption that the commonalities between the languages can be used to trace them back to a single parent. Are we to believe that the Germanic languages came from a broken home? Using a methodology which is based solely on attested features produces some different results.

Witold Mańczak used a method of comparing attested lexical items rather than morphological and phonological features. He compared an excerpt from the Gothic version of the Bible to modern translations of the Bible in the other Germanic languages and found that by percentage of occurrence of lexical items, the Gothic work more closely resembled Modern High German than it did Modern Swedish ( Mańczak 5). If the Goths were supposed to have come from Sweden, he argues, then Gothic should resemble modern Swedish more than it does Modern High German. This leads him to believe that the actual breakdown of the three groups should be North, Middle and South, with Gothic being South, rather than North, East and West. He operates under the assumption that the morphological and phonological systems of languages are unstable and are not useful in demonstrating genetic relationships between languages. He points out that Gothic has more inflectional (morphological) features in common with Old Church Slavonic, a member of the Slavic language family, than it does with English which is supposed to be in the same Germanic family (7). He goes on to demonstrate that the number of lexical convergences between Gothic and English is greater than that of those between Gothic and Old Church Slavonic:

In total there are 93 lexical convergences between Gothic and English and 74 lexical convergences between Gothic and Old Church Slavic [in his sample]. That is to say that the unanimous opinion according to which linguistic kinship depends on grammatical structure is false. In reality, not phonetic or inflectional features but roots allow us to determine linguistic relationship, with the restriction that roots are to be counted not in dictionaries but in parallel texts. (9)

He uses examples of reconstructed roots (although as stated, the method is not dependent on reconstruction) to show that the root is stable while the grammar is not: "Examples of the type *sūn-u-s> Engl. son or Russ. syn, *es-ti> Engl. is or Span. es, *bher-e-si> Engl. bear show that not the ending, not the thematic suffix, but the root is the most stable part of the word" (10). Mańczak has received a lot of criticism for his ideas, in fact, the article from which I took these examples was a rebuttal to J.E. Salmons' criticism of Mańczak's original work. Even as a novice in the field of linguistics, I could think of several reasons for criticizing his theories, but that is not the point here. The point is to demonstrate that if a different methodology is used, a different result will be achieved. Deciding whether a methodology is valid or not cannot be based on evidence, because there are no attested forms which capture the languages at in-between stages of development. The choice of validity is ultimately a leap of faith. A methodology which takes as much information as possible into account must be chosen and worked with, and if the answers produced go against the grain, it cannot be assumed that it is because the answers are wrong. It could be that our understanding of them is wrong, or it could be that the beliefs that have been accepted in the past were wrong. People change their minds about theories in this field all the time. It's when people close their minds off completely and refuse to analyze things anew that we cease to learn anything. This would allow us to perpetuate fallacies indefinitely, ultimately constructing a world of academic lies.

I believe that the only way to accurately assign genetic relationships between the languages is to take historical, archaeological, and linguistic data and combine them somehow. It is not enough that these areas should support each other, but they should each come to the same conclusion independently. If we arrive at the same answer having come from three different directions, it is that much more likely that our answer will be correct. Again, there will be no real way of checking, until science bestows upon us the gift of time travel, but at least it will be a more accurate picture than the one we have now.

In attempting to synthesize the historical information with the archaeological and linguistic information, we find that Tacitus and his fellow historians came fairly late in the game. If the Goths truly came from Sweden, thus making North Germanic and East Germanic closer to each other than either to West Germanic, and Tacitus places the Goths on the Vistula, then the migration from Sweden must have taken place before Tacitus' time. If this trek had already been made, and the innovations common to North and East Germanic came into being while the languages were still in contact, the split between North/East Germanic and West Germanic must have taken place long enough before Tacitus' time for the innovations to have developed and then the split between North and East Germanic to have occurred. I don't know exactly how much time that would take, but the development from Old Saxon to New Low German, with about a thousand years between the two, shows that it can take a very long time for languages to develop. The North and East Germanic reflexes of sharpening for example, must have taken a couple of hundred years to come about; then there must have been another couple of hundred years put between Gothic and Old Norse to separate them far enough to be considered different groups. Thus, Gothic and Old Norse underwent a lengthy period of combined development separate from the West Germanic languages, followed by another, possibly lengthier period of independent development. The problem is that the history does not address the period in which this could have taken place, while the archaeology doesn't tell us anything about the linguistic developments.

Looking back further, without the benefit of written history, we find that the spread of agriculture across Europe started in the south and went north. The earliest archaeological evidence of farming in Europe was found in Greece and dates to around 6500 BC (Renfrew 242). Farming sites in France and Germany do not surface until over 1000 years later. If the Indo-European languages slowly spread with the agriculture from south to north across Europe, developing linguistically along the way, this should produce the same sort of dialect continuum in the Germanic languages that we see today. Since Gothic is the oldest attested source of a specific Germanic language that we have (Wulfila's Bible translation was done in the fourth century AD), and is also closest to the reconstruction of Proto-Germanic, it might follow that Mańczak's assertion is actually correct: Gothic is a South Germanic language and did not come from Sweden. If so, that would leave us with a south to north dialect continuum which parallels the spread of agriculture. Thus, by combining one linguistic method, in this case a controversial one, with archaeological evidence, we come up with something that might fit. This is a case where a conclusion not only goes against the grain of most modern linguistic thought, but also flies in the face of accepted history, for if it is correct, then the migration of the Goths from Sweden to the Vistula never took place. It is also a case where two forms of evidence point in the same direction, south to north. I am not saying here that this is necessarily what happened, but this is the type of methodology which needs to be used if we are ever to determine what did happen.

I pointed out above that the historical information comes long after the supposed split of the Germanic languages from their original source. The linguistic information also comes too late. By the time we start getting written examples of Germanic languages the three main groups have long since diverged and taken their different paths. If reconstruction (comparative linguistics) is a valid methodology, we can use it to take up some of the slack in the time differential, but as I mentioned before, reconstruction is guesswork. Although guessing might produce the correct answer, there is no way to prove it even in a case where it may seem fairly straightforward.

There is one example where archaeological, historical and linguistic evidence converge. This is a helmet hoard found at Negau in Slovenia. The hoard has been dated to around 100 BC. One of the helmets in the deposit has writing on it which appears to be in a Germanic language. It is written in what looks to be a North Etruscan alphabet, and the text has been variously translated as <Harigast dedicates this> or <This is dedicated to Harigast>. The helmet may be associated with the historically attested movements of the Cimbri, who won a battle against the Romans at Noreia, near Negau, around this time. This helmet, if it is from the Cimbri, could give us a clue as to what their language was. However, with all the reconstruction that has been done of the Germanic languages, linguists still cannot determine exactly what language this is. They cannot even agree on a translation. History says that the Cimbri were there, the linguistic evidence discovered by archaeology shows that Germans were there, but because we cannot identify the language specifically as Cimbric we cannot say for certain that the hoard came from the Cimbri. So even here, where all three types of evidence seem to converge, we are still left without an answer.

In this paper, I have attempted to show that a theory claiming to show the genetic relationships between the Germanic languages cannot be based solely on linguistic information. The Stammbaum theory finds little support in the archaeological record, and as I pointed out, if one uses a method which is different from that used in the genesis of the Stammbaum, one produces different results. This fact alone should be enough to cast some doubt on the methodology used in the Stammbaum theory. There is a lot of debate over what constitutes a valid methodology in linguistics, but I think that the methodology will ultimately be validated by producing a correct answer. This may sound like a paradox, but I believe that as long as there is debate over methodology, we will not find the answer to our problem (how the Germanic languages are actually related to one another), and when we do find an answer, the argument will stop.

We will never be able to make more historical evidence regarding this time period, but it is possible that more information will turn up of its own accord. Unfortunately, since there is no evidence that the Germans were writing much of anything, let alone their memoirs, it is unlikely that we will ever receive more accurate historical information about the Germans than that which we already have. Hopefully I have demonstrated that this information cannot be taken as entirely accurate.

Every day archaeology produces more data, which I suppose could either be seen as taking steps towards an answer or adding pieces to the puzzle. I firmly believe that there will come a day when we will have a definitive answer as to how the Germanic languages are related. That answer will come not from linguistics, not from history nor from archaeology. It will come from a method which combines all three of these disciplines and uses them not merely to support one another, but to show that, from no matter how one approaches the problem, the same solution presents itself. That will be the day when the guns of academic conflict will finally fall silent and a new era of interdisciplinary cooperation will dawn. The artillery will rust, its cobwebs blowing in the breeze of knowledge.