Featured post

Gender in the Merovingian World

Tuesday 31 May 2011

Archaeology and Migration: Rethinking the Debate

[This is the text, more or less as given, of the key-note lecture I gave to the conference on on The Very Beginnings of Europe? Cultural and Social Dimensions of early Medieval Migration and Settlement (5th-8th centuries) that took place in Brussels on 17-19 May. The argument is essentially that by concentrating on debating the existence or scale of migration we are missing the really important questions, which concern why migration took place in the first place.  I also argue that the terms of the debate need radical rethinking, to  place fifth-century migration in a longer-term perspective looking at both sides of the divide between the 'Roman' fourth century and the 'migration-period' fifth century and to see the Roman and 'barbarian' regions as interlinked parts of the same world rather than as two antagonistic, opposing, confronted worlds.  Doing this will not only allow a better understanding of the migrations themselves (as I suggest by a quick look at the North Sea regions) but also permit a more politically responsible contribution to modern debates on migration.]

1. Introduction and …
It course a great honour for me to open the proceedings of such an international important conference on the origins of Europe by setting out my own perhaps rather idiosyncratic views on migration in what I prefer to call the post-imperial period, the fifth and sixth centuries, and I would like to thank Dries Tys and the organisers for their invitation.
2. … Historiography: Polarised debates
I must express my sincere thanks for this opportunity, because it may very well be that what I am about to say will not please very many people very much!  “A plague on both your houses!” cries Mercutio, famously, in Romeo and Juliet, and in some ways that will be the Leitmotif of what I have to say this evening, although with luck – unlike Mercutio – I will not find myself dying from stab wounds shortly afterwards, no matter how unhappy some of you might be with the general thrust of my argument.
I say this because although the study of these centuries around the North Sea has seen the production of ever more, and ever-better-quality, data of all sorts, new and old, they still seem to me to be being forced into interpretative frameworks which have not developed with the same rapidity, or, indeed, at all.  In some specific cases we seem to be going backwards, in our interpretations, almost as fast as techniques of recovery and analysis are going forward.  This retrograde movement is ironically veiled by a smoke screen laid down by appeal to the latest scientific developments.  I also want to say, very forcefully, that – politically – this is no time for a decline in the sophistication of our interpretations of the historical phenomena of migrations, their causes and consequences.  This is much more than an academic issue; it is one of social and political responsibility.
A good case study to illustrate what I mean is the analysis of settlement and house styles around the North Sea.  Traditionally, changes on rural settlement sites and the new forms of building that appeared in the fifth and sixth centuries were ascribed to movement from Germania into the former provinces of the Roman Empire.  These changes are well known: the spread of the Grubenhaus, the appearance of post-built halls with confronting doorways, and so on.  The similarity in building styles between lowland Britain (or England) and the North Sea areas of Germany and the Netherlands cannot be denied.  However, in the 1980s, as the popularity of migration as an explanation declined, people looked more closely at the settlement data without the existence of migration governing their analyses.  It was noticed that post-built houses had many of the same features, in terms of the ratio of length to breadth, etc., as some buildings on Romano-British sites.  A classic article by Simon James, Martin Millett and others suggested that there might have been a fair amount of Romano-British input into the development of the Anglo-Saxon hall.  ‘How Saxon', it was asked, 'is the Saxon house?’  The absence from Britain of the long house, the Wohnstallhaus, so familiar from North-West mainland Europe, was noted, as was the more disorderly arrangement of early Anglo-Saxon settlements when compared with the lay-out of, say Feddersen Wierde in the fourth century.  With these points in mind, some authors began to question the extent to which there had been an Anglo-Saxon migration at all.  The ‘no migration’ argument finds its most extreme form in a popular volume written by the Bronze Age specialist Francis Pryor, to accompany a television series.  This gave voice to the idea that perhaps the whole idea of the Anglo-Saxon migration was an historical myth.  Certainly some younger British Anglo-Saxonists liked to express this interpretation, often linking it to the alleged dominance or tyranny of a set of documentary historical narratives.
The posing of these sorts of question must be seen as important for the development of the discipline, and many of the points raised were either valid, or at least needed to be raised to make people look more closely at the evidence.  Nonetheless, the revisionist interpretation came in for formidable critique from Britain’s leading authority on early medieval settlement forms, Professor Helena Hamerow, who pointed out, notably, that the Wohnstallhäuser disappear on fifth-century mainland sites too and that fifth-century settlements in the Anglo-Saxon homelands also lost their planned lay-outs in the fifth century.  Thus the opponents of the migrationist reading of the settlement data were not comparing like with like.  Fifth-century settlements in lowland Britain were much more like their contemporaries on the east shore of the North Sea than the critics had imagined.
As I want to stress throughout this paper, the movement of people across the North Sea in the fifth and sixth centuries is not in doubt.  One does not need the written data to know about that.  The linguistic change that occurred in what is now England after the fall of the Roman Empire is impossible to account for without migration, and the same is true of a range of archaeological or material cultural features: certain forms of artefact and the cremation burial rite.  But it is also important to underline what I just said: it is impossible to account for these changes – fully or with any real degree of persuasiveness or conviction – without population movement; I did not say that population movement is what, in itself, explains them.
The revisionist interpretation of the Anglo-Saxon settlement data rapidly went from reasonable arguments, that the new house types were more of a hybrid than people had believed, to wild denials that a migration had even taken place.  Similarly, arguments in favour of the input of Anglo-Saxon migrants into socio-cultural change, usually (it must be said) rather less subtle even in their moderate form, soon reached the point of explaining all change in terms of migration and incorporating various extra, less intellectually worthy, elements.  The measurement of the width of doorways was alleged to show that houses were ‘Anglo-Saxon’ on both sides of the North Sea.  I have to say that measuring doorways to prove the migration of an ethnic group, seen in fundamentally primordialist terms, does not, to me, represent a great theoretical advance from measuring skulls to the same ends, such as went on a century ago and alas still seems to take place in French archaeology.
Similar arguments are visible elsewhere within settlement archaeology.  It was once proposed that the first phase of the famous Anglo-Saxon high status settlement site at Yeavering was British.  The counter argument essentially involved finding parallels for the various building features, concluding that more were found on ‘Anglo-Saxon’ sites than elsewhere and thence declaring the phase to be ‘Anglo-Saxon’.  In 2011, one really must wonder whether this is a valid means of establishing the ethnicity of a site’s inhabitants.  The discussion just referred to implies – essentially – that if one can determine where material cultural traits originate, geographically (or where most of them do), then that will tell you the geographical roots of the people concerned.  In turn, runs the implication, the geographical roots of the people concerned will provide their ethnic identity: Saxons or Britons.  When this reduces to a simple numerical exercise one really has to wonder whether it will withstand any sort of critical scrutiny.
Other examples multiply, especially when one moves outside England.  The spread of Anglo-Saxon settlement in what is now Scotland has, for example, been measured by the distribution of Grubenhäuser.  Yet, these are now known from sites right up the North Sea seaboard of Scotland as far as the Mounth.  Similarly, although such ‘Sunken-Featured Buildings’ are known in Gaul from the third century and are increasingly being discovered across Roman and post-imperial Europe, and are not known universally, either through time or region, in Germania, these buildings are still somehow claimed to represent unproblematically the presence of ‘Germanic’ newcomers, wherever these newcomers may have come from and whether or not they ever used such buildings in their homelands.  I have heard two objections to this: that the early Gallic examples do not have the ‘right’ number or combination of post-holes and that whatever archaeology might say, Tacitus demonstrates that the Germani had Grubenhäuser in his day…  Leaving aside just how worrying – indeed in my view just plain frightening – the perpetuation and persistence of such crude Germanism is, amongst scholars as young and otherwise intelligent as the origin of that last remark, the debate sums up some principal problems with the discussion of migration around the end of the Roman Empire.
It raises a number of crucial issues.  The first is the polarisation of the argument.  Either migration occurred and it explains every observable change in the archaeological record, or it did not occur, or had no explanatory role.  Note too that frequently – even if one does not wish to talk either side in this debate – one still finds oneself assigned to one or other camp.  Thus, for saying that various cultural forms are not evidence of migration I have been accused of denying that there were migrations.  Anyone with a university post really ought to have the intelligence to be able to distinguish between the argument that a piece of evidence does not prove that a migration took place, and the argument that no migration took place.  The first argument might be employed as an element in making the second argument but there is no reason at all why it even has to imply that the second argument is being made.  It is a shame that many eminent archaeologists have not been able to distinguish between these two arguments but evidently they have not.

The second issue is the domination of the debate by a historical master-narrative, even in counter-migrationist works.  The framework of the debate continues to be: did this thing referred to in later written sources happen, or did it not?  Crucial is the way that the archaeological evidence is simply not allowed to speak for itself, especially when what it seems to be saying contradicts a set of predetermined ideas. 

This is especially visible in looking at burial archaeology.  The introduction of the rite of furnished inhumation is still regarded as an index of the appearance of ‘Germanic’ barbarians within a region, regardless of the fact that the archaeological record provides absolutely no prima facie support for the idea.  Frans Theuws, Sebastian Brather and myself have been making this point for a long time and it might be that it is beginning to gain ground, but one only needs to look at the circular arguments and utter illogic that pervade the defences of the traditional position to see how reluctant archaeologists are to abandon it.  It is time to consider afresh other aspects of the material cultural record, like the Grubenhäuser, using the archaeological data similarly to see what they say on their own terms.  When we do so, we should not reject their testimony if they appear (superficially) to contradict the historical master-narrative. 
A third problem revealed by the discussion of the Anglo-Saxon debate on the origins of settlement and building types is its attitudes towards ethnicity.  This all too often remains untouched by the decades of research into the nature of ethnicity by anthropologists, sociologists and historians.  It seems still to be believed that archaeology provides some way into discovering the ethnicity of the people whose remains, funerary or otherwise, skeletal or cultural, are uncovered in the course of excavation.  In 2011, it is surely time to admit that it does not.  Objects do not have an ethnicity and the plotting on distribution-maps of cultural features – ritual, artefactual or in terms of building design – even where a geographical and chronological point of origin is revealed, does not yield any clear insight into the ethnicity of the people involved.  While I want to stress that this is a point that ought to have been learnt by now, I also want to stress – equally clearly – that this does not mean that archaeology has nothing to say about the archaeology of the early Middle Ages, as seems (inexplicably) to be the fear of traditionalists such as Volker Bierbrauer and Gianpietro Brogiolo.  Nor does it even mean that archaeology has nothing to say about ethnicity or ethnic change.  It certainly does not mean that ethnicity was unimportant in the period, which is a conclusion that has wrongly been drawn from the lack of a direct correlation between the archaeological record and historically-attested ethnic groups. 
The whole ethnicity debate is a manifestation of just how deeply archaeology of this period is permeated by the documentary historical master-narrative, incidentally and ironically a master-narrative of which most historians are nowadays sceptical.  Ethnic identities such as Frank, Lombard, Saxon, Goth, Frisian or Alaman could not and could never emerge from the archaeological record on its own, no matter how many computers the data were put through.  To talk in terms of a ‘Frisian’ brooch or a ‘Frankish’ buckle automatically, and by that act alone, subordinates the archaeological record to the documentary, evaporates any claim at all to be using pure archaeological argumentation or to be following any rigorous ‘Cartesian’ approach to the data.  However, we can still read, defences of H.J. Eggers’ 1950s argument that archaeology’s most important role in understanding this era is through the identification and study of ethnic groups. Eggers famously argued that he was avoiding ‘Mischargumentation’ but using historically-attested ethnic groups to name any grouping at all in the archaeological record is impossible without Mischargumentation.   Saying this does not represent the intrusion of some sort of new-fangled form of post-modernist theory into early medieval archaeology; it is simply the application of logic and argumentative rigour.  
Related to this is another key problem raised by the case study I mentioned: the persistence, even if often merely implicit, of Germanism, the idea of some kind of shared Germanic ethos or identity that transcends things like time and space.  Thus, if Germani from a particular area, who have never hitherto used burial with grave-goods, settle in a specific part of the Roman Empire and furnished inhumation is found there, then that custom must be a sign of the presence of these Germanic newcomers, because other peoples within the enormous area occupied by speakers of Germanic languages, extending from Scandinavia to the Ukraine, used this rite at some point (whether or not contemporaneously or even, on occasion, in a prior era) and the rite can therefore be adjudged ‘Germanic’.  The same goes for other material cultural features such as the Grubenhäuser
The fact that this approach stems ultimately from Graeco-Roman ethnographic chauvinism, turned round in the German renaissance and then, especially in the nineteenth-century process of German reunification, is well-known.  Less widely-appreciated, but well-documented, is the fact that the term ‘Free Germany’ or Germania Libera, so frequently encountered in the literature, originated in the twentieth century.  The uses to which the idea of a pan-Germanic history, culture and ethos were employed in the middle of the twentieth century are generally recognised, even if the ways in which they are essentially continued beliefs and political agendas that were common enough in the nineteenth are not.   And yet, in spite of the modern, politically contingent nature of these formulations being established facts, the idea that one group of germani can be treated interchangeably with another, and that there were fundamental, shared aspects of ‘germanitas’, of a ‘Germanic’ culture and ethos, continue to be encountered.  Thus the term ‘Germanic’ continues to be employed as though it has any meaning or analytical worth at all outside the sphere of linguistics.  Historians now argue strongly that it does not.

3. A conflict of theories? 

The polarised sides in the argument over migration have deployed their own bodies of theory.  Those who minimise the importance of migration have tended to turn to theory about ethnic identity and its mutability developed by social anthropologists, perhaps refining this with theories of practice, most obviously that developed by Pierre Bourdieu.  This lays great stress on human beings as active agents in historical change, through the interplay of all sorts of different political identities, of which those based on ethnicity are only one aspect.  The ‘migrationist’ camp has, by contrast, deployed so-called ‘migration theory’.  The problem here is that one body of ‘theory’ by no means stands in opposition to the other.
‘Migration theory’ in some ways isn’t theory at all, or at least not theory on the level of Bourdieu’s or Giddens’ structuration, or their analogues.  It is rather a set of observations about how migrations work.  This is not to imply that it is not of significance.  It is of considerable importance, although in a different way from those in which it has been employed thus far.  The issue is that Migration Theory has simply been adduced to show that migration could have taken place, even on a large scale, in the early Middle Ages.  Once the migrants have moved, however, one is still faced with the same old problems about how the newcomers could bring about cultural and political change or about how the ethnicity of what, even on a maximalist view, could only represent a small minority of the post-imperial population of a former imperial province, might come to be adopted by the remainder, and for that one needs to resort to the other types of theory, just mentioned, about socio-cultural interplay.  Or one assumes that migration suffices to explain all of these changes on its own.  And indeed that is usually what happens.  Once having adduced Migration Theory – and in so doing evidently thinking that one has countered one’s opponents’ use of Bourdieu and the rest – the argument progresses no further in explaining cultural, ethnic and political change.
Putting together all of what I have said thus far, one can perhaps see where my argument is going.  On the one hand we have the production of ever larger quantities of data, of good quality, and on the other we have the persistence of fairly old-fashioned, polarised ways of thinking about the period and indeed about the role of migration in particular.  Thus it seems to me that this data is being hammered into a framework which it will not and cannot fit, and that therefore neither side in these debates is capable of carrying the day.  Neither side is capable of producing a really convincing explanation.  This is why in my view so little real progress has been made in the interpretation of the archaeology, even as our knowledge of the nature of the archaeological data has progressed astonishingly.  In a negative reading, one could return to my Mercutio quote and my ‘plague on both your houses’: in other words both sides are wrong.  In many ways I could defend this position, but a more politic or emollient reading might be that both sides are right – as far as they go, or if we reconfigure the questions correctly.  Or that both sides have important things to bring to the table.  Frequently, as I read it, the traditional migrationist viewpoint is based in the better knowledge of archaeological data, but the opposite viewpoint frequently has the more advanced and more subtle ways of reading those data.
It is, therefore, time to rethink the debate and the framework.  What I want to do in the remainder of this lecture is to suggest other ways in which the framework needs to be adjusted to get us out of long-standing and unfruitful arguments, and then suggest a way towards a new model.

4. Migration and archaeology

The base line of my argument is acceptance of the fact that migration happened.  There is no place here for migration-deniers!  In some cases the archaeological and linguistic evidence makes that abundantly clear.  The change of the language from either Brythonic or (with more likelihood in my opinion) low Latin to Old English in what is now England and lowland Scotland, or from Gallic and low Latin to Old High German along the Rhine is impossible adequately to explain without invoking migration.  The spread of a cremation ritual from the North Sea coastal regions of Germany to England, or the appearance of various artefact forms there with direct ancestors in what we think of as the Saxon homelands: all these things cannot be fully explained without invoking the movement of people – even if we must admit that this was not the only factor involved.
But the fact is that normally migrations are very difficult to detect archaeologically.  Migrants frequently leave no material cultural trace.  Often they adopt the culture of their new homeland very rapidly.  We know full well from a range of sources that many, many thousands of barbarians from Germania and elsewhere crossed into Roman territory, throughout the whole period of the late Republic and the Empire.  Yet these immigrants left almost no archaeological trace of their presence.  Often the only archaeological traces of their movement are in the opposite direction.  Thus the Saxons who joined the Roman army are only visible archaeologically from the fact that after returning home they were cremated wearing their old uniform, with its buckles and brooches, or had these items buried with them. 
The Anglo-Saxon migration to Britain is one of the few migrations that can be demonstrated independently of historical evidence.  Indeed it is more or less the only post-imperial migration in from barbaricum into the Empire that can be so demonstrated, with the possible exception of some short-distance movements across the Rhine and Danube.  Ironically, most of the population movement attested to by the late antique archaeological record is in completely the opposite direction – from the Empire to barbaricum.  Further, much of the evidence cited as archaeological proof of migration is no such thing.  Furnished inhumation is one example; I think that the Grubenhaus will turn out to be another, once more rigorous levels of argument are adopted.  That, however, does not mean that migration did not happen.  When looked at in comparative context it is, if anything, the Anglo-Saxon case that is the unusual and unexpected one.  There is no especial reason why we should expect migration to show up at all in the archaeological record.  As I just said, four centuries of migration from barbaricum into the Empire are more or less archaeologically invisible.
5. Ethnicity
One reason why this is the case is that ethnicity – so often the object of archaeological investigation – is itself archaeologically invisible.  It is a shame that in 2011 this is still a point that needs to be made. 
Early medieval historians, since the 1960s and thus for nearly half a century, have adopted and adapted ideas from social anthropology about the ways in which ethnic identities are mutable and not a fixed ‘given’.  Analyses of ethnicity reveal that there is no fixed element is that defines an ethnic group.  For every case where such a grouping was said to be defined by a belief in common descent, or shared religion, or language, or whatever, there was another where the opposite was the case.  The only constant is that ethnicity is a simple matter of belief.  People think of themselves as belonging to one group and think of other groups (for whatever reason) as different.  Trying to find an innate (or primordial) factor, allowing us to identify past people as members of an ethnic group, other than what they said they were (at particular times), is a pointless task.  The implications of this for the archaeology of post-imperial Europe are obvious. 
Furthermore, anthropological studies of ethnic groups in Africa and south-east Asia revealed quite clearly that the links between material culture and ethnic identity were very vague indeed.  Sometimes they were the diametric opposite of those assumed in traditional archaeology; artefacts associated with one group were actually used in another to stress other types of social difference, such as age-grades.  What people said were the distinctive traits of their ethnic group ran contrary to what could actually be observed in practice.  One can find late antique parallels for these features.  These problems must be borne in mind when reading archaeological interpretations of ‘Britons’ and ‘Saxons’, ‘Franks’, ‘Alamans’ or ‘Gallo-Romans’.  They raise practical obstacles to accepting such ethnic identifications which are all but insurmountable. 
The other crucial lesson of modern social anthropology (and closer study of the written evidence from mainland Europe) was that ethnicity can be changed.  In other words, over time, a family that at one point thought of itself as, say, Catuvellaunian Britons could come to see itself as Roman.  That same family might, with the passing of further centuries, eventually consider itself to be East Anglian or English. The reasons why people change their identity are complex but one important factor is political and social advantage.  It is thus not difficult to envisage how post-imperial people might adopt new identities in order to maintain or improve their social standing in a world dominated by warrior élites who claimed a barbarian ethnicity: Frankish, Gothic, Burgundian or Saxon. 
Putting all these ideas together we can conclude that, even if we can identify the geographical origins of a custom, type of object or feature of a building, this would not necessarily provide a sure guide to the ethnicity of the people who used them.  Even if we could plausibly link a material cultural feature to an ethnic identity, it would not necessarily mean that all such people came, originally, from the same area.  ‘Saxons’, for example, could include immigrants from all over the Saxon homelands of north-west Germany but also people of native Romano-British descent and perhaps people of other origins as well.  These conclusions are vitally important. 
They have not, however, gone unchallenged.  It has been objected that people cannot just pick and choose whatever identity they want.  There are constraints on this, such as whether the group into which membership is sought will accept them.  Furthermore, ethnic identity is not just a matter of political advantage; even minority identities can exert a powerful affective force, holding back change.  Simple ‘straight swaps’ are in fact not very clearly attested in the evidence from this period.  Actually, however, none of these good points ultimately negates the general thrust of the broadly constructivist/situationalist position vis-à-vis ethnicity.
Nonetheless, recently, as well as generally unconvincing appeals to cemetery data to support the counter-revisionist position, genetic evidence has been adduced to prove large-scale immigration, particularly the study of modern DNA.    Other, more sophisticated analyses of skeletal data have also been developed which might also be able to shed light on the migration.  These include the analysis of the stable isotopes in the tooth enamel of the excavated skeletons of this period and ‘ancient DNA’, extracted from bone samples from early medieval cemeteries.  Techniques have now advanced to a stage where this data is considered capable of furnishing usable conclusions. 
One might point to numerous general methodological problems involved in the study of modern DNA evidence.  For me, though, the crucial issue is that it fundamentally misses the point.  Examination of its underlying assumptions is revealing.  One assumption is that migration is something that happened in discrete periods, so that, for example, the fifth and sixth centuries are often known as the period of the migrations (an appellation long criticised).  Thus, in this view, demonstrable population-mixing can be dated to such specific blocks of time.  Yet migration is a constant of human existence.  I’ve already mentioned that people were moving from barbaricum east of the Rhine into the Roman Empire for centuries before 400.  People moved, within the Empire, on a large scale too.  And of course people have continued to move and to marry the inhabitants of other areas ever since.  The similarities between the DNA of England and Germany might result from migration from Germany in the late antique period but it might stem from such movement at many other times and, indeed, from movement in the opposite direction.  Once again, movement from the Empire to barbaricum, amply demonstrated in the archaeological record, is excluded from reasoning because it does not fit a model derived from problematic written sources.  Thus this use (or misuse) of DNA is driven by a particular, and a particularly crude, reading of history and its results chosen to fit this story rather than to examine it. 
Another, perhaps even more serious, problem concerns the movement from DNA to conclusions about ethnic or political identity.  DNA chains, like bronze belt buckles, do not have an ethnic identity.  One way in which one counters the points made against the mutability of ethnicity is by stressing that it is multi-layered.   It is deployed (or not) in particular situations as the occasion demands.  A person’s DNA will not give you a sense of all of the layers of that person’s ethnicity, or of which s/he thought the most important or even if s/he generally used a completely different one, or of when and where such identities were stressed or concealed. 
Let me illustrate this.  A male Saxon immigrant into the Empire in, say, the fourth century, would – one assumes – have DNA revealing the area where he grew up, but he would probably increasingly see himself, and act, as a Roman.  His Saxon origins would have no part in his social, cultural or political life, and even less for his children, if he stayed in the Empire.  If he returned home with all the cachet of his imperial service, it might well have been his Roman identity that gave him local status.  However, if a distant male relative of his moved into Britain a hundred and fifty years later, his DNA might be very similar but, in complete contradistinction, this man might make a very big deal of his Saxon origins for they would, or could, propel him to the upper echelons of society.  DNA tells us nothing about any of this. 
What is pernicious about this use of genetic data is that it is essentialist.  It views a person’s identity as one-dimensional and unchanging, and it sees that dimension as entirely derived form that person’s biological and geographical origins.  In short, it reduces identity to something very similar to nineteenth-century nationalist ideas of race.  Everyone sane knows that people moved from northern Germany to Britain in the fifth and sixth centuries.  In that sense, these expensive analyses tell us nothing we did not already know. 
However, in their implicit reduction of identity to a form of race, and by masking all the other contingent and interesting aspects of cultural interaction and identity-change, they not only risk setting back the understanding of this period by more than a century but provide pseudo-historical and pseudo-scientific ammunition for present-day nationalists, xenophobes and racists.  Many of the same general points can be levelled at other analyses of such things as the isotopes from teeth, which are far less exclusive in the geographical zones revealed than one might hope – again, the part of the map chosen is often determined by the historical story the analyst wants to tell.

6. Migration and explanation

All these points are profoundly significant and cannot be ignored.  Allow me to reiterate.  Archaeology does not and cannot in itself reveal ethnicity; genes do not and cannot reveal ethnic identity; we should not expect migration necessarily to show up in the archaeological record; yet there is no need to doubt migration happened in the post-imperial period, or even that it in some cases, such as the North Sea it happened, over time, on some scale. 
The important issue, implicit in what I have just said, is the relationship between migration and explanation.  The example I gave of the two Saxon migrants to Roman Britain makes the point clear.  The question that must be asked is why migration and ethnic change played a different role in the history of the fifth century than in that of the third or fourth.  As I have said, thus far the debate seems to be polarised into two arguments: migration happened, and therefore it explains everything that we can see happening; or it didn’t happen on any scale and was therefore unimportant.  Neither argument can be made to fit the data in any sophisticated or really convincing fashion.  We need new questions: a new agenda.
In the current political situation it is, furthermore, extremely important to establish this agenda.  Moving from the demonstration of migration to the explanation of the downfall of the Roman Empire through appeal to migration can be highly irresponsible, at best.  Southern German newspapers have, for example, reviewed a museum exhibition set out in the traditional migrationist paradigm as showing how ‘the Roman Empire fell because it could not keep out the immigrants’ and that we can learn from this.  When a British historian places an argument that the Roman Empire fell because of the immigration of large numbers of barbarians next to arguments that the end of Rome was the end of civilisation and that we need to take care to preserve our own civilisation, when another British historian writes sentences saying ‘the connection between immigrant violence and the collapse of the western Empire could not be more direct’, and especially when the arguments of both involve considerable distortions of the evidence to fit their theories, one cannot help but wonder whether these authors are wicked, irresponsible or merely stupid.  Are they setting themselves up as ideologues of the xenophobic Right or have they simply not realised the uses to which such careless thinking and phrasing can be put?  I have already drawn attention to the worrying implications of the use of genetic data to study this period, and they do not stop with implications about monolithic ethnic identities based upon genetics.
If we accept that it is fundamentally misguided to start looking at the archaeological record primarily in terms of whether or not it demonstrates migration, broad new interpretative horizons are opened up for us.  In particular, new horizons are presented and, simultaneously, ways of contributing in more humane fashion to current debates about immigration, if we break down the set of binary oppositions that bedevil the study of the period: Romans against barbarians; Saxons against Britons; (especially importantly) the Roman empire against barbaricum; the Roman period against the Migration period; Britain versus the mainland of Europe; the Saxon migration moving simply from east to west.  As a means of demonstrating this, I want to end this lecture by briefly suggesting some ways of recasting the study of the North Sea in this period.
6. The Channel/North Sea Cultural Zone: A different framework
My key point is that we should start seeing the late and post imperial North Sea in the same way as archaeologists have long become used to seeing it in the seventh century and later: that is to say as a zone of cultural exchange and interaction, not as a political border.  The domination of the historical narrative means that we are accustomed to see the North Sea as a zone for travel principally in an east-west direction, first by Saxon raiders and then by migrants to Britain.  We have also become used to a separation between the study of the interaction of the two sides of the North Sea (in this limited fashion) and the study of Northern Gaul.  To understand this period, we need to look at all the areas bordering on the North Sea, not just those which the historical sources lead us to think were involved in migration, whether as destination or as points of departure.  This will change our perspective importantly.
One of the most important things that we must do is break down the idea that we must see the Roman and barbarian worlds as in a binary opposition.  The links between the two were dense.  Hostile military confrontation was only one element, and probably not the most common.  The lands east of the Rhine and north of the Danube where saturated with Roman influences that took all sorts of forms and went far beyond the simple importing of Roman goods, enormously important though this was.  We must remember that the exchange of goods and cultural of influences between the Empire and barbaricum in the fourth century went overwhelmingly from the Empire to barbaricum. The network of relations determined the nature of politics, depending upon the links that rulers could make with the Empire.  Roman cultural ideas also dominated.  Alamannic chieftains may have produced imitations of the brooches used as badges of office in the Roman army; as mentioned Saxons used the elements of their uniform as key elements of funerary display.  Barbarians were long accustomed to modelling their own jewellery on Roman exemplars.  And so on.  Large numbers of barbarians served in the fourth-century Roman army.  Many went home, as we have seen, when their period of service was up.  Clearly traders plied the routes not only along the North Sea coast as far as Scandinavia, but also along the amber routes along the Elbe. 
By the fourth century, then, Germanic-speaking barbaricum was very much socially, economically and culturally, the hinterland of the Roman Empire.  It depended upon links with the Empire for all sorts of social and political stability.  Similarly a good case can be made that the late Roman Empire itself depended upon the existence of the barbarian world for its political and social stability, even if in very different – political and ideological – ways.  These were not, therefore opposed, separate worlds, but intricately interwoven areas of the same world.  It is also very important to underline that in the fourth century individuals were travelling north from the Empire to barbaricum to at least as great an extent as they were moving from barbaricum to the Empire, even if they were only going home.  Ideas went with them of course and cultural influences going north were by far the most important.  We only need to look at burials to see this.  Note too that contacts and cultural influences seem to have spread from Roman Britain to the Saxon homelands, and indeed it might be that an official settlement of Saxons took place in Britain in the late fourth century, which would increase the ties between the two regions, but they were also dense between the north of Gaul and the Saxon homelands.  The areas around the North Sea, on both sides of the frontier, form a single interlinked world, even if politically and culturally divided. 
This is of huge importance if we want to use Migration Theory really to understand what went on in the fifth and sixth centuries.  This theory has stressed the importance of the flow of information in both directions, to and from the land of destination.  This continued to be the case after the Roman Empire’s final crisis.
The way that North Sea barbaricum, northern Gaul and Roman Britain were interlinked is made especially clear by the fact that all three suffered profound crisis at the same time, around 400.  The crisis in Roman Britain might require some nuance in terms of the speed of collapse but it cannot be denied that a dramatic crisis took place leading within a couple of generations to the collapse of town life, of the hitherto prosperous villa-system, a monetary economy, organised industry, long distance trade and so on.  At the same time, northern Gaul suffered a sharp decline in its towns, one or two of which were abandoned as in Britain, the final collapse of a villa-system and dramatic economic decline.  This synchronicity is acknowledged all too rarely, given the lack of contact between British and French archaeologies and this means that explanations rarely take account of the similarity of the changes in the two areas – although the explanations must differ to some extent.  What is even more remarkable, though, is the way that these transformations within the Empire are contemporary with very similar structural changes along the North Sea coast of Germany, in the Saxon homelands.  There too, we see the desertion of settlements and changes in the location of cemeteries, as well as the appearance of the new furnished burial ritual and the concomitantly increased use of the funerary ceremony as a means of cementing or restoring a family’s local social position.  It is against this background that we also have to see the appearance of very similar-looking settlements in all three areas, with post-built halls and Grubenhäuser. 
Such a way of thinking by-passes old questions linking new types of building to single, concrete ethnic types and their arrival.  Seen thus, what we have are new developments that respond to similar circumstances through the interplay and interchange of cultural influences.  This allows us also to incorporate the appearance of similar building forms in different areas, not linked by the same process of migration, and the early appearance of Grubenhäuser in northern Gaul.  It allows migration and it permits migration to have a role in the spread of ideas but it does not restrict the explanation either to migration alone or to indigenous development.  Migration theory is important here in that it has stressed that movement during migration is rarely one-way.  Migrants return home, either permanently or, as scouts, to bring new migrants and show them the way.  Thus even during the fifth-century migrations we should not envisage movement across and around the North Sea as one-way.  Indeed the written sources themselves describe movements in both directions.
Migration, then, must be retained as an element of the description and as a part of the explanation for specific cultural forms, but the important point is to view the North Sea world as a interlinked whole, which moves the main problem of explanation elsewhere.  Migration should be seen as a symptom of change but not its cause.  What produced the migration?  What produced the crisis that we can see taking similar forms in various areas around the north-western fringes of the Roman Empire?  Here we must return to close linkage of the regions on both sides of the frontier.  They shared, in admittedly different ways, the fact that social stability relied upon a particular style of Roman imperial government, based on the frontier.  In northern Gaul the state had harnessed the rural landscape to the supply of the military and bureaucratic personnel resident in the region.  In a different way, much of the prosperity and stability of Britain was related to its supply of grain and other raw materials to the Roman army.  The imperial court’s location on the frontier served other purposes in distributing and redistributing patronage and maintaining the involvement of the north-western aristocracy in imperial governmental structures.  That in turn was probably vital to maintaining the local pre-eminence of aristocrats in the area.  Simultaneously, this government closely watched the frontier and maintained a balance of power between different barbarian groups.  Gifts were important, particularly to the leaders of barbarian groups away from the frontier and those around the North Sea coast seem to have relied heavily on the control of trade and thus access to prestigious Roman imports.
Thus, when the imperial government was removed to Italy from 381, definitively from 395, and took its eye not only off the management of patronage in the north-west but also off the maintenance of frontier policy, it is not difficult to see why crisis ensued, resulting equally in the Great Invasion of 406 and the British usurpations of the same year.  The archaeological changes we can detect in all of the regions under consideration can easily be understood as manifestations of this crisis.  In the convoluted politics that followed, and the social and political crises on both sides of the North Sea, a matrix of ‘push’ and ‘pull’ factors, such as Migration Theory has discussed, can be envisaged, creating new relationships across the North Sea and movement of people, particularly from the Saxon homelands into Britain.  It is worth remembering that the Saxons only appear in the third century, like many other groups, and that earlier groupings like the Frisians, Angles and Jutes disappear, only to reappear in our period.  It looks as though the Saxon confederacy fractured and fragmented, and some groups who lost out moved into the former Empire, in a manifestation of a dynamic that is attested as far back as Caesar’s day and earlier.  The importance of taking a wide perspective – and not looking at migration between regions in isolation – is that this process was probably also linked to the appearance and short-lived political dominance of the Thuringians on the Elbe valley.
In Britain especially and in Northern Gaul local social competition created a context within which barbarian warrior-leaders could create opportunities for themselves, very much within frameworks whose outlines had existed since the fourth century.  Thus the crisis around the North Sea produced push factors from barbaricum and pull factors within the provinces, but all ultimately stemmed not from the previous antagonistic separation of these worlds but their intimate symbiotic connections.

7. Conclusion.

The framework I have sketched out is subtle enough, and flexible enough, to accommodate a range of interpretative factors, geographical and chronological dynamism and variability.  These incorporate the data as they stand without having to resort to explaining away data that do not fit, and they do not force people into one or another of two mutually exclusive, opposing camps.  This framework also allows the archaeological evidence to speak for itself without adopting a confrontational attitude towards the study of the written sources and what they have to tell us.  Finally, it should promote more sensitive understanding of the challenges of modern migration and its processes, rather than erroneously positing ‘us’ Europeans as the civilised Romans opposed to ‘them’, the threatening foreign outsiders, whom the French Right at least are already accustomed to calling ‘les Nouveaux Barbares’.  Thus, although this might have seemed to be a paper that set out to annoy everyone, it is in fact an attempt to propose a framework within which a new more intellectually subtle and sophisticated consensus might be possible.