guy halsall-libre (1)

8/10/2019 Guy Halsall-libre (1)

1/13

AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134 -3184

Ar queolo ga yTe rr ito rio Medieval 18, 20 11. pp. 15-2 7 I.S.S. N.: 113 4 -3 184

SUMMARY

This article responds to recent work by Michel Kazan-ski and Patrick Prin, defending the ability of archaeo-logy to recognise ethnic identity in the burial recordof the early Middle Ages. After summarising the mainoutlines of their argument, it takes the components of

their hypothesis in turn and subjects them to analy-sis. This analysis is based around the archaeologicalevidence and what it can and cannot say without theintrusion of preconceptions drawn from a (usually old-fashioned) reading of historical sources. After finding

the argument wanting even on its own terms, thearticle concludes by looking at the nature of ethnicityitself and whether it is likely to leave such obvious andstraightforward traces in the archaeological record.

Key words: Ethnicity, Burial, Archaeology, EarlyMiddle Ages, Western Europe

RESUMEN

Este artculo pretende responder al reciente trabajode Michel Kazanski y Patrick Prin, que defiende lacapacidad de la Arqueologa para reconocer la iden-

tidad tnica en los registros funerarios altomedievales.Tras resumir las lneas principales de su argumentacin,somete a anlisis cada postulado de sus hiptesis.Dicho anlisis se basa en la evidencia arqueolgica yen lo que sta puede o no apor tar, sin introducir pre-concepciones extraidas de una lectura (generalmenteanticuada) de las fuentes histricas. Tras encontrar elargumento deficiente, incluso en sus propios trminos,el artculo concluye planteando la naturaleza de laetnicidad en s misma, y si es verosmil que deje tanobvias y directas huellas en el registro arqueolgico.

Palabras clave: Etnicidad, Enterramiento, Arqueolo-ga, poca altomedieval, Europa Occidental

1 The time taken to research, write and present this paper was funded by the award of a Major Research fellowship to the author byThe Leverhulme Foundation. Travel to Vitoria was funded by the Universidad del Pas Vasco/ Euskal Herriko Uniber tsitatea. I shouldalso like to thank Juan Antonio Quiros Lopez, for his kind invitation to give the lecture upon which this paper is based and for hishospitality in Vitoria in November 2010.

* Department of History, University of York, United Kingdom

This paper responds to two recent articlesby Michel Kazanski and Patrick Prin (KAZANSKI/PRIN 2008; 2009), which make an extended andsustained attempt to make a case in favour ofarchaeology's ability to recognise and identifyethnic identity, particularly in cemeteries. In

Britain at least, this would not be a fashionableposition to take, since the publication of SinJones' monograph on the topic (JONES 1997),even if one can argue that in Anglo-Saxonarchaeology its implications have not yet beenfully internalised. In the archaeology of mainland

Europe, however, it is remains a much morerespectable stance and indeed seems currentlyto be suppor ted by one might call a "counter-revisionist" scholarly offensive. Patrick Prin'sknowledge of the archaeological data fromFrance, particularly the burial record, pertaining

to the Merovingian era, is second to none;indeed one wonders whether it will ever bematched. Similar ly, Michel Kazanski has an unri-valled empirical knowledge of metalwork and ofthe archaeology of the East Germanic-speakingregions of late antiquity. For all these reasons,

Ethnicity and early medieval

cemeteries1

Etnicidad y cementerios altomedievales

Guy Halsall *

DOSSIER

Archaeologyand

ethnicity.

ReassessingtheV

isigothicnecropoleis

JuanAntonioQuirsCastillo(ed.)


2/13

16 AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134-3184

Guy Halsall

these publications deserve to be taken seriouslybut the ideas they express must be subjected toclose scrutiny. This is a mark of the respect withwhich this work deserves to be considered.

The argument in the longer and moredetailed piece (KAZANSKI/PRIN 2008) can besummarised as follows:

Kazanski and Prin take their methodologi-cal starting point from work published by H.-J.Eggers (1950), which claimed to derive its streng-th from the avoidance of the Mischargumen-tation (mixed argumentation) which avowedlycharacterised earlier work. Instead, it allegedlytreated the different bodies of evidence his-

torical, archaeological, linguistic or onomastic,etc. - separately and on their own terms. Theseconclusions are then compared to produce anoverall theory. This looks exactly the same asthe "multidisciplinary" methodology used in myfirst book, on the Merovingian Region of Metz(HALSALL 1995). It aims at the same advantagesand at avoiding the same pitfalls. Ironica-lly, however, I adopted that methodology toavoid pitfalls in work carried out using Eggers'methodology! There is a link between the

claims made in these articles and those whichPrin is accustomed to make (not untypicallyin French academic practice), of Cartsianisme:that is to say a radical scepticism, making noprior demands on the evidence; everythingmust be demonstrated through reason. Theseassertions of methodological rigour and purity("purification regressive") must be subjected toclose examination.

Mischargumentation, an alleged mix ofarchaeology folklore, linguistics and historythrown together in an ad hoc fashion, was whatpost-war archaeologists like Eggers claimed,not incorrectly, had lain behind the Germanist,nationalist works of Gustav Kossinna. Kossinna'swork, of course, had been popular with theNazis and had underpinned some of Hitler'sclaims to territory, in France and in the SovietUnion (FEHR 2002). German archaeologists wan-ted to distance themselves from this. Similarly,Kazanski and Prin argue that Kossinna's ideashad seen archaeological cultures as simple

reflections of ethnic groups equated with

peoples or nations - in too monolithic a way.Their work, they state, is based on quite diffe-rent premises. They also claim that it would beunrealistic to expect homogenous or monocul-tural archaeological manifestations of the barba-

rians planted on Roman soil and make sensiblestatements about the fluidity of ethnic identityand the mixed and changing composition ofsupposedly ethnic groups. This takes accountnot only of the work of Reinhard Wenskus butalso of his successors such as Herwig Wolframand perhaps even of Walter Pohl (WENSKUS1961; WOLFRAM 1988; 1997; POHL 1998). So far, onemight say, so good. However, whether, or towhat extent, these fine sentiments are reflectedin Prin and Kazanski's actual conclusions needs

to be scrutinised.

The two authors argue that the accul-turation of barbar ians on Roman soil wasineluctable and demonstrate this through theexample of the Visigoths. In the course of a30-year wandering across Europe, by the timethe Visigoths arrived in Gaul in 412, wherethey were tasked with the repression of theBagaudae and formed a kingdom (KAZANSKI/PRIN 2008:188) they had lost their material

culture. This disacculturation led to a rapidacculturation in Aquitaine and explains whythe Visigoths left no archaeological traces there.When forced into Spain they developed, bycontrast, a national material culture, and in thiswere helped by their contact with the Ostrogo-thic army of Widimer. This is an argument thatPrin has made before (PRIN 1993).

Kazanski and Prin then discuss a series ofcriteria that are relevant to the definition ofethnicity:

1: Funerary practices: They claim (KAZANSKI/PRIN 2008:191) that burial practices arestrictly linked to religious belief in tradi-tional societies and thus deeply rootedwithin ethnic groups. They are also linkedto social factors. All that said, Kazanski andPrin nevertheless conclude that it wouldbe impossible to distinguish, archaeologically,a Barbarian who was perfectly integrated inRoman society or a Roman living in barbari-

cumand buried according to local practice.


3/13

17AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134 -3184

Ethnicity and early medieval cemeteries

2: Ethnic costume: This is a key pillar of Kazan-ski and Prin's argument but it is developedmostly with regard to female costume, aswe shall see shortly.

3: Ethnic weapons: Males were trained inthe use of weapons from early boyhoodonwards, and so, say Kazanski and Prin(2008:195-6), particular weapons can beidentified as ethnic markers. The examplethey use is that of the francisca, which is(they say) is found throughout the Frankishprotectorate.

4: Traditional Female Costume: As mentioned,this is the key suppor t of the argument

(KAZANSKI/PRIN 2008:196-9). According toPrin and Kazanski, in traditional societiesthese costumes are sacral ised and regu-lated. Vague reference is made to thework of ethnographers in support of thispoint, but it is nevertheless claimed to bealmost a universal rule, proved over andover by anthropologists (KAZANSKI/PRIN2008:196; 2009:150). Against the back-drop ofa claimed ethnic costume for east and westGermanic women, they then discuss a series

of tombs, where the brooches are of thewrong type, or where the right broochesare worn in the wrong place as examplesof acculturation.

Grave 756 at Vicq, for example, wore a Visi-gothic buckle and a pair of bow brooches atthe shoulder but also wore, at the chest twosmall local zoomorphic brooches (KAZANSKI/PRIN 2008:196-7 & 195, fig.22).

On the other hand, grave 140 at Nouvion-

en-Ponthieu wore two Visigothic broochesbut at the waist rather than at the shoulders,where they should have been (KAZANSKI/PRIN 2008:197-8).

Explanations relating to the dead belongingto the second generation of immigrants areadduced (KAZANSKI/PRIN 2008:198).

5: Hand-made pottery. This is claimed not tobe an object of commerce, but made bywomen in the settlement and therefore a sign

of ethnic identity (KAZANSKI/PRIN 2008:198-9).

6: Germanic Animal Art. Kazanski and Princlaim that this has specifically pagan andtherefore barbarian significance (KAZANSKI/PRIN 2008: 199-201).

On the basis of these points, Kazanski andPrin move on to four case studies:

1. The presence of West Germanic barbariansin northern Gaul in the late 4th and early5th centuries (KAZANSKI/PRIN 2008:201-207).

2. Eastern Barbarians in Gaul at the same time(KAZANSKI/PRIN 2008:207).

3. Common Prestige fashions among barba-

rian warriors in the second half of the fifthcentury (KAZANSKI/PRIN 2008:207-209).

4. Germanic minorities in Gaul in the late fifthand early sixth centuries (KAZANSKI/PRIN2008:209-12).

So much for Kazanski and Prin's argument,which takes originates in work by Kazanski(KAZANSKI 1997). As stated, it is based upon avery thorough knowledge on the one hand of

Merovingian archaeology in France and, on theother, of the material culture, particularly themetalwork, of the cultures from around theBlack Sea and Danube areas and their presencein Gaul. As far as empirical awareness of datais concerned I cannot hope to equal theseauthors. However, the logical and methodolo-gical coherence of the arguments they present,in support of the idea that ethnicity can bedetected through the archaeology of Gaul atthe time of the Migrations, can be examinedmore closely.

Serious engagement with these ideas is amark of respect for the work (and knowledge)of their authors. This makes it all the moredisappointing that Prin and Kazanski do notdeal with the growing literature on early medie-val cemeteries that rejects their model and itsantecedents. This work is simply ignored. Inthis heading I could include Sebastian Brather'smonumental Ethnische Interpretation in derfrhgeschichtliche Archologie (BRATHER 2004a),

or Bonnie Effros writings on the supposedly


4/13

18 AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134-3184

Guy Halsall

conservative dress of barbarian women (EFFROS2004) or Philipp von Rummels discussions of theshortcomings of ideas about barbarian dress(VON RUMMEL 2007; see also von RUMMEL 2010), orother Freiburg School studies casting doubt on

the geographical origins of key classes of object(GAUSS 2009), or my own or Frans Theuws stu-dies of the late Roman and Merovingian ceme-teries of Gaul (HALSALL 1992; 2000a; 2010:131-67;THEUWS 2009; THEUWS/ALKEMADE 2000), and soon. It is possible to read between the linesand to see these pieces implicitly as partof a growing counter-offensive by traditionalistarchaeologists against new readings of the exca-vated data, prompted mainly by the publicationof Brathers book (BIERBRAUER 2004; BROGIOLO/

CHAVARRA ARNAU 2010; VALENTI 2009), but onewould have preferred a closer engagement withthe specific arguments proposed in the newerworks, rather than a simple restatement ofthe old views, and the application of an unjustdamnatio memoriaeon revisionist work.

In British archaeology, the approach takenwould usually be to address the nature ofethnicity and whether, theoretically, such formsof identity would or could be identifiable in

the archaeological record. Adopting that lineof argument would, however, leave us in aposition (for reasons we shall encounter atthe end) with little to say about Kazanski andPrin's works other than simply to restate animportant methodological difference betweenBritish and much of mainland European earlymedieval archaeology. Instead, therefore, it hasbeen felt more profitable to examine, in depthand on their own terms, the arguments of thisrecent and detailed defence of the traditionalviewpoint. Other problems with the projectof detecting ethnicity in the cemetery evidencewill then be discussed.

To what extent does Kazanski and Prin'sargument live up to the claims of methodolo-gical purity made for it? One obvious pointmust be made at the very outset, and cannotbe made too forcefully: an object does nothave an ethnicity. This is perhaps an insultinglyobvious point, but how many times do we readin archaeological literature (not just in the work

of the two authors under discussion) about a

Visigothic belt buckle, or a Lombard brooch?At the 2010 International Medieval Congressin Leeds Philipp von Rummel, was asked whatwe should call belt buckles or brooches if notGothic or Vandal or whatever. He replied by

drawing attention to the fact that no one hasany difficulty in talking about pottery withoutusing ethnic terminology. An amphora is Spa-nish or Eastern Mediterranean, a fine warebowl is African Red Slip or a drive sigillepalochrtienne, or whatever. There is noreason why we cannot use such general termsfor brooches as well, or (perhaps better) justdescribe them in terms of their principal fea-tures (as Anglo-Saxon archaeologists do, withtheir "Great Square-Headed" Brooches &c.).

One really must wonder how much clearerthe archaeology of the Vlkerwanderungszeitwould become if all these superfluous ethnicterminologies were abandoned.

Indeed assigning any ethnic name toarchaeological evidence is quite impossible onarchaeological grounds alone. Nospecific ethnicidentity of anysort can eversimply emerge fromthe archaeological record on its own, whetherthat record be studied through artefact design,

or from distribution maps or charts of percen-tage frequencies (as, e.g. in SIEGMUND 1998; 2000).Such an interpretation can only ever arisethrough the application to the archaeologicaldata of a series of assumptions derived fromwritten sources. In other words, the ethnicinterpretation of material cultural data cannever, everresult from looking at archaeologyalone, and taking it on its own terms. Put ano-ther way, noethnic interpretation of archaeo-logy can everclaim to be Cartesian, or to beusing pure archaeological reasoning. To assignany of these names to an object immediatelycontaminates the archaeological evidence withthe influence of an historical narrative. Indeed,a number of Kazanski and Prin's examples aremade entirely in accordance with one particularview of the period's history.

Let us take, for example, the case of theVisigoths in Aquitaine and Spain. The first thingthat needs to be said is that the whole problemis driven indeed the problem is created by

the historical narrative. Without a historical


5/13

19AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134 -3184


record that told us that people called Gothscame to Aquitaine in the second decade ofthe fifth centur y, there would be no problem,there would be no absent or invisible Gothsto explain. The second problem, and it is one

to which I will return, is that it assumes thatthe historical record, in the form of historically-attested ethnic identities, would be simply,passively reflected in the archaeological record,and that the variability in that record will auto-matically reflect the variability or oppositionssuggested in the written data.

The next issue with this case study concernsWidimer and his army. For the last twenty yearsPatrick Prin has used this Ostrogothic army

as the explanation upon which to hang theappearance of an apparently Gothic materialculture in Spain (PRIN 1993). Unfortunately,there is little or no evidence for this army.Widimer is not attested in any contemporarysources: only in Jordanes Geticafrom the middleof the sixth century. A Billimer mentionedby Paul the Deacon has been suggested tobe the same man but neither source is verytrustwor thy on these matters. A Wittimerappears in two letters of Ruricius of Limoges

who might or might not be the same man, butthey say nothing about his arrival from Italy oranything that would confirm Jordanes story(HALSALL 2007:278-9; 2010:70). Even in the late,unreliable stories we have, there is insufficientevidence to say what became of this army. Itis nowhere said that it ever went to Spain. Alltold, this example is about as far away as onecould possibly get from being an example ofa Cartesian approach to the archaeologicalevidence, without being influenced by datafrom other sources.

A second instance can be found in the cita-tion of the franciscaas a diagnostically Frankishweapon. Isidore famously, and incorrectly, saidthat the Franks drew their name from the fran-cisca(Isidore, Etymologiae 18.6.6)but the associationbetween the Franks and this weapon actuallygoes back no further than the middle of thefifth century, and Sidonius Apollinaris. Accountsof the fourth-century Franks make no mentionof the weapon and the archaeological record

of franciscasequally does not go back earlier

than the fifth century. Moreover the franciscais found overwhelmingly in Gaul and veryrarely in barbaricum. There is quite abundantevidence that the axe was in use within theRoman army (HALSALL 2010:134-5). Looked at

in purely archaeological terms one would arguethat the franciscawas a weapon that appearedin Gaul in the fifth century and was occasiona-lly, usually later, found beyond the Rhine. Thearchaeological record, when set alongside amore critical reading of the documentary andepigraphical sources, suggests that the franciscawas a weapon used by the very late Romanarmy in Gaul and that the Franks adopted itfrom their service in those armies. Whetheror not one accepts that, the interpretation of

the francisca as diagnostically Frankish couldnot emerge from the archaeological recordon its own.

Indeed, in many cases the archaeologicalevidence is not being taken on its own termsat all. The example of Germanic barbariansin Late Roman northern Gaul, claimed to bean assured case of archaeology showing thepresence of an intrusive ethnic group in Gaul(KAZANSKI/PRIN 2008:201), is a good illustration.

Almost none of the standard interpretation ofso-called federate graves in northern Gaul inthe fourth century would emerge from a purelyarchaeological reading (HALSALL 1992; 2000a;2010:131-67). Were this evidence generated ina prehistoric context, as I wrote nearly twentyyears ago (HALSALL 1992:201), no one wouldever find in it any evidence of a migration.Almost all of the material culture found in theburials is of Roman Gallic origin; the rite itselfis basically the standard rite of Roman Gaul,but with more grave-goods; the rite is actuallyquite different from that used in the barbarianterritories; etc. The traditional argument findsits strongest support in a series of brooch typesburied with some of the women in these gravesbut when one consults the distribution maps ofthese objects one finds that it is almost exactlythe same as that for other items of metalworksuch as belt buckles and other belt appliqus,or of Roman pottery and metal vessels, which,as no one is in any doubt about, were produ-ced in Gaul and exported beyond the Rhine

to Germania. This alone begs the question of


6/13

20 AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134-3184

Guy Halsall

why one interpretation is followed for sometypes of artefact and not others with the samedistribution and, frequently, similar elements ofdecoration. Close inspection of the artefactstoo suggests that they were being manufactu-

red in Gaul, exported to Germania, and copiedthere, as had long been the case with Romanjewellery. It is clear to me that only the intru-sion of a pre-determined historical narrativehas led to this evidence being read as evidenceof barbarian immigration into Gaul. There aremany, many other illogical arguments and self-contradictions in the traditional argument whichI have discussed at length elsewhere (HALSALL2010:131-67).

Similar cases from other areas can be addu-ced. For example, recent work has suggestedthat some of the brooches used as evidencefor the presence of eastern Germaniin the westin the fifth century are not imports from theeast at all (GAUSS 2009). When one looks atdistribution maps one can indeed join the dotsto produce a migration from the Danube toGaul or Spain (KAZANSKI/MASTIKOVA/PRIN 2008),but one need only do that if one has decidedin advance (on the basis of non-archaeological

sources) that that has to be the direction ofmovement. Why not fromSpain tothe Danu-be? Or, more plausibly and as has apparentlybeen argued recently, from a Mediterraneanproduction centre and then distributed in twodirections, east and west, to Gaul or Spain andto the Danube and further east?

Sometimes a historical narrative is adduced,without worrying about the fact that actuallyit is not attested in any actual written sources!Like Widimers army, such is the case, with theargument that certain brooch types foundin northern Gaul, which have some generalsimilarities with others found on the Danube,represent the presence of East Germanic sol-diers in the region. These brooches are items offemale apparel, so it is argued that these womenare the wives of the (archaeologically invisible)soldiers. I have already drawn attention to theproblem with assuming an east-west movementbehind the distribution map. No matter thatno written source mentions the presence of

East Germanic soldiers (let alone their wives)

in northern Gaul. A story is composed onthe basis of the political history of the period,which has East Germanic federates arriving innorthern Gaul to fight in the armies of theRoman king Syagrius. Note too that the very

nature of Syagrius even the reality of his exis-tence as a king of the Romans, in opposition toinvading Franks cannot be securely proven onthe basis of the written evidence! Therefore,the written sources are not being subjected toclose scrutiny, either. At every turn, wheneverone looks into the details of the approachand the arguments deployed, we could not befurther from a methodology which relied uponthe strict, rigorous, pure analysis of separatebodies of evidence on their own terms before

the comparison of conclusions at a higher level.In actual fact, this is Mischargumentation at itsmostmixed! Indeed, mixtae confusaeque, to usea phrase of Gregory of Tours.

In fact, in an appendix to the 2008 article,Kazanski criticises R. Hachmann, one of thepioneers of Eggers' methodology, praised at thestart (KAZANSKI/PRIN 2008:185-6), for not makingScandinavian archaeology fit the story providedby Jordanes' Getica(KAZANSKI/PRIN 2008:212-3).

Never mind that just about every scholarlyanalysis of the Geticahas rejected its accountof the Goths Scandinavian origins (HEATHER1991; GOFFART 1988:20-111; 1995) This does notseem to demonstrate a very deep commitmentto the principles of regressive purification!

Similar incoherence emerges when thearguments other premises are examined. Letus start with the idea of burial ritual as a markerof ethnic identity. Of course, in theory, the wayone disposes of the dead, bound up as it iswith ideas of cosmology and religion, might beexpected to be an area where traditions werestrictly guarded. It is therefore frequently saidthat burial is a very conservative element ofsocial practice. And yet, in terms of its archaeo-logical traces, it is anything but conservative.Changes in burial practice come thick and fastin antiquity. At least tenchanges in methods ofplacing the dead took place in lowland Britainduring the first seven and a half centuries of theChristian era (HALSALL 2000b:261). Between the

time when unaccompanied inhumation, wra-


7/13

21AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134 -3184


pped in a shroud or in simple costume, becamenormal in Europe sometime in the latter halfof the first millennium (ZADORA RIO 2003)andthe revival of cremation in the early twentiethcentury, burial does look very conservative

across most of Mediterranean and westernEurope, in terms of its archaeological remains,but one need only consult other records, aboutmourning, funerals, commemoration or eventhe above-ground markers or gravestones tosee that burial in fact continued to be a dynamicarea of social expression.

The alleged evidence of Germanic migra-tion into northern Gaul in the four th centur yagain stands as a useful lesson. Here, the cus-

tom employed in these supposedly intrusiveburials is completely different from that usedin the alleged incomers homeland. This isexplained as evidence of acculturation but, forthis to carry any weight, archaeologically, onewould need to see communities crematingtheir dead without accompanying goods andthen gradually adopting the host populationsrites. Instead, were we to assume that thesearethe graves of incomers, something for whichI see no good evidence, what we would have

here would be communities abandoning theirancestral funerary customs (those supposedlyclosely guarded, conservative markers of ethnicidentity) immediately that they were over thefrontier. Another of Kazanski and Prin's argu-ments in favour of acculturation is that burialswith belt buckles alone are Roman whereasthose with the belt buckles andmore grave-goods (weapons for mean and suites of jewe-llery for women) are those of the immigrants,because the latter have included more of theirtraditional customs (KAZANSKI/PRIN 2008:191).Yet actually noneof this custom of burial withgrave-goods is traditionally Germanic at all. Thebarbarians between the Rhine and the Balticor the North Sea Coast cremated their dead,sometimes including dress-adjuncts or otherobjects, often not, sometimes in a cremationurn, sometimes with no container at all indeedoften without container orgoods, making themarchaeologically all but invisible.

This brings us on to the key support for

Kazanski and Prin's position, which is the alle-

ged conservatism of female dress. It is oftenargued, against the empirical evidence for thenorthern Gallic production of the material inthese burials, that the way it is used revealsthat these burials are of immigrant Germani

(SCHMAUDER 2003: 279-80, n.31). Allegedly, thebrooches are used in the traditional fashion ofWest Germanic Trachtor costume. Two bro-oches are used at the shoulders, either to fastena Peplos dress, or probably more plausibly, topin a shawl over a dress. The problems with thisargument are many. Most impor tantly, as I havejust mentioned, the burial record of the areaswhence these alleged immigrants are supposedto have come is overwhelmingly formed bycremation (the Frankish homeland famously

being more or less blank on distribution maps).This means that we have very little evidenceabout how brooches were worn by the womenof the Germani. Indeed most of it is furnishedby the burials under discussion (e.g. BHME1974:161), making the argument more logicallyproblematic! A second problem is that, for allthe supposed immutability and conservatism ofsacralised female costume, the archaeologicalrecord reveals great variability in the numbersand positioning of brooches, the presence and

absence of other artefacts, and so on. It isoften forgotten that Roman women also worejewellery. Although the brooch had droppedout of use, temporarily at least, by the middleof the fourth century, it had been common andindeed sometimes used in exactly the sameways, up until the third century (FEHR 2008:89-97). One must ask why fashion only explainsthe Romans' discarding of the brooch, but nottheir readoption of it; why immigration onlyexplains the brooch's reappearance and notits disappearance; and above all, why Romanfemale costume, in being subject to fashionlike this, was less sacralised and conservativethan Germanic women's dress. In fact, though,when looking at late Roman Gallic burials theimplicit assumption is that Roman Trachtwasmore immutable than Germanic because thearchaeologically-revealed diversity of femalegraves supposedly shows variability and accultu-ration by Germanic women, whereas they can-not be Roman women because (it is implied)Roman women were not allowed to adopt new

items or otherwise change their dress!


8/13

22 AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134-3184

Guy Halsall

The third problem for the Kazanski-Prinhypothesis concerns the very nature of thissupposedly North-West Germanic costumewith its pairs of brooches at the shoulders. Itis actually a pair of problems. The first is that

it runs completely against the argument thatwearing two brooches at the shoulders is thetradit ional East, not West, Germanic femalecostume, so that burials with brooches at theshoulders can be argued to be of fifth-centuryEast Germanic immigrants (KAZANSKI/PRIN2008:201). What the slightly earlier burials fromnorthern Gaul show is that such a mode ofemploying brooches was already known in thatregion. Indeed the other half of the problemis that a rigorous examination of the data

suggests that not only was it not exclusive toeastern Germani, it might indeed have been alate Roman provincial Gallic fashion.

This idea of fashion only a descriptionrather than an explanation, to be sure never-theless produces yet more reasons to questionour authors methodology. When one looks atthe archaeological record one sees only hugevariety, through time and place, such as doesnot accord with the assertion of deep con-

servatism in dress. Indeed the notion is morethan slightly undermined by the invocation ofacculturation. In Gaul the general lesson wouldindeed seem to be of the general acculturationof the Franks into the structures of fifth-centurynorthern Gallic society at the same time astheir political, ethnic identity began to be widelyadopted. This nevertheless casts some doubtupon the degree to which female costume iseither as conservative and regulated as is beingproposed, or as directly, intimately linked to anethnic identity.

When I have discussed these problemswith Patrick Prin 2he has explained that hesees the first generation of immigrants keepingclosely to their traditional dress but subsequentgenerations adopting more and more items ofthe fashions of the host population. This is aninteresting idea, but to explore it would require

close scrutiny of all aspects of the burial, notjust the grave-goods and their date. It wouldrequire us to examine the deceaseds age, forone would expect, were Prin's model correct,that old women would retain their traditional

costume into the second generation of burials.One might expect children in the first genera-tion to be buried in traditional dress by theirparents, but what of adolescent and young adultwomen who died during the first generation,who would otherwise have become the moreacculturated women of the second generation?How does this transition play out, and how isit reflected in the archaeological record? Thehypothesis requires sophisticated thinking andmodelling, not simply mapping onto different

chronological phases. It should also be saidthat where the earliest intrusive graves on a sitedo not fit the traditional model, the argumentis sometimes deployed that these women hadpartly acculturated before they arrived at theplace where they died (KAZANSKI/PRIN 2008:198;2009:157)! Of course, if one does not think thatthey are immigrants in the first place (I havealready suggested that the empirical groundsfor this assumption are weak) this is not a veryconvincing argument, even if it is convenient.

These points lead on to two further issues.One is that female costume, as revealed in thecemeteries, not just of Merovingian northernGaul but in some parts of the Roman frontierprovinces, in Anglo-Saxon England, southernGermany, and northern Italy, in fact varies sig-nificantly according to the age of the deceased(e.g. BARBIERA 2005; BRATHER 2004b; 2008; CLARK

2000). My study of the Frankish cemeteries ofLorraine reveals that children did not usuallyreceive items related to gendered costume(HALSALL 1995:254; 1996). Most jewellery (theessential elements of traditional, ethnic costu-me) is found with teenagers and young adults,and women older than their twenties areincreasingly rarely interred with these ar tefacts.This alone must make a purely ethnic readingof the costume much too simplistic. It is notto deny that there might be ethnic significance

2 I should underline that M. Prin has always been most friendly, supportive and willing to discuss these issues. I want to make it clearthat, although we hold diametrically opposed interpretations, that opposition is founded in no personal animosity.


9/13

23AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134 -3184


in the nature of adolescent female costume. If,however, the nature of dress and its adornmentchanged through the female life-cycle, as I havejust suggested, then this seriously questionsthe model of a progressive abandonment

of traditional ethnic costume through time,generation by generation, according to idea ofacculturation.

This indeed raises one of the most seriousproblems of all with the traditional point ofview: why the variability observable in thearchaeological record need have anything at allto do with ethnicity. Kazanski and Prin ack-nowledge this point (KAZANSKI/PRIN 2008:191)but they do not allow it to obstruct their

argument. All sorts of other dimensions of anindividuals identity can come into play in theconstruction of the burial record. I have alreadymentioned, and discussed in detail elsewhere(HALSALL 2010:289-412), the role of gender andage. Kazanski and Prin mention religion andsocial factors as being involved in the establis-hment of a burial rite and its archaeologicallyobservable (KAZANSKI/PRIN 2008:191) featuresbut they do not pause to consider how thesedifferent dimensions might work together as

ultimately and primarily ethnic, rather than(as I would see it) cross-cutting each otherand making the ethnic interpretation moreincoherent.

This in turn leads me to the problem thatunderlies all such traditional ethnic readings,and that is that they ignore the processesbehind the creation of the archaeologicalrecord itself, seeing it simply (as mentioned)as a passive reflection of reality. Whateverelse one might say about British archaeolo-gical theory in its current state, in its post-processual phase in the 1980s it did bring tothe foreground the idea that the formationof archaeological evidence is a deliberate andmeaningful activity, founded upon active choi-ces, designed to create information as well asconveying it to an audience. One must always,therefore, ask whypeople chose to bury theirdead in this way. The problem with the ethnicreading is that for many of the ethnic groupsknown to us it is clearly the case that they didnotgenerally bury their dead in the particular

style that archaeologists have pinned on them.There are no better examples than the Gothsthemselves. The followers of Theodoric, howe-ver, minimalist a view one might want to take,must have numbered very many times more

than the fifty or so archaeologically knownOstrogothic graves in Italy and the Balkans(BIERBRAUER 1994). The Goths of Aquitaine andSpain patently did not bury all of their deadin a particular, Gothic style. Indeed they did notbury their dead in that way when they werein the Balkans, and even before 376, to judgefrom the ernjachov/Sintana-de-Mure culturethey did not have a single burial rite in any case,but a mix: cremations and inhumations of allsorts, found within the same cemeteries (for

useful survey, see HEATHER/MATTHEWS 1991:59-69).So, even on the best-case scenario (where oneactually accepts the ethnic import of the rite),these ethnic groups only buried someof theirdead a small minority in a particular way. Soone must ask, again, why? There must clearly have been some reason, other than simpleethnicity, that led some people to distinguishsome of their dead from the great majority.In other words, even where ethnicity mightbe an acceptable descriptionof the meaning of

objects, it is rarely a satisfactory explanationfortheir deposition.

Here lies, in my view the solution to theproblem or non-problem, as I would prefer-of the archaeological invisibility of the Gothsin Aquitaine. The explanation cannot simplylie in the Goths' lack of any Gothic meta-lwork. As the dominant force in the regionsurely they could simply have forced Romancraftsmen to make some. Or they could asthe Kazanski-Pr in hypothesis suggests withregard to other situations- have worn localproducts in accordance with the Gothic Tracht.The simple absence of the right metalworkcannot explain the abandonment of a rite.At this point it must, however, be said that, asthe Gothic inhumation rite was actually onlycreated in later generations (in Spain), Prin'stheory about the Aquitanian Goths reversestime in arguing about the non-appearanceof something that had in fact not yet beendeveloped! For Kazanski and Prin a Goth isalwaysa Goth and will (or should) alwaysdo


10/13

24 AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134-3184

Guy Halsall

what she or he is attested as doing at somepoint in Gothic history (regardless of when orwhere). This is only one instance where, inspite of claiming to believe the opposite, theydo in fact treat ethnic identities and cultures

as unchanging and monolithic.

Where an ethnic or political identity is displa-yed in burial with grave-goods, the crucial thing isthat it is displayed to an audience for a particularreason (HALSALL 2010:203-60). Therefore onemust ask whythe Goths would necessarily haveburied their dead in a costume that proclaimedtheir Gothicness in Aquitaine in the fifth century.Migration is not something that automaticallyshows up in the excavated record; indeed it

is very often archaeologically invisible. I haveargued repeatedly that furnished inhumation(with grave-goods) is essentially a sign of socialcompetition of some sort (HALSALL 2010:203-60). In earlier sixth-century northern Gaulishcemeteries, in a very fluid society with few or norigid class distinctions and few means of securinglocal pre-eminence beyond royal service, wholecommunities seem to have participated in thecompetitive grave-goods ritual, as they did inlowland Britain (HALSALL 2010:278-84). In other

areas, such as Ostrogothic Italy, lavishly furnishedgrave seem to be concentrated in the urban fociof the realm and may demonstrate a claimedGothicness to an audience of other membersof the aristocracy competing for royal favour(HALSALL 2007:336-8). There is not a blanketexplanation for all burial rituals with grave-goods.One must look at what sorts of individual isbeing buried, in what numbers within what sortsof cemeteries and with what types and quanti-ties of object (HALSALL 2008). But the display ofgrave-goods is transient by its very nature andtherefore requires both the bringing together ofan audience to see it and the existence of a sym-bolic language rendering the message intendedby the ritually-deposited objects comprehensi-ble to that audience. All this points, inexorably,towards political competition of some sor t.

Indeed, the so-called Gothic cemeteriesof Spain lie generally along the fringes of thekingdom: not just in the northern Meseta, as iswell-known, but also in the south, around the

Byzantine enclaves, and on the Frankish border

in Septimania. Furnished burials are also wellattested on the Basque frontier where thepresence of that political border must surelybe part of the explanation. In other words, onthe fringes of political authority, where claims

to local power might be contested betweenindividuals or groups asserting the backing ofdifferent political forces, and in situations likethose of the sixth century, where political andmilitary power were often based on ethnicideas, we might expect the meaning of objectsplaced with the dead to have some ethnicimport - but in a very different way from thatenvisaged in traditional readings.

This brings me to my concluding points. I

have taken the Kazanski-Prin argument on itsown internal merits, to show that the thesis isnot very satisfactory even by its own lights, usingthe sorts of empirical archaeological methodo-logies that it claims to espouse. From there,my argument has led us, bit by bit, to overalltheoretical problems, which cast serious doubton the whole project.

The first is that the relationship betweenmaterial culture and ethnic identity is very

problematic. A classic anthropological studyfrom many years ago showed that one couldquite easily compile a list of features, of lan-guage, of dress or hairstyles, or other featuresthat people said distinguished their group fromothers, or which distinguished other groupsfrom theirs. And yet, in practice one foundthat these features were either rarely if everobserved in use or, where they were, onefound that they did not distinguish one groupfrom another (MOERMAN 1969; POHL 1998 makessimilar points about the early Middle Ages). Anotherstudy, from East Africa, showed that age-gradeswithin one particular society adopted materialculture associated with a neighbouring groupto distinguish themselves from the age groupsbelow and above them (LARRICK 1986). It isdifficult to begin to imagine the havoc that thiswould play with any attempt to read ethnicidentity from the distribution map of artefacts!Yet we can see similar things within our lateantique evidence, where Roman soldiers andaristocrats adopted items of costume which

are held to be barbarian what I have termed


11/13

25AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134 -3184


barbarian chic (HALSALL 2007:110) or wherepeople within barbaricum, women as well asmen, used imported Roman material cultureto show their high status (HALSALL 2007:57-58).Indeed, around 400 some people in the north

of Germania adopted the Roman inhumationrite in order to distinguish themselves from theirfellows (BEMMANN 1999; KLEEMANN 1999). Thisdid not make them Romans by birth, althoughfor all we know some of these individuals, if theyhad served the Roman Empire, might well havestyled themselves Romani. In Gothic Italy or onthe margins of the Gothic kingdom of Spain,individuals might very well have styled themsel-ves Gothiwithout being descended from peoplewho had crossed the Danube in the 370s

or 380s. It was a claim to power and status.

That, ultimately, is the point. Ethnicity is astate of mind, with no necessary correlationto things which are objectively measurable ,whether material, biological or genetic. This willalways make attempts to read off monolithicethnic identities, or even the interplay betweenmonolithic ethnic identities (which is what isat stake in acculturation arguments), highlydubious. More pertinently, perhaps, ethnicity

is itself a complex dimension of an individualsidentity, existing in several layers which canbe adopted or highlighted, abandoned, playeddown or concealed. Early medieval people didnot have to see themselves as eitherRomans orFranks, as eitherGoths orSueves. An inhabitantof sixth-century Spain, who took up arms andattended the army using an assertion of Gothicidentity as a means to acceptance within thismilitary-political group, was not thereby pre-cluded from having Roman, provincial or civitasidentities as well, which he might have used atother times in other circumstances. None ofthese groups was monolithic in itself: Romansself-identified by their civitas, a very importantand much neglected level of post-imperialidentity: there were different groups within theFranks, there were political regional groupings,by kingdom or by Roman province, which havemost of the features of ethnicity (Neustrian,Austrasian, Aquitanian or Provenal).

In a sense we have come full circle, because

it may be that, as with the military associations

of barbarian ethnic identities, we can proposethat some objects in graves weapons mighthave conveyed that identity to an audience.This archaeological reading would fall foul ofmost of the strictures set out at the beginning

of this paper, being a reading of material cultureentirely in the light of documentary sources.However, this reading of the documentarysources and its application to material cultureis somewhat more subtle (and indeed moregrounded in the written data). It might bethe case that cer tain types of brooch, used inparticular ways with particular types of people,in particular contexts, didhave an ethnic con-notation, so that a Jutish brooch in Englandmight have implied that the wearer claimed a

Jutish identity.

An important caveat for this point, though,is that it only remains a suggestion, which canonly be made in a particular context. It cannotbe taken as a general rule, such as that peoplewith weapons are always Franks, whereverthey are found and in whatever context: thatis plainly untrue. It also implies nothing biolo-gical, genetic or exclusive about the claim beingmade. Indeed this suggestion has the fluidity of

our modern understandings of ethnicity. Thus,although having the appearance of come roundin a circle, we end with a very different unders-tanding of the relationship between materialculture, and ethnicity from that with which westarted. The argument moves forward, as in aspiral and in so doing I think that it opens upour cemetery evidence to much more inter-esting and less constricting readings.

BIBLIOGRAPHY

ISIDORE, Etymologiae: Isidori Hispalensis Episcopis Etymo-logiarum sive Originum libri XX, ed. Lindsay, W.M., (2 vols.;Oxford, 1911).

BARBIERA, Irene, (2005). Changing Lands in ChangingMemories: Migration and Identity dur ing the Lombard Inva-sions (Florence).

BEMMANN, Jan, (1999). Krpergrber der jungerenrmischen Kaiserzeit und Vlkerwanderungszeit ausSchleswig Holstein. Zum Aufkommen einer neuen Bes-

tattungss itte im berregionalen Vergleich. Studien zurSachsenforschung13:5-45.


12/13


13/13

27AyTM 18, 2011pp. 15-27 I.S.S.N.: 1134 -3184


LARRICK, R., (1986). Age grading and ethnicity in thestyle of Loikop (Samburu) spears. World Archaeology18:269-83.

MOERMAN, M. (1968). Being Lue: Use and abuses ofethnic identification in HELM, J., (ed.), Essays on the Problemof Tribe Proceedings of the 1967 Annual Spring Meeting of

the American Ethnological Society, , (Seattle), pp.153-69.

PRIN, Patrick, (1993). L'arme de Vidimer et la questiondes dpts funraires chez les Wisigoths en Gaule et enEspagne (Ve-VIe sicles) in Franoise VALLET & MichelKAZANSKI (ed.), L'Arme romaine et les barbares du IIIeau VIIe sicle(Paris), pp.411-23.

POHL, Walter, (1998). Telling the Difference: Signs ofethnic identity, in Walter POHL & Helmut REIMITZ(ed.), Strategies of Distinction. The Construction of EthnicCommunities, 300-800(Leiden), pp.17-69.

SCHMAUDER, Michael, (2003). The relationship betweenFrankish gens and regnum: a proposal based on the

archaeological evidence, in H.-W. Goetz, J. Jarnut & W.Pohl (ed.), Regna and Gentes: The Relationship between lateAntique and Early Medieval Peoples and Kingdoms in theTransformation of the Roman World(Leiden), pp.271-306.

SIEGMUND, Frank, (1998). Alemannen und Franken:Archologisch berlegungen zu ethnischen Strukturen inder zweiten Hlfte des 5. Jahrhunderts, in Geuenich, D.(ed.), Die Franken und die Alemannen bis zur "Schlacht bei

Zlpich"(Berlin), pp.558-80.

SIEGMUND, Frank, 2000.Alemannen und Franken. (Berlin).

THEUWS, Frans, (2009). Grave-goods, ethnicity and therhetoric of burial rites in late antique northern Gaul in

Ton DERKS & Nico ROYMANS (ed.), Ethnic Constructsin Antiquity: The Role of Power and Tradition (Amsterdam),pp.283-319.

THEUWS, Frans; ALKEMADE, Monica, (2000). A kind ofmirror for men: sword depositions in Late Antique nor-

thern Gaul in Frans THEUWS & Janet L. NELSON (ed.),Rituals of Power. From Late Antiquity to the Early Middle

Ages(Leiden), pp.401-76

VALENTI, Marco, (2009). I villaggi altomedievali in Italia, in

Juan Antonio QUIROS CASTILLO (ed.),The Archaeology ofVillages in Europe (Documentos de Arqueologia e Historia1: Bilbao), pp.29-55.

VON RUMMEL, Philipp, (2007). Habitus Barbarus: Klei-dung und Reprsentation sptantiker Eliten im 4. und 5.

Jahrhunder t (Ergnnzungsbnde zum Reallexikon serGermanischen Altertumskunde 55; Berlin)

VON RUMMEL, Philipp, (2010). Gotisch, barbarisch oderrmisch? Methodologische berlieferungen zur ethnischenInterpretation von Kleidung, in Walter POHL & MatthiasMEHOFFER, M., (ed.), Archaeology of Identity/Archologieder Identitt (Forschungen zur Geschichte des Mittelalters17)(Vienna), pp.51-77.

WENSKUS, Reinhard, (1961). Stammesbildung und Verfas-sung. Das Werden der frhmittelalterlichen Gentes(Colog-ne)

WOLFRAM, Herwig, (1988). History of the Goths (trans.of revision of 2nd German edition; Berkeley).

WOLFRAM, Herwig, (1997). The Roman Empire and itsGermanic Peoples(Berkeley).

ZADORA RIO, Elisabeth, (2003) The making of church-yards and parish territories in the early-medieval lands-cape of France and England in the 7th-12th centuries: Areconsideration. Medieval Archaeology47:1-19

guy halsall-libre (1)

Documents