The Archaeology of Celtic Art

D.W. Harding

• Publisher: Routledge
• Year: 2007
• eBook ISBN: 9781134264636
• Edition: 1
• Topic: Art
• Subtopic: Archaeology

More wide ranging, both geographically and chronologically, than any previous study, this well-illustrated book offers a new definition of Celtic art.

Tempering the much-adopted art-historical approach, D.W. Harding argues for a broader definition of Celtic art and views it within a much wider archaeological context. He re-asserts ancient Celtic identity after a decade of deconstruction in English-language archaeology.

Harding argues that there were communities in Iron Age Europe that were identified historically as Celts, regarded themselves as Celtic, or who spoke Celtic languages, and that the art of these communities may reasonably be regarded as Celtic art.

This study will be indispensable for those people wanting to take a fresh and innovative perspective on Celtic Art.

1
DEFINITIONS, MATERIAL AND CONTEXT

Few topics in archaeology have spawned as many perceptions and misconceptions as Celtic art. An Internet search for ‘Celtic art’ immediately offers patterns of ‘Celtic’ interlace and knot-work, elements of later Celtic art in fact derived from Mediterranean or Germanic origins, or images of high crosses of ninth-century date or later and related icons of the early ‘Celtic’ church. For coffee-table books a dust-jacket depicting the Gundestrup cauldron is considered representative, notwithstanding the fact that it was almost certainly of Thracian manufacture, and discovered in northern Jutland, well beyond the limits of Celtic Europe. In academic publications, Celtic art is generally synonymous with the La Tène ornamental style of the pre-Roman Iron Age, but even this equation should not pass unqualified. Since in recent years the concept of Celts and Celtic as an ethnic descriptor has itself been questioned, it seems appropriate now to re-define and re-assess what we mean by Celtic art.

History of research

The identification of distinctive styles of ornament on Iron Age metal-work as ‘Celtic’ has its origins in the mid-nineteenth century, and is particularly associated with John Kemble, whose work was published posthumously in 1863, together with contributions from R. G. Latham and A. W. Franks, under the title Horae Ferales. Franks used the term ‘Late Keltic’ to describe objects such as the Battersea shield and horse-gear from the Polden Hills hoard, in contrast to earlier material of the Bronze Age that was also regarded as Celtic on the basis of contemporary studies of human craniology. ‘Late Keltic’ was still being used at the end of the century, notably by Arthur Evans (1890) in his report on excavations at Aylesford in Kent, but fell out of use thereafter with the decline in fashion of craniological and ethnic correlations. Notwithstanding his later Aegean interests, Arthur Evans was a pioneer in the study of Celtic art, his Rhind Lectures in Edinburgh of 1895 anticipating Paul Jacobsthal by nearly half a century in recognizing the classical influences on the early La Tène style. On the Continent, cultural and even chronological identifications of some of the classic assemblages were still more tentative, with early discoveries of chieftains’ burials in the Rhineland being assigned to the Roman period, while related finds were alternatively attributed to Teutonic times. One of the pioneers in the field of La Tène studies, Ludwig Lindenschmit, had classified finds from the site of La Tène on Lake Neuchâtel, but had not identified them as native Celtic, and likewise believed that objects like the Durkheim torc were Etruscan imports. In 1871, de Mortillet recognized metal-work at Marzabotto near Bologna as similar to material from the Marne, and inferred that here was the archaeological evidence for trans-alpine Gauls of the documentary sources. As late as 1889, Adolf Furtwängler published the Schwarzenbach bowl as the product of a workshop in the vicinity of Massilia on the analogy of east Greeks on the Black Sea producing high-status metal-work for Scythians. In effect, the equation between La Tène metal-work and Iron Age Celts was not fully established until Joseph Déchelette’s Manuel d’Archéologie préhistorique, celtique et gallo-romaine was published immediately prior to the First World War.

After the publication in 1944 of Paul Jacobsthal’s Early Celtic Art, the equation between La Tène art and Celtic art was effectively taken as read, and it is only in the past decade that this been seriously challenged. Unfortunately that challenge has been in the context of a wider ‘deconstruction’ of the ‘myth’ of the Celts, promoted more vigorously among English archaeologists than among their ‘Celtic’ neighbours in Britain and Ireland, and equally not so widely canvassed in Continental Europe. Judicious re-appraisal is unlikely to proceed while the polarized rhetoric of the Celticity debate still rages. It is self-evident that Celtic art studies in modern times owe a fundamental debt to the magisterial work of Paul Jacobsthal. In acknowledging this debt, however, we should recognize that Early Celtic Art adopted the perspective of a classical archaeologist, whose interest in the art of the European Iron Age had been triggered by the Celtic embellishment of the Klein Aspergle kylix (Pl. 3), noted while the author was studying Greek vases in Stuttgart in 1921 (1944, vi). Accordingly, the strength of Jacobsthal’s perception was his appreciation of the various stylistic influences from classical art that impacted especially upon the earlier phases of north-alpine La Tène art. Because of the vicissitudes of late 1930s Europe and the war years, Early Celtic Art was produced under extremely difficult circumstances. But it has to be acknowledged that it often reads more like a scholar’s notebook than a research synthesis, and its catalogue could hardly be described in contemporary terms as user-friendly.

Like many scholarly landmarks, however, Early Celtic Art has both stimulated further study and impeded it by imposing a framework and terms of reference that now need to be challenged. Jacobsthal concluded that ‘Celtic art is an art of ornament, masks and beasts, without the image of Man’ (ibid., 161). In effect, his Early Style, Waldalgesheim Style, Sword Style and Plastic Style are not art styles, but ornamental styles, as Jope evidently recognized when he referred to the La Tène ornamental style in Britain (1961a). Social anthropologists would not define art so narrowly, and would certainly include a range of artefacts whose role was not solely utilitarian, whether explicitly ornamented or not (Layton, 1991). It was the restricted interpretation of Celtic art as synonymous with La Tène ornamental styles that presumably caused Jacobsthal to dismiss Celtic art in Spain (1944, v). The impact of the La Tène ornamental styles in the Hispanic peninsula was, as we shall see, minimal. But if Celtic art is alternatively defined in terms of the range of weaponry and defensive armour, personal ornaments and accessories to ceremonial or ritual activities, for example, all of which from documentary sources appear to be fundamental to Celtic society, then the evidence from South-Western Europe seems as mainstream to the study of Celtic art as is the La Tène art of Central and West-Central Europe. We should surely pay homage to Jacobsthal’s achievement; but after more than sixty years it is time that the theoretical framework of Celtic art studies was reviewed, and that some of the fundamental assumptions of study were challenged.

Most studies of Celtic art since Jacobsthal have been concerned primarily with discerning a sequence of ‘styles’ and their inter-relationships. In effect, though the principal contributors – Jacobsthal, Martyn Jope, Paul-Marie Duval, Otto-Herman Frey, Miklós Szabó and others – were archaeologists, their approach to Celtic art has been substantially from an art-historical viewpoint. This approach is important, and should not be deprecated simply because it is now less fashionable than socio-economic or cognitive reconstruction. Vincent Megaw recognized the need to set the study of Celtic art in the context of Celtic society, and has contributed significant papers to which the present study is indebted. Accordingly, this treatment of the subject will attempt to evaluate Celtic art not just in terms of stylistic developments over time and space, but in the context of Iron Age society, as far as it can be reconstructed from the evidence of archaeology. What was the role, symbolic, ritual or social, of ornamented metal-work and sculpture? What does it tell us of the technological skills and status of jewellers or armourers in Celtic society? Were there ‘workshops’ and ‘schools’ headed by master craftsmen, and, if so, did they operate under princely patronage or in a commercial market environment? What was the nature of the long-distance connections that are manifest in stylistic influences? Do these reflect population movements, movement of craftsmen, trade or diplomatic exchange? And how does this high-status expression of Celtic art compare with decorative arts in more mundane media, like pottery, wood or textiles? How might the role of art objects that survive archaeologically have functioned in the context of non-tangible art forms such as oral poetry, song and dance? And, finally, are there significant discernible changes over time or between different regions of Europe in the role of art in society?

Celtic ethnicity, Celtic languages and ‘Celtic’ art

The first questions that should be addressed in a book that incorporates the phrase ‘Celtic art’ in its title are whether the term ‘Celtic’ is justified, and in what sense is it being applied? Chapman (1992) cast doubt on the belief that Celts in Iron Age Europe existed as an ethnic group at all. Collis (2003) was more qualified in his critique, noting that Caesar’s identification of the inhabitants of his third part of Gaul, who were known as ‘Celtae in their own language, but “Galli” in ours’ (de Bello Gallico, 1, 1), might endorse the concept of a Celtic ethnic identity. He also cited among others the case of the Romanized poet Martial, who in the first century ad claimed to be half-Celtic and half-Iberian. However inadequate or confusing the sources may be, the ancient writers evidently thought of Celts as an ethnic identification. The real problem therefore is the correlation of ancient ethnic Celts with Celtic languages, on the one hand, and with any coherent set of archaeological material, on the other.

The earliest usage of the term ‘Keltoi’ by ancient writers is by Hecateaus and Herodotus in the late sixth and fifth centuries bc, in reference to one of the recognized groups of barbarian neighbours of the Greeks. Herodotus’ grasp of European geography and his understanding of ethnography may have been tenuous, but it is important that the recognition of Celts as an ethnic identity, however ill-defined or imprecisely located in Central and Western Europe, pre-dates the appearance of the La Tène culture in the mid-fifth century bc. Since it is likely that the emergence of the Celts considerably pre-dated their first impact upon Greek historians or geographers, there is a case for believing that Celts in Continental Europe existed from at least the later Bronze Age.

Later classical sources are by no means consistent in their references to Celts for various reasons. The problem is compounded by the various usages in Greek and Roman sources of the terms Keltoi, Galatai, Celti and Galli. The fact that tribal groups are identified by Caesar among the Celtae or Galli, the Belgae and Aquitani, for example, suggests that there may have been a hierarchy of levels within which the native communities identified themselves, and that ‘Celt’ was therefore almost certainly a supra-tribal and perhaps supra-regional descriptor. In this case it seems possible that Caesar’s fundamental division of Gaul into three parts mistakenly equates entities at different levels. ‘Belgium’ plainly included a dozen or more tribal groupings, as did ‘Aquitania’, so these would appear to be ‘middle-order’ entities. Their contrasting by implication with ‘Celtic’ Gaul might suggest that neither Aquitania nor Belgium were Celtic, but if Caesar was unaware of a ‘middle-order’ designation for the rest of Gaul, he might have resorted to the ‘supra-regional’ name as shorthand for ‘the rest’.

The absence of references to Celts in the very partial documentary record, either for Britain or for other regions of Continental Europe, particularly east of the Rhine, is no guarantee that the inhabitants of those regions were not part of the wider Celtic community. Collis’ (2003) preference for regarding France west of the Rhine as the probable Celtic heartland in part derives from the fact that, through Caesar, this is where Celts are most clearly located, and in part from the fact that the documentary sources point most clearly to these regions as the homeland of Celtic migrants of the early fourth century into Italy. Yet this is not to say that regions east of the Rhine were not also Celtic from an early date, even though the surviving documentary sources are more equivocal. Caesar’s distinction between Gauls and Germans along the Rhine (as opposed to Teutonic Germans of Northern Europe), quite evidently was a red herring introduced by him for political reasons. Strabo (Geography, IV, 4, 2; VII, 1, 2) was in no doubt that Gauls and Germans were related by kinship, and explained that the Romans called the Germans ‘Germani’ (L. germanus = true, genuine, as a natural brother) to emphasize that they were blood brothers of the Gauls. Accepting the historical migrations of Gauls into South-Eastern Europe as originating west of the Rhine, then plainly the situation in north-alpine Central Europe may have been affected by this phase of expansion, but there must be a strong possibility that people of Celtic ethnicity and speech occupied Europe east of the Rhine from a much earlier date. For Strabo, at any rate, Celtica at the supra-level extended north of the Alps to the mouth of the Rhine and to the Pyrenees and the Ocean in the west.

The equation of ethnic Celts of antiquity with Celtic languages has aroused equal controversy. It is true that the group of Indo-European languages now known as ‘Celtic’ have only been so designated since George Buchanan’s pioneer work of the sixteenth century, being more widely adopted from the early eighteenth century. It is equally self-evident that much of Victorian and modern ‘Celtomania’ has no sound scholarly foundations in ancient history or archaeology. Yet however the language group is designated, it is clear from linguistic, epigraphic, numismatic and place-name evidence that by the early Roman Empire it covered a wide region of Central and Western Europe, including the Hispanic peninsula, northern Italy, Britain and Ireland. In the absence of evidence for wholesale population incursions of the late pre-Roman period to account for such linguistic super-strata, it seems reasonable to regard this as the language group of the various communities whose archaeological material culture has been systematically identified by archaeologists as early Iron Age or even later Bronze Age (Harding and Gillies, 2005). Across the territory covered by the proxy-map of Celtic languages, there is plainly no uniformity of material culture, though there may be common elements. It is certainly not co-terminous with the La Tène distribution, nor with that of Hallstatt or the Urnfield series before that, though these Central European cultures certainly fall within the putative Celtic zone, and the Urnfield distribution is perhaps closest of any to a pan-European phenomenon. But in Atlantic Europe in particular there are sizeable regions, such as northern and western Britain, southern Ireland, western France and the Hispanic peninsula, where Urnfield, Hallstatt and La Tène material culture made minimal impact. While we might share Collis’ view (2003, 195) therefore that ‘there is likely to have been some feeling of common identity across Europe, at the level of a shared language’, there can be little doubt that regional patterns of material culture must indicate some quite striking differences, notably between Central Europe, on the one hand, and the Atlantic seaboard, on the other.

How far Celtic languages can be projected backwards in time is much more contentious. Renfrew (1987) saw the emergence of Celtic languages in Europe as an indigenous development from a much earlier introduction of Indo-European with the first farmers, but this view scarcely takes account of the complexity of the evidence, and has not gained widespread support among linguists. In any event, we should not expect patterns, linguistic, archaeological or ethnic, to have remained immutable over centuries. Nor should we necessarily therefore expect close correlations between archaeological and linguistic distributions, or linguistic and ethnic distributions, any more than we would now expect the level of correlation between archaeological cultures and ethnic groups that was firmly envisaged by Gustaf Kossina or Gordon Childe in the earlier twentieth century.

If we are prepared to accept ‘Celtic’ as a language group, variants of which were widely spoken by Iron Age communities in Central and Western Europe, and even that the concept of Celtic ethnicity, however ill-defined in the classical sources, implies a measure of commonality of identity between neighbouring groups, what should the term ‘Celtic’ mean in the context of Celtic art? Most studies of Celtic art since Jacobsthal have been principally focused on La Tène art, a term that conventionally includes insular British and Irish metal-work, even though diagnostic or typical types of Continental La Tène are really relatively poorly represented here. Yet in contrast to the general pattern of Iron Age material culture, in which Britain is decidedly peripheral to Central Europe in the number and range of definitive types, in the field of ornamented metal-work from the third century bc onwards at least the British inventory is as spectacular as anywhere in Celtic Europe. In Ireland not only are key La Tène types such as safety-pin brooches represented by barely three dozen known examples, compared to a thousand or more in the Duchcov hoard from the Czech Republic alone, but even those few are of distinctive insular types, quite without parallel in Continental Europe. Other so-called La Tène types, such as Y-pendants and spear-butts, are likewise not at all characteristic of Continental La Tène, to the extent that one might question how La Tène the Irish assemblage actually is. Yet the La Tène in Ireland would normally be regarded as an important sub-group within the overall family of Celtic art.

Apart from distinctly regional sub-groups like the Irish La Tène, there are other areas of Atlantic Europe that might well qualify as ‘Celtic’ on the basis of linguistic or allied evidence, but where the impact of La Tène material culture is minimal or non-existent. Ireland south of a line from Dublin to the Galway Bay presents a particular problem that will be discussed in due course. But south-western France and the Hispanic peninsula beyond the Iberian zone are regions where La Tène or La Tène-related types are relatively few, and where ornament of material artefacts is not nearly as prolific or distinctive as in the La Tène tradition. In Spain, apart from the area of documented Celtiberians, there are regions to the south-west and north-west where place-names and allied evidence suggest the presence of Celtic speakers. These regions too, therefore, will need to be considered if we are to justify the title of ‘Celtic’ art beyond simple convenience and convention. If, then, there are regional populations that were Celtic-speaking but not characterized archaeologically by a La Tène culture, we should question reciprocally whether all bearers of La Tène culture were necessarily Celtic-speaking. Self-evidently the exclusive equation of Celtic identity with La Tène material culture is mistaken, but could the exclusive equation of La Tène with Celtic also be erroneous?

Finally, in this section, we should consider the chronological limits of Celtic art. Most Continental studies conclude with the Roman Empire, which effectively brought an end to the La Tène art style. In Britain, on the fringes and beyond the Roman frontiers, and in Ireland, by contrast, the ‘long Iron Age’ extends well into the first millennium ad. Though any elements of continuity from earlier Celtic art into the ‘Pictish’ period in Northern Britain, or into Early Christian art in Ireland, need to be carefully scrutinized, nevertheless these communities were Celtic-speaking, and may legitimately be included in the broader discussion of Celtic art here proposed. Indeed, consideration of the composition, context and potential meaning of these later styles may prompt questions relevant to the study of earlier Celtic art.

Materials and techniques

Attempting to define an art object in the context of later prehistoric societies is likely to be contentious. A flint axe, a bronze pin or a pottery vessel may be technically accomplished and aesthetically pleasing to handle, but we would not necessarily regard them as art objects. Yet many bronze pins or brooches that presumably served a utilitarian function as dress accessories may conform to a form and style not dictated by function alone that consciously or unconsciously identified the individual or community that made them, or satisfied the social or ritual conventions that governed their use. More elaborate objects may be ornamented in a manner that permits the identification of recurrent motifs and images, the arrangement of which according to conventions that might be compared to the rules of grammar constitutes a particular ‘style’ in the sense used by Jacobsthal. ‘Art’ is plainly not synonymous with ‘ornamentation’, but may be implicit in the object itself. It is probably an anachronistic coincidence that the Neuvy figurines (Figure 10.4A) should appeal to modern aesthetic taste, but they and other artefacts, like the boar images from across Iron Age Europe, must surely have had significance as art or cult objects to contemporary communities, or sects within those communities. Associations, as in graves or hoards, should be informative, and while we may legitimately be concerned with detailed analyses of individual items, it is important not to overlook their associations in order to evaluate significance or ‘meaning’ in context.

The media of Celtic art were various, though metal-work predominates in the literature because it was the medium of high-status artefacts of the greatest technical competence. Bronze and iron are frequently found in combination on objects such as scabbards or parade armour that ...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• List of Colour Plates
• List of Figures
• Preface
• 1. Definitions, Material and Context
• 2. ‘An Art With No Genesis’: Later Bronze Age and Hallstatt Origins
• 3. The La Tène Early Styles: Origins and Influences
• 4. The La Tène Developed Styles
• 5. The Art of the Swordsmith
• 6. The La Tène Later Relief Styles
• 7. Insular British Art to the Roman Conquest
• 8. La Tène and Non-La Tène In Ireland
• 9. South-West Europe and the Celtiberians
• 10. Later Styles and Romanizing Influences
• 11. Later Insular Art In Britain and Ireland
• 12. Conclusions: Archaeology and Celtic Art
• Bibliography