The University of Sydney

The Celtic Question

Archaeology Honours Thesis

Harriet Donnelly
309228204
Men cling passionately to old traditions and display intense
reluctance to modify customary modes of behaviour, as
innovators at all times have found to their cost.

V. Gordon Childe (1951)

Page | 2
Abstract

The identity of the “Celts” has played an integral role in the understanding of the
Iron Age and the more recent socio-political history of Europe. However, the
terminology and attitudes which have been in place since the 19 th century have
created a field of research characterised by assumptions about a ‘people’ and a
culture. Study of the “Celts” has been conducted in three main areas - genetics,
linguistics and material culture from the archaeological record. Through the
reassessment of these three fields, substantial divergence in the patterns and trends,
as well as the highly regional nature of the evidence has been revealed within the vast
interconnected trade and communication network that developed in Iron Age Europe.
As a result the unitary phenomenon identified under the term “Celts” is actually that
network. “Celtic” should be redefined as the label for that trade and communication
network, not as a label for a group, culture or people, enabling the establishment of
new identities for the regional populations of the European Iron Age.

Page | 3
Acknowledgements

I would like to acknowledge the support and contributions of my wonderful

network of colleagues and professors, family and friends during the writing of this
thesis. I cannot thank you all enough for your help and support.

First and foremost to my supervisor, Professor Roland Fletcher, thank you for
everything. I wouldn’t have been able to finish this without your guidance, many
words of wisdom, enthusiasm and energy at every turn. Thank you for constantly
reminding me that there is a life beyond honours, you helped me look forward and put
this hurdle into perspective when it was very much needed.

Secondly, I would like to express my thanks to my partner and all my family and
friends, archaeologists and non-archaeologists alike who have listened to my rants and
raves. You have endured coffee catch ups filled with thesis talk and helped me carry
piles of books from libraries across Sydney. I especially want to thank my
grandmother, Luise di Corpo, for your unwavering support through reading and re-
reading the words of this thesis. To those who have taken this path with me, nothing
can be said to show how much gratitude I have towards you all.

Finally, to my parents. Mum and Dad if anyone can claim the dedication of this thesis,
it is you. The fact that I was crazy enough to embark on this quest comes down to
knowing you were there with me. From reading through all the information at the
university open day where I became fixated on archaeology to supporting me through
everything along the way, I couldn’t have done it if it wasn’t for the amazing people
that you are. From daily phone calls, words of love and encouragement, you’ve given
me the courage to travel this path and seek all the rewards that lie at its end.

Thank you.

Page | 4
Declaration

The work in this thesis is the result of research carried out under the Department of
Archaeology at the University of Sydney. No part of this thesis has been submitted in
any form for another degree or diploma at any university or other institution of
tertiary education. All external sources and work by others, whether published or
unpublished, have been acknowledged in text and in the list of references.

October 2013

Page | 5
Context

List of Figures 7
Chapter One - Introduction 8
Chapter Two - Background to the Celtic Question 14
Terminology and Classical Literature 14
Political Associations 16
Theoretical Approaches 19
Conclusion 22
Chapter Three - A Celtic People? Genetic Studies and the “Celts” 24
Approaches to Genetics in the Debate about the “Celtic” Question 24
Paternal DNA 27
Maternal DNA 30
Conclusion 33
Chapter Four - The “Celtic” Languages and Linguistic Development 35
Phonological Linguistic Analysis and Language Movement 37
Place-name Analysis 39
Inscriptions 42
Conclusion 44
Chapter Five - Material Culture: the Archaeological Record 46
Art 47
Fibulae 55
Trade and Interaction 61
Conclusion 67
Chapter Six – Integration and Conclusions 69
Conclusions 74
References 76

Page | 6
List of Figures

Figure 1 Comparative Y-chromosome table 27

Figure 2 Y-chromosome relationships in Europe 29
Figure 3 European Correlation Maps 32
Figure 4 Distribution of -brigā names of ancient Europe 40
Figure 5 Distribution of -dūno- names of ancient Europe 41
Figure 6 Distribution of -duro- names of ancient Europe 41
Figure 7 Distribution of -mag- names of ancient Europe 42
Figure 8 Mindelheim and Gündlingen swords 50
Figure 9 Dragon-pair ornamentation designs and European distribution 52
Figure 10 Distribution map of Kastav type and variant fibulae 56
Figure 11 Distribution of Fibulae 57
Figure 12 Distribution of Kessel Fibulae 58
Figure 13 Distribution of Gallische Bügelfibel 58
Figure 14 Distribution of Nauheim Fibulae 59
Figure 15 Iberian Horseman Fibulae distribution 60
Figure 16 Summary of maritime relationships between Britain and Gaul 63
Figure 17 Spread of Oppida in European Iron Age 65

Page | 7
Chapter One – Introduction

Since Grahame Clark dispelled the traditional Celtic invasion model of the British
Isles in 1966 (Cunliffe, 2008, 55), the field of Iron Age Europe has seen a dramatic
increase in the resources and data available for analysis. However, despite this, there
has been little change in the attitudes and approaches used; the key assumption being
that the Celts constitute a “people”. The ultimate aim of those interested in the Celts
also remains unchanged; to find their origin. This goal has become the focus of
scholars in all fields relating to Celtic studies, from the traditional archaeologists
studying material culture through to the geneticists applying recent DNA analysis to
the question of Celtic origin. There has also been an increasing interest in the Celts
from the public with television shows such as Richard Rudgley’s 2006 The Celts and
Neil Oliver’s 2012 A History of Celtic Britain.

Current research into the Celtic origins can be divided into three main areas; the
genetic analysis of DNA throughout Europe, Celtic languages and linguistic
development, and the study of the material culture itself using stylistic, chemical and
distribution analysis to examine the dates, trade routes and other interactions of the
Iron Age communities referred to as the Celts. This division into the genetics,
linguistics and material culture is a useful one which enables the conclusions and
conflicts within the field to be more easily examined and understood. The Celts are
also divided along the geographic line between the Isles, comprising of England,
Scotland, Wales and Ireland where they are called the Insular Celts and the rest of
mainland Europe, the Continental Celts.

For those unfamiliar with the study of the Celts and their origins there are questions
which naturally come to mind. ‘Who are the Celts?’ and ‘When did they live?’ are two
of the more prominent queries. To answer these questions is complicated. There is no
universally accepted answer and nor is there any clear way of indicating which concept

Page | 8
of the Celts is being referred to in scholarly research. How can the question of Celtic
origins be sought if there is no conclusive definition of Celticity? The simple answer is
it cannot. Without understanding what we are looking for, and what we are dealing
with, there will continue to be confusion, error and inconclusive understanding of the
Celts.

The conventional motivation for finding an origin for the Celts comes from a range
of academic, social and political sources and, as a result, the importance of answering
the question “Who were the Celts?” and dealing with the serious issue of “Were the
Celts even Celtic?” cannot be underestimated. From an academic standpoint increased
understanding of the Celtic identity and the definition of the Celts will help reassess
the way we treat the Celts and the way in which the relationships between pre-historic
European people are understood. In turn this understanding will have repercussions
for our knowledge of trade networks, movement of materials, and the factors in the
cultural development across the European Bronze, Iron and Dark Ages. Socially, there
are many people today, predominantly in the British Isles and Ireland, as well as
migrants from these areas who claim Celtic heritage. For them this is a defining
feature of their identity and currently the connection between modern and ancient
Celtic identity is taken for granted. Greater study into the history and characteristics of
presumed Celtic society will allow the putative link between the identity of the so-
called ancient and modern Celts to be better understood. Finally, there is a significant
political involvement in the Celtic debate. Following the rise of nationalism in the
aftermath of the French Revolution, the Celtic identity was used by some countries to
create a national identity that was distinct from dominant England. Scotland, Wales
and Ireland used the history of the Celts and famous Celtic artefacts to prove they
were not English or Anglo-Saxon by heritage. For Ireland, the Book of Kells and the
Tara brooch were key artefacts used to justify and validate a Gaelic history for Ireland
in the same manner that Christian Thomsen had used archaeological material and his
Three Age system to identify a separate material history for Denmark. Separate
histories assisted Scotland, Ireland and Wales in creating their own nationalistic

Page | 9
identities and encouraged a political movement for each of them to be viewed as
independent, separate nations instead of being under the political rule of England.

A more extensive knowledge of the presumed Iron Age identity of the Celts and
their relationship with other peoples is of particular interest from a political
standpoint, as the understanding of the Celts as a separate entity has been the
foundation of national identities and political structures in Europe today. Collis (2011,
195-204) presents a concise explanation of the significance of the Celtic identity as a
basis for historical, political and nationalist standpoints and highlights how strongly the
study of the Celts has been interwoven into this political agenda. It is important to
note that previous research has made very little distinction between the concept of
Celtic identity from the Iron Age and the identity of modern Celtic communities. In
order to answer the questions surrounding whether or not there is a Celtic identity, it
is necessary to both understand the history of the term Celt and to recognise what has
led to those definitions. It is equally vital to approach all aspects of a presumed Celtic
identity without the pre-conceptions inherent in the commonly used terms and to
evaluate the evidence from the three main areas of genetics, linguistics, and material
culture to identify what arguments are actually demonstrable.

From this point on the use of the word “Celt” will refer only to a presumed
aggregate population of common ancestry and cultural identity extending across
Europe. The Celts can be called, without question, a European Iron Age people who
are not of Classical origin. In this definition, there is recognition of the geographic and
temporal boundaries within which we know the Celts operated. This is not to say that
they did not extend beyond these limitations, only that they are known to have existed
within them, nor does it predefine to whom the label “Celt” should actually apply in
the 1st millennium BC. There is also clear evidence from the works of Classical scholars
that the Celts were not seen as being part of the Classical world. Terms such as
“barbarian” reveal the distinction that the scholars made between themselves and the
Celts. Martial wrote in the 1st century AD discussing his own identity as half Celt and
half Iberian, coming from the town of Bilbilis in central Spain (Collis, 2011, 23). While

Page | 10
this reveals a shift in the concept of owning Celtic identity, it also highlights the way
that identity was viewed as being segmented and distinct. Being a Celtic Iberian did
not make that individual a Celt who lived in Iberia. This was understood as belonging
to two separate identities, much as someone today might consider that they belong to
two nationalities or to a cultural identity regardless of which passport or passports
they hold. Furthermore, the distinction which is often made between the Insular Celts,
being those of England, Ireland, Scotland and Wales, and Continental Celts implies
recognition of two individual groupings of the Celtic people. These labels should be
cautiously used with reference only to the geographic space and not to indicate any
level of collective grouping. Indeed we should be wary of the degree to which we
create and allocate labels to aggregate identities. Despite their collective treatment in
the majority of the studies of origins, the Insular Celts cannot confidently be said to
share a single origin, culture or language. Much as the Celtic identity operates today,
with ‘Celticness’ being implicit in personal identity instead of overtly claimed as a
national or ethnic label, the definitions that have become prevalent in the study of the
ancient Celts are riddled with the same implicit universality of culture, language and
identity (James, 1999, 17) despite numerous problems of partial overlap, non-
correspondence and chronological ambiguity.

Traditionally the evidence has been arranged into the pre-existing construct of the
Classical accounts, an approach which pre-dates archaeology as an established
discipline. While this would have originally seemed to take advantage of the recorded
history from the Classical world and allow the pieces of European Iron Age history to
be quickly linked, the result has been a reluctance to let go of the perceptions and
divisions set up in the Classical accounts, and, more alarmingly, the need to be able to
attach new research to the existing framework to leave very few loose ends or
conflicting lines of research. As will be further examined, this approach fails to
recognise the conclusions that the evidence is producing which distinctly lead in
opposite directions and consequently work from other fields is used to fill gaps in the
research of the discipline concerned.

Page | 11
 In order to identify the trends and arguments that are emerging from the data, the
three fields of genetics (Chapter 3), linguistics (Chapter 4) and archaeology (Chapter 5)
should be dealt with individually before being compared and evaluated against the
others. Analysis of these fields will first be established against the historical
development of Celtic studies and the varying theories. Each academic field will then
be covered in turn. Each is multifaceted in nature with multiple strands of research to
further complicate the analysis, as even within each field there can be conflict. In
chapter 3 genetic research is mainly focused along the female/male divide, with data
coming either down the female line in mitochondrial DNA, or through the male line in
the Y-chromosome DNA. The field of linguistic study, in chapter 4, can be divided into
the analysis of place-names and the tracing of linguistic development and movement,
both focusing on individual languages and on comparative analysis between multiple
languages. Finally, analysis of the material culture of the European Iron Age
populations that are referred to as “Celts” in chapter 5 can be divided a number of
ways. Some studies have focused on an object based distinction, looking at the art
styles, fibulae designs, raw material trade or settlement structure. Alternatively, other
studies have been arranged around the geography of Europe, following the geographic
distinction already discussed between the Insular and Continental Celts. Most studies
use a combination of the two, making the study of the material culture highly
specialised and highly regional. The field is also elevated by strong attitudes and
decided opinions.

Once an understanding of the trends in each field has been established without the
influence of the other fields, similarities and conflict will be more easily identified
between them. Identity, as has already been discussed, can be two-fold: the identity
as recognised by an external party and identity as understood by those to whom it
relates. There are several ways in which we establish the identity of ourselves and
each other. Religion and nationality are frequently the larger criteria, however
language, gender, political and sexual affiliations also play a role. These smaller criteria
are often contributing factors, or are intertwined with the larger criteria, for example
language is often entangled with nationality. Genetics or family histories are also

Page | 12
related, although this can be a much more complicated identity contributor. Renfrew
(1993, 19) outlines the division of identity into the examination of genetic, linguistic
and cultural diversity. This parallels the divisions that already exist in the study of the
“Celts”, enabling the conclusions of each field to be aligned with the corresponding
marker in the conceptual framework of identity as discussed by Renfrew.

Page | 13
Chapter Two - Background to the Celtic Question

Terminology and Classical Literature

The term ‘Celt’ has a long and complex history. As Cunliffe explains in Europe
Between the Oceans, examination of the Classical sources indicates a shift in the
meaning of the word Celt (Cunliffe, 2011, 354). The Greek scholars such as Hecataeus
and Herodotus in the 6th and 5th century BC understood the Celts to be people in
central continental Europe, inland from Massalia and with the source of the Danube in
their territory (Collis, 2011, 16). The term is then used by various scholars to indicate
different people, groups and subgroups until in the 1 st century AD Silius Italicus uses
Celtae as synonymous with Galli (Collis, 2011, 22). While this transition may give the
impression of a changing understanding of the word ‘Celt’, the prioritisation of the
Classical sources has led to a hugely disproportionate significance being placed on
their definitions and terminologies. The history of Celtic studies has been greatly
affected by the terminology used in the Classical texts. Historically, archaeology and
the study of past societies developed as a private hobby of individual collectors. This
reached its height in the 18th and 19th centuries when the Classical Mediterranean
societies were still considered the ideal. As a result the writings of the Classical
scholars were revered and treated as definitive. Any artefacts, ancient settlements or
other aspects of material culture that was discovered were placed into the verbal
categories of the Classical works. While it is recognised today that the use and
transition of terminology in the Classical texts relates to the attitudes of the Greek and
Roman societies to a far greater extent than the internally understood identity of the
Celtic people themselves, the repercussions of the historical attitudes to the Classical
sources has been far reaching and can still be seen today.

Up until the 1960s, the predominant model of Celtic expansion followed the model
of an invasion from the central Celtic zone outwards into Spain, the British Isles and

Page | 14
western Europe (Cunliffe, 2008, 55). Based on the discovery of Celtic art and decorated
materials in the Hallstatt region, the Hallstatt culture became the iconic
representation of the ancestral Celts. The art style which informed this culture spread
from the central parts of France through to modern day Hungary (Cunliffe, 2011, 309).
The traditional view saw this culture, and the successive La Tène culture in central
Europe, expanding outwards throughout the surrounding territory. This movement of
“ideas” and the materials of the culture into the territories introduced the Hallstatt
and La Tène to a wider Europe. While this model is no longer accepted in its simple
form in the academic community, Hallstatt and La Tène are still treated as the central
point for much of the iconic Celtic material culture, and therefore the start of the
Celtic culture.

Studies relating to the Celts use a wide variety of time periods and boundaries as
their frameworks. This variation is partially dependant on the field, as well as being a
product of the views of the researcher and the subject matter of the work. Iron Age
Europe is perhaps the most commonly used label, there is a great deal of variation in
the time that this is meant to cover, and it does not provide the level of context that
enables a thorough and well informed analysis of the relationships between the
people, materials and geography of the Celtic debate. The Iron Age in Europe is
generally considered to have begun between the 8th and 6th century BC and this is
where the majority of the studies begin. However, there is very little indication of
what came earlier and whether the Celtic identity existed before this time. James
discussed the first known Celts as those first referred to by the Greeks as Keltoi around
800-600 BC (1999, 26). Collis (2011, 13) places the original mention of these people in
the 6th century BC as does Champion et al. while focusing more on the art and
established Hallstatt culture (1984, 298). By contrast, Chapman (1993, 1) prefers to
indicate sometime in the first millennium BC instead of assigning a specific century and
Cunliffe (1997, 39) narrows the window to a period between 1300 BC and 400 BC.
When seeking to understand society, culture or people and the interactions they had,
knowledge of the environment in which they existed is crucial. The 7th and 6th
centuries BC, which can be treated as the earliest boundary of Celtic identity, is when

Page | 15
the Celts as a named group appear, regardless of this name being externally imposed
rather than adopted internally. Expanding the range of study to include a brief
understanding of the preceding Bronze Age will allow comparison and insight into the
approaches that have been taken in studying the people of Iron Age Europe. The Iron
Age in Europe is often viewed as continuing through until the 4 th century AD (Laing and
Laing, 1995, 12). This end of the chronology is more complex. Following the expansion
and conquest by Rome into central and western Europe, the Iron Age in Europe was
replaced by Roman occupation in a gradual and partial manner. The fall of Gaul to the
Roman Empire under Emperor Augustus between 15 and 12 BC predates the conquest
of Britain in AD 43 under Emperor Claudius (Wells, 2001, 65). By contrast, due to
limited Roman activity in Ireland, the historic period and the end of the Iron Age do
not occur in Ireland until 400 AD. This ending of the Iron Age in Europe has a large
influence on how Celtic identity has come to be understood and how the interactions
and relationships between various peoples would have operated in different parts of
Europe. Once the Romans arrived, increasing interactions and shifting political and
hierarchical dynamics significantly changed the social and political environment. For
this reason, answers to Celtic identity should be sought in the period up until Roman
invasion when the people of Iron Age Europe were independent of external rule and
when their own decisions and activities can be related to the nature of their identity.
These time parameters will exclude analysis of the Picts in modern day Scotland.
However, a brief discussion of the way in which conclusions about Celtic identity might
impact on understanding the Celts will assist in considering the repercussions for more
recent post-Roman decisions and debate about Celtic identity.

Political Associations

The study of the Celts has been approached using many different fields, theories
and ambitions. The political environment is the first theoretical construct through
which the Celts were viewed. Following the events of the French Revolution, late 18 th
century Europe saw a rapid movement towards established national identities and

Page | 16
borders and it is through this push for nationalism that the modern identity of the
Celts, or those who identify as Celtic today, began to emerge. The Irish nationalist
movement is the strongest example where Celtic identity was used to create a
distinction between the Irish and the British Empire. At the end of the 19 th century, the
existence of two national focus areas can be identified; first the Gaelic revival which
had a predominantly cultural agenda, aiming to increase the use of the Gaelic
language and spreading an increased interest in the Art Nouveau style inspired by the
Book of Kells (Collis, 2011, 198). Originally a passive movement aimed at revitalising
Irish culture under the dominant rule of the British, the historical and political nature
of the movement changed when Owen Jones described the designs of the Book of
Kells as Celtic in his The Grammar of Ornament linking them to the popular Tara
Brooch discovered in 1850 (Jones, 1856). Based entirely on the artistic styling of old
artefacts, a Gaelic history was created for Ireland using the Tara Brooch and Book of
Kells in the same manner that Christian Thomsen used the archaeological study of
styles and materials to create a history for Denmark in the Three Age system (Collis,
2011, 80). The second focus for the 19th century nationalism movement was the use of
Celtic characteristics in an active political role rather than for passive cultural interest.
This focus began to gain momentum after the 1922 Union of Ireland when the Gaelic
language became a symbol of separation from Britain. The work to establish a Celtic
cultural link and separate history for Ireland through the politicisation of the Gaelic
Revival was not universally supported by those in the Irish Nationalist Movement.
Isabella M. Tod, a prominent feminist who became involved in the nationalist
movement, rejected the Gaelic account of history in Ireland as fundamentalist and for
the narrowing and indifference which she believed it would bring to her multicultural
nation (Biagini, 2007). The political use of the Celtic history in Ireland is highly selective
and does not take into consideration the diversity of the archaeological record. The
knowledge that the Celtic art styles and material extended beyond the Irish, Scottish
and Welsh regions was already well documented by the end of the 19 th century,
specifically through the discovery of the Battersea Shield in 1863 in London. The
consequences of these links and the indication they provided of a wider and more
diverse ancient Celtic history were simply ignored. The differences between the
attitudes of the Gaelic revival movement which highlighted a Gaelic history without

Page | 17
excluding the links with other regions, and the excluding nationalistic aim of some
areas of the nationalist movement have directly resulted in the confused modern
Celtic identity which exists today. Cooney (1995) sees the failure of archaeology to
address these volatile areas as a failing of the discipline, which has combined with
“reluctance on both sides of the border to engage in debate about the nature of links
between archaeology and politics”. In Scotland and Wales, the same approach to
Celtic identity can be seen, with a desire to separate from the English being a
fundamental political motivation. The failure to make a distinction between the
modern and ancient Celtic identity is significant, especially as there has been
recognition that there is little to imply that any concept of national identity existed
prior to its construction in the 18th century (Laing, 1979, 1). This acknowledgement
should carry with it a warning about accepting the link between the two Celtic
identities as it underlines the impact that contemporary social constructs have had on
our understanding of identity and social groupings as well as intra- and inter- group
relationships.

While the history of materials and languages linked to the ancient Celts in the
modern Celtic regions cannot be disputed, it is important to remember that the
associations between the material archaeology and the modern Celtic identity have
been established by using the definitions of the term “Celt” that have already been
discussed. Furthermore, the historical associations and use of many well know “Celtic”
artefacts result in a third dimension to their study. It is necessary to understand the
different approaches that can be taken to understanding identity and how this shifted
with the introduction and spread of nationalism. In his work Nationalism in Ireland
(1995), Boyce defines Cultural Identity as being “felt by members of a group who
either have or have had a distinct or relatively autonomous existence, and who have
shared a recognisably common way of life.” By contrast he outlines National Identity
as being “felt by members of a group who define their culture as the national one, and
their group as the true and legitimate inheritors of the national territory, of the
homeland” (Boyce, 1995). Keeping this distinction in mind and using these theoretical
frameworks while reassessing the evidence in the “Celtic” identity debate will create a

Page | 18
level of clarity and play an important role in distinguishing between the modern Celtic
identity and that of the ancient Celts. This clarity will further assist in placing these
trends against the present theories to help answer the questions of Celtic identity,
group distinctions and intergroup relationships. Through closer analysis of these terms
and the identity of the Celts it will be possible to re-examine the place and relationship
that modern Celtic identity has in comparison to that of the ancient Celts and to assess
whether the term “Celtic” and the ancient populations of Europe to whom it is applied
are not quite what they seem.

Theoretical Approaches

The most prominent theory on the “Celts” at present is the concept of the Atlantic
zone first presented by Barry Cunliffe as a reaction to the traditional east to west
movement idea. Throughout the preceding Neolithic and Bronze Ages, Cunliffe (2010,
20) identifies a complex water based trade network. This network can be
demonstrated, according to Cunliffe, through the archaeological distribution of the
megalithic tombs and the Bell-Beaker culture (2010, 24-31). Cunliffe (2010, 22) divides
this period into three stages of development along the Atlantic coastal region;
establishment of connectivity from 5000 to 2700 BC, escalating mobility 2700 – 2200
BC, consolidation 2200 – 800 BC. Cunliffe associates the clear archaeological record
with the results of place-name analysis by Patrick Sims-Williams as mapped by
Oppenheimer (2010, Fig 1.1, 17), and suggests that this provides an Atlantic origin to
the “Celts” of Iron Age Europe. The assumption of this association is problematic on a
number of levels, primarily due to issues from the field of linguistic study that will be
discussed in depth later. These linguistic concerns make the dating and the reliability
of the density of the “Celtic” place-names a substantial issue in Cunliffe’s theory of
“Celtic” Atlantic origin. It can be argued that the Atlantic zone as a source for pre-Iron
Age trade, interactions and social groupings did have a significant influence on the
relationships and movement of objects, people and ideas as the Bronze Age gave way
to the Iron Age. However by the 8th century, shortly before the first records of the

Page | 19
“Celts” by external sources and evidence for the “Celtic” languages in Tartessian
inscriptions in the Iberian Peninsula, these networks appear to have collapsed
(Cunliffe, 2010, 33). The events occurring in wider Europe may provide an insight into
the changing nature of trade and possible causes for the decreasing prominence of the
Atlantic trade networks. Around this period the Greek and Phoenician societies were
beginning to expand, and with them exploration of the Mediterranean and the
beginning of new trading and communication links (Cunliffe, 2011, 270-316). As the
Iron Age progressed, Cunliffe suggests that the networks that existed in the northern
Atlantic zone between modern France, Britain and Ireland also started to break down
(Cunliffe, 2010, 33). The isolation he suggests for Ireland from 500 BC until the 1st
century BC was shared to a lesser extent by Britain from the 4th to 1st centuries BC
(Cunliffe, 2010, 33).

Recently Peter Wells (2004, 31-32) has applied a different theoretical framework to
the concept of the “Celts” - the Tribal Zone theory. This theory, arrived at by Ferguson
and Whitehead, discusses the effect that the existence of a more complex society has
on the less complex societies with which it is in close proximity to (Wells, 2004, 31).
The tribal zone itself is the area which is affected by what Ferguson and Whitehead
call the ‘state’, though not under its direct control or administration (2000, 3). The
impact of this proximity is argued as being the changes in the social structures within
the tribal zone and frequently the emergence of new tribal groups (Ferguson and
Whitehead, 2000, 3). The definition of ‘tribe’ which Ferguson and Whitehead employ
is that used by Haas (1990). He states: “In the simplest terms, a tribe is a bounded
network of communities united by social and political ties and generally sharing the
same language, ideology and material culture. The communities in a tribe are
economically autonomous and there is no centralized political hierarchy.” (Ferguson
and Whitehead, 2000, 28: Haas, 1990, 172). There are two models of tribalisation that
Haas (1990, 173-174) presents: first the traditional model, as discussed by Service and
Sahlin, requires the consolidation to be achieved through coordination rather than
conflict which ultimately results in integration and the formation of new social
boundaries around the new tribe. The second model which Haas (1990, 173-174)

Page | 20
presents as an alternative specifies that consolidation and integration are a separate
process to the creation of social boundaries. Instead the tribalisation process can
occur as a reaction to social and environmental factors or risks, of which warfare can
be an example, and that these factors are the driving motivation behind the
cooperation and locally forged alliances. While Haas, Ferguson and Whitehead have
predominantly examined this theoretical transitional process in the context of warfare
and in the geographic and historical context of the North American tribes, Wells (2004,
113-114) argues that the nature of tribal societies as well as the dynamics between
more complex and less complex societies can be applied to the European Iron Age.

Throughout the history of archaeology and the study of the “Celts” different
theoretical frameworks have been applied in an attempt to explain the events and
relationships of past populations. In contrast these recent discussions now allow the
effect of the preceding ages and developments, as well as those contemporary with
the “Celts”, to be compared in the context of the evidence from the three fields of
analysis. Both theories can be seen to have initial support. The strength of Iberian
language evidence and genetic association with the Isles lend themselves to the
Atlantic theory (Cunliffe, 2010, 33-35), while the increasing appearance of tribal names
in the Classical sources throughout the Iron Age and well documented similarities in
Celtic art tend towards the tribalisation process argued for by Wells (2004, 113-114).
Most importantly, both theories are currently discussed using the assumptions and
strong pre-determined associations which are found throughout Celtic studies. By
reassessing the use of this terminology and applying the theories to the evidence
without the assumptions and pre-determined views that the terminology brings with
it, the value of these theories and their importance to understanding Iron Age Europe
can be explored.

Page | 21
Conclusion

The background and history of “Celtic” analysis is complex and multifaceted.

Traditional sources are fraught with complications of reliability, accuracy and
contradictions. However, it is important to consider the works of scholars from the
Classical world as insightful without them being unquestioned. The appearance in
Classical works of the terms Keltoi and Galli has resulted in the continuation of these
labels in reference to the population of Iron Age Europe outside the Classical world
and as a result the terminology used in the discussion of the “Celts” is out-dated,
inaccurate and encourages the assumptions and pre-determined conclusions which
have occurred throughout the three fields of research until now. The “Celtic” debate
concerns the interpretation of regional variation in the throughout the Iron Age across
Europe. Recognition of the changes resulting predominantly from Roman domination
first on the Continent and then the Isles is necessary in approaching the later stages of
the Iron Age and the social phenomena of this period. Ever since the early 19th century
when the discipline of archaeology emerged, debate has struggled to find common
ground for any discussion on the “Celts”. Major theoretical frameworks which were
taught as accurate and accepted by the academic community have been removed
entirely and replaced by newer, continuously contested theories. Increasingly the
theoretical frameworks presented today are not as narrow in their view, often
incorporating interpretations of wider social networks and timeframes than were
previously applied in the migration theory. The Atlantic zone theory presented by
Cunliffe (2001: 2010: 2011) encompasses both the Bronze Age and Iron Age, while the
tribalisation process applied to Iron Age Europe and the “Celtic” debate by Wells
(2004) has been taken from work by Ferguson and Whitehead (2000) and Haas (1990)
looking at social structures and interactions as well as wider time periods.

The final area surrounding the “Celtic” debate requiring reassessment is the
entanglement between the studies of the ancient people of Iron Age Europe with the
modern concept of “Celtic” identity as it applies to people today. While the
importance and significance of the identity of those who consider themselves

Page | 22
“modern Celts” should not be undermined or disputed, it is necessary to recognise
that this identity is based on an artificial and constructed view of the ancient “Celts” to
avoid confusion. Furthermore, the modern identity has evolved from a political debate
where the idea of a “Celtic” ancestry predominantly concerned language and the
artistic design of ancient artefacts. The genetic identity, although it will in later
discussions be seen to be consistent with pre-Iron Age people, was not known to and
has been neither a focus nor a concern in the nationalism movements. The ‘cultural’
focus of the nationalism movements surrounding the “Celtic” regions arose
predominantly from the language which provided a point of difference from the
English and adopted the “Celtic” label for this identity from the name that had been
assigned to the language group.

The historical environment from which the modern discussion of the ancient “Celts”
emerges is one dominated by tradition, confusion and assumption. While important in
the understanding of current approaches, the background highlights the many issues
that persist in this field of research and the extent to which the assumptions of the
past are restricting the development and advancement of “Celtic” studies. This in turn
has a profound effect on the way that we understand those societies and what we
mean by the term “Celts” as we apply it to the Iron Age. The reassessment of this
question in the following chapter through the fields of linguistics, genetics and
archaeology will aim to highlight and remove these historical restraints.

Page | 23
Chapter Three - A Celtic People? Genetics and the “Celts”

The introduction of genetic research into the study of the “Celts” is a relatively new
development, with the past decade seeing an increase in analysis specifically focused
on tracing genetic origins in Europe. These genetic studies use DNA from the present-
day inhabitants of different regions to look for genetic similarities and differences
which indicate genetic relationships and divergences between regions. When a gene
has been present in a particular area for a longer period, there will be more mutations
of that gene found in the genetic makeup of the people in that region and its
surrounding areas (Sykes, 2006, 141-151). The genes associated with the two sexes;
the Y-chromosome and the X-chromosome, are passed down from the earliest
ancestors and it is the variation and mutations in these chromosomes that is analysed
to understand the relationships and movement of genetic material across Iron Age
Europe. All genetic studies use the Y-chromosome to trace the male DNA paths and
mitochondrial DNA to trace the female paths, with X-chromosome linked
microsatellites, tandem repeat sequences associated with the X chromosome
(Goldstein et al., 1995), being used to further trace the female story (Wilson et al.,
2001). What appears to be clear across the field of genetic analysis at the moment is
the focus on the genetic story of the Insular Celts as opposed to that of the
Continental Celts, with key researchers Stephen Oppenheimer and Bryan Sykes leading
this trend.

Approaches to Genetics in the Debate about the “Celtic” Question

Sykes (2006) and Oppenheimer (2007) evaluate the influence of subsequent

cultural interactions and migrations into the Isles in order to identify the composition
of current DNA and therefore how much it represents the DNA of the Celts. From
these two studies, it has been found that the contribution to the DNA of the Isles from
the Celts has been minimal; approximately 40% from the Vikings in Orkney and

Page | 24
Shetland (Sykes, 2006, 192-194), 20% from the Anglo-Saxons (Sykes, 2006, 286), 2%
from the Normans (Sykes, 2006, 286) and a sparse trace of the Romans (Sykes, 2006,
254; 286). This means that the DNA of the Isles has not changed significantly since the
Late Bronze Age and Early Iron Age. However, the importance of genetic analysis
particularly arises from the insight that can be gained regarding the movement of DNA
in the preceding Neolithic and Bronze Ages.

Since the early papers on the genetics there has been an increasing move away
from small sample group studies. While the specific approach enables detailed genetic
understanding of various ethnic and cultural groups of specific regions, it fails to allow
for the greater connections and establishment of links that could lead towards a
resolution of the debate about the origins of the “Celts” in genetic terms. In a
significant paper for the comparative study of DNA, McEvoy et al. (2004) took the data
and results from 39 previous studies between 1991 and 2004 and added a further 200
Irish mitochondrial DNA samples unrelated to any used in previous sample groups
giving them a larger data range than any prior study. McEvoy et al. (2004) concluded
that neither mitochondrial DNA nor Y-chromosome DNA indicated a strong link to
Central Europe, suggesting that the origin of Insular Celts does not lie in the Central
European Iron Age (McEvoy et al., 2004, 699). Instead, the affinities of the Insular
Celtic regions are with the Atlantic zone, which McEvoy et al. consider to include the
area from northern Spain to northern Britain (2004, 699). This association is one which
is supported by all genetic studies which are indicating a stronger link between the
DNA of the modern inhabitants of the Isles and those of the Iberian Peninsula than
between the Isles and Central Europe (Wilson et al., 2001: McEvoy et al., 2004: Sykes,
2006: Oppenheimer, 2007; 2010). While this appears to suggest a clear case for the
Atlantic theory, the field of genetic studies is however more complex and other
analyses have yielded much more challenging results. McEvoy et al. also identify
exceptions to this general trend; Brittany and Cornwall demonstrate the Atlantic
affinity in Y-chromosome DNA only, not in mitochondrial DNA (McEvoy et al., 2004,
696). However, while the results of each study currently do display similarities, it is
important to be aware of the limitations which affect genetics studies. In the study by

Page | 25
McEvoy et al. (2004), in order to ensure that all the data from previous work as well as
the extra 200 samples used could be compared, the DNA analysis was limited to
positions 16030-16394 on the chromosome (McEvoy et al., 2004, 694). This limitation,
while necessary to enable comparison between different sets of data led to reduced
sample size and restricted the data range of the study. There were also restrictions on
the groups used for analysis with minority sample sources such as the Western Isles,
Orkney and the Isle of Skye not included (McEvoy et al., 2004, 695). While this was
done to prevent possible skewing of the data, it has meant that significant pockets of
genetic data have not being included in this overall comparison and discussion. The
importance of a wide data range and the inclusions of smaller genetic groups can be
seen in Wilson et al. (2001) where analysis reveals the position of the Orkney Isles on
the genetic spectrum in comparison to the Scandinavian and Basque populations and,
more importantly, to other parts of the Isles. In this study the significant differences
between the Orkney DNA and that of Wales and Ireland can be seen, highlighting the
variation and changes in genetic movement that can occur in a relatively small area
(Wilson et al., 2001).

Until recently, the attitude has been that the understanding and knowledge already
gained concerning the Celtic groupings or tribes in the Isles, such as the Picts, was
extensive in comparison to other historical groups. Agreeing with McEvoy et al. (2004)
that there is a significant affinity between Insular DNA and the Iberian Peninsula, Sykes
significantly concludes that the Pictish people of Scotland were of the same genetic
make-up as the other Iron Ages people of the Isles (Sykes, 2006, 282). McEvoy et al.
(2004, 697-698) observed that Scotland had a very strong affinity with the Atlantic
zone as indicated by localised haplotypes. However, they did not recognise the
significance of this link. The suggestion that the Picts are the same people, genetically
speaking, as the Insular Celts contrasts to Bede and Buchanan’s implied alternate
origins or migration path (Collis, 2011, 29; 39). Such a revelation highlights the
importance of genetic analysis in our understanding of past peoples. Wilson et al.
(2001) recognised the shift in archaeological study from settlement of the Isles by

Page | 26
invaders to viewing the development as a result of cultural exchange through trade
networks which perhaps involved only restricted, if any, movement of population.

However, there is one area of analysis which is providing some interesting results
among genetic studies and that is the division between the behaviour of the male and
female DNA. Wilson et al. (2001) first examined this division using the Y-chromosome
analysis to look at the paternal history and the mitochondrial DNA as well as X-linked
microsatellites to follow the maternal path. While this division is a naturally occurring
one when looking at the DNA to trace the origins of the Iron Age people, the
difference in the female and male stories was not expected, nor had previously been
considered by those studying the “Celts”.

Paternal DNA

Wilson et al. (2001) compared the Y-chromosome DNA of different parts of the Isles
to several parts of mainland Europe. From the Isles they had samples from Orkney,
Ireland and Wales, and from continental Europe samples came from Norway,
Friesland, Basque, Turkey and Syria. With Basque representing one end of the genetic
spectrum and with Turkey and Syria at the other, Wilson et al. (2001) demonstrate the

Figure 1: Comparative Y-chromosome table taken from Wilson et al. (2001)

Page | 27
similarity in the genetic patterns from Wales and Ireland and the Basque region. With
a P value of 0.96 and 0.75 respectively, these are strongly positioned in the statistically
similar range (Wilson et al., 2001). While P values cannot provide definite figures for
the degree of similarity, these P values indicating similarity can further be used to
suggest there has been limited genetic change in any of the regional groups keeping
them closely linked. Further studies by McEvoy et al. (2004) continue to show results
which support the paternal similarities between Ireland and the western part of the
Isles and the Atlantic fringe areas of the Iberian peninsula. This connection has been
especially clear through the modal haplotype markers SNP and STR (McEvoy et al.,
2004, 694). McEvoy et al., like Wilson et al., suggest that this historical similarity is
quite old and has remained relatively undisturbed since the Palaeolithic (McEvoy et al.,
2004, 694).

Taking a different approach to the genetic history of Europe, and specifically the
Isles, Bryan Sykes (2006) provides a range of different genetic Clan groups and
compares the strength of each genetic signature. Referring to his previous research,
Sykes traces both the maternal and paternal genetic story of the Isles back to common
ancestors, or the founder of that particular clan grouping (Sykes, 2006, 151). For the
paternal DNA, Sykes identifies five clans, while each clan is identified by a letter, Sykes
refers to them instead using a corresponding male name; Oisin, Wodan, Sigurd, Eshu
and Re (Sykes, 2006, 151). Following the conclusions of previous studies, Sykes again
finds that there is a strong Atlantic trend with Oisin appearing as high as 80% in
Ireland, 60% in Shetland and Orkney and 86% in mid-Wales (Sykes, 2006). In Scotland
the strongest affiliations with Oisin are found on the eastern side in the Grampian and
Tayside regions with a higher frequency, and in England by contrast there is the lowest
appearance of Oisin sitting at approximately 51% in East Anglia (Sykes, 2006, 325). The
general trend of the apparent “Celtic” genetic marker for the Y-chromosome is of an
increased density with a movement north and west from south-eastern Britain.
Without the comparison to the other parts of Europe, this Oisin affiliation
demonstrates a strong male genetic marker in the “Celtic” regions of Ireland and the
British Isles. The strength of the male Iberian connection in the Isles can be

Page | 28
demonstrated through one particular chromosome from Oisin, the Atlantis
chromosome, which is dominant across the Isles and has very strong affiliation to the
Iberian Peninsula (Sykes, 2006, 334). It is important to note, however, that due to the
appearance of haplogroup R1b3f in Iberia and not in Ireland, the arrival of the male
Iberia DNA into the Isles must have been no later than approximately 3,000 – 2,000
years ago (McEvoy et al., 2004).

Figure 2: Y-chromosome relationships in Europe (Oppenheimer, 2007)

Page | 29
Maternal DNA

The other side of the DNA story is that of the maternal path, studied through
mitochondrial DNA, also called mtDNA, and through X-chromosome linked
microsatellites. The female DNA has a much more complicated history, with all
researchers agreeing that it does not have as strong an affiliation with the Atlantic
fringe as is represented in the male Y-chromosome DNA. In Wilson et al. (2001) the
placement of Wales and Ireland on the relative scale produced by the mtDNA is much
more central than had been seen in the Y-chromosome analysis and is considered to
represent a greater degree of gene flow from other European regions than has been
seen in the male story. Wilson et al. (2001) then jumps to the statement that such
mediation of the female gene flow must have occurred in the Neolithic or Bronze Age
period, suggesting that there was at some stage a significant movement of female
DNA only. While McEvoy et al. (2004) agree that there has been a greater movement
of mtDNA, the time period indicated for this is much greater, covering any time since
the postglacial re-expansion of the Isles. Furthermore, while Wilson et al. (2001) and
McEvoy et al. (2004) subscribe to the opinion that the story behind the mtDNA is more
complex than that of the Y-chromosome DNA, both studies conclude that there
remains a greater affinity across the genetic story as a whole with Iberia than central
Europe in the recent genetic past of the Iron Age. However the older Neolithic and
Bronze Age genetic movement is where the significant difference in the genetic
pathways is evident.

Returning to the clan based genetic structure, Sykes examines the maternal DNA
pattern across the British Isles. Helena is dominant across all of the Isles with 45.3% in
Scotland, 45.7% in England and 46.3% in Wales is identified as being the predominant
early genetic signature (Sykes, 2006, 341-343). While Sykes does not provide exact
statistical data for Helena in Ireland, the distribution maps for the maternal clans
clearly indicate that again Helena has a significant presence with more markers than
any other clan in Ireland (2006, 343). The genetic signature for Helena has been traced
to southern France; however the movement of these genes to the Isles is estimated to

Page | 30
have occurred approximately 10,000 years ago in the Palaeolithic and Mesolithic
(Sykes, 2006, 138; 330). Later genetic introduction to the Isles was made by Jasmine
and Tara originating from Syria and Northern Italy respectively. Both appear to have
taken a coastal path via the Mediterranean on their journeys (Sykes, 2006, 138; 256).
Distinctly, Jasmine can be seen to have split into two subgroups around the Balkans
with one group taking an overland route and the other travelling by sea (Sykes, 2006,
256). While the mutations characteristic of both branches are found in the Isles, the
seafaring characteristics are the only ones present in Ireland and along Scotland’s west
coast (Sykes, 2006, 256). There is also a strongly Atlantic affiliation for the Tara genes
with very limited presence inland (Sykes, 2006, 257). Sykes concludes that this later
seafaring movement occurred sometime in the Neolithic and that the maternal story
of the Isles has remained mainly undisturbed from approximately 6,000 years ago
(Sykes, 2006, 331). The old and continuous nature of the maternal DNA indicates that
the “Celtic” people of Ireland and the Isles are not the same as the “Celtic” people of
Europe whose culture appears to have spread outwards from Hallstatt and La Tène
during the Iron Age (Sykes, 2006, 332).

By including analysis of X-chromosome linked microsatellites to the mitochondrial

DNA, Wilson et al. (2001, 5082-5083) demonstrate asymmetry in the genetic stories of
the two sexes. Their conclusions are in polar contrast to Sykes (2006) suggesting that it
was the female population and DNA of the Isles which was significantly altered by at
least one cultural revolution from the Bronze Age (Wilson et al., 2001, 5082-5083).
Oppenheimer (2007; 2010) posits that there were two movements from Western
Europe and that they did not necessarily influence or involve the other (Oppenheimer,
2007, 99). Through discussion of the geographic environment preceding the
movement to the Isles, Oppenheimer (2007) argues that following the last Glacial
Maximum the Isles were a blank canvas on which subsequent living DNA could be
situated (Oppenheimer, 2007, 99-100). This places the movement of DNA to the Isles
as a pre-Neolithic occurrence, a temporal proposition shared by Sykes (2006).

Page | 31
 Figure 3: European correlation maps of mtDNA variation (A), Y-chromosome diversity
(B) and classical gene frequencies (C) as taken from McEvoy et al. (2004)

Interestingly, Sykes identifies significant myths which might correspond to the

maternal genetic story in the Isles. The arrival of the Iberian genes in the Isles occurred
around the time that farming was introduced, and this connection is said, without
evidence or justification, to be reflected in the myths of Brutus and the Milesians
(Sykes, 2006, 332). This predominantly Atlantic zone connection for both the maternal
and paternal side is clear in all genetic studies. However the significant distinction
between the two stories is the complexity of genetic material as well as their time of
arrival to the Isles. Despite opposition from Wilson et al. (2001) the maternal DNA
appears the oldest with relatively recent additions from the Iberian region. By contrast
the paternal genetic story in the Isles is much more recent and displays the “Genghis
effect “, where there is greater reproduction by the successful and powerful members,
resulting in higher than average numbers of their genes in the younger generations,

Page | 32
contributing in the Iron Age to the spread of the Iberian genes and preventing further
mutation (Sykes, 2006, 336).

The end of the Iron Age with the arrival of the Romans in 43AD, starts a new
chapter in the DNA of England. Later, the Vikings would also leave their mark on the
Isles. It is from the early centuries AD that there begins to be a separate story between
the eastern and western coastlines of the Isles, however this separation is not one
which impacts on our approach to addressing the “Celtic” question.

Conclusions

Putting aside the issues of selective data use and assumptions from the traditional
historical viewpoint, the most significant problem remaining is how the results are
interpreted. Despite demonstrating the maternal story of the Isles is much older than
any knowledge of the “Celts”, Sykes concludes that the Isles share a common Celtic
ancestry which has been unaffected by later genetic arrivals from the Romans and
Vikings (Sykes, 2006, 338), a distinction that is addressed by Oppenheimer in his
conclusion to The Origins of the British (2007, 470).

While there is a consensus that the strongest affinity in modern Isles DNA is to the
Iberian Peninsula, there are many as yet unanswered questions which arise from
genetic studies. What happened to produce the different maternal and paternal
stories? Did trade and therefore cross-settlement marital ties influence the movement
of DNA? Furthermore, there are interdisciplinary questions, for example does mapping
genetic movement provide answers to linguistic movement or cultural movement?
Does it indicate both? Does it indicate neither? While the developments and
understanding that DNA studies have offered to investigations into the origins of the
Insular Celts have a significant amount to give, they cannot and must not be taken as a
definitive answer. There is also a blank remaining in regards to Continental Europe

Page | 33
which can at the moment only be shown to be different from the Isles. It can be
assumed that this would also mean that there is a genetic difference between the
Iberian Peninsula and the rest of mainland Europe. However as McEvoy et al. (2004)
demonstrate in an MDS plot of mtDNA much of the mainland of Europe is quite closely
related, although there are significant variations in different parts of some countries
such as Portugal which is split into the North, Central and Southern regions. The plot
shows that while the Basque region lies on one end of the spectrum, the countries
which are often said to be the centre of the ancient “Celtic” culture; Germany, Austria,
France (ancient Gaul), are clustered nearby to Ireland, Wales, England and Brittany.
Does the map distribution by McEvoy et al. (2004) simply represent a greater genetic
movement in these countries following the European Iron Age? Or does it indicate that
using the Basque population as indicative of the Iberian Peninsula genetics has
narrowed the differences between the other European genetic material? As McEvoy et
al. (2004) state “some level of Iron Age immigration into Britain and Ireland could
probably never be ruled out by the use of modern genetic data”. However the location
from which this migration may have occurred remains divided between East and West
with a stronger ling along the Atlantic seaboard from much earlier than the presumed
“Celts”.

Page | 34
Chapter Four - The Celtic Languages and Linguistics Analysis

If an origin of modern “Celtic” study is to be sought, it could very arguably be

looked for in the study of the Celtic languages. While there does not appear to be any
evidence for an Iron Age division in the European social structure based entirely on
language, our initial understanding of Celtic identity was constructed using language
characteristics as the defining criteria. The nationalism movement of the 18th and
19th centuries saw the growth of linguistically defined identities for both recent and
ancient peoples, which led to the modern concept of the present-day Celts and the
inevitable intertwining of past and present identities.

The concept and study of a Celtic language family were forged before the
nationalism movement became fully fledged. Across the 17th and early 18th centuries
Paul-Yves Pezron and Edward Lhuyd studied and pursued a relationship between what
they each saw as being the Celtic languages of the Celtic people. Paul-Yves Pezron first
used the references in Classical works to construct his history of the Celts. As outlined
by John Collis (2011, 48), Pezron’s work centred on the Breton language as the last
remaining remnant of the pre-Roman languages spoken in ancient Gaul. From here,
Pezron proceeded to construct a direct line of descent from the Gomerians to the
Celtae (Collis, 2011, 49). Not only did he provide a history for the Celtic languages, he
provided a line of human descent that carried this language across the sea from Gaul
into Britain introducing the Celtic languages to the Welsh (Collis, 2011, 49). This line of
movement from central Europe across through modern day Germany and France to
the British Isles fits nicely with the traditional archaeological view of a central
European Celtic homeland. While Perzon recognised the relationship between the
languages of modern Welsh and Breton, suggesting this indicated a common parent
language of the continental Gauls or Celts, it was the uptake of Pezron’s work by
Welsh scholar Edward Lhuyd that popularised the Celtic label. Lhuyd secured the
languages of Breton, Welsh, Cornish, Irish and Scots Gaelic as Celtic which started the
modern “Celtic” identity of this region. In his 1707 Glossography which was the only

Page | 35
part of his Archaeologia Britannica to be published, Lhuyd identified the modern Celtic
languages not just as descendants remnants from the ancient languages as Pezron had
done, but as current and actively spoken languages (James, 1999, 45-46). In doing so,
Lhuyd provided a platform for linguistic work that was not, and could not be made,
British. By the end of the 18th century, the idea of a Celtic identity had taken root,
showing up in a range of works by cultural and academic figures from William Stukeley
to Sir Walter Scott (James, 1999, 47). It is at this point that the Celtic linguistic identity
was taken and expanded over several centuries to form the politicised modern Celtic
identity that is claimed in Ireland, Scotland, Wales and other parts of Celtic speaking
regions today.

However, the concept of a Celtic identity for the speakers of the “Celtic” languages
that Lhuyd popularised is not new. Through a systematic study of the descriptions and
phrases used by ancient scholars, Collis (2011) produced a Venn diagram to explain
and highlight the complexity of Celtic identity as it was referred to during the
European Iron Age (Collis, 2011, 101, Fig. 47). The concept of Gaul and Celt as
identifying labels is one of the most marked and telling in Celtic studies. These two
terms appear both in contrast and in parallel through the Classical texts as previously
discussed and continue to divide and perplex modern scholars. Over the centuries of
scholarly discussion, the following definitions have been produced; some Gauls are
Celts, all Celts are Gauls, all Gauls are Celts, No Gauls are Celts, no Celts are Gauls and
some Celts are Gauls (Collis, 2011, 101, Fig. 47). Where the distinction between two
groups is blurred, it is understandable that historical scholars were at a loss to find a
consensual explanation or definitions for the ‘Celts’ and that such variation has carried
through to today. Collis suggests that one of the reasons for the confusion arises from
the transition in names and redefining of groups over the centuries, leading to more
complex naming systems and, what appears to modern scholars, the unexplained
beginning of new clans or groups (Collis, 2011, 104, Fig. 48). It is here that Classical
literature finds its link back into the linguistic arguments of Pezron and Lhuyd. The
linguistic study has developed down two main paths; traditional phonological linguistic

Page | 36
analysis of language age and movement, and, the cross referencing and analysis of
place-names.

Phonological Linguistic Analysis and Language Movement

Forster and Toth (2003) significantly dispelled the concept of P-Celtic and Q-Celtic
branches as developed by Lhuyd in the 18th century (James, 1999, 46). Lhuyd
suggested that P-Celtic reached Britain and Q-Celtic reached Ireland (James, 1999, 46).
In contrast, Forster and Toth (2003, 9082-9083) suggest that Celtic, following a split
from Gaulish, reached the Isles as a single language before it branched out into the
two subsequent groups. However, whether this earlier split from Gaulish had entirely
occurred prior to the movement of Celtic into the Isles or whether the languages were
severed by the move is unclear from their study. Forster and Toth (2003, 9082) give
the approximate date for the movement across the Channel as being 3200 B.C. ± 1500
years. However, while Forster and Toth (2003) present a well-considered argument,
they recognise that a limited and partial picture of the relations and movements of
Indo-European languages has resulted from the exclusion of certain Indo-European
languages such as Hittite and Tocharian from their study, as well as the fragmented
knowledge of some of the languages (Forster and Toth, 2003, 9084). Their conclusions
recognise that further phylogenetic studies may discredit their own results as easily as
they may confirm them (Forster and Toth, 2003, 9084). What Forster and Toth (2003)
fail to acknowledge is that the limitations of their study are not only due to the limited
languages that were included. Their study covers the characteristics and language
features of only 29 items of the original 35 compiled (Forster and Toth, 2003, 9080).
While the fragmented nature of some languages may be the reason for such a limited
sample size, the use of multiple variables, including verbs, nouns, phrases as well as a
range of cases and tenses, severely limits the reliability of the patterns that their work
has identified and the depth of the conclusions drawn from those patterns. Overall,
without follow up work and an increased understanding of the fragmented languages

Page | 37
leading to greater data sample sizes, Forster and Toth’s study is one to be treated and
used with the utmost caution.

Koch, by contrast, has chosen to focus on one particular language and to place it in
context instead of producing an overview and study of the relationships of all Celtic
languages (Koch, 2009; 2010). Koch studies the inscriptions written in Tartessian, a
language discovered in the south-west of the Iberian Peninsula and which can be
dated to the 1st millennium B.C. (Koch, 2010). Covering 700-500 B.C., 95 inscriptions
comprise the corpus of the Tartessian language (Koch, 2010, 198-202). There appears
to be a strong Phoenician connection with Koch finding no correlation between
Tartessian and the indigenous languages of the Iberian Peninsula (Koch, 2010, 205).
The main basis of the Celtic claim appears to arise from a commonality in the linguistic
patterns in Tartessian and those found in early Welsh and Irish poetry as well as
principles of modern Gaelic languages appearing in incomplete forms during the
transition from the Phoenician systems to those used in Tartessian (Koch, 2010, 205).
In contrast to the earlier dates of Forster and Toth (2003), Koch posits that Celtic
language reached Ireland through the ‘professional classes’ in 1200-600 B.C., a date
that is consistent with that proposed by Koch in 1991 (Koch, 2010, 209). This is a much
later date than Forster and Toth calculated (2003), even if their error of 1500 years is
taken into account, which makes 1700 B.C. their outside date. This contrast in dates
could perhaps be explained by Forster and Toth’s (2003) suggestion of a movement to
Britain and Koch’s (2010) discussion of a movement to Ireland. However the split of
the Celtic language into Q-Celtic and P-Celtic is not consistent with a separate
movement of language into Ireland. At present it would seem that the main role of
Tartessian, until greater knowledge and study in Celtic linguistics can be achieved, is to
demonstrate the link between the Isles and the Iberian Peninsula in a linguistic
context. Furthermore, the language of Tartessian suggests greater interaction with
non-Indo-European speakers in the Mediterranean and northern Africa (Koch, 2010,
209).

Page | 38
Place-name Analysis

Moving away from the study of specific languages and their relationships, Sims-
Williams’ (1998; 2006) has applied the known locations of Celtic place-names to the
issue of the origins of the Insular Celts. Through the application of place-names in
Europe, Sims-Williams data has been produced in map form by Oppenheimer which
shows a largely Western European dominance in Celtic language (Cunliffe, 2010, 17,
Fig. 1.1). The study of place-names including particular affixes not only provides a
distribution for the languages as a whole, it also provides insight into the survival of
particular affixes across Europe. Parsons (2010, 176-181) used distribution maps by
Sims-Williams to demonstrate the spread of place-names using the affixes -brigā, -
dūno-, -duro- and -mag- (Fig. 4 - 7). The results of this study indicated distinct
distribution areas and differing densities for each affix (Parsons, 2010, 176). The affix –
brigā- is the only affix in his work occurring in high frequency south of the Pyrenees.
This limited evidence through place-name analysis in the Iberian Peninsula is in direct
opposition to the evidence for Tartessian in the same area (Parsons, 2010, 176). The
study of place-names, which requires an understanding of the languages spoken in
each region to enable the correct identification of syntactic, phonetic and
morphological features, raises multiple issues. Any study that aims to look at the
languages of the Celts, especially the Insular Celts, must face the reality that this was a
world pre-dating widespread writing and recording, making any record of place-names
subject to multiple obstacles. Names are often recorded by external societies, with the
records being made years later following word of mouth, or after the original names
have been subject to Roman, Viking or Norman contact. Furthermore, the spread of
language and change of political environments and relations over time may have
resulted in the changing of Celtic names to other linguistic families and left no trace of
Celtic place-names. Parsons recognises the further issue of chronology in using place-
names as more than evidence of settlement or contact, pointing out that the
distribution of the very different -brigā- and -dūno- could be used to argue for a longer
chronology of Celtic languages in the geographic area where each affix is found
(Parsons, 2010, 177). While arguments for chronological variation can be made by

Page | 39
comparing affix distributions, through common meanings and geographic overlap,
such an approach does not have room for the possibility of varying contemporary
dialects or multiple words for one concept or form. Parson’s criticism of Helmut Rix,
who uses such justifications when ordering the chronology of the affixes, highlights
this weakness and limitation when studying place-names without correlation with any
other method of linguistic, or indeed archaeological work (Parsons, 2010, 177). Most
place-name analysis is carried out in the western part of mainland Europe and the
British Isles; however the trail of Celtic names is much more wide spread, with
Noviodunum, a name which occurs multiple times in Gaul, also appearing in Dacia near
the Black Sea (Parsons, 2010, 176). The failure of place-name analysis to encompass
the true extent of Celtic languages, and the reliance on other fields to contribute to a
chronology of Celtic languages, is perhaps the best example of the convoluted and
inter-disciplinary nature of the study of the Celts. It highlights the need for greater
study and flexibility from the traditional academic approaches and theories that have
been so influential since the early studies by Lhuyd and his contemporaries.

Figure 4: Distribution of -brigā names of ancient Europe

(Parsons, 2010, Data from Sims-Williams)

Page | 40
Figure 5: Distribution of -dūno- names of ancient Europe
(Parsons, 2010, Data from Sims-Williams)

Figure 6: Distribution of -duro- names of ancient Europe

(Parsons, 2010, Data from Sims-Williams)

Page | 41
 Figure 7: Distribution of -mag- names of ancient Europe
(Parsons, 2010, Data from Sims-Williams)

Inscriptions

The academic approach often used for the study of the Celtic languages uses
patterns and inter-language relationships to posit the path ways and developments
that these languages would have followed. The examination of the dateable evidence
for Celtic languages and their presumed existence throughout the European Iron Age,
mainly in the form of inscriptions, highlights the complexity not only of language
study, but Celtic studies as a wider discipline. The first inscriptions which can be
assigned a date appear in southern Gaul written in Greek script in the 3 rd century BC
(Wells, 2001, 107), however the presence of the Greek colony at Massalia, modern
Marseilles, around the 6th century BC would imply that a system of writing was
available and the local people exposed to it much earlier (Russell, 1995, 3). Later

Page | 42
inscriptions occurring in the 1st century BC contain Celtic names which can be aligned
with those mentioned by Caesar. Evidence suggests that the use of Greek script and
Celtic languages was a common practise throughout the region, with name inscriptions
on local pottery in Manching and the blade of an iron sword found in Port, Switzerland
demonstrating the interactions between the Celtic tribes and Classical people in the
south (Wells, 2001, 108). The Romanisation of the Cisalpine Gaul region in the 2nd
century BC saw the recording of Leptonic, the Celtic language in the area, although
some argument has been made for evidence of this language in the 6th century BC
which would make it the oldest evidence for a Celtic language in the region (Russell,
1995, 5).The transition from the use of Greek to inscriptions in Latin as on the Coligny
calendar from the late 2nd century BC, found in Ain, France, is widespread by the 1 st
century BC with Latin inscriptions appearing on the coins of the Boii in central Europe
(Wells, 2001, 108). These inscriptions demonstrate the transition from Greek-Celtic
relations to Roman-Celtic relations; however, more importantly, they suggest an
expansion of these relationships and the use of foreign script to record in Celtic
languages.

Renfrew discusses the development of the “Celtic” languages in Europe and follows
Tovar’s wave model which suggests centres which branched out from the oldest Indo-
European languages (1998, 246-248). Renfrew and Tovar posit that these new
secondary centres, which were interspersed with more ‘conservative’ regions, were
the source of language development which then spread further outwards (Renfrew,
1998, 246-248). The dating of the oldest inscriptions to the 6 th century for both
Tartessian, and perhaps Leptonic, supports this notion of plural centres of
development for “Celtic” Indo-European languages in central and western European.
The development of the “Celtic” language in the central and western parts of Europe
and the distribution of the inscriptions, whether using Greek or Latin text, suggests a
distinction between the people of the Celtic languages and those of the Germanic
languages, though Wells (2001) cautions against over simplification of these patterns.
While there are outliers from the normal geographic range of both language groups,
any attempt at forming simple boundaries is prevented by comparison with the

Page | 43
distributions of Celtic place-names, which, while similar, does not support the same
line of distinction (Wells, 2001, 108).

Conclusion

Linguistic study at this point does not have a conclusive argument as to the origins
of the Insular Celts. However what it does suggest is that any question of a linguistic
origin is far more complicated than merely pinpointing a geographic locale from which
the languages of the Continental and Insular Celts spread. The issues of dating further
highlight the problems already apparent in place-name analysis and it remains vital
that the limitations of the evidence continue to be considered, whether they be a lack
of data or inaccurate representation or distribution. The use of the discovered
inscriptions to study and map language through pattern projections may result in a
skewed representation of the spread as well as frequently changing patterns as new
evidence comes to light. At present there appear to be two distinct points from which
the Celtic language inscriptions are dispersing. First the oldest sources appear to be
those of Tartessian in the Iberian Peninsula and, secondly, the evidence in central
Europe appears to spread outwards towards the Atlantic coast with the transition
from interactions with the Greeks to those with the Romans. The distribution of the
inscriptions and the discussion of “Celtic” languages is predominantly a western
European pattern with movement in a general north-westerly direction from either
point of dispersal. While the traditional models focus on the origin of the language,
the evidence only provides a suggestion of the development of different languages in
different regions well after the arrival of the Indo-European languages into Europe.
The search for a single chronological spread may not, at these later stages of Indo
European language development in Europe, be the only option when variation in
dialects and the different pathways of Celtic language are considered. At this stage
more work needs to be done to broaden the area of study, as well as to cross-check
the data with other fields of research to provide a greater number of definite dates.
However, perhaps the largest concern with linguistic study, is the hangover of the

Page | 44
terms used to describe the language and the different branches of the language which
can be seen to have an extended impact on the way data is treated, most notably
Celtic as a term for the family of languages and P-Celtic and Q-Celtic to indicate
language in specific geographic regions. These terminologies and pathways reinforce
the idea of a single people or society, and do not recognise or consider the wide range
of different dialects which studies in Britain suggest for mainland Europe and the Isles
towards the end of the Iron Age. The expansion of the “Celtic” languages is a distinctly
Iron Age phenomena with concurrent development contributing to the increasing
variation between languages. By contrast the genetic patterns for the Iron Age are
consistent with those of the preceding millennia. The divergence between the two
lines of development reveal that the genetics and linguistics do not refer to a single
integrated entity, and do not constitute a basis for any “Celtic” group, people or
culture.

Page | 45
Chapter Five - Material Culture: the Archaeological Record

The final field of enquiry that needs to be examined in the assessment of Celtic
debate is that of archaeology. Predominantly, the study of the archaeology of the
“Celts” has revolved around the weaponry and metal artefacts that were left behind as
the material representation of the phenomenon, with specific focus on the designs
and embellishment that were used. This field grew out of the 18th and 19th century
discussions surrounding the identity of the pre-Roman inhabitants of the Isles (Collis,
2011, 73). The conclusions that were reached concerning “Celtic” ancestry have been
sustained over several centuries to take an unquestioned pride of place in the
established concept of “Celtic identity”, as has already been discussed.

Archaeology in 19th century Europe was a relatively new and expanding discipline,
with the concept of chronological frameworks developing rapidly for both the study of
the historical period and increasingly for use in prehistory (Collis, 2011, 73). Further
development resulted in the introduction of the Palaeolithic and Neolithic labels by Sir
John Lubbock in 1863 and the widespread adoption by archaeology of artefact
typology (Collis, 2011, 74). It was during these developments, along with a rush of
funds for collections and excavations to use archaeology for national purposes, that
the excavations of Hallstatt in Austria began under Johann Georg Ramsauer in the
1840s (Collis, 2011, 74: Cunliffe, 1997, 28). The use of the label “Celtic” by Ramsauer
for the graves he uncovered in the Hallstatt cemetery was a direct application of the
academic labels and chronological frameworks that had already been established
(Cunliffe, 1997, 28). As excavations continued over the next hundred and fifty years,
these labels remained in place, unquestioned and helping to create and define the
concept and identity of a population that was presumed to cover the majority of
mainland Europe.

Within the field of archaeology there are distinct aspects of the material culture

Page | 46
that have created their own sub-fields for research. The “art” of the artefacts has
become one of the dominant areas of study; however there is also research on
artefact distributions as well as settlement types.

Art

The most significant question when looking at Celtic art is definition. What is meant
by the term Celtic art? The beginning of Celtic art definitions followed two distinct
lines of development. Still identifying the features of Celtic art when he died, the work
of John Kemble, published post-mortem by his colleagues, suggests the most
distinctive feature is that of a ‘trumpet shape’ or ‘trumpet scroll’ (Collis, 2011, 82). This
curved triangular shape had been mentioned by Kemble at a lecture in 1857 (Collis,
2011, 80). Augustus Franks, one of Kemble’s colleagues who helped complete his
work, identified this shape as ‘Late Celtic’ on the basis that it was not Roman, Saxon or
Viking and did not correspond to the geometric designs that had been identified from
the Bronze Age (Collis, 2011, 82-82). The second developing theory which influenced
the definition of Celtic art was the understanding that specific artefact types may be a
material representation of certain populations, and possibly their migratory
movements (Collis, 2011, 84). This understanding emerged from the recognition by
Emile Désor and Gabriel de Mortillet in 1871 that some artefacts found at the Etruscan
site of Marzabotto by Count Gozzadini were distinctively different to the rest and had
a closer association to their own finds from La Tène and northern French burials
(Collis, 2011, 85).

Both paths merged and the division of the European Iron Age into the Hallstatt and
La Tène periods began in 1872 with Hans Hildebrant, commencing the view of La Tène
as iconic of the “Celts” (Cunliffe, 1997, 32). Otto Tischler soon introduced the
subdivision of these two periods with Early and Late Hallstatt in 1881 and Early,
Middle and Late or I, II and III La Tène periods in 1885 (Cunliffe, 1997, 32). For the La
Tène period this division into three parts resulted from the artefact styles that were

Page | 47
seen to represent that section of the La Tène era. Tischler identified certain
characteristics of fibulae and swords as being iconic of the three stages of the La Tène
material culture (Cunliffe, 1997, 32). Subsequent divisions using a letter numbering
systems (Cunliffe, 1997, 32-33) resulted in Reinecke’s structure in 1902 where the
early Bavarian artefacts, which predated Tischler’s division, was labelled La Tène A,
and the later periods B, C and D corresponded to Tischler’s Early, Middle and Late
labels (Cunliffe, 1997, 33). Reinecke also finalised the division of Hallstatt into A and B
for the late Bronze Age and C and D for the early Iron Age (Cunliffe, 1997, 33). A
slightly different approach was suggested in 1910 by Déchelette who divided his La
Tène into three geographical groups; Celtic, Germanic and Insular (Collis, 2011, 90).
Few significant changes have occurred to the chronological and ‘cultural’ markers
since the early 20th century, with the majority of works using one or other of these La
Tène dating systems. It is important to note, however, that the division of time periods
in Ireland and the Isles has not seen the same attention or division, with the labels of
“Late Celtic” or “Early British” remaining despite attempts to link the Insular to the
Continental (Cunliffe, 1997, 33).

Since these early approaches to Iron Age “Celtic” art there has been a high level of
consistency in the understanding of how it is defined (Garrow and Gosden, 2012, 49),
and as a result the study has been predominantly concerned with the Hallstatt and La
Tène styles and their perceived manifestations in the Isles (Harding, 2007, 262).
Recently this narrowness has been challenged with Harding (2007, 6) arguing for the
overall acceptance of Celtic art as ornamentation from the regions identified as having
a linguistically Celtic background. This approach to a definition brings back the many
issues associated with the use of linguistics as an indicator of culture or background.
However, the foundation of Harding’s argument that Celtic art cannot remain centred
on La Tène styles is one that should be more widely recognised for its potential
importance in future analysis. As the definition of “Celtic” becomes increasingly
unclear, the removal of traditional frameworks in the study of Iron Age art is a crucial
step in disposing of the preconceived ideas and assumptions inherent in the study of
the “Celts”. Celtic art, or Iron Age European art, needs to be rethought, widening it to

Page | 48
include the ornamentation outside the Classical societies, but not restricted to those
areas commonly referred to as “Celtic”.

The engravings and designs on the artefacts from the central European sites of
Hallstatt and La Tène have built the foundation for the analysis of Iron Age “Celtic” art
through a stylistic and typological approach. There remains however, no overarching
approach for dividing or labelling the styles, with many scholars having separate labels
and stages for the ornamentation throughout the Iron Age. There are some common
divisions based around features with Waldalgesheim, Vegetal and De Navarro style II
all referring to the use of continuous tendril patterning form approximately 350 BC
(Megaw and Megaw, 2005, 11). There is also a consensus that there is a transition
throughout the Iron Age from the geometric patterns of Hallstatt art onto the softer
tendril patterns before the appearance of stylised human and animal forms and finally
the influence of Roman occupation and the shift into Christian Celtic art (Megaw and
Megaw, 2005, 11). While this stylistic approach, with some variation, is one of the
most common approaches to a discussion on Iron Age “Celtic” art, there is a significant
geographic component that provides some insight into the understanding of art and
ornamentation in Iron Age Europe. The majority of stylistic chronologies and
typologies make a distinction between the styles on the Continent and on the Isles. De
Navarro in his stylistic divisions placed Insular Iron Age “Celtic” art into two groups; IV
which covers the period before 100 BC, and V including the 1st centuries BC and AD,
an approach which has also been used by Ian Stead and, more recently, Martyn Jope
(Megaw and Megaw, 2005, 11). While there is no agreed way in which to divide and
label the different periods in Iron Age “Celtic” art, there is a clear consensus that the
Insular Iron Age “Celtic” art is distinct from the developments happening on the
continent. The placement of the Insular Iron Age “Celtic” art in separate stylistic
groupings ensures that there is always a geographic divide in any stylistic or
typological analysis. This links to the second manner of discussing Iron Age “Celtic” art
which is using a geographic system (Megaw and Megaw, 2005, 10).

Page | 49
 Figure 8: Mindelheim and Gündlingen swords (Collis, 1989)

Where there is a geographic focus on the study of Iron Age “Celtic” art, it is most
common for this to be a focus on the art in the Isles as has been used by Megaw and
Megaw (2005). There is also a comparative analysis of different regions from
Continental Europe; however this geographical division generally appears within the
context of the stylistic discussion such as is used by Harding (2007). The use of
geographic regions as a classification follows and helps to highlight the regional
variation that can occur within artefact or ornamentation types. There are distinctive
geographic boundaries which occur in the archaeological record itself. The geographic
divisions within the Sword Style of the La Tène period are a clear representation of the
merging of these two methods of analysis. First discussed by Jacobsthal, the Sword
style was derived from the swords dating to the mid La Tène in modern Hungary
(Harding, 2007, 93). Harding’s (2007) discussion of the style reveals that the
archaeological record naturally divides into four distinct groupings; Hungary,
Switzerland, Britain and Ireland, with each region distinct in regards to style,
manufacture and technical characteristics. In terms of the stylistic grouping, all swords

Page | 50
can be linked to the Waldalgesheim style (Harding, 2007, 93), also called vegetal or De
Navarro II (Megaw and Megaw, 2005, 11). However, there remain some significant
divides along the geographical boundaries. First, the major geographic divide of Insular
and Continental Europe can be seen in the use of material. During the middle La Tène
period, Harding identifies the Hungarian and Swiss swords to be predominantly made
of iron, while bronze swords are prevalent in the Isles (Harding, 2007, 95). The
manufacture of the swords also reveals the Irish artefacts to be shorter with a
tendency to be less than 50cm, while those from the Continent and Britain can be
upwards to 75cm in length, a discrepancy that Raftery and Jope argue reflects the
fighting style of the different regions (Harding, 2007, 113). The second division which
demonstrates strong geographic tendencies is the ornamentation of the swords and
scabbards. Those which come from the Insular region tend to have the entire surface
of the sheath decorated, while their Continental counterparts will demonstrate more
restricted ornamentation appearing in panels, with further characteristic placement of
the ornamentation in the Swiss sheaths to below the scabbard mouth and rarely down
the plate (Harding, 2007, 93; 106). The Hungarian sheaths also have distinctive
placement, with a highly symmetrical, diagonal placement ornamentation, although
there is still debate as to whether these features are sequential or contemporary
(Harding, 2007, 96; 99). While the Hungarian and Swiss sword sheaths demonstrate a
more reserved use of ornamentation, the Irish sheaths are densely decorated with
repetitive patterns which can often give the impression of greater complexity
(Harding, 2007, 113).

Despite the strong evidence for distinct geographic trends in manufacture

processes and stylistic choices, there is also evidence among the La Tène swords for
the expansive trading network and movement of ideas that is increasingly
characteristic of the European Iron Age. Dragon-pairs (Fig. 9) are one of the most
common pan-European designs, which has been found from southern England through
to Transylvania as well as south of the Alps and even with one example in the Iberian
Peninsula, south of the Pyrenees (Harding, 2007, Fig 5.5 102). Though there is a higher
density of these finds in the central European mainland, particularly to the east

Page | 51
 Figure 9: Dragon-pair ornamentation designs and European distribution
(Harding, 2007, Data from de Navarro and Stead (A) and from Stead and Ginoux (B))

between the middle Danube region and the Tizsa in modern Hungary (Harding, 2007,
104), the spread of this design in three different types is significant in revealing the
wide distribution range and similarity that could be achieved in the European Iron Age.
Further evidence for the span of designs can be seen in two features common
between the swords from Ireland and Eastern European regions; the use of open-ring
chapes and the triple dot signature (Harding, 2007, 93;117). The exclusion of Ireland
from the distribution of the dragon-pair style and the isolation of certain features to
the ends of the geographic spectrum suggests that while patterns and ideas may have
moved vast distances across Iron Age Europe, the manufacture methods and

Page | 52
ornamental designs employed in each region were selected specifically, or are
representative of carefully chosen and established trade links. The long standing view
that the Irish scabbards represent a distinct localised tradition further supports the
selective and reasoned used of specific motifs.

The final approach that can be used in a discussion of Iron Age “Celtic” art styles is
to place them in their social context and view them as a reflection of different social
aspects in an attempt to understand any meaning that the ornamentation may have
communicated to the pre-historic population (Green, 1996, 11). This approach is not
as concerned with the typological and chronological understanding that has
dominated the study of Iron Age “Celtic” Art until now. It is also an increasingly
common and significant style of analysis with scholars such as Green (1996) who
focuses on the more abstract aspects of society including gender, class and spirituality,
and Garrow and Gosden (2012) who use different locations of deposition such as
burials, hoards and settlements to look at Iron Age “Celtic” art in a social context.
Through this approach Garrow and Gosden (2012) also highlight the other issue in the
definition of “Celtic” art - the concept of art itself. The debate surrounding the
definition of the word ‘art’ is a very complex one and while this discussion does not
seek to resolve that debate, it is important to understand its influence on the analysis
and study of the “Celts” and their ‘art’. Today the concept of art is often seen as a
representation of the emotional and sometimes irrational, as a contrast to the
scientific (Garrow and Gosden, 2012, 38). Garrow and Gosden emphasis the
connection art, religion and magic would have had in prehistory, and through into the
Roman period (2012, 38). The distribution of specific motifs and manufacturing
methods is not simply an indication of trade. Green (1996, 17; 31) argues that art took
the place of writing in the pre-literate societies of the European Iron Age, holding a
significant place in the representation of rank and relationships and well as religious
symbolism. Green (1996, 120-121) sees the wide spread use of the palmette and lotus
designs as clear evidence of this citing their frequent use with human head
iconography as well as the wide distribution and cross-cultural appearance as
suggestions of their spiritual significance. Common themes and associations with these

Page | 53
symbols throughout history and across the geographic boundaries are strong evidence
for the communication of ideas from different regions, and the origin of these symbols
from external social groupings also suggests they were deliberately adopted by the
Iron Age European people (Green, 1996, 119). While much of the symbolism of the art
is conjecture arrived at through understanding contemporary views, commentary by
Roman scholars following their later arrival and co-habitation with the Iron Age
people, and some assumptions; the correspondence between the human face and
both foreign and native vegetal designs to the life of Iron Age people is clear. Green
(1996, 122) draws on these connections to argue spiritual importance for the
representation of native plants, yew berries and mistletoe.

Each of the approaches used in the analysis of the ornamentation of non-Classical

Iron Age artefacts have issues of analysis which must not be overlooked,
predominately the clarity of argument through inadequate terminology, presumed
meanings, and concerns for the reliability of analysis resulting from limited data
sources in particular areas. There are however, some trends which can be found across
all approaches of analysis which demonstrate clear features of social structure and
operation in Iron Age Europe. The geographic divisions, both on the large scale
between the Isles and the Continent, and on a smaller scale between regions now
represented by separate nation states, highlight the localised nature of ornamentation
and manufacture across Iron Age Europe. While on its own this indicates a segregated
and autonomous system of existence instead of a centrally managed and operated
social structure, it is the relationship of this trend to those that emerge from other
areas of archaeological study, and then other fields of “Celtic” study, that will enable
an extensive reappraisal of the evidence and terminology in order to reassess the
phenomenon labelled as the “Celts”.

Page | 54
Fibulae

An artefact continued from the Bronze Age into the beginning of the Iron Age
through Hallstatt and La Tène, fibulae are found in relatively high numbers compared
to other artefacts including swords. The decoration and ornamentation of the bow is
parallel to the designs found on other “Celtic” metalwork (Green, 1996, 81), making
them a marker of chronological change through the use of “Celtic” art styles as
previously discussed as well as manufacture and localised choice. They have also been
adapted throughout the Iron Age across the whole of Europe, which results in large
amounts of data that can be analysed to identify geographic trends and distributions.
The study of fibulae is characteristic of the localised site or region based style of
archaeological analysis. Due to the amount of data, the placement of fibulae in burial
contexts and the analysis of the fibulae as items of “Celtic” art, the majority of works
which discuss fibulae are focusing on a localised area. The Insular (Garrow and
Gosden, 2012: Hattatt, 1985: Kilbride-Jones, 1980) and Continental (Collis, 1989:
Harding, 2007: Alexander, 1965: Roes, 1965) divide is the first localisation. Beyond
this, there is further specialisation ranging from central European river networks of
Marne-Mosel and Rhine Danube (Lorenz) to the Baltic regions (Blečić Kavur, 2009). The
product of such specialisation is that the results of these studies are placed into the
greater debate concerning the trade networks and “Celtic” art styles; however there
has been no fibulae focused study across the whole of the European Iron Age. This is a
mammoth task, and cannot be achieved in the space here, however an overview of
the results of in-depth studies and the trends that they have identified across the
different regions and time period periods can be highlighted, as well as the beginnings
of a cross-Europe perspective focusing on the distribution of fibulae.

Fundamental to the study of Iron Age fibulae is the variety of labelling that is used
across different regions and time periods. This variegated use of labels and
characteristics results in highly complex systems of identification, such as Hull’s 300
type classification for ancient British brooches, the use of the same terminology for
multiple styles, such as the Polden Hill label, and finally the application of different

Page | 55
labels and stylistic names to one brooch or fibulae (Hattatt, 1985, 3). While some
labels are unique to specific regions, such as the Kastav type fibulae (Fig. 10) from the
Caput Adria or head of the Adriatic (Blečić Kavur, 2009), others are identified in a
much wider region such as Nauheim fibulae (Hattatt, 1985, 20).

Figure 10:Distribution map of Kastav type and variant fibulae (Blečić Kavur, 2009)

While the issues and confusion which arise from multiple labelling systems would
require a thorough study focused on reassessing all currently known Iron Age fibulae,
there is an alternative option for understanding the spread and manufacture in the
Iron Age. Distribution maps have been created from different studies covering the
region in which specific styles have been found (Fig. 10 - 15). These as a whole do not
cover the entire region of the European Iron Age which suggests that the area in which
the fibulae are made and traded is contained. Using the distribution maps created for
different types of fibulae from a range of sources, areas of higher density can be
identified. In Figure 11, these areas act as an indication of higher frequency for each

Page | 56
type of fibulae shown, with each area of higher density separate from the others
thereby demonstrating distinct centres for the different types represented. The spread
of the fibulae from the areas of highest density represents the area of influence of
each style. This appears to have occurred through the movement of artefacts via trade
and exchange. The areas of higher density likely represent one of two things; the area
of manufacture for that fibulae style or the area from which the trade and exchange of
that style begins. In some cases these could be the same place, however with the
emergence of the oppida and the centralisation of trade there may be a difference in
the location of the two. The term oppida is borrowed from the accounts of Caesar and
roughly translated from Latin to mean “towns” (Collis, 1989, 150). The result of the
survey of the fibulae distribution maps is that each type of fibulae emerges with an
area, or in some cases, areas of higher density and therefore centres for distribution or
manufacture.

Figure 11: Distribution of Fibulae (Müller and Steuer, 1994, 94)

Page | 57
 Figure 12: Distribution of Kessel Fibulae (Roymans, 2004, 121)

Figure 13: Distribution of Gallische Bügelfibel (Müller and Steuer, 1994, 102)

Page | 58
 Figure 14: Distribution of Nauheim Fibulae (Müller and Steuer, 1994, 53)

While the general trend emerging from the styles of fibulae is a region specific
distribution, there is wider spread for later fibulae designs, such as the Dux and
Nauheim fibulae. While Nauheim distribution is clearly a pan-European trend (Fig. 14),
there is however, some level of inaccuracy regarding the Dux type such as the
examples supposedly from Wallingford, England which Stead argues were probably
attained without knowing their original provenance (Stead, 1984, 53-54). The
expansive distribution of the Nauheim type fibulae is due to the change in the
production methods to being cast as a single piece and designed for mass-production
(Wells, 2012, 109). The change in the distribution and frequency is evident through the
comparison of the previous fibulae distribution maps (Fig. 10 - 15) and the distribution
map of the Nauheim fibulae (Fig. 14). This shift in the production and distribution
coincides with the introduction of the oppida and the extensive, market driven trade

Page | 59
 Figure 15: Iberian Horseman Fibulae distribution (Cunliffe, 1997)

networks that will be discussed in the next section. Furthermore, many of the areas
identified on the Nauheim distribution map are in locations of oppida or trading
centres, such as Massalia on the Mediterranean coast. Through the study of fibulae
type distribution from different parts of the continent and isles, the localisation of
manufacture and distribution throughout the European Iron Age can be highlighted.
While this changes in the last two centuries BC, the majority of the period is
characterised by this style of manufacture and distribution. Combined with the
localised nature of art styles, the distribution of fibulae represents the choices made
by the people of different regions regarding the production and ornamentation of
fibulae. An understanding of the wide spread network of trade and interaction which
was well established at the same time as the fibulae production demonstrates that
local patterns of distribution and preference were still prevalent.

The distribution of artefacts traditionally considered to represent the presumed

“Celtic” culture is generally more limited south of the Pyrenees, where a distinct
pattern of ornamentation and artefact manufacture can be observed (Harding, 2007,

Page | 60
189). Harding (2007) has argued for this variation to be considered part of the “Celtic”
group, although distinct from the more traditional La Tène definition. One trend that is
found across Europe and continues into the Iberian Peninsula is the production of
localised fibulae of which the horseman fibulae mapped by Cunliffe is a prime example
(1997, Fig 113, 142). Such distributions demonstrate that the southern Atlantic region
does have a distinct archaeological record in regards to the artefacts themselves,
significantly through not conforming to the pan-European trends of dragon-pairs (Fig.
9), Nauheim fibulae (Fig. 14) and Mindelheim and Gündlingen swords (Fig. 8). This
record demonstrates that the Iberian Peninsula, just like the rest of the European
mainland, has distinct regional, localised ornamentation and distributions.

Trade and Interaction

Cunliffe’s work on pre-Iron Age Europe highlights the expansive trade and
interaction that was taking place across the Continent, as well as between the
Continent and the Isles, especially along the Atlantic fringe (Cunliffe, 2010: 2011). The
most striking conclusion of his research in relation to the Iron Age is the collapse of the
Atlantic fringe trading networks at the junction of Bronze and Iron Age (Cunliffe, 2010,
33). Cunliffe states that this break down not only reduced the relations between the
Isles and the Iberian Peninsula, but also those between Britain, France and Ireland, a
state of limited interaction that remained until the influence of the Roman Empire
began to emerge and be felt in the more northern and western parts of Europe in the
1st century BC (Cunliffe, 2010, 33). However, the evidence for common ornamentation
through the art, the manufacture of fibulae and Iron Age round-house distribution
would suggest that this break down of communication was not as extreme as Cunliffe
has suggested.

The archaeological record provides some more insight into the relationships
between different parts of Iron Age Europe, and it is this that needs to be examined
and assessed along with the trends from the artefacts in order to understand how the

Page | 61
trade networks of the Bronze Age were affected by the transition into the Iron Age
and by the increasing power of the Romans across the Continent.

Renfrew outlines the two types of trade that can occur and which are observable in
the archaeological record of the Iron Age: the trade of materials which have a limited
distribution in the natural world and the trade of products which are made in a limited
area due to the availability of knowledge or other factors which make that area the
most efficient for production (1979, 28). In the study of Iron Age trade, which centres
on the tracing of materials and products from the Classical world, both types can be
observed. As discussed by Cunliffe (2011), the increasing prominence of the Greek,
Etruscan and Phoenician societies moved the trade networks of the early Iron Age
away from the Atlantic, and focused them on the Mediterranean region. Evidence of
the trade relationships between the Mediterranean regions and the Iron Age
populations in central and northern Europe can be traced through the distributions of
Greek amphorae along the coast and river networks. The Greek settlement of Massilia
on the southern French coast was a central point for trade up the Rhone River
(Milisauskas, 1978, 270). This trade connection continued from 600 to 500 BC and
branched in two directions; towards the Rhine and towards Austria-Bohemia
(Milisauskas, 1978, 270-271). The amphorae indicate the trade of wine; however it is
also likely that the Greeks brought finished products, glass and corals to the early Iron
Age populations in these regions and received copper, tin, amber, gold, slaves as well
as maybe iron and salt (Milisauskas, 1978, 271). Interestingly, there seems to be a
trend of processed materials or products being traded by the Greeks and with natural
materials in return representing both aspects of the trade that Renfrew discussed. This
pattern is continued when the trading power of the Greeks as represented through
the appearance of their products in this region is replaced by the Etruscans and their
products around 500 BC (Milisauskas, 1978). The shift in power also saw an increasing
level of trade with the eastern regions of central Europe as shown in the source of
materials found in burial mounds at Dürrnberg, Austria (Milisauskas, 1978, Table 9.6,
273). The materials included jet, tin, copper, silver, gold, amber, coral and glass with
the majority of materials coming from the eastern regions including Bohemia,

Page | 62
Slovakia, Poland and the Dinaric Alps as well as the Mediterranean coastal regions. The
spread of these materials is apparent in specific items such as the Basse-Yutz flagon
found in Yutz eastern France (Caygill, 1999, 53) and the Witham Shield from the River
Witham in England (Jope, 1971, 61). The bronze flagon dates to late 5th century BC and
has coral inlaid in its base (Caygill, 1999, 53). The shield dates to approximately the
3rd century BC and contains three red coral embellishments (Jope, 1971, 64). The
source of the coral for both artefacts is the Mediterranean, with the enamel for the
Basse-Yutz flagon coming from Asia Minor (Caygill, 1999, 53: Megaw and Megaw,
2005, 29).The importation of this coral from the Mediterranean demonstrates the
range of the trade, even after the end of the Etruscan trade around 400 BC
(Milisauskas, 1978, 272).

Figure 16: Summary of maritime relationships between Britain and Gaul

(Cunliffe, 1984)

Page | 63
 In the later stages of the Iron Age there is an increasing level of complexity in the
trade relationships as well as changes in the location of trade. This occurs in the
increasing movement away from the Atlantic connections towards the continent
before the expansive trade networks are all brought together again as is demonstrated
by Cunliffe’s geographic and chronological mapping between Britain and Gaul (Fig. 16)
(Cunliffe, 1984, 18). As the social structure changes following the appearance of
oppida in the 2nd and 1st centuries BC (Wells, 2002, 343-344). The spread of the
oppida was extensive, stretching from southern England to Hungary (Cunliffe, 2011,
372: Haywood, 2001, 47). The appearance of the oppida and Caesar’s stay in one
(Cunliffe, 2011, 373) demonstrate a significant shift in the structure of the social
system in Iron Age Europe. The introduction of a coinage system in the 4th century by
mercenaries was refined and the use of cooper and silver alloys began at the same
time that the oppida grew (Wells, 2002, 344). By the final century BC, the oppida ere
centres of commerce, trade and luxury that was equal to that of the Mediterranean
markets, having adopted many of the systems of weighing, exchange, recording and
mass production from the Romans (Wells, 2002, 345). Intriguingly, the oppida coincide
with much of the region conventionally labelled as “Celtic” and they clearly have a key
role in trade, perhaps suggesting that the “Celtic” phenomenon is a trade
phenomenon rather than a specific “people”.

Buildings such as the oppida are not simply arenas for the exchange to occur, but
solid traceable indicators of the movement of styles and products. While portable
items such as the fibulae previously discussed can provide insight into the trade
networks, built structures cannot be moved at later dates, or imported. The agendas
for their construction must be taken up and reproduced by the people in that area.
Before the appearance of oppida across Europe, the establishment of hilltop defences
and hillforts are documented throughout the landscape. Across Europe there is a long
history of hill defences which date from the second millennium BC and reappear from
at least 800 BC in the Isles (Dyer, 2001, 10) and the late Bronze Age on the continent
(Collis, 1989, 37). The establishment of hillforts occurs in central Europe around the
period of Hallstatt C in the 7th and 6th centuries BC (Collis, 1989, 79-81). By contrast

Page | 64
this development occurs much later in the Isles, with larger hillforts being established
approximately 400 BC, although few appear to have been permanently occupied
(Dyer, 2001, 10-11). Hillforts and oppida are seen as cross European phenomena and
demonstrate the similarities between the built social systems and the expansive
network of ideas and communication that existed across the continent and isles.

Figure 17: Oppida Distribution in late Iron Age Europe (Cunliffe, 2011)

By contrast, the form of domestic housing does not indicate that a single form of
social space associated with a single cultural “people” was present. At the most basic
level of house shape the round-house is a localised, regional construction. There has
been a general consensus that houses from the European Iron Age followed the
tradition of being round in the Isles and rectangular on the continent (Harding, 2009,
14-15). This trend has been reinforced by many studies; with the exception of Iberia
however, following the observation by Hodson (1964) that the only other location
where the round-house style he has studied in the Woodbury Complex could be found
was in the north-west of the Iberian Peninsula. In the 1990s it became clear that the
prevalent view of rectangular houses as a Continental phenomenon, with the

Page | 65
exception of Iberia, was not as clear cut as it had been presented by Hodson (Harding,
2009, 15). The excavation of round-houses in Normandy changed the geographic
spread of the round-house to cover the Atlantic zone which included the Iberian
Peninsula, western coast of France and the Isles (Harding, 2009, 15).

Just like the study of the Celts, the study of housing shape in the European Iron Age
has issues associated with terminology and simplification. The term ‘round-house’ is
used in archaeology to identify any building with a circular plan in contrast to those
with a linear layout (Harding, 2009, 27). Harding’s (2009) reassessment of round-
houses highlights the variations that can occur in their construction from material to
number of entrance ways. He makes two significant conclusions. First, he identifies the
association between the Iron Age round-houses and the previous constructions of the
Neolithic stone circles and the Bronze Age round-houses (Harding, 2009, 144-145).
Secondly, Harding concludes that the round-houses of the European Iron Age are, as
has been the general consensus, an Atlantic zone phenomenon (Harding, 2009, 297).
The combination of these two conclusions highlights the localised traditions that
increasingly emerge from reassessment of the European Iron Age. The expanse of
trade networks and communication across the Continent and the Isles make it
plausible, if not probable, that the people living in different parts of the continent
would have been aware of the round-house and rectilinear styles elsewhere and so
could have adopted them in either direction. This is supported further by Caesar’s
report of commonality between the houses he observed in Gaul and those in Britain
(Harding, 2009, 15), but the prevalent impression is primarily of long term regional
stability. The round-house tradition, especially in the Isles has an ancestry back to the
Bronze Age at least 800 years before with the Deveral-Rimbury settlements (Harding,
2009, 125) suggesting strong regional continuities rather than significant movements
of societies with new kinds of spatial behaviour, even if in some areas there were
some population movements.

Page | 66
Conclusion

The field of archaeology is the largest in the study of the “Celtic” debate due to the
abundance and range of data that is available. However, it is also the one which has
suffered most from the assumptions and pre-determined conclusions that have arisen
from the use of old, out-dated terminology and the reliance in the 19th and much of
the 20th century on the Classical sources. The analysis of “Celtic” art styles has been
one of the longest in archaeology and the variation in labels, groupings and
conclusions have had significant implications for the further study of the
archaeological record. Emerging from the study of artefacts from Hallstatt and La
Tène, the study of the ornamentation in the European Iron Age has clung on to the
names of Hallstatt and La Tène and this has resulted in confusion between not only
the styles and time periods, but also in the divisions across Europe with labels such as
La Tène I, II and III as well as A, B, C, D (Champion et al., 1984, 298).

The analysis of distribution of ornamentation style as well as fibulae type highlights

some common trends across the archaeological record which are also present in the
shape used in housing construction. There is a significant variation in style which
produces a pattern of local use and manufacture. While the scale of this localisation
for the housing structure is greater than is observed in the ornamentation and fibulae,
the localisation that it represents occurs along a clear geographic divide. This trend of
localisation across artefact and settlement type is set against an expansive and
relatively recently established extensive Iron Age trade network linking down to the
Mediterranean. Portions of that network, such as along the Atlantic coast, had been in
existence at least since the Bronze Age. The nature of the trade network and its extent
in the Iron Age can be seen through the Greek, Etruscan and Roman materials which
have been found along the trade paths. This extensive communication network is in
stark contrast to the localised patterns emerging from the artefacts, although there
are some indications of this network in the spread of the dragon-pairs design on
swords and scabbards (Harding, 2007, 102). The Nauheim type fibulae (Fig. 14) is also
a representative of this pan-European trade and interaction. However it is also a

Page | 67
marker of the changes in social structure that may have resulted from the interaction
with Mediterranean societies. While the shift in the 2nd and 1st centuries BC towards
the use of oppida as trading nodes was accompanied by the increasingly wide spread
use of coins (Wells, 2002, 344) as well as increasing Roman interaction and dominance
over regions of Europe, there is no evidence to suggest that the change was anything
more than an increasing development of communication, trade and social links
between the different groups and existing peoples of Europe. A single integrated
“people” is not apparent.

Page | 68
Chapter Six - Integration and Conclusions

The concept of identity as presented in Chapter 1 outlined two ways in which it

could be approached. The first was the view of identity as seen by an external
observer and by those whose identity is under discussion. The second is an analysis of
the components of identity using Renfrew’s division of genetics, language and cultural
diversity (Renfrew, 1993, 19). Under the first division, the external view is well
documented through the accounts of scholars and military figures from the Classical
world. However, the concept of an identity as seen by those to whom it refers is much
more complex in regards to the “Celtic” debate , as the “Celts” did not leave written
records and the Classical view has therefore become dominant. The alternative path,
through the general trends and patterns which have emerged from the three fields of
genetics, language and archaeology, allows another understanding of this complexity
to be achieved. The conclusions from genetics, are in dramatic contrast to the
traditional assumptions about the presumed “Celtic” people. First, the majority of the
genetic story is much older than the Iron Age, dating back to the Bronze Age and even
into the Neolithic over 2000 to 4000 years earlier. The continuous nature of this DNA
demonstrates that there had been no significant migration of people into the Isles
during the Iron Age.. This removes the genetics from the discussion of any “Celtic”
identity as the genetic work which has been done, predominantly focusing on the Isles,
highlights the age and the consistency of the DNA, thus affirming that there is no
“Celtic” ethnicity or genetic grouping. The second trend that can be identified in the
genetic field is the existence of two distinct stories for the paternal and maternal DNA.
The maternal DNA, as shown by Sykes, is very old in the Isles. While the routes that
were taken by the genetic Clans Jasmine and Tara are via the Mediterranean or
overland across central Europe, the age of the DNA in the Isles indicates that this
Continental link is much older, arising from the introduction of agriculture to the
European continent. The male DNA, by contrast, highlights the Atlantic connection.
The final genetic migration that has been dated using haplogroup R1b3f (McEvoy et
al., 2004) corresponds with the decline of Bronze Age Atlantic trading routes as

Page | 69
outlined by Cunliffe (2010: 2011: 2001). The variation that is evident in the genetics
highlights both the geographic and temporal differences between the sexes. The
maternal DNA is much older with a cross-European pattern, while the paternal DNA is
younger, potentially with the last movement as recent as the end of the Bronze Age,
with an Atlantic coastal affiliation. While the exact reasons for the movement of the
DNA may not be known, it is likely that the female movement corresponds to the
expansion of agriculture while the male trend follows the Neolithic and Bronze Age
trade and communication network of the Atlantic. The divergence between the
movements of the two DNA stories indicates that the movement was not a mass
migration by a combined social group. Despite some analytic and methodological
issues, the genetic field as a whole has identified two trends which contradict the
concept of a “Celtic” ethnicity or people. The correlation of the genetics in the Isles on
the paternal side to the Iberian Peninsula through the use of Basque as a comparison
further strengthens the evidence against “Celtic” genes. The Basque region, with a
language and culture which are seen as unique and separate from any presumed
“Celtic” culture, acts as a marker against which any potential association across
“Celtic” regions can be considered. The result is a decidedly conclusive divide between
the DNA of the Isles and the Continent. Both the age of the DNA in the Isles and the
lack of maternal and paternal DNA correspondence reveal that the events prior to the
European Iron Age did not lead towards the emergence of a single social system or
cultural group of related people. The genetics reveals that the any shift in identity
from the European Bronze Age to the Iron Age was cultural or social.

Linguistics is the first scholarly field to have established a “Celtic” label through the
linguistic grouping of the “Celtic” languages by Pezron and Lhuyd. The western
European trajectory of the languages as revealed through inscriptions and place-name
analysis would appear at first to correspond to the western trend of the male genetic
story. However, the genetic correspondence with the Basque DNA reveals a link
between two populations with varied language systems. This western trend in the
fields of genetics and linguistics does not represent any correlation that indicates a
group or ‘people’. The linguistic geography of the European Iron Age is divided into a

Page | 70
complex network through assigned language names, with the oldest dated being
Tartessian and Leptonic, and through the suffix distributions in the studies by Sims-
Williams. While the relationship of the P-Celtic and Q-Celtic labels to the actual spread
of the languages is an old model, the wave model discussed by Renfrew argues for the
development of different language branches in centres which have themselves
diverged from the older Indo-European languages. The continual branching and
development of languages created a complex network of dialects. As language is one
of the criteria for establishing identity in the Renfrew scheme, the variation in
language represents a variation in identity. This is both from the perspective of the
observers, as is clear from the tribal and language groups in Classical works, but also
from the standpoint of those to whom that identity relates. The divided and regional
language system would have been a contributing factor to the geographic variation in
the concept of identity as it applied to the people, groups and ‘tribes’ of the European
Iron Age. While the division of the linguistic geography is not in itself a new
proposition in the study of “Celtic” languages, the degree to which these variations,
dialects or languages could be inter-comprehensible has often been overestimated,
presenting an image of a much more universal language system than is suggested by
the linguistic evidence.

In opposition to the western association that has emerged from the genetic and
linguistic reassessment, the material culture of the European Iron Age is more
widespread. The treatment of artefacts still follows the 19 th century identification of
an expansion from a central European location identified in the sites of Hallstatt and
La Tène. The divergent evidence of continuity in the round-house tradition from the
Bronze Age in the Atlantic zone and the Isles as well as the development of complex
trade networks expanding outwards from the Mediterranean suggest that this
persistent notion of a central origin for the spread of a “Celtic” people as marked by a
material culture is still influenced by the terminology of the presumed “Celts” with a
single culture and identity. The complexity of the archaeological record, and the
extensive interconnectivity which it reflects, do not support the conclusion of a
“Celtic” population group. The regional distribution patterns that can be identified in

Page | 71
the location of ornamentation designs, fibulae types and the shape of domestic
structures each represent the spread of a particular design or physical representation
of the identity of that region. The existence of rare pan-European phenomena,
including the dragon-pairs on swords and scabbards, combined with the finds of Greek
amphorae are an indication of some extensive movement of ideas and communication
throughout the European Iron Age. Further complexity in social systems and trade
relationships is embodied in the construction of the oppida and the coinage system
and also relates to the increasing power and influence of the Mediterranean cultural
world. Despite the expanding universality in the social structure of the European Iron
Age, the growth in the number of ‘tribal’ names or group labels which appear in the
Classical sources suggests that the Romans were aware of variations and differences in
the identities of these regions. The evidence requires that the people of Iron Age
Europe be approached in a new manner with the understanding that the social system
and network in the Iron Age is much more complex and integrated but also more
segregated than has previously been acknowledged. Just as the term “Celtic” has been
applied to the languages of this time period in Europe to identify similarities between
separate, regional languages, so likewise the term “Celtic” has actually been used to
subsume the various similarities and interconnections between the diverse cultural
populations of the European Iron Age.

It is through these developments at the end of the pre-Roman period that the
tribalisation theory introduced to the “Celtic” debate by Peter Wells (2004), which
identifies a change in social structure as the impact of interaction with more complex
societies, becomes relevant. In the European Iron Age, the development of hillforts
and then oppida, as well as the dramatic shift in trade networks and production from
localised to pan-European is characteristic of this change in social structure. The
introduction of new products including coins also significantly changed the way that
exchange was organised and the minting of coins in specific regions acted as a marker
of identity by reinforcing the separation of territory. This distribution of coinage as an
indicator of social divides can clearly be seen in Gaul where there are distinctive
distributions of coinage systems on either side of the Seine in the territory of the

Page | 72
Belgic and Armorican Gauls (Nash, 1984, 105). As the interactions and trade networks
expanded and the influence of the Mediterranean societies on the Iron Age European
tribes increased, the development of more complex art, production and social
structure followed. The events of the European Iron Age correspond to the features
both of the tribalisation process and the trade interconnection process but not to the
expansion of a “people”.

It is necessary to end the reliance on and use of the term “Celtic” as a pan-European
label for the people of the European Iron Age. Furthermore, the clear association of
particular styles and specific regions suggests that the use of artefacts to express the
individuality and identity of these ‘tribes’ is even more important and extensive than
previously understood. The continued use of the round-house in the Atlantic regions in
stark contrast to the rectangular style in central and eastern Europe shows distinct and
sustained differences between the behaviour of regional populations even though the
trade networks indicate that exposure to other designs and approaches was occurring.
Wells’ tribalisation process provides a framework to understand the changes in the
social structure and trading networks of the European Iron Age. However, the
interaction with the complex societies of the Mediterranean and the combining of
previous trade networks that resulted from this had one major implication for the
future terminology concerning Iron Age identity; the foundation for the out-dated
term “Celts”. This term is a construction of various grouped networks in the Iron Age;
the grouped languages, the grouped ornamentation, the grouped fibulae and oppida.
Within each of these there was variation and geographic localisation, but these
differences have all been put together to construct a “Celtic” culture or people.

However what the label “Celtic” actually encompasses is not a single culture, people
or language. Rather it comprises, and is the label for, of the development of the pan-
European trade networks and interactions in the Iron Age. The adoption of Greek
ornamentation and equipment designs by the people north of the Mediterranean, the

Page | 73
establishment of oppida and coinage systems as well as the use of Classical texts for
local inscriptions are all a product of the interaction between different social groups,
or tribes, and the reaction to the influence from the more complex societies in the
Mediterranean. The rising complexity of the ornamentation and manufacture styles
which contributed to the diversity seen in the division of the La Tène period into sub-
phases are the result of increasingly localised expressions of identity prior to the
successful intervention by the Romans in the late 1st millennium AD.

Conclusions

The implications for this reassessment of the term “Celtic” are considerable. In
Chapter 2, one of the most sensitive areas for any reassessment of the people of Iron
Age Europe and the associated labels was identified as the political connections and
the modern identity of “Celtic” people in several regions of Europe and those who
have migrated to different regions of the world. What this reassessment of the “Celts”
has revealed is that there are no links between the ancient and modern populations of
those regions labelled as “Celtic” today. This is not to say that there is no connection
between the modern inhabitants and those in the Iron Age. There has been a
preservation of the languages and a revival of the ornamentation designs found in that
region, although the preserved ornamentation is treated as a singular body rather
than representative of different regions and groups. However the terminology and
generalisation that have reduced the complex social divides of the European Iron Age
to the incorrect form of a single group minimises the preservation of any cultural
aspect that is not incorporated into the generalised or politicised revival of the 19th
century view of European Iron Age culture. Furthermore, the wide variety of dialects
and designs that existed in the Iron Age cannot be presumed to have contributed
towards a single concept of identity across Europe at that time. Any such connections
are unclear and fragile.

Page | 74
The genetic analysis has demonstrated that while there is a continuous genetic history
in the Isles, this is not a specifically “Celtic” genetic history; rather it is a Neolithic or
Bronze Age link. There is no concept of “Celtic” in a genetic sense and therefore no
“Celtic” people or ethnicity. The social and cultural identity of the modern “Celts” is
constructed through the combination of language and a generalised idea of pan-
European styles. While there are small links between the past and the present, the
concept of identity is understood and applied differently today than in the Iron Age,
with nationalism changing the way that the factors of identity are treated. The holistic
identity approach of the modern “Celts” is in stark contrast to the regional and ‘tribal’
identity that the archaeological record has provided for the different groups in the
European Iron Age. The continuation of the genetic profile in the Isles make the
modern “Celts” the guardians of an ancient DNA which long predates the Iron Age,
and of languages which have developed from the old Indo-European “Celtic” language
branch. The significance of the modern “Celtic” identity for the development of
nationalism and preservation of those “Celtic” languages and the enhancement of an
interest in, and awareness of, the history of Europe, especially in regard to the Iron
Age, should not be taken lightly. What this does not alter, however, is the nature of
the past in the 1st millennium AD where the term “Celtic” actually refers to the
extended communication and trade network that developed in the Iron Age, not to
any kind of people or more vaguely to a trans-European culture.

Page | 75
References

ALEXANDER, J. 1965. The Spectacle Fibulae of Southern Europe. American Journal of

Archaeology, 69, 7-23.
BIAGINI, E. F. 2007. British Democracy and Irish Nationalism 1876-1906, Cambridge,
Cambridge University Press.
BLEČIĆ KAVUR, M. 2009. Universal and Original. Transformation of style in the North-Adriatic
Region. In: TIEFENGRABER, G., KAVUR, B. & GASPARI, A. (eds.) Keltske Študije II:
Studies in Celtic Archaeology: Papers in honour of Mitja Guštin. Montagnac: Éditions
Monique Mergoil.
BOYCE, D. G. 1995. Nationalism in Ireland, London, Routledge.
CAYGILL, M. 1999. The British Museum: A-Z companion, London, The British Museum Press.
CHAMPION, T., GAMBLE, C., SHENNAN, S. & WHITTLE, A. 1984. Prehistoric Europe, London,
Academic Press.
CHAPMAN, M. 1993. The Celts: The Construction of a Myth, Basingstoke, The Macmillan Press
Ltd.
CHILDE, V. G. 1951. Man Makes Himself, New York, The New American Library.
COLLIS, J. 1989. The European Iron Age, London, B.T. Batsford Ltd.
COLLIS, J. 2011. The Celts: Origins, Myths & Inventions, Stroud, The History Press.
COONEY, G. 1995. Theory and Practice in Irish Archaeology. In: UCKO, P. J. (ed.) Theory in
Archaeology: A world perspective. London: Routledge.
CUNLIFFE, B. 1984. Relations between Britain and Gaul in the First Century B.C. and Early First
Century A.D. In: THOMPSON, S. M. F. H. (ed.) Cross-Channel Trade between Gaul and
Britain in the Pre-Roman Iron Age. London: Thames and Hudson.
CUNLIFFE, B. 1997. The Ancient Celts, Oxford, Oxford Universtiy Press.
CUNLIFFE, B. 2008. A Race Apart: Insularity and Connectivity. Europa Day Conference.
University of Oxford, Oxford: The Prehistoric Society.
CUNLIFFE, B. 2010. Celticization from the West: The Contribution of Archaeology. In:
CUNLIFFE, B. & KOCH, J. T. (eds.) Celtic from the West: Alternative Perspectives from
Archaeology, Genetics, Language and Literature. Oxford: Oxbow Books.
CUNLIFFE, B. 2011. Europe between the Oceans 9000 BC - AD 1000, New Haven, Yale
University Press.
CUNLIFFE, B. W. 2001. Facing the Ocean: the Atlantic and its people, 8000 BC - AD 1500,
Oxford, Oxford University Press.
DYER, J. 2001. Discovering Prehistoric England, Princes Risborough, Shire Publications.
FERGUSON, R. B. & WHITEHEAD, N. L. (eds.) 2000. War in the Tribal Zone: Expanding States
and Indigenous Warfare, Santa Fe: School of American Research Press.
FORSTER, P. & TOTH, A. 2003. Toward a Phylogenetic Chronology of Ancient Gaulish, Celtic,
and Indo-European. Proceedings of the National Academy of Sciences, 100, 9079-9084.
GARROW, D. & GOSDEN, C. 2012. Technologies of Enchantment? Exploring Celtic Art: 400 BC to
AD 100, Oxford, Oxford University Press.
GOLDSTEIN, D. B., LINARES, A. R., CAVALLI-SFORZA, L. L. & FELDMAN, M. W. 1995. An
Evaluation of Genetic Distances for Use With Microsatallite Loci. Genetics Society of
America, 139, 463-471.
GREEN, M. 1996. Celtic Art, London, Orion Publishing Group.
HAAS, J. 1990. Warfare and tribalization in the prehistoric Southwest. The Anthropology of
War. Cambridge: Cambridge University Press.
HARDING, D. W. 2007. The Archaeology of Celtic Art, Abingdon, Routledge.
HARDING, D. W. 2009. The Iron Age Round-house: Later Prehistoric Building in Britain and
Beyond, Oxford, Oxford University Press.
HATTATT, R. 1985. Iron Age and Roman Brooches, Oxford, Oxbow Books.

Page | 76
HAYWOOD, J. 2001. The historical atlas of the Celtic world, London, Thames & Hudson.
HODSON, F. 1964. Cultural grouping within the pre-Roman British Iron Age. Proceedings of the
Prehistoric Society, 30, 99-110.
JAMES, S. 1999. The Atlantic Celts: ancient people or modern invention?, Madison, The
University of Wisconsin Press.
JONES, O. 1856. The Grammar of Ornament, London, Day and Son.
JOPE, E. M. 1971. The Witham Shield. The British Museum Quarterly, 35, 61-69.
KILBRIDE-JONES, H. E. 1980. Zoomorphic Penannular Brooches, London, Thames and Hudson.
KOCH, J. T. 2010. Paradigm Shift? Interpreting Tartessian as Celtic. In: CUNLIFFE, B. & KOCH, J.
T. (eds.) Celtic from the West: Alternative Perspectives from Archaeology, Genetics,
Language and Literature. Oxford: Oxbow Books.
LAING, L. 1979. Celtic Britain, London, Routledge & Kegan Paul.
LAING, L. & LAING, J. 1995. Celtic Britain and Ireland: Art and Society, London, The Herbert
Press.
LORENZ, H. Regional Organisation in Western Early La Tene province. In: CHAMPION, T. C. &
MEGAW, J. V. S. (eds.) Settlement and society : aspects of West European prehistory in
the first millennium B.C. Leicester: Leicester University Press.
MCEVOY, B., RICHARDS, M., FORSTER, P. & BRADLEY, D. G. 2004. The Longue Durée of Genetic
Ancestry: Multiple Genetic Marker Systems and Celtic Origins on the Atlantic Facade
of Europe. The American Journal of Human Genetics, 75, 693-702.
MEGAW, R. & MEGAW, V. 2005. Early Celtic Art in Britain and Ireland, Princes Risborough,
Shire Archaeology.
MILISAUSKAS, S. 1978. European Prehistory, New York, Academic Press, Inc.
MÜLLER, R. & STEUER, H. (eds.) 1994. Fibel und Fibltracht, Berlin: Walter de Gruyter GmbH.
NASH, D. 1984. The Basis of Contact between Britan and Gaul in the Late Pre-Roman Iron Age.
In: MACREADY, S. & THOMPSON, F. H. (eds.) Cross-Channel Trade between Gaul and
Britain in the Pre-Roman Iron Age. London: Thames and Hudson.
OPPENHEIMER, S. 2007. The Origins of the British, London, Robinson.
PARSONS, D. N. 2010. Tracking the Course of the Savage Tongue: Place-names and Linguistic
Diffusion in Early Britain. In: CUNLIFFE, B. & KOCH, J. T. (eds.) Celtic from the West:
Alternative Perspectives from Archaeology, Genetics, Language and Literature. Oxford:
Oxbow Books.
RENFREW, C. 1979. Problems in European Prehistory, Edinburgh, R. & R. Clark Ltd.
RENFREW, C. 1993. The Roots of Ethnicity: Archaeology, Genetics and the Origins of Europe,
Rome, Unione Internazionale degli Istituti di Archeologia Storia e Storia dell'Arte in
Roma.
RENFREW, C. 1998. Archaeology & Language: The Puzzle of Indo-European Origins, London,
Pimlico.
ROES, A. 1965. Continental Quoit-Brooches. The Antiquaries Journal, 45, 18-21.
ROYMANS, N. 2004. Ethnic Identity and Imperial Power: The Batavians in the Early Roman
Empire, Amsterdam, Amsterdam University Press.
RUSSELL, P. 1995. An Introduction to the Celtic Languages, New York, Longman.
STEAD, I. M. 1984. Some Notes on Imported Metalwork in Iron Age Britain. In: MACREADY, S.
& THOMPSON, F. H. (eds.) Cross-Channel Trade between Gaul and Britain in the Pre-
Roman Iron Age. London: Thames and Hudson.
SYKES, B. 2006. Blood of the Isles: Exploring the genetic roots of our tribal history, London,
Bantam Press.
WELLS, P. S. 2001. The Barbarians Speak: How the conquered peoples shaped Roman Europe,
Princeton, Princeton University Press.
WELLS, P. S. 2002. The Iron Age. In: MILISAUSKAS, S. (ed.) European Prehistory: A Survey. New
York: Kluwer Academic/ Plenum Publishers.

Page | 77
WELLS, P. S. 2004. Beyond Celts, Germans and Scythians: Archaeology and Identity in Iron Age
Europe, London, Gerald Duckworth & Co.
WELLS, P. S. 2012. How Ancient Europeans Saw the World: Vision, Patterns and the Shaping of
the Mind in Prehistoric Times, Princeton, Princeton University Press.
WILSON, J. F., WEISS, D. A., THOMAS, M. G., BRADMAN, N., GOLDSTEIN, D. B. & RICHARDS, M.
2001. Genetic evidence for different male and female roles during cultural transitions
in the British Isles. Proceedings of the National Academy of Sciences of the United
States of America, 98, 5078-5083.

Page | 78