1. Introduction

Textbooks on European prehistory published during the last seventy-five years confirm that ever since the first theories were formulated at the beginning of the nineteenth century, the Celts have played a crucial role in studies of European protohistory. The Celts have even been accused of taking over Europe during the Iron Age, through the cultures of Hallstatt and La Tène. Any conclusions drawn by European historiography about the Celts have been based on the image conjured up by Classical sources and the archaeological and cultural evidence collected at La Tène. It was only during the mid 1980s that scholars began to debate this issue and started to break away from the conventional established image of the Celts (Collis 2003; Ruiz Zapatero 1993, 2001). Traditionally, theories surrounding the Celts practically excluded the Iberian Peninsula from the equation, since archaeological finds from the Hallstatt and La Tène cultures were rare in this area, and did not provide enough evidence for a cultural scenario comparable to that of Central Europe. However, this is not the only reason for such an oversight: north of the Pyrenees Spanish archaeological studies were practically unknown, and Spain was manifestly under-represented in forums and European publications on the subject. This combination of factors explains why European Celtic studies either exclude the Iberian Celts, or barely acknowledge their existence, let alone display any depth of knowledge of these people.

Thus, it is now more important than ever to study and publicize information about the Celts in Iberia. Only in the last decade, and especially during the last few years, have Iberian Celts been included in introductory texts about the Celts, for example in Cunliffe's excellent book The Ancient Celts (1997: 133-144), and in his illuminating and highly recommended brief overview entitled The Celts: A Very Short Introduction (2003). Other works include Kruta's ambitious textbook Les Celtes: Histoire et Dictionnaire (2000: 316-333) or Collis' most recent book The Celts: Origins, Myths and Inventions (2003: 122-123, 177-180), a lucid study that goes beyond a simple appraisal of Celtic historiography. Comprehensive works such as The Historical Atlas of the Celtic World (Haywood 2001: 44-45) place equal emphasis on the Iberian Celts and their cousins from other areas. This increased scholarly interest in the Celts of Iberia is also observable most recently in Raftery's Atlas of the Celts (2001: 48-49, 113).

However, unfortunately the situation continues in general as before: many studies persist in ignoring the Iberian Celts, despite the publication, albeit limited, of Spanish works in English, French and German (Almagro-Gorbea 1991a, 1994, 1995a, 1995b, 1995c; Almagro-Gorbea and Lorrio 1991, 1993; Álvarez-Sanchís 2000; Arenas 2001-2002; Burillo 1991; de Hoz 1992a, 1992b; Lorrio 1997b, 1999; Marco 1998; Ruiz Zapatero 1996; Sánchez Moreno 2001; Tovar 1986) and the international conferences that have focused specifically on the subject of the Iberian Celts (Almagro-Gorbea and Ruiz Zapatero 1993; Almagro-Gorbea et al. 2001; Cousquer et al. 1999).

This essay has two key objectives: Firstly, to provide a general overview of the latest research on the Iberian Celts, especially in the last fifteen years, and thus serve as a point of reference for other monograph-length studies. Secondly, to present a frank appraisal of research that reveals the polarization that is taking place in a few cases in Spain between pro- and anti-Celtic positions, which have in some cases a political dimension (Díaz Santana 2002; López 2003; Ruiz Zapatero 1996, 2003). This is comparable to the state of research in other European countries.

First and foremost, we have included a brief historiography of Celtic research carried out in Spain (Ruiz Zapatero 1993, 1995-97, 2003), which we consider absolutely fundamental when approaching contemporary research and seeking answers to a number of questions (Collis 2003: 11). Referring to past history and the development of the ideas we are working with is crucial when undertaking critical research. Secondly, we will examine the image of the Celts between the late third and first centuries BC according to Classical sources, historical linguistics, and archaeology in order to assess what is defined as Celtic and what criteria are applied. We believe that it is extremely important to explore the concept of the Celts from the perspective of ancient history, palaeolinguistics and archaeology, and also to give equal weight to independent theories, which must then be compared with the traditional sources. Such a task is by no means straightforward. We must proceed with extreme caution, always bearing in mind that the objects of comparison are fundamentally different. Thirdly, we will tackle the complex issue of the origins of the Iberian Peninsular Celts by studying the origins of the Celtiberians as a possible key to unlocking the process of Celticization of the rest of the Peninsula. Fourthly, we will define the different Peninsular archaeological groups that can be classified as "Celtic", and briefly analyze historical sequences along with the problems of how to identify their Celticity from an archaeological perspective. Fifthly, we will introduce and examine the model of the Celticization of Iberia, and we will lay particular emphasis on the need to proceed in this field with a greater critical awareness in order to produce interpretative models of the archaeological information. Sixthly, we will explain why there should be joint multidisciplinary research into the Celtic issue, in which disciplines such as ancient history, palaeolinguistics, archaeology, history of art, anthropology, population genetics, history of religion or ethnology, mythology and folklore studies should all be taken into consideration together. We would, however, like to highlight the very real danger of what John Collis has termed the continuous circular argument (2003: 12). In other words, each discipline must assess its own data without entrusting it to the interpretation and use made of it in other subject areas. Finally, we will attempt to demonstrate the importance of the Iberian Celts in the context of Iron Age Europe, and we will even suggest that the unique Iberian vantage point can offer new and interesting approaches to research into the Celts north of the Pyrenees.

As Cunliffe suggests, the Celtic debate reached its zenith at the dawn of the twenty-first century (2003: 6). A recent article by Sims-Williams summarizes the wide range of opinions, which run the gamut from celtomania to celtoscepticism (1998). The debate is heated and at times even inappropriate, but all sides would agree that the complexity of this issue is intriguing. There are many different approaches with different types of data to be analyzed. As Cunliffe has summed it up so well, "it is a rich mix of ingredients, but what we cannot do is to fling them all into one pot and expect a perfectly formed Celt to emerge. Each of the different categories of evidence has to be considered within its own critical parameters to separate fact from wishful thinking and to distil out what it has to offer to the debate. Whether, in this process of deconstruction, Celts and Celticism will vanish altogether remains to be seen" (2003:7). This has been our aim in this article (Fig. 1).

Figure 1. The Iberian Peninsula: main geographical features.

2. The history of research

The first studies on the Iberian Celts date as far back as the scholarly tradition of the fifteenth to eighteenth centuries, which focused on identifying cities mentioned in Classical sources, including Numantia. However, it was not until the first quarter of the nineteenth century that W. von Humbolt (1821) was able to identify a few Celtic toponyms in the Iberian Peninsula that came from literary sources. During the second half of the century, F. Fita (1878) and E. Hübner (1893) greatly added to the collection of documents by focusing on names based on epigraphic evidence. At the same time, the first non-coin finds of Celtiberian epigraphs written in the Iberian alphabet were made. Despite various attempts from the sixteenth century onward to decipher these texts using numismatic documentation, this task was not achieved until 1925, by M. Gómez Moreno. The late nineteenth century saw H. d'Arbois de Jubainville (1893-94) publish several seminal studies on the subject. He was the main exponent of the Ligurian theory, according to which this Indo-European race of people had colonised the West before the Celts arrived. D'Arbois de Jubainville began an assessment of Peninsular Celtic elements primarily based on Classical literary sources and onomastic documentation.

Alongside these studies, important archaeological activity began to take place in Peninsular Celtic areas, in particular during the first few decades of the twentieth century. Excavation was begun in locations such as Numantia, particularly the Iron Age burial grounds located in the Upper Tagus, Upper Jalón and Upper Douro Rivers that, according to written sources, were the nucleus of Celtiberia (Lorrio 1997a: 16-18). Archaeological work began in the city of Numantia in 1803, and was taken up again at the beginning of the second half of the nineteenth century. However, it was the excavation work funded by Kaiser Wilhelm II and carried out by A. Schulten between 1905 and 1912 that revealed the city and the Roman camps that formed Scipio's border to the international community. The results of these excavations were published in four volumes between 1914 and 1931 (Schulten 1914, 1927, 1929, 1931). The first of these volumes contained the first ever analysis of Celtiberia, in which Schulten (1914: 7-11, 281-290) included a collection of literary sources on the Celtiberians. He used literary sources to create a highly personal reconstruction of the process of ethnogenesis of the Celtiberians, which would become the basis for subsequent studies by Bosch Gimpera. According to Schulten (1914: 98-99; 1920: 108-111), the Celts probably gained eventual control of the whole of the Meseta region. His evidence to support this theory was the geographic distribution of the toponyms ending in -briga, which were later absorbed into Iberian cultures. Therefore, the Celtiberians could have been Iberians who settled in the land of the Celts, which would directly contradict the traditional theories suggesting that the Celtiberian people developed as the result of a Celtic invasion of the land of the Iberians. A sign that these two populations intermingled can be seen in the presence of Celtic elements among Celtiberians, such as the names of Celtiberian nobility (Schulten 1914: 246).

From 1915 onwards, P. Bosch Gimpera wrote a series of studies broaching the issue of the Celts in the Iberian Peninsula. He used Schulten's invasion theories - which were largely based on Classical texts and linguistic data - as his basis, and attemped to shore them up using archaeological evidence. He linked the Celts with the recently discovered Urnfield Culture in the northeast of the Peninsula and suggested that various invasions took place. Bosch Gimpera was in some respects the "creator" of the Hispanic Celts, at least from an archaeological point of view (Fernández-Posse 1998: 40). In his work Two Celtic Waves in Spain (1943), Bosch Gimpera distinguished two waves of invasions. He placed the first around 900 BC and linked it with the Urnfield Culture from the southern Germany, which according to him penetrated the Peninsula via Catalonia, where a process of naturalization of the invaders took place up until the mid seventh century BC. He defined the second wave as being made up of groups of Hallstatt peoples from the Lower and Middle Rhine, who arrived in the Peninsula in several stages between 650 and 570 BC via the western passes of the Pyrenees, which is why their influence was fully felt in the Meseta region. The Belgae were the last Celtic group to arrive in the Peninsula around 570 BC, pre-dating the first La Tène culture objects, which were the result of trade relations. The Belgae settled in the northern Meseta and in the Ebro River Valley, bringing with them elements that would give rise to the so-called "Post-Hallstatt Culture".

From the 1930s onwards, M. Almagro Basch wrote several articles (1935, 1939, 1947-48) in which he defended his conclusions, which were in direct conflict with Bosch Gimpera's theories. He developed his ideas more fully in his publication La invasión céltica en España (1952), which was included in R. Menéndez Pidal's Historia de España. In this work, Almagro restated his theories on the Indo-Europeanization of the Iberian Peninsula, and suggested a single Celtic invasion that started ca. 800 BC and took place slowly and gradually, against the background of the Celtic Urnfield culture.

From the 1940s onward, archaeological research into the Celtic question lapsed into a state of relative torpor (Díaz-Andreu 1993), which added to the difficulty in relating Hispanic material to that found in the rest of Europe, and led to a marked decline in new theories and interpretations. Several German archaeologists provided the impetus for a change in direction. The conclusions drawn from their work identified the historical Celts with the La Tène culture, in line with the hypotheses of Central European Celtic research (cf. Ruiz Zapatero 1993: 49). This new research avenue can be clearly seen in the work of E. Sangmeister (1960), in which he tried to clarify the value of the Celtic contribution in the Iberian Peninsula. For Sangmeister, Hallstatt D represented a new cultural phase in southeastern Germany and northeastern France, bringing together surviving elements of the Urnfield Culture and reviving elements of the Tumulus Culture and other typically Hallstatt elements. He could not determine precisely which of these components would have brought with it the Celtic language, which, according to Sangmeister, was the only and definitive argument that could justify referring to Celts in the Iberian Peninsula.

Evidence provided by the field of linguistics supports the idea of several invasions. Following Gómez Moreno's deciphering of the Iberian alphabet (1925), Caro Baroja (1943) identified Celtic elements in certain inscriptions written in Iberian that came from Celtiberia. The geographical zone of Celtiberia was defined according to where coins having certain word-endings were found. However, it was Tovar who in 1946 detected several fundamental traits of the Celtiberian language, which was thus finally classified as belonging to the Celtic languages. This pioneering work was followed by a number of important publications: Lejeune (1955), Schmoll (1959) and Tovar himself (1948, 1949, 1950, 1955-56, 1961, among other works). For Tovar (1971: 18-19) there were two Indo-European linguistic strata: one was 'pre-Celtic', as attested in Lusitanian, a "more archaic [language] in some elements than Celtic, which could represent the remnants, although in an evolved state, of the early Indo-European invasions in the West". He defined the other stratum, Celtiberian, as an "archaic Celtic dialect" (Tovar 1971: 18 ff.). For Untermann (1961), on the other hand, there was only one Indo-European invasion of the Iberian Peninsula of Celtic origin responsible for the dialectal differences found in this region. In the early 1970s, the presence of a Celtic language in the Iberian Peninsula was definitively confirmed with the discovery of the first Botorrita Bronze, considered one of the most important texts in the whole of the Celtic continental world (Beltrán and Tovar 1982; Eska 1989; de Hoz and Michelena 1974; Meid 1993, 1994: 7-28).

From the mid 1980s on, research on the Celts in the Iberian Peninsula was revitalised by renewed efforts that approached the issue from various disciplines. Prior to this, research had almost entirely been confined to linguistic studies. Greater knowledge and understanding of the Late Bronze and Iron Age material cultural in the central and western regions of the Iberian Peninsula have allowed us to move forward in the interpretation of the origins of some of the main peoples thought to be Celtic by Graeco-Latin authors, against the background of the Celticization of the Iberian Peninsula. Research had practically ground to a halt following the seminal theses by Bosch Gimpera and Almagro Basch, and apart from laudable attempts by Sangmeister (1960) and Schüle (1969) this question had not been revisited. Lenerz-de Wilde (1981) and Stary (1982) took up the reins of interpretation and followed the Central European tradition of Celtic research already present in Sangmeister's work. They tried to demonstrate that Peninsular Celts were La Tène Celts, despite the fact that the distribution of discoveries relating to La Tène culture in the Iberian Peninsula did not coincide with the Indo-European linguistic areas. Subsequently, Lenerz-de Wilde (1991, 1995) proposed her invasionist theories based on her belief that there was evidence from the sixth century BC onward for contact between the Iberian Peninsula and the Hallstatt Culture. Later on, Celtic groups arrived in the Peninsula via routes that had already been established through previous trade contact via the Ebro River and through to the Meseta plateau. This would certainly explain the discoveries dating from the fifth century BC in the Meseta that demonstrate the links between this region and the Hallstatt and La Tène cultures in Central Europe. These finds include certain types of brooches, La Tène swords and their characteristic sheaths, objects that appear during the fourth century BC, and a rare belt hook, which was doubtless made in the La Tène culture area (Lenerz-de Wilde 1995: 538). However, Kalb (1979), in his study on the Celts in Portugal, did not uphold the idea that La Tène style discoveries made in Portugal constituted archaeological proof of a Celtic culture, and stated that this "term is not suitable to describe unequivocally an archaeological context" (Kalb 1993: 155). As Untermann (1995a: 20) pointed out, even the discovery of Latenian objects cannot be linked to the linguistic Celticization of the Iberian Penisula, since "Linguistics demands a considerably earlier date for the first settlement of speakers of Celtic languages".¹

It was not until M. Almagro-Gorbea began to develop his research from the mid 1980s onward that the issue was given a new dimension (Almagro-Gorbea 1986-87, 1987, 1991a, 1991b, 1992, 1993; Almagro-Gorbea and Lorrio 1987). In the first place, it is difficult to maintain that the origin of the Peninsular Celts can be linked to the Urnfield culture. Research since the 1970s has confirmed that the latter culture originated from the other side of the Pyrenees, and also that this people settled throughout the northeast of the Peninsula, a zone that does not coincide with that occupied by the historic Celts nor with Celtic linguistic evidence (Ruiz Zapatero 1985). Furthermore, there was a smooth uninterrupted transition from the Urnfield culture in the northeast into the Iberian culture, whose epigraphic findings correspond to a language - Iberian - which is not Celtic and indeed not even Indo-European (de Hoz 1993b; Untermann 1990; etc.).

Figure 2. Origin of the Celtiberians and process of Celticization in the Iberian Peninsula (After Burillo 1998, following Almagro-Gorbea 1993 data).

Therefore, Almagro-Gorbea (1987, 1992, 1993, 1994) sought a new interpretation that would determine the origin of the Celts documented by written sources, by tracing their material culture, their socio-economic structure and their ideology in the Iberian Peninsula as interrelated parts of the same cultural system (Fig. 2). He sought the roots of the Peninsular Celtic world in its 'proto-Celtic' substratum (Almagro-Gorbea 1992, 1993), which was preserved in western regions of the Peninsula, although in the transition from the Late Bronze Age to the Iron Age this zone expanded to cover the area from the Atlantic to the Meseta. This phenomenon is attested by the existence of ideological elements (such as initiation rituals for 'fraternities' of warriors, archaic deities, etc.), linguistic evidence ("Lusitanian" and anthroponyms and toponyms P-) and common archaeological finds (weapons cast into the water, round houses, the absence of "fortified settlements", etc.), as well as a primitive social organization, which appears to be linked to the Atlantic Late Bronze Age, but whose characteristics are similar to historical Celts, and therefore allow a connection to be made. Thus, even if we do not rule out the movements and migrations of populations, the formation of Peninsular Celts would probably have come about due to the in situ evolution of this culture, in which processes of acculturation, particularly of the Tartessian and Iberian world, would have played a defining role, even constituting a key element to understanding the personality of the Peninsular Celts.

According to Almagro-Gorbea (1993: 146-150), the Celtiberian culture emerged out of a 'proto-Celtic' substratum, which would explain the cultural, socio-economic, linguistic and ideological similarities, as well as the progressive assimilation of the 'proto-Celtic' substratum into the Celtiberian culture. This process of Celticization allows us to understand better the heterogeneity and clear personality of Celtic Hispania within the Celtic world.

In a recent study, Almagro-Gorbea (2001: 95) proposed that the origins of the Celts could be traced back to the third millennium BC, and sought to find the initial roots of the formative process that would eventually give rise to the Celtic people in the Bell Beaker culture. This remote origin would certain explain the wide dispersion of the Celts throughout western Europe, as well as the variability of the different Celtic peoples, and the existence of ancestral traditions. The Celts described in Classical sources and known to us through archaeological remains would, therefore, have been the result of a long process of progressive or "cumulative" Celticization, presumably explaining the cultural variety which, although they may have all spoken similar languages and held similar beliefs about life and its values, made them stand out from other people in the ancient world (Almagro-Gorbea 2001: 98). However, there is an undeniable presence in the Eastern Meseta of ethnic contributions from the Ebro valley, approximately dated to the eighth and seventh centuries BC. Such contributions have allowed us to assess the importance of the Urnfield culture in the development of the Celtiberian world - possibly in the form of small infiltrations of "settlers" (Arenas 1999: 246) - stemming from its introduction of a new socio-economic model into local populations as well as possibly an Indo-European language, the direct predecessor of the Celtiberian language attested in inscriptions. This proposal (Lorrio 1997a: 260-261, 372; Ruiz Zapatero and Lorrio 1988, 1999: 26, 34) allows for the indigenous cultures - which did not disappear along with Cogotas I, a characteristic culture of the Late Bronze Age in the Meseta (for an opposing view, see Arenas 1999: 246) - to play an active role in the process of interaction with the Urnfield model. It was this interaction that gave rise to the ancient Celtiberian world.²

3. Iberian Celts: literary sources, linguistics and archaeology

The main sources that allow us to build up a picture of the Celts in the Iberian Peninsula are the following: Classical texts, linguistic and epigraphic evidence, and archaeology. Mythology and folklore should also be included here, since these disciplines provide evidence of the endurance of certain traditions that are allegedly Celtic in origin, although the value of these traditions for the purpose of Celtic studies is yet to be determined.

1. Firstly, we will analyze the information provided by Classical Graeco-Roman authors (Figs. 3-5), who approached the description of the Iberian Peninsula from different perspectives and had diverse motivations and interests. Only a few of these descriptions refer to the presence of Celts, and thus paint a picture of the Celtic World as seen "from the outside" in which mistakes, personal interests and manipulation of the information are all present to a certain degree (Champion 1985: 14-17).
   
   Furthermore, it is important to bear in mind that specifically Graeco-Roman historians and geographers created some of the terms used to refer to Celtic peoples in Iberia. Moreover, the reasons that led to the usage of these terms are in general unknown, which without doubt makes it difficult to understand the groups defined as Celts. This is certainly the case with the term Celtiberi, which must have been used to refer to a people that the authors believed to be a mixed group (Untermann 1983, 1984). This term appears in accounts by Diodorus, Appian and Martial, for whom the Celtiberians would have been Celts mixed with Iberians, although other authors such as Strabo saw the Celts as the more dominant group in this blend. There is an extremely interesting reference by Pliny (3, 13) to the Celtici in Baeturia, who are classified as Celtiberians because of their rituals, language and the names of their towns. This sheds light on some of the defining factors for a Graeco-Roman author when it came to classifying people into concrete ethnic categories. Although Pliny makes no reference to their material culture, we know that when Classical authors refer to aspects such as weaponry, they are showing the differences or similarities observed between the peoples they describe (Lorrio 1997a: 189-196).
   
   
2. Linguistic evidence includes epigraphs written in indigenous languages and also the field of onomastics, mainly known through Latin inscriptions. The discovery of inscriptions written in indigenous languages in the Peninsula, as well as the wealth of onomastic records that have survived, allows us to define two important linguistic areas with a reasonable amount of clarity: one linguistic area is not Indo-European and was found in areas of the south and the Levant; the other is Indo-European and was dominant in the center, north and west of the peninsula (Fig. 6).
   
   
3. Archaeological records are not easily reconciled with the two above sources, and this lack of correlation has led the fields of archaeology and linguistics to go their separate ways. Archaeology must function independently, mainly on the complex issue of the formation of the Peninsular Celtic world. On this subject, literary, linguistic and onomastic evidence - despite their undeniable value - can only be of limited use owing to the impossibility of determining exactly how far back such phenomena date.

3.1 Written Graeco-Roman sources

Despite the wealth of information provided by Classical authors on the subject of the Iberian Peninsula, there are very few references to the Celts.³ The oldest sources simply describe the coastal areas of the Peninsula, mainly the South and the Levant, of which the Classical authors had first-hand experience. References to inland areas are a great deal more general and often vague. As a consequence, any information about the location of these peoples is far too imprecise to be of much value. Mostly, the authors simply made a cursory acknowledgement of their presence, and located them sometimes near cities or other presumably non-Celtic peoples, occasionally explaining their presence as a result of geographical accident.

Traditionally, one of the most ancient sources on the Iberian Peninsula is thought to be the Latin poem Ora Maritima (Mangas and Plácido 1994), written at the end of the fourth century AD by Rufus Festus Auienus, although some scholars believe that it presents the story of a voyage to Massalia from the sixth century BC, with a few later interpolations. Despite its controversial nature, a certain passage from the Voyage (129-145), which is on the whole excessively obscure, has been interpreted as the most ancient recorded reference to the Celts (Rankin 1987: 2-3; Schulten 1955: 36-37; etc.), along with allusions to a series of other peoples such as the Cempsi, Saefes and Berybraces who are difficult to characterize (195, 485).

Leaving aside the Ora Maritima for the moment, the first reference to the Celtic world was made by Hecataeus of Miletus (ca. 500 BC), who defined Narbo as a Celtic city (as well as Nirax, but the location of this city is uncertain). He placed the Greek colony of Massalia in the land of the Ligurs, close to the land of the Celts. Herodotus (2, 33; 4, 49) is credited with the first certain reference to the presence of Celts in the Iberian Peninsula, and the oldest evidence of the ethnonym Keltoi. In the fifth century BC, Herodotus indicated that the source of the river Istro (now known as the Danube) was in the land of the Celts, which stretched beyond the Pillars of Hercules, and bordered the land of the Kynesios (or Kynetes), who were thought to be the most western tribe in Europe.

Later on, there were repeated mentions of the presence of Celts in the Iberian Peninsula. Ephorus (in Strabo 4, 4, 6), ca. 405-330 BC, believed that the Keltiké occupied most of the Peninsula, as far as Gades (today's Cádiz). Eratosthenes (in Strabo 2, 4, 4, p. 107), ca. 280-195 BC, seemed to confirm the fact that the Celts reached the southwest of the Iberian Peninsula. He believed that border regions of Iberia were inhabited as far as Gades by Galatae, a term that was doubtless used as a synonym for the Celts. The references to Celts further inland are reflected in another passage of Pseudo-Scymnus (162 ff.) attributed to Ephorus, according to which the source of the river Tartesos, now the Guadalquivir, began in the land of the Celts. Diodorus (25, 10) confirmed the presence of Celts in the south of the Peninsula. He indicated that when Amilcar arrived in the Peninsula in 237 BC, he had to face Tartessians and Iberians who fought alongside the Celts of Istolacio.

A rapid increase in information regarding Iberia occurred in the late third century BC, and particularly during the following two centuries, when the Peninsula became of increasingly strategic interest for Rome. Information came pouring in, not only about the geography of the area, but also about economic, social and religious structures. This information boom provides a much more complete overview of the Peninsular Celts, and allows us to pinpoint their settlements more exactly.

During the wars with Rome and the period directly after that, written sources modified the concept of Celtica, and applied it to areas to the north of the Pyrenees⁴, although this did not mean that there were no express references to the existence of Celtic populations in the Iberian Peninsula. In-depth analysis of the works of Polybius, Poseidonius, Strabo, Diodorus, Pomponius Mela, Pliny the Elder and Claudius Ptolemy sheds light on three distinct areas located in the center and the west of the Peninsula, in which the presence of Celtic peoples is explicitly indicated: the eastern Meseta, the northwest, and the southwest of the Iberian Peninsula. Clearly, this does not exclude the possibility of other areas that are not mentioned by these sources.

3.1.1. The Celtiberians

Figure 3. The main pre-Roman populi in the Iberian Peninsula (second and first centuries BC) (After Alvárez-Sanchís 2003, following Untermann 1992).

The first of the areas mentioned above corresponds to the inland regions of the Iberian Peninsula, where the Celtiberians would have been located (Fig. 3).⁵ This group was expressly considered by several authors to be Celtic (Posidonius, in Diodorus 5, 33; Strabo 3, 4, 5; Martial Epigr. I, 55, 8-10 and X, 63, 3-4; Isidorus Ethym. 9, 2, 114). From the late third century BC onwards, Graeco-Roman literary sources began to provide the earliest information on the Celtiberians and Celtiberia, and gave accounts of the names of Celtiberian peoples and their location. This information, which is at times contradictory, not to say blatantly imprecise, allows us to define the theoretical Celtiberian territory in two ways (Fig. 4):

Figure 4. Celtiberia: towns and populi (third to first centuries BC) (After Lorrio 2001).

1. By using Strabo's descriptions of Celtiberia, or the use of the nouns "beginning" or "end" when referring to this area, such as Clunia, Celtiberiae finis (Plin. 3, 27), Segobriga, caput Celtiberiae (Plin. 3, 25), or Contrebia, caput eius gentis, referring to the Celtiberians (Val. Max. 7, 5, 5),
   
   
2. By analysing the peoples thought to be Celtiberians and examining the location of cities linked to them, or by looking at peoples who occupied neighbouring territories but who were not Celtiberians. However, Classical sources and contemporary historiography do not coincide on these issues, mainly because of the different ways the Graeco-Latin authors used the terms "Celtiberian" and "Celtiberia". There is no single unanimous opinion regarding the links between peoples that could be considered Celtiberians, which include the Arevaci, Pelendones, Lusones, Belli and Titti, and occasionally Vaccaei, Carpetani, Olcades, Lobetani, and even more distant groups such as Oretani, Bastetani, Bastuli or Celtici. There is similarly no shortage of authors who reject the ethnic content of the term altogether, and take it to refer to all inhabitants of an extensive area of the inland Peninsula.

The first reference to Celtiberia is made within the context of the Second Punic War, in Polybius' narration of the siege of Saguntum (3, 17, 2) in the spring of 219 BC. From this date onwards, information about Celtiberians and Celtiberia is plentiful and varied, since the Celtiberians were one of the key players in the various wars and battles that took place throughout the second and first century BC, which culminated in the destruction of Numantia in 133 BC. Celtiberia also played a vital role in the Sertonian Wars (Salinas 1996: 27-37).

The geographical concept of Celtiberia underwent a process of evolution following its first appearance in texts. It began as a rather generic concept, obvious in the most ancient literary testimonies, such as Poseidonius, for whom the Pyrenees separated Gaul from Iberia and Celtiberia (in Diod. 5, 35). Clearly, these initial concepts were vague and contained blatant errors. However, there was also a narrower definition that referred to the eastern Meseta and the right side of the Middle Ebro Valley (for an opposing viewpoint, see Gómez Fraile 1996, 1999, 2001). This concept was largely due to an increased understanding and knowledge of the ethnic complexity of the Peninsula. Such an overview was offered by authors such as Strabo, Pliny or Ptolemy, although Strabo included information that referred to both the extensive and narrower definitions of Celtiberia, which has been put down to his use of sources from different periods (Capalvo 1996: 52-53).

Strabo, who wrote around the time of Christ's birth, offered a very broad concept of Celtiberia using information provided by Polybius and Poseidonius. Many of the largest rivers of the Atlantic basin started there, such as the Douro, the Tagus, the Guadiana and even the Guadalquivir, as well as the Limia and the Miño, although Poseidonius believed this last example began in Cantabria. The easternmost point of Celtiberia was located in the Idoubeda - the Iberian Mountains - although Strabo himself placed a few Celtiberian cities even further to the east. According to Strabo (3, 4, 13), Celtiberia was divided into four territories, of which he only made reference to those inhabited by the Arevaci and Lusones, although Polybius (35, 2) and Appian (Iber. 44, 48-49, 50, 61-63 and 66) revealed that the other two corresponded to the Belos and Belli and Titthi tribes. A little further on, Strabo (3, 4, 19) indicated that some believed that there were five areas. Several candidates were proposed as inhabiting this fifth zone, including the Vaccaei, who were considered Celtiberians by Appian (Iber. 50-52, 53-54) although in general they appeared in sources as two separate peoples. In all probability, this fifth section was inhabited by the Pelendones, whom Pliny (3, 26) described as Celtiberians.

Pliny's writings reflected the administrative situation in Hispania following the reforms introduced by Augustus; in his descriptions of Hispania Citerior he only referred to the Arevaci (3, 19; 3, 27) and Pelendones (3, 26; 4, 112) as Celtiberians. However, in the second century AD, Ptolemy, in his description of the province of Tarraconensis, treated the Arevaci (2, 6, 55) and Pelendones (2, 6, 53) as separate from the Celtiberians (2, 6, 57), whom he considered another ethnic group, and placed all these peoples on an equal footing (see García Alonso 2003: 295-297, 301-310, 326-343).

Analysis of the literary sources reveals an enormously complex Celtiberia whose geographic scope and ethnic composition are difficult to define and substantially changed during the process of the Roman conquest and subsequent Romanization. It is important to remember that over three centuries passed between the oldest references to the Celtiberians and Ptolemys' writings. During this period, the wars and battles and later on the administrative reforms surely had a marked impact on the Celtiberian lands.

The fact that the term "Celtiberia" did not originate there makes it even more difficult to assess, as do the frequent contradictions found in its usage and application in literary sources, which can sometimes but not always be explained by chronology. As for the term Celtiberi, it appears to be used to refer to a people considered to be a mixed group (Untermann 1983, 1984). Diodorus, Appian and Martial certainly used it in this way, and believed that the Celtiberians were Celts mixed with Iberians, although for other authors such as Strabo, there was a definitive domination of the Celtic component in this mixture. Even if not all scholars believe that the concept "Celtiberian" refers to an ethnic unit (Gómez Fraile 1996: 172; Koch 1979: 389), it is equally important to study aspects of the indigenous features that could have been passed on to visitors to the area, such as customs and languages, since they could form the basis of a proven identity (Burillo 1998: 121-146). The Celtiberians could be considered an ethnic group insofar as they included subordinate ethnic units, just as the Gauls or the Iberians did, but on a smaller scale, without evidence for centralized power or even a political hierarchy (Burillo 1993: 226).

In short, the true significance of the terms "Celtiberian" and "Celtiberia" in the different contexts in which they were used is unknown, although it is likely that they had ethnic and purely geographical connotations. It has been suggested that the term "Celtiberian" could have been used to refer to the "Iberian Celts" (Tovar 1989: 83), even though, as we have already pointed out, the Celtiberians were not the only Celts on the Peninsula. Possibly this term was simply used to highlight the personality of these people within the Celtic world (Ciprés 1993a: 57).

The Berones were another Celtic population (Str. 3, 4, 5) that inhabited the area now known as La Rioja, on the right bank of the Upper Ebro River (Tovar 1989: 77-78). According to Strabo (3, 4, 12), they settled to the north of the Celtiberians, were neighbours of the Cantabrian Coniscos, and had been part of the "Celtic immigration". He stated that they inhabited the city of Varia. Ptolemy (2, 6, 54) also mentioned this city, which he calls Vareia, along with Tritium Megallum and Oliba.

3.1.2. The Celtici in the southwest

Many sources concur in pointing to the presence of Celtic populations in the southwest of the Peninsula. Strabo (3, 1, 6) echoed Poseidonius when he mentioned the Keltikoi as the main inhabitants of the region located between the rivers Tagus and Guadiana, approximately where the Alentejo (Portugal) stands today.⁶ There were a few Lusitanians living among the Celtici who had been transferred there by the Romans from the right banks of the Tagus River.⁷ They shared the same "gentle and civilized" character of the Turdetani. Strabo put this down to the fact that they were neighbouring populations, and Polybius proposed that they were related, "although the Celtici are less [civilized] because they generally live in hamlets" (Str., 3, 2, 15). The Celtici from Guadiana had blood links with the Galician Celts, since there had been large-scale migration to the northwest of these Celts along with the Turduli (Str., 3, 3, 5). Their most famous city was Conistorgis (Str., 3, 2, 2), which, according to different sources, belonged to the Cunetes or Conii (App., Iber. 56-60). Similarly, Strabo (3, 2, 15) indicated that the Celtici established colonies, such as Pax Augusta.

Pliny (4, 116), writing in the mid first century AD, placed the Celtici in Lusitania, and stated that the inhabitants of the oppidum of Mirobriga "are called Celtici" (Plin., 4, 118). Ptolemy (2, 5, 5), in the second century AD, provided a list of Celtici cities in Lusitania.

In addition to the territories already mentioned, the Celts settled in Baeturia,⁸ an area located between the rivers Guadiana and Guadalquivir (Plin., 3, 13-14). According to Pliny (3, 13), Baeturia was divided "into two parts and contained among a few minor peoples: the Celts, who lived next to Lusitania, and the Turduli...". For Pliny (3, 13), the Celts that inhabited Baeturia were Celtiberians - who had come from Lusitania - as demonstrated by their rituals, language and the names of their peoples, known in Betica by nicknames. In Celtic Baeturia, which can be located in the basin of the river Ardila (Berrocal-Rangel 1992: Fig. 2), he mentioned a few cities whose names coincided with a Celtiberian city, such as Nertobriga Concordia Iulia or Segida Restituta Iulia.⁹ Pliny (3, 13) also included some other cities in Celtica, located outside Baeturia, including a few of them in the hills near Cádiz and Malaga, and to the south of the river Guadalquivir.

The Celticization of Baetica is evidenced by toponyms of cities that are considered Celtic (Tovar 1962: 360 ff.), such as Segida Augurina (Plin., 3,10) or Celti, in the Conventus Hispalensis (Plin., 3, 11).¹⁰ Ptolemy (2, 4, 11), in the second century AD, listed the cities belonging to the "Celtici of Baetica": Arucci, Arunda, Curiga, Acinippo and Uama. A few of these coincided with those mentioned by Pliny, although he included Seria, Segida and Nertobriga among the towns belonging to the Turdetani (Ptol., 2, 4, 10).

The presence of Celts in Betica - indicated more or less explicitly in texts spanning the period from the most ancient sources up until Flavius Philostratus II (vit. Apoll. 5, 2, 166), ca. 200 AD - and their links with the Celtiberians are supported by the ostensible location of the Ultima Celtiberia (Liv. 40, 47) - that was conquered by Sempronius Gracchus in 179 BC - in the Ulterior province (Capalvo 1994, 1996: 107-116; Pérez Vilatela 1989: 258, 1993: 428; for an opposing perspective, see Ciprés 1993b: 282 ff.), because toponyms of the cities of Munda and "the city that the Celtiberians call Certima" have been identified as modern-day Monda and Cártama, in the province of Málaga.

3.1.3. The Celtici in the northwest

Figure 5. Suggested distribution of the main populi and civitates in the northwestern Iberian Peninsula according to Roman sources (After González Ruibal 2003).

The northwest of the Peninsula is the third area where Graeco-Roman geographers and historians explicitly indicated the existence of Celtici populations,¹¹ all of which settled in Gallaecia Lucensis. By analysing certain passages from Strabo (3, 1, 3; 3, 3, 5), Pomponius Mela (3, 10-11; 3, 13) and Pliny (3, 28; 4, 111), we can confirm that a whole series of populations were grouped under the umbrella term Celtici, including the Neri, the Supertamarci, whose existence has also been confirmed through epigraphs (Albertos 1974-75), the Praestamarci and possibly also the Cileni, and the Artabri, who according to Mela (3, 13) were celticae gentis. These Keltikoi, according to Strabo (3, 3, 5), were related to the Celtici from the Guadiana region that migrated with the Turduli as far as the Limia River, where it would seem that the two groups separated. The Artabri then continued their journey north without the Turduli.

The Artabi, or Arrotrebae settled in the area surrounding cape Nerium, which could well be modern-day Finisterre; the Neri were probably neighbours of the Artabri, and lived on the west coast of the Peninsula (Mela 3, 11); the Supertamarci and the Praestamarci were linked to the Tambre River, respectively settling on the north and south banks of this river. The Cileni settled further to the south (Fig. 5).

3.1.4. The Gauls of the northeast

Finally, we must mention the passages from Classical sources that refer to the presence of Gauls in Hispanic territory. With the exception of the reference made by Eratosthenes to the Galatae (in Str., 2, 4, 4), a term which should be interpreted as a synonym for Celts, the scanty information available points to much later dates, from the late third century BC onwards. Thus, much of this presence can be interpreted as infiltrations by groups of Gauls who came from the north of the Pyrenees.

In Livius' narration of the events that took place between 214 and 212 BC (24, 41), he referred to the death in combat of Moenicoeptus and Vismarus, two reguli Gallorum who were allied with the Carthaginians. The booty was mostly made up of spolia plurima Gallica: a great number of golden torques and bracelets. The names given to these rulers originated from the words for them from outside the Peninsula (Albertos 1966: 158, 253; Tovar 1977: Footnote 15). However, they could be put in relation to the Celts from the Meseta according to Schulten (1935: 85).

There are better records of the presence of Gauls in the northeast. Owing to the geographical proximity of this area to Gaul, contact would have been particularly intense, as the toponomy (see below) and the archaeological record (Almagro-Gorbea and Lorrio 1992: 414) demonstrate. The context for Caesar's famous quotation lies here (Bell. ciu. 1, 51). In the year 49 BC, he indicated the arrival at his camp (situated opposite the city of Ilerda) of a contingent made up of Gallic cavalrymen and Ruthenian archers, accompanied by over 6,000 men together with their servants, women and children.

In short, Classical sources indicate the presence of Celts in the Iberian Peninsula from at least the fifth century BC onwards. From the third century BC on, they locate Celtic groups in three well-defined areas: the center, the southwest, and the northwest. However, the relationship between the various populations and individual characteristics were far from certain.

3.2. Linguistic and epigraphic evidence

The value from an onomastic perspective of the large amount of epigraphs written in Latin that have been found in the Iberian Peninsula will be discussed later. Furthermore, a series of epigraphic testimonies in indigenous languages have also been found, and their distribution is restricted to a narrower geographical area than the one indicated by literary sources or by onomastics.¹² There are basically two epigraphic areas relating to Indo-European languages (Fig. 6):

1. The Celtiberian zone,¹³ defined by the discovery of texts in an archaic Celtic language, some of them written using the Iberian alphabet - which had been adapted from Iberian at some point in the second century BC - and some in the Latin alphabet, dated to the first century BC, although in a few cases the texts are dated after the birth of Christ. There are different types of epigraphic documents: hospitality tessera, religious rock inscriptions, coin inscriptions, burial inscriptions, ceramic graffiti, etc. There were also extensive public documents, such as the first and third Botorrita Bronzes. Most of these discoveries were made within the eastern Meseta and the Central Ebro Valley, the area identified as Celtiberia by Classical sources, also including adjacent territories.
   
   
2. From the west of the Peninsula there is a small number of inscriptions written in the Latin alphabet in an Indo-European language that is different from Celtiberian, which has been named Lusitanian owing to the geographical distribution of the discoveries. These inscriptions are later in date, from the first few centuries AD. Although most researchers argue that these texts exhibit Indo-European language features that differ from Celtic (Gorrochategui 1987; Schmidt 1985; Tovar 1985; Villar 2001: 118), it has also been proposed that the language is linked to the Celtic sub-family, possibly representing a Celtic dialect that is different from Celtiberian (Untermann 1987).

A brief mention should be made of the southwest or Tartessian inscriptions (which are mostly from mortuary contexts dated between the seventh and sixth centuries BC). They were initially thought to be related to a non-Indo-European language, but subsequently the possibility arose that they were written in a western Indo-European language, and more specifically a Celtic one (Correa 1985, 1989, 1990, 1992; Untermann 1995c, 1997: 165-166, 2001: 196 ff.). However, difficulties in deciphering these texts have cast some doubt upon this interpretation.

The co-existence of several Indo-European languages, some of which are Celtic but others perhaps not, was a generalized phenomenon that reflected the immense complexity of the Indo-European Peninsula when the Romans arrived, complicated even further by the presence of this Mediterranean superpower.

The information provided by the documents written in indigenous languages that have been found is enhanced by the study of onomastics, known to us through the works of Graeco-Roman authors, and above all by studying epigraphs. Epigraphs have been found written in indigenous languages - either in the Iberian or Latin alphabets - and also in the form of Latin inscriptions, which are much more numerous, and which date back to the Imperial period (Albertos 1983: 858 f.). Onomastic evidence¹⁴ mainly consists of anthroponyms, toponyms, theonyms and the names - formed from anthroponyms - of the supra-family kinship groups or clans (González 1986; Pereira 1993; among others).

Figure 7. Main Celtic place-names in Iberia: toponyms ending in -briga (After Albertos 1990).

Toponomy is of great interest owing to its conservatism, a characteristic that makes it more reliable than personal onomastics, which are much less stable. Toponyms ending in -briga are the most widespread onomastic element in Celtic linguistics (Albertos 1990; Villar 1995: 153-188). There are good records of this suffix, which is clearly Celtic in origin, and means "fortified place" (from the Old Irish brig, genitive breg meaning 'hill'; see Marstrander 1913 Dictionary of the Irish Language), in the continental Celtic territories, although it is found with much greater frequency in the Iberian Peninsula (Fig. 7). The geographical distribution of such toponyms defines a wide area that encompasses the central and the whole western part of the Peninsula, including the southwest, where they co-exist with Tartessian epigraphic findings. This fact, as with the Gallic toponyms ending in -dunum, perhaps should be interpreted as later evidence of Celticization (Rix 1954). This is particularly the case with toponyms found at the most westerly points of the Peninsula, which could possibly be linked to the expansion of Celtiberia. Usually, toponyms ending in -briga contain other Celtic components, but there is also an important presence of non-Celtic indigenous prefixes, mainly to be found in the west, such as Conimbriga. This is even more evident when associations with Latin names are brought into the equation, such as Caesarobriga, Augustobriga, Juliobriga or Flaviobriga that demonstrate that this toponym was still being widely used long into the Roman period.

Another group of toponyms include the suffix Seg- (Untermann 2001: 201-204) whose etymology can be explained by the Celtic word seg meaning 'victory' (Old Irish seg, see Delamarre 2001: 228-229; Evans 1967: 254-257); they are occasionally linked with toponyms ending in -briga, such as Segobriga.

If toponyms that include Seg- and -briga allow us to define an area of Celticization, or perhaps even Celtibericization, a different meaning should be given to toponyms ending in -dunum, meaning 'hill, fortress' (Delamarre 2001: 130), which occur within a much narrower area in the Iberian Peninsula, but are extremely frequent in the rest of Celtic Europe, and can most probably be linked to Gallic groups to the north of the Pyrenees. This toponym is clearly of a later date, mainly from the Roman period. Since its geographic distribution is concentrated in the northeast of the Peninsula, it can be linked to other historical, archaeological and toponomical evidence recorded in this area and interpreted along the same lines.

The study of personal onomastics is extremely interesting, even though it is less reliable than toponomy because it is less stable and is subject to the mobility of individuals. The Romans played a crucial role in this phenomenon, both directly and indirectly: not only did they relocate villages and people (Str., 3, 1, 6), but also the wars against Rome affected the movement of individuals. We now have at our disposal a complete anthroponymic corpus that allows us to study this aspect with a certain degree of accuracy.¹⁵

Figure 8. Some Celtic personal names in Iberia: A, Ambata/-us, Ambatio. B, Camala/-us and related (After Abascal, unpublished).

There seems to be a characteristic anthroponymy common to Indo-European areas defined by the distribution of toponyms ending in -briga. This anthroponymy, generally Indo-European although not always clearly Celtic (Albertos 1983: 860-861; de Hoz 1993a: 367-369), provides a certain feeling of homogeneity, as, for example, in the case of the anthroponym Ambatus (Fig. 8A). Nevertheless, there are concentrations of anthroponymic series, which, on occasion, are clearly dominant in a certain area: Cloutius, Clutamus or Camalus (Fig. 8B), in the west of the Iberian Peninsula, Boutius, Tancinus, Tongetamus or Pintamus, in the Central Lusitania, etc. Therefore, it would not be entirely inaccurate to talk about Lusitanian or Lusitanian-Galician personal onomastics, regardless of the fact that often their distributions overlap, thus making it difficult to pinpoint the precise geographical locations of the towns mentioned in literary sources.

Ethnic anthroponyms such as Celtiber and their variants are perhaps of even greater importance. These anthroponyms helped to define negatively a purely Celtic and/or Celtiberian area. It is thought that these names were used to describe non-native individuals, to categorise them and indicate their origin. They are of greater interest for the following reason: in addition to contributing to the definition of what is not within the Celtic area (hence 'define negatively'), they also highlight areas to which Celtic people emigrated. In the case of Lusitania and Vettonia, this emigration must have taken place on a fairly large scale.

Figure 9. Supra-family organization structures (After González 1986, modified and enlarged).

In order to obtain a better understanding of Celtic society it is crucial to understand the concept of their clans or supra-family organization structures (Albertos 1975; González 1986; González and Santos 1993; de Hoz 1986: 91-98, Pereira 1993), which included the descendants of a certain individual by whose name they were identified. These family structures were normally recorded using the plural genitive form of an adjective derived from an anthroponym, and often appeared in indigenous onomastic formulas, both in Latin and Celtiberian inscriptions. A typical formula would be: Lubos (the individual's name), of the alisokum (name of the family group), son of Aualos (father's name), of Contrebia Belaisca (place of origin). Despite the large number of individuals who were named in this way, repetition of a certain combination was relatively infrequent. When this did arise, the two individuals were either blood relatives (father, son or brother), or lived in separate areas located very far away from each other. This system allows us to put together extended family units - grouping up to four generations at a time (de Hoz 1986: 97-98) - which form very small groups. The geographical distribution of these "plural genitives" encompasses the area known as Celtiberia according to Classical sources, and areas in the center, to the north of the Tajo River, as well as in Cantabria, with an important concentration in Asturias (Fig. 9).

Figure 10. Celtic deity names in the Iberian Peninsula (After Olivares 2002).

The west of the Iberian Peninsula had its own set of anthroponymic and linguistic features (see de Hoz 1994; Untermann 1994; etc.). It also had its own exclusive theonymy, with deities such as Bandua, Reue, Cosus and Nabia (García Fernández-Albalat 1990; Olivares 2002; Prosper 2002; Untermann 1985b, 1994: 34 ff.; Villar 1993-95, 1994-95, 1996). There are records of these deities in areas of modern-day Portugal, Galicia, the western border of the Meseta region, and in Extremadura.¹⁶ Regardless of the differences in theonyms, there are structural elements that, according to Olivares (2002: 257), link the religion practiced in Lusitania and Galicia with the rest of the Celtic areas of the continent, both pantheons including the same number of gods, who in turn possessed similar qualities. Recently it has been suggested that the names of certain indigenous deities such as Bandua, Reue, Quangelus, Trebaruna, Nabia, Arentius and Arantia are specifically Lusitanian, and could be used to define the boundaries with other populi (Olivares 2000-2001). The presence of a genuinely Celtic theonym, Lugus, is significant. The geographical distribution of this theonym spread throughout Celtiberia and Gallaecia lucensis (Almagro-Gorbea and Lorrio 1987: Map 7; Marco 1986, 1987: 59-61; Tovar 1981). Olivares (2002: 257-258) proposed that, despite the different names given to Celtiberian and Lusitanian-Galician deities, suggesting the existence of a certain cultural difference, "the most characteristic elements of their gods were the same in both areas, thus showing that they shared the same religious and ideological universe" (Olivares 2002: 258) (Fig. 10).¹⁷

However, the study of onomastics allows us to properly identify Celtic Hispania. This geographical area is defined by the presence of toponyms ending in -briga, and by a characteristic Indo-European anthroponomy that enables us to distinguish between different regional groupings, which can on occasion be extremely important. This is certainly the case with anthroponyms having an ethnic content. On the other hand, the "supra-family organization structures" are concentrated in the central and northern Peninsula and are absent in the west, which had its own exclusive theonymy.

3.3. Archaeological evidence

Information provided by archaeology is essential in our reconstruction of society in the centuries immediately preceding the beginning of the Christian era, which is partially known thanks to the information provided by Graeco-Roman historians and geographers. Moreover, archaeological records allow us to analyze the foundation of the Peninsular Celtic world and determine the cultural processes that led to its development and later expansion. Over the last few decades, we have gained increasing knowledge about the Late Bronze Age and Iron Age in the Iberian Peninsula, and consequently have been able to interpret material culture that in theory corresponds to the Celts much more accurately, and detect links with other related fields such as linguistics or religion (Almagro-Gorbea 1992, 1993; Almagro-Gorbea and Lorrio 1987). Archaeology enables us to approach this process with a certain degree of chronological accuracy, as opposed to other disciplines such as linguistics, which in theory should adhere closely to the dates provided by the written documents themselves.

During the first millennium BC, at least one group of Hispanic Celts, through their contact with the Tartessians or the Iberians, assimilated elements that came from the Mediterranean; these elements included weaponry, the potter's wheel, urbanism and the alphabet. The extent of this process of assimilation meant that this group of Celts developed a material culture that was clearly distinguishable from the Central European Celts of Hallstatt and La Tène. This would certainly shed some light on why they are difficult to understand using traditional approaches, and would also justify the mixed nature of these people - Celtic and Iberian - alluded to by Classical authors when they refer to Celtiberians (Diod., 5, 33; App., Iber 2; etc.).

This archaeological evidence also enables us to analyze the process of ethnogenesis of the Celtiberians, despite the difficulties in applying this term before it was actually coined by the Graeco-Roman authors. Clearly, this area is extremely interesting, since it is one of the few cases in which all the analyzed sources point to their Celtic nature, at least during the Wars with Rome. Thus, the appearance of certain material culture elements, settlements, funeral rites and socio-economic structures, which were characteristic of the Celtiberian world throughout its evolutionary process, enable us to pinpoint when and where that process began: around the seventh or sixth centuries BC in the highlands of the eastern Meseta (Lorrio 1997a: 261). Therefore, as we will see later on, the use of the term "Celtiberian" is appropriate when referring to the archaeological cultures located in the areas of the Upper Tagus, Upper Jalón and Upper Douro Rivers, in the eastern Meseta and in the Iberian Mountains, from the formative stages onwards, which occurred around the aforementioned dates (see Fig. 15). The continuity shown in the cultural material (Lorrio 1994a-b, 1997a: 258, Fig. 110) of this area of the Meseta enables us to link archaeological evidence with historical or ethnic evidence, given its highly peculiar nature within an area that almost completely coincides with Celtiberia, as defined by the Classical authors. The Celtiberians were considered a Celtic people. Celtiberian was spoken in this area, at least during the historical period, and is the only language that has been identified as unequivocally Celtic in the whole of the Iberian Peninsula (see Fig. 11).

On the other hand, the discovery of elements that could be considered Celtiberian in areas that are not strictly within Celtiberia could be viewed as an indication of the Celtibericization, and therefore Celticization of those areas. This would not have required large-scale ethnic movements but could have been an intermittent process that had a cumulative effect, with a few dominant groups imposing themselves, local migrations or even the acculturation of the substratum (Almagro-Gorbea 1993: 156, 1995b). If this theory is correct, then the geographic dispersion of a few elements like horse fibulae (Almagro-Gorbea and Torres 1999) or certain typically Celtiberian weapons such as bi-globular type daggers (Lorrio 1997a: Fig. 8) could be interpreted as signs of this expansion and therefore the process of Celticization, which is also attested in the distribution of the ethnic anthroponyms Celtiber and their variants, or of toponyms ending in -briga. From a linguistic point of view, this process is evident from the appearance of texts written in Celtiberian outside of the theoretical Celtiberian zone. Most of these texts have been found within the Meseta, but some have been discovered in more remote areas such as Extremadura.

Clearly, this evidence does not negate the possibility that there were other Hispano-Celts besides the Celtiberians - a fact that literary sources appear to confirm with regard to the Berones - or that this process of Celtibericization occurred in areas where there had previously existed a Celtic component that is difficult to identify. A particularly complex picture begins to emerge in relation to those ethnic groups whose formative process has been determined by archaeology, and whom no Classical author considered specifically Celtic, or whose language is unknown, as is the case with Lusitanian, whose Celtic character is far from being universally accepted. If we view the Celtic world in this way, the variability within time and space makes it impossible to view it as uniform and "simple". Recent discoveries and knowledge have shown that the reality of this world was extremely complex.

4. The genesis of "Celtic"

Research into and knowledge of the Celtiberians is crucial to understand the Celts of the Spanish Peninsula. Firstly, the Celtiberians played an important role in the context of Celtic Iberia. Secondly, disparate sources such as literature, linguistics, onomastics and archaeology - which deal with aspects such as society, religion, language or material culture - all point to the fact that that they were a Celtic people, a fact which ethnonymy also confirms. Current research and thinking places the ethnogenesis of the Celtiberians around the seventh or sixth century BC, or perhaps even as far back as the eighth century BC. This process is fundamental to our understanding of the Iberian Peninsular Celts themselves.

4.1. Ethnogenesis and archaeology: the case of the Celtiberians

Traditionally, archaeological studies of the Celtiberians have been restricted to what we could term the Historical period, which spans the period from the late third to the first century BC. Ever since Bosch Gimpera's time, the previous culture was called the Post-Hallstatt period, or simply the Second Iron Age. This approach to Celtiberian archaeology remained unchanged until Almagro-Gorbea (1986-87, 1987) began to alter perspectives on this question. Quite rightly, he highlighted the need to extend this analysis even further back in time to the sixth century BC because from then on there was a basic continuity in archaeological finds that means we can justifiably talk of a Celtiberian culture. This same conclusion has also been drawn from research into the continuity of customs surrounding the arrangement of graves in a number of cemeteries from the sixth century BC up to the Roman conquest (Lorrio 1997a: 258, Fig. 110). This cultural continuity can also be discerned in towns and socio-economic structures. Therefore, the term "Celtiberian" can justifiably be used as an archaeological category from the sixth century BC onwards (Fig. 11). However, this continuum in material culture from the First Iron Age into the Roman era does not mean that the Celtiberians lived throughout this period in exactly the way they were portrayed by Greek and Roman authors writing between the third and first centuries BC (Ruiz Zapatero 1999).

Figure 11. Development of the Celtiberian culture (After Lorrio 1997a).

The difficulty lies in how to establish the origin of three different cultural components - ethnonyms and their ethnic background, material culture and language - without there being a strict correlation between them (Fig. 12). Hence, an ethnonym found in Classical texts does not mean that the people it designated would have recognized this name as referring to them. Moreover, this clearly does not imply that, before they came into contact with Roman culture, they did not recognize their own ethnic status by assigning a generic name to themselves. The ethnonym corresponds to a period in history, but we cannot neglect the possibility that before this time there was already such a term that classified them as an ethnic group. If this is the case, then it would have evidently been different from the name bestowed upon them by the Romans, since the term Celtiberi was created by the Romans to refer to an ethnic group with mixed cultural characteristics - Celtic and Iberian - which were expressly indicated by the Graeco-Latin authors themselves. Even though literary sources do not talk specifically about each of the ingredients contained within the mélange, we could hazard a guess that the Celtic components would include their language, religion and city names, all of which were arguments used by Pliny to characterize the Celtici in Baeturia as Celtiberians. The characteristics of the Iberian people, on the other hand, would include their urbanism, the defensive systems of their towns and cities and a large portion of their material culture.

It is simply impossible to search for the origin of an ethnonym at a time when most societies were illiterate. However, although it is difficult, we believe it is possible to look for the concept of Celtiberian ethnicity. The ethnic reality reflected by writers in antiquity has a few clear archaeological correlations despite the fact that links between ethnicity and material culture are complex, and finding them in archaeological records is complicated (Emberling 1997; Hall 1997; James 1997, 1999; Jones 1997). However, this does not deter us from examining this issue in relation to the Peninsular Iron Age (Lorrio 2000; Ruiz Zapatero and Álvarez-Sanchís 2002). The most serious problem is not the exploration of links between ethnicity and material culture when we know that there were ethnic realities; it is far trickier to use archaeological evidence to discover if such realities actually existed before they were mentioned in written texts. We do not know if there were ethnic groups who would recognize themselves as such before historical Celtiberia was defined. Even if they did exist, clearly these prehistoric or "ancient Celtiberians" would not have corresponded exactly to the Celtiberians of the Roman period, at least from the point of view of the ethnic reality. However, at the same time, there is no doubt that the former immediately preceded the latter. We find a similar situation with the Celtiberian language. Clearly, we cannot confirm anything with great certainty before the date of the most ancient inscriptions, around the mid second century BC. We can merely conjecture that in previous centuries they must have spoken a language that we could call Proto-Celtiberian, a term we could apply, for ease of reference, to any linguistic state prior to the Celtiberian we know through epigraphs (de Hoz 1993a: 392, Footnote 125). Another issue involves the origins of this 'proto-Celtic' language, and when it arrived in the eastern Meseta and the right banks of the Middle Ebro River.

It is only possible to trace material culture back along a continuous line to the early sixth century BC. Therefore, archaeology confirms that the basic elements of Celtiberian cultural reality were established from this phase onwards. These basic elements included ceramic and metallurgic traditions, typical weaponry, fortified settlements, and mortuary rituals - although these were quite diverse - and economic organization. Of course, this archaeological continuity does not imply that there was ethnic or even linguistic continuity. But certainly, if there are no sharp breaks in the archaeological record, it is difficult to prove the arrival of new populations and a new language at any point between the sixth century and the late third century BC. Therefore, we must attempt to recreate a historical process that would allow the existence of new ethnic and cultural contributions without a detectable change in the material culture record. Given the complexity of this possible process, the simplest and most coherent hypothesis would be to imagine that the formation of the Ancient Celtiberian ethnic group(s) and the adoption of a Celtic language coincided with the beginnings of an archaeological culture from the Ancient Celtiberian phase (ca. 600-450 BC). This hypothesis would then pre-suppose the existence of "prehistoric" Celtiberians who were different from the Celtiberians of the Roman period, yet somehow the same, at the dawn of the historical process during which this category evolved.

The archaeological record not only provides information about Celtiberia and the Celtiberians from the historical period, it also enables us to examine the process of the formation and evolution of Celtiberian culture and territory, a phenomenon that dates back centuries before the presence of the Romans in the area. Analyses of dwellings and burial grounds, as well as weaponry and crafts in general, have enabled us to establish the cultural sequence of the Celtiberian world (Lorrio 1997a: 258, Fig. 110). A general periodization has been established for this culture, based on burial records and various Celtiberian cultural manifestations. The proposed periodization comprises four phases that cover the period from the eighth to the first centuries BC: a formative period about which we only have sketchy information and for which we have reserved the term Proto-Celtiberian (ca. eighth/seventh century to mid sixth century BC); an initial or Early Celtiberian phase (ca. mid sixth to mid fifth century BC); a development or Middle Celtiberian phase (ca. mid fifth to late third century BC); and a final or Late Celtiberian phase (late third to first century BC). The aim of this is to bind the complex Celtiberian reality into a continuous sequence that can be applied to the whole of the Celtiberian territory.

4.2. Looking back: The origin of the Celtiberians and the 'proto-Celtic' culture

Terms such as Urnfield culture, Hallstatt, Post-Hallstatt or Celtic have frequently been used in an attempt to establish a connection with European archaeology, thus more or less accepting the invasionist theories which relate the formation of the Celtiberians to the arrival of successive waves of Celts from Central Europe (see Lorrio 1997a: 16-24). This thesis was defended by Bosch Gimpera (1932), who postulated the existence of various invasions in an attempt to combine the historical and linguistic sources with the archaeological evidence. This resulted in a constricting line of Spanish archaeological research, in view of the difficulty of correlating these cultures with those of the Iberian Peninsula, while the idea of successive invasions was not confirmed by the archaeological evidence (Ruiz Zapatero 1993: 45-47).

The linguists upheld the invasion hypothesis even though they were unable to offer any information on the chronology of these movements or the routes by which the invaders supposedly arrived. The oldest language to be considered pre-Celtic included Lusitanian, a language that some researchers consider a Celtic dialect, whilst the most recent is the so-called Celtiberian language, which was by that time fully Celtic.

The restriction of the Urnfield culture to the northeast of the peninsula, linguistically an Iberian area (non-Celtic and not even Indo-European), and the absence of that culture in Celticized areas, made it necessary to reconsider the invasionist theories. Even if these theories accepted the notion of a single invasion by the Urnfield culture, this still would not explain the phenomenon of the Celticization of the peninsula.

Throughout this entire process of investigation, linguists and archaeologists have worked separately. Archaeologists have tended either to look for exogenous elements that would prove the invasionist theory, or, even if they have not gone so far as to deny the presence of Celts in Iberia, to at least restrict the use of the term Celts to the evidence provided by the linguists and by epigraphs, regardless of the information provided in literary sources. In fact, the difficulty of correlating the linguistic data and the archaeological evidence has led the two disciplines to go their separate ways, which has made it difficult to obtain a holistic perspective, since a linguistic hypothesis cannot be fully accepted if it does not acknowledge the archaeological evidence, which in turn cannot be explained without a coherent evaluation of the linguistic information.

An alternative interpretation has been proposed by M. Almagro-Gorbea (1992, 1993, 2001). He takes as his starting point the difficulty of maintaining Bosch Gimpera's or Almagro Basch's theory that the origin of the Hispanic Celts can be related to the Urnfield culture, which did not spread beyond the north-eastern quadrant of the Peninsula (Ruiz Zapatero 1985), but who gave rise to a people whose culture and language were both Iberian, utilizing a language which was neither Celtic nor, apparently, Indo-European. In Almagro-Gorbea's view, their origin must be sought in the Indo-European 'proto-Celtic' substratum preserved in the western regions of the Peninsula, which existed in the Late Bronze Age, at the start of the first millennium BC. He considered this culture to be 'proto-Celtic' because its characteristic elements were Celtic. These elements were more archaic than the Central European Celts, since they were derived from Atlantic traditions that were almost certainly introduced during the period of the Bell Beaker culture. According to Almagro-Gorbea (2001: 101), this would explain the affinities that run along the Atlantic Celtic world, stretching from the British Isles to the Iberian Peninsula, whose origin and characteristics cannot be explained by Celtic invasions from Central Europe, hence it seems more appropriate to define them as "Indo-European". Such elements include the practice of offering weapons to water sources, creating rock altars, and leaving the bodies of warriors who had died in combat to be devoured by vultures. They could be linked to Indo-European ancestral religious rituals such as those described in the Lusitanian inscription at Cabeço das Fraguas, which refers to a sacrifice comparable to the Roman practice of souveturilia or the Hindu sautramani. They are also similar to those referred to by Strabo when he talks of hecatombs and human sacrifices (Str. 3,3,6-7; Liv. Per. 49). The deities, specifically those of the western Peninsula with names such as Bandua, Cosus, Navia, Reve or Trebaruna, displayed etymology and characteristics that could be considered Celtic.

For Almagro-Gorbea (1993: 146-152; 2001: 102), the Celtiberian culture emerged from this 'proto-Celtic' culture, although it was refined by the arrival of the Urnfield culture from the Ebro Valley to the Iberian Mountains, the future Celtiberia, which happened around the ninth or eighth century BC. This would explain the cultural, socio-economic, linguistic and ideological similarities between these two cultures, and the progressive assimilation of the 'proto-Celtic' substratum by the Celtiberians. From the sixth century BC on, the Celtiberians began spreading further afield, absorbing the Atlantic 'Proto-Celtic' substratum - or perhaps we should refer to it as an Indo-European culture, for the aforementioned reasons - and exploiting these affinities. Thus, the Iberian Celts emerged from a long and complex process of Celticization, which did not exclude ethnic movements such as those occasionally referred to by the literary sources. However, the influence of one dominant population, as proposed by invasionist theories, is no longer considered to be a crucial element in explaining the emergence and development of the Iberian Celts. This process, which can be tracked through the changes registered in archaeological findings, indicates that the Celtiberians evolved, and spread out over the peninsula, which led to the acculturation of other substrata, based on the Celtiberians' socio-cultural organization of warrior elites. According to Almagro-Gorbea (2001: 110), this process would have intermittently unravelled over a long period of time, and would have varied in different areas, depending on the substratum, the period, the intensity and the duration. Hence, the Celticization of Iberia could be conceived of as a cumulative process that developed "mosaic-style", which would certainly explain the personality and diversity of the different Celtic populations in the Iberian Peninsula (Fig. 13).

Figure 13. The "mosaic model" for the Celticization of Iberia from the Celtiberian nuclear area (Drawn after Almagro-Gorbea data, pers. comm.).

However, our deeper knowledge and understanding of the Late Bronze Age and the Iron Age in the Iberian Peninsula has meant that we have been able to determine the undeniable influence of the Urnfield culture from the northeast on the formation of the Celtiberian world. This phenomenon is confirmed by burial rituals, important traditions of ceramics and metallurgy, and certain fortified settlements. The ancient Celtiberian world emerged out of the interaction between the socio-economic model imposed in the eighth and seventh centuries BC by the northeastern peninsular Urnfield culture, and the indigenous culture, which played an important role in this process (Ruiz Zapatero and Lorrio 1999: 26, 34).¹⁸

There is evidence of ethnic contributions originating in the Ebro Valley and the highlands of the eastern Meseta (Arenas 1999: 246; Lorrio 1997a: 260-261, 372; Ruiz Zapatero and Lorrio 1988, 1999: 26, 34).¹⁹ Therefore, the possibility that these infiltrating Urnfield groups may have brought with them an Indo-European language should not be rejected, even if their true role in creating the Celtiberian world has yet to be determined.

Be that as it may, the trans-Pyrenean origin of the Urnfield groups of the northeast would seem beyond all doubt; the penetration, at least it its initial phases (which can be dated around 1100 BC) of human groups, which were of little importance in demographic terms, is generally accepted (Maya and Barberá 1992; Ruiz Zapatero 1985, 1997). In view of the continuity of the material culture in the northeast through the first millennium - if the Indo European character of this contribution is accepted - a possible explanation of the continuing dominance of an Iberian language at a late date is what Villar (1991: 465-466) calls "failed Indo-Europeanization". According to this theory the Indo-European languages of the northeast, probably in the minority, must have waned as those who spoke them were culturally and linguistically "Iberianized". That at least some of the Urnfield groups spoke an Indo-European language of a Celtic or 'proto-Celtic' type seems to be a very credible proposition, as the case with Lepontic would later confirm: a Celtic language spoken in the north of Italy from at least the sixth century BC linked with the Golaseca culture that shared its roots with an Urnfield group culture, the Canegrate culture (de Hoz 1992b; de Marinis 1991). Thus, it could be postulated either that "the migratory Urnfield groups were so small that they did not succeed in imposing their own language on the peoples of the substratum, or the Iberian transformation largely erased the Indo-European linguistic features that had hypothetically been assumed by the natives" (Maya and Barberá 1992: 176).

However, Almagro-Gorbea (forthcoming) does not believe that the solution is to trace the origins of the Iberian Celts back to the Urnfield people or that the classic theory proposing that the Celts arrived on the peninsula with the cultures of Hallstatt D or La Tène is tenable. These theories do not explain the early arrival of Celts in western Iberia. According to Almagro-Gorbea (1994 and forthcoming), the best solution would be to "link the Celts with a wide, fluid and polymorphous Atlantic Bronze Age 'proto-Celtic' culture".

5. The Iberian Celts: a multidisciplinary interpretation

After all the different elements relating to the issue of the Celts in the Iberian Peninsula have been analyzed, we can conclude that the reality of the Celtic community in this area was a highly complex issue that included the following features:

1. The homogeneity of Celtic Hispania as opposed to Iberian Hispania, based on characteristic onomastics, which were Celtic or at least Indo-European in origin.
   
   
2. The existence, in Celtic areas of the Peninsula, of disparate linguistic evidence. Some of this is clearly Celtic, such as Celtiberian, an archaic Celtic language, recorded in a series of epigraphic documents in the Iberian or Latin alphabet. But there were also other languages that are more difficult to define, at times Celtic and at times Indo-European, like Lusitanian. Furthermore, Celtic names have also been found in inscriptions written in Tartessian.
   
   
3. The extraordinarily complex picture painted by literary sources, which vary in both content and usefulness. These sources describe the nature of Celts such as the Celtiberians and Berones. At other times, we can draw our own conclusions based on the use of the ethnonym Keltikoi or Celtici, which is applied to several populations from the southwest and northwest of the Peninsula, or based on their links with the Celtiberians, as is the case with the Celts of Baeturia. Similarly, it is worth highlighting the cultural proximity between the Celts and other populations in the central and western Peninsula, as is possibly the case with the Vettones, Vaccaei, Cantabri or the Astures, which confirms the linguistic and onomastic evidence.
   
   
4. The presence of Celts, according to literary sources, does not appear to be limited exclusively to the areas in the central and western Peninsula, and they are described in other areas such as the south, which are generally considered to be Iberian. Pliny even mentions a Celtic population in Betica. The northeast seems to be a different case, where the presence of Celts appears to be linked to influences from Gauls from the north of the Pyrenees, as indicated by their toponyms ending in -dunum, the names of some of their rivers, such as the Gállego or the Gallo, or the existence of a pagus gallorum, among other things.
   
   
5. Finally, material culture enables us to paint more individualized portraits of some of the "Celtic" archaeological groups, by recording certain elements, generally objects of prestige with considerable ideological value such as horse collars, or bi-globular type daggers, whose geographical distribution coincides with Celtic Iberia as defined by other documentary sources analyzed. As for their origin, we could justifiably defend the Celtic nature of those archaeological groups whose origin dates back at least to the dawn of the Early Iron Age at the turn of the sixth century BC, and which remain continuously up to the period of the wars with Rome. These are groups whose nuclei are situated in areas where at a later date historical Celtic populations are known to have lived, and in which there is also evidence of a Celtic socio-political organization and religion, as well as linguistic proof that a Celtic language was spoken.

In general, the first four elements discussed chronologically take place quite late, between the late third and first centuries BC, or even later than that. This evidence provides a relatively uniform panorama, which perhaps is not such a far cry from the process of Celtibericization described by literary sources and recorded in the archaeological record, which could conceal the prior Celticity of some of the areas of the center and west of the Peninsula.

Only archaeological records allow us to examine the stages immediately prior to this period with any certainty. Although some literary sources refer to this period with clarity, most of them are much vaguer in their accounts, and we are unable to determine how far back the onomastic and linguistic elements described can be dated. Presumably, some of the elements described that were present when the Romans conquered Hispania and during the subsequent Romanizing process already existed in a more developmental stage during these earlier periods. Some researchers believe that their origin could date back to the beginning of the Iron Age, around the seventh or sixth century BC, and they may even be older than that. The Celtiberian zone is the most appropriate area to analyze the process of formation of the populations described by Graeco-Roman historians. However, new theories proposed by Almagro-Gorbea, for whom the origins of the Celts lies somewhere at the dawn of the Bronze Age towards the second half of the third millennium BC, raise new questions, and demand a complete reexamination of studies of Celtic populations in Iberia.

6. "Celtic" archaeological groups (Fig. 14)

6.1. The Celtiberian group

According to the literary sources, Celtiberia was a large area in the interior of the Iberian Peninsula (see below). An analysis of the cemeteries, and also the weapons and the material culture, has made it possible to establish the cultural sequence of the Celtiberian world, from at least the sixth century BC to the Roman Conquest (Fig. 15) (Lorrio 1997a).

Figure 14. "Celtic" archaeological groups in the Late Iron Age of Iberia.

The Early Celtiberian period (ca. 600-450 BC) is documented in the highlands of the eastern Meseta and the Iberian Mountains, in the regions of the Upper-Douro-Upper Tagus-Upper Jalón Rivers. The new Celtiberian culture emerged in this area. This new archaeological group could be defined by the following elements: 1) the appearance of the first hill-forts (locally called castros, small settlements in high areas with natural defences and man-made defensive structures); 2) the establishment of the first cremation cemeteries (Burillo 1990); and 3) a whole range of ceramics and new metallic objects, many of which were forged from the new metal, iron. Such objects had not previously existed in local pottery and metallurgy traditions (Burillo 1991; Ruiz Zapatero and Lorrio 1999). The distribution of this Celtiberian group in the highlands of the eastern Meseta may have been linked to the possibility of raising cattle there, the availability of salt resources and the exploitation of iron ores in the region (Lorrio 1997a: 264-270).

A settlement pattern appears to have emerged at this early stage that would continue to be used right through to the end of the Celtiberian period (Fig. 15). This pattern was based on the almost exclusive use of castros, which were distributed all along the river valleys on hilltops, hillsides and in river breakwaters (Arenas 1999; Burillo 1995; Lorrio 1997a: 65-71; Ruiz Zapatero 1999). Even if we only know about them from surface data, there must have also been a few open settlements on the plains. However, the characteristic habitat was certainly the hill-fort. The earliest hill-forts were very small (0.2 - 0.5 hectares), rarely bigger than one hectare (Lorrio 1997a: 67). The natural defenses of their settlements were further improved by several methods: they sometimes dug out defensive ditches, or set in a belt of closely sunk stones (chevaux-de-frise), but they almost always built dry-stone walls (Romero 2003). We have very limited information on their internal organization, but it would seem that the most prominent features were those found at La Coronilla (Cerdeño and Garcia Huerta 1992): the houses were rectangular and built of stone, which would remain the model of a Celtiberian home until the Late period (Cerdeño and Juez 2002). The adjoining walls were closed to the outside to form a rampart, and doors opened onto the interior of the hill-fort. The largest settlements had scarcely over a few hundred inhabitants. There was no settlement hierarchy, and the communities were small, fairly homogeneous and self-sufficient (Ruiz Zapatero 1999: 109).

Figure 15. From ancient Celtiberian groups to Celtiberia (After Ruiz Zapatero and Lorrio 1999).		Figure 16. Tombs with stelae in straight rows in the cemetery at Centenares (Luzaga, Guadalajara). (Photo: Museo Cerralbo).

The first Celtiberian cremation cemeteries were located close to settlements, mostly at a distance of between 150 meters and just over 300 meters (where both were in clear view of the other), and generally near rivers or streams (Lorrio 1997a: 111-114), a fact almost certainly linked to the symbolic liminal function represented by the course of rivers in the Celtic world. The bodies were incinerated in the cemeteries, although the precise locations of the ustrina are difficult to pinpoint, and the cremated bones, adornments, weapons and other personal belongings were placed in urns, and buried in graves. Sometimes the ashes were buried without an urn, in containers or bags made out of organic material. Furthermore, in some burial sites, the location where an urn was buried was marked with a stone stele (Lorrio 1997a: 125, Fig. 50). Early excavations have provided us with knowledge of a ritual that arranged the tombs in straight lines, like streets, as in the cemeteries of Aguilar de Anguita and Luzaga (Fig. 16) (Lorrio 1997a: 114-117). The inequality of the grave goods or funeral offerings reveals the incipient social differentiation practised by these peoples, as well as the presence of a "warrior" class, whose tombs can be distinguished by the weapons contained therein (Fig. 17). This early social elite was buried with relatively austere items: long spearheads, curved knives, and double-sprung fibulae (brooches). During this early period, there were no swords or daggers in the graves (Lorrio 1997a: 262).

Figure 17. Evolution of the funerary equipment in Celtiberian cemeteries (After Lorrio 1997a).

During the Middle Celtiberian period (ca. 450-225/200 BC), the Celtiberians secured and stabilized the areas settled in the Early period, and began to occupy new lands such as the right bank of the Middle Ebro Valley (Capalvo 2001). The settlements grew in size (some of them between two and five hectares) and number, and the number of cemeteries increased (Fig. 18). All of this indicates sustained demographic growth and systematic occupation of territory. The continuity between this period and the Early phase bears witness to the fact that the communities were increasingly tied to the land, and had begun to adopt a form of territorialization. The majority of the hill-forts had a central road or space. The houses were rectangular, with adjoining walls and were grouped into blocks. New defensive methods were implemented such as double walls, angled walls that date as far back as the siege warfare of Hellenistic times such as Guijosa (Guadalajara) (Lorrio 1997a: 90, Fig. 28,1), and rectangular towers and ditches are more widespread than in the previous period. These defensive methods would be fully developed in the Late period, such as the ditch built at Contrebia Leukade (La Rioja), which was particularly spectacular (Fig. 19). It was excavated into bedrock, and its perimeter was nearly 700 meters, with a width of between seven meters and nine meters, and a depth of eight meters. The volume of rock excavated was around 40,000m³, which involved an enormous amount of collective work (Hernández Vera 1982: 122).

As previously mentioned, there was a large increase in cemeteries, something which indicates that the communities had increased in size. The number of tombs in some of these burial sites gives us an idea of this (Fig. 20): approximately 5,000 in Aguilar de Anguita, nearly 2,000 in Luzaga, 1,200 in Gormaz (Soria), and over 800 in Osma and Quintanas de Gormaz (Soria). Some others were smaller: Almaluez, Alpanseque (Soria), and Arcobriga (Saragossa) had 300 tombs and La Mercadera (Soria) only 100 (Álvarez-Sanchís and Ruiz Zapatero 2001: 67-70). Unfortunately, since most of the burial sites were excavated in the first few decades of the twentieth century the records are inadequate, and regrettably there are only a few sites available to us now that have not been excavated. The obvious inequality exhibited by the grave goods bears witness to a process of clear social differentiation. During this period, we see the appearance of aristocratic tombs with very complete panoplies (Lorrio 1997a: 275-283, 314). The "wealth" of the tombs can be measured by the number of objects left as offerings as well as the amount of work required to construct them. This allows us to establish the order of social ranks. Those who were buried with numerous weapons and horse harnesses belonged to the social elite, and would have been mounted warriors or equites. The next highest rank consisted of warriors who were buried with expensive weaponry including swords, knives, lances, shields, and bronze breastplates. The tombs of the third highest rank to be buried with weapons included more modest grave goods, only lances and shields, and occasionally a sword. Naturally, most of the graves belonged to the general population, who were buried without weaponry and with varying degrees of tomb "wealth". We have very little knowledge of the internal organization of the burial sites, since only those that have been excavated recently have been surveyed properly, as at Carratiermes (Argente et al. 2000), and El Raso (Fernández Gómez 1997). At least two types of cemeteries can be discerned (Lorrio 1997a: 144): in the first type there is a wide range of wealth, and a pyramidal structure with a small number of aristocratic tombs at the pinnacle. Despite the aforementioned problems with earlier excavations, the majority of burial sites seem to fall into this category. The second type of cemetery is much less common and can be categorized as "fairly wealthy" to "very wealthy" with a high percentage of tombs containing weaponry, for example La Mercadera (44%) and Ucero (35%). There is a great deal of variation between Celtiberian cemeteries, and it is tempting to link this with possible funeral customs of different Celtiberian groups with distinct identities (Lorrio 1997a: 146).

Figure 20. Demographic estimates for Celtiberian cemeteries: living community size estimated from mortuary population. In black, correction factor of 10%; in white, 20%. Each figure represents ten individuals. (After Ruiz Zapatero and Álvarez-Sanchís 1999).

Figure 21. Celtiberian pottery with polychrome painting from Numantia (Soria). (After Jimeno et al. 2002).

The Late Celtiberian period (ca. 225/200 - first century BC) was undeniably a period of transition and profound change, largely as a result of the "clash" against Rome, which would eventually lead to the fall of the Celtiberians. As a result of native evolutionary processes, the Celtiberian communities adopted more urban ways of life. Thus, cities began to emerge that were over 10 hectares and sometimes even 20, with public buildings, monumental architecture, and sometimes a grid layout. At times these cities emerged via a process of synecism, as in other areas of the Mediterranean. The Classical sources (Liv. 40, 33 and 47; Str., 3, 4, 13) mention all the settlement size categories, from the smallest villages to the big cities, and classify them as follows: towers (turres or pyrgoi), villages and castles (vicos castellaque), large villages (megalas komas) and cities (urbs, poli, civitas or oppida). However, it is difficult to reconcile this information with the archaeological remains that have been found (Burillo 1998: 213-214; Lorrio 1997a: 319-321). The hierarchical order of cities in the Middle Ebro Valley has been established according to which cities minted silver coins (Burillo 1998: 237, 294-297). A further innovation was the use of writing by the Celtiberians, first on coins from the mid-second century BC (Domínguez 1988), and later in other materials (Rose 2003) including the laws written on bronze to be publicly displayed, such as in Contrebia Belaiska (Saragossa) (Beltrán Lloris 2001). For this they first used the Iberian and later on the Latin alphabet. In cemeteries weapons seem to disappear from the tombs, which is a sign that funeral ideology had changed along with the criteria for defining social status. Precious metal objects, collected in family treasure troves or community depositories and not in the graves themselves, replace weaponry as a mark of social status (Delibes et al. 1996). A strong Iberian influence can be observed in bronze and pottery objects, such as the ones from Numantia (Figs. 21-23), which gives them a definitive personality within the Celtic world to which they belonged, as their stylistic and ideological elements attest. The development of a truly monumental architecture should also be mentioned (Fig. 24).

Figure 24. The horreum building at Contrebia Belaisca (Saragossa) from the north. At the front, the colonnade and shafts of columns and capitals indicating that it was abandoned. (After Beltrán Lloris 2001).

The Celtiberian communities of the second and first centuries BC were linked via the oppida, which acted as central units (Burillo 1998: 210). The cities had effective control of the surrounding land, and also housed socio-political institutions such as the Public Assembly, and the Councils of Elders or Nobles, which are referred to in the sources as Senates (Fatás 1987). Above the cities were the ethnias or populi, which according to Classical authors made up the Celtiberian collective. Although references to the populi do not always agree, it seems that the Celtiberian collective was made up of the Arevaci, the Lusones, the Belli, the Titthi and probably the Pelendones. The relationships between the Celtiberian ethnias varied, and included specific alliances at the time of undertaking military action (Burillo 1998: 225 ff.).

Classical sources and epigraphy also provide information about other institutions such as the hospitium, clientship and devotio (Ramos Loscertales 1924, 1942; Salinas 2001b). The hospitium involved the acceptance of a stranger by a family group or a particular community. The official recognition of this action came in the form of the hospitality tessera, two sheets of bronze (or silver in exceptional circumstances) with inscriptions in Iberian or Latin, which were joined together to form one single piece. They included geometric or animal shapes, or the image of two hands joined together. As for content, some had short inscriptions that referred to one of the participants, whereas other inscriptions were longer, mentioning both parties involved in the pact (Sánchez Moreno 2001; Tovar 1948). Clientship was a device for the purpose of integrating people of different social standing that involved mutual obligation; in general the patron offered protection and material compensation in exchange for services from the clients, which included military services. This institution existed between members of the upper social ranks and groups of lower social status, as well as between ethnias or populi. The devotio was a formula that contained a strong religious and ideological component, whereby the warrior would swear to protect the life of his superior and even give up his own life in order to do so.

Figure 25. The Last Day of Numantia (133 BC). Painting by Alejo Vera (1881) at the Diputación of Soria. Partial view.

When the Romans arrived in the Meseta in the second century BC, the Celtiberians were expanding into several peripheral areas, as they had been doing during previous centuries, spreading outwards from the eastern Meseta. This process was particularly intense towards western regions of the Iberian Peninsula, where genuinely Celtiberian objects such as horse collars and bi-globular daggers have been found. These objects have been discovered as far as the southwest (Celtic Baeturia) and provide archaeological confirmation of the migrations of Celtici indicated by the written sources as well as linguistic and epigraphic evidence (Berrocal-Rangel 1998).

The Celtiberian resistance to the Roman occupation is exemplified by Rome's blockade of Numantia. In the summer of 133 BC this heroic Celtiberian town was destroyed by the Roman general Scipio after nearly a year of siege (Fig. 25) (Jimeno et al. 2002: 90-91). The example of the Numantians has been widely used in modern and contemporary "public use of Spanish history" (de la Torre 1998).

6.2. The Vetton group

The Vettones were one of the most important populi in Celtic Hispania (Álvarez-Sanchís 1999; Salinas 2001a; Sánchez Moreno 1996, 2000). Roman writers located them in the gently undulating plains and mountains of the western Meseta (Fig. 26), especially between the Tormes, Douro and Tagus Rivers (Roldán Hervás 1968-69). After the disappearance of the Late Bronze Age Cogotas I culture, there followed a period about which knowledge is still patchy (Álvarez-Sanchís 2000). This era was characterized by castle settlements such as Los Castillejos de Sanchorreja in the mountainous areas, and by open settlements in the land next to the Douro River that were similar to those of the Soto culture (from the Early Iron Age in the Middle Douro). The communities from the mountain areas were tied to the land, an unknown phenomenon up to that time, and received imports and products from the Tartessians in the south of the Peninsula, who in turn were influenced by trade with Phoenician merchants who had established themselves in Andalusia's coastal regions. Throughout the fifth century BC an important transformation occurred in the western regions of the Meseta: firstly, some of the most important nuclei were abandoned, and secondly, there was a drastic decrease in the exchanges that had arrived from the south, after the fall of the Tartessians.

Figure 26. The cultural area of the Vettones in Western Meseta with main settlements and the boundaries with other populi according to Classical written sources (After Álvarez-Sanchís 2001).

At the beginning of the fourth century BC, a series of changes came about that have been attributed to a process of Celticization whose origins should be sought in the Celtiberian zone. These changes were as follows: an increase in fortified settlements, many of them built following new designs and on a much bigger scale than before; an expansion of the ritual of cremation and cemeteries with weapons; and lastly, innovations in material culture such as the widespread use of iron metallurgy and the first ceramics made on a potter's wheel. This archaeological group has been named Cogotas II and also "Culture of the Verracos (boars)" owing to the discovery of characteristic crude zoomorphic sculptures made out of granite, an abundant material throughout the mountainous areas of this region (Álvarez-Sanchís 1994, 1999: 215-294; López Monteagudo 1989). The great reorganization of the social landscape, begun in the fourth century BC, would continue into the second century BC. According to historical references (Roldán Hervás 1968-69), the Vetton territory coincided with the region occupied by Cogotas II or the "Culture of the Verracos". Therefore we can deduce that this group could be identified with the early Vetton period, following the same argument used above for the Celtiberians (Sayas and López Melero 1991).

Land seems to have acquired a new economic value, and in the case of the Vetton people, pasture fields were particularly prized given their inclination to raise cattle. This enables us to understand the intense occupation of optimal pasture areas such as the Amblés Valley or the Yeltes and Agueda Rivers in the Salamanca region (Álvarez-Sanchís 1999: 115-122). There was a great deal of compartmentalization of land between communities that lived in large hill-forts (which could be better termed oppida), and the verracos sculptures were used as landmarks to divide up pasture areas.

Figure 27. Stone walls and chevaux-de-frise at the hillfort of Las Cogotas (Ávila) (Photo: A.J. Lorrio).

Figure 28. Oppida and settlement sizes in the Vettones area (After Álvarez-Sanchís 1999).

The most characteristic elements of the Vettones were the hill-forts or oppida, their cemeteries, the famous zoomorphic sculptures, certain types of ceramics (such as those decorated with a combing technique), and certain weapons and metallic adornments (Ruiz Zapatero and Álvarez-Sanchís 2002). The Vetton settlements were mostly new in design and layout, and were located on hilltops, in river breakwaters, meanders, and on valley slopes. They were between 20 and 80 hectares in size with good defenses (Fig. 27) consisting of moats, stones sunk into fields (chevaux-de-frise), and stone walls, sometimes with solid bastions (Fig. 28) and several walled enclosures which are typical of the Vetton area (Cabré 1930; Cabré et al. 1950; Fernández Gómez 1986 Vol. I; Ruiz Zapatero and Álvarez-Sanchís 1995). There was no hipodamic urbanism, and the housing was usually made up of free standing square or rectangular units adapted to the topography of the interior of the hill-forts. In a few cases, houses were erected outside the walled enclosures, and Classical sources refer to the existence of an outlying area when Hannibal conquered Salmantica in 220 BC. On the whole, the internal organization was quite different from the Celtiberian urban model (Ruiz Zapatero and Álvarez-Sanchís 2002). In the oppidum at Ulaca (Ávila), a possible sacred or ritual area has been documented that contains two structures carved into the rock: a sanctuary with a sacrificial altar, and a sauna (Ruiz Zapatero and Álvarez-Sanchís 1999). The sanctuary is a rectangular space carved out of the rock that was probably covered. Next to this a large cliff was used to create access to the upper part with a double set of steps, a type of platform with several intercommunicating hollows through which liquids could flow (Álvarez-Sanchís 1999: 310). The sacred nature of the structure can be discerned owing to the parallels with the Roman world, the most expressive example of which is the Portuguese sanctuary at Panóias (Rodríguez Colmenero 1999), with Latin inscriptions that refer to the animal sacrifices carried out there. The victims' entrails were burned in hollows or casks, and the blood was poured into other cavities to flow down to the lower part of the rock, whilst the deities were worshipped. The sauna at Ulaca is a rectangular structure, also cut into the rock, which has three areas: a wide antechamber, a small chamber with two benches on each side, and the oven or fire chamber. It has been linked with the initiation baths of warriors, such as the "pedras formosas", the famous saunas of the "Cultura Castreña" or Castro Culture in the northwest (Almagro-Gorbea and Álvarez-Sanchís 1993). It would certainly confirm Strabo's (3, 3, 7) reference to steam baths used by the inhabitants of the areas surrounding the Douro River.

The best-known Vetton burial sites were excavated in the first decades of the previous century: Las Cogotas (1,613 tombs) (Cabré 1932; Castro 1986; Kurtz 1987) and La Osera (2,230 tombs) (Baquedano 2001; Cabré et al. 1950). A few recent excavations have broadened our knowledge of the Vetton funeral culture, even if far fewer graves have been found in these contexts: El Raso (123) (Ávila) (Fernández Gómez 1986 Vol. II, 1997) and other burial sites in Extremadura. Vetton cemeteries have a series of peculiar characteristics (Álvarez-Sanchís 1999: 172): 1) the use of cremation ritual; 2) tombs with urns in graves, sometimes with small protective rocks and occasionally with stelae (Las Cogotas), and in small burial mounds weighted down with stone (La Osera, La Coraja); 3) the location of cemeteries opposite to and near the settlement gates (150-300 meters); 4) cemetery proximity to streams; 5) concentrations of graves in well-marked areas with empty spaces between them (four in Las Cogotas and six in La Osera), which have been interpreted as the mortuary plots of family or kin-groups. As far as we know, there are no cemeteries at the westernmost tip of the Vetton territory, perhaps because they have not yet been discovered, or more probably because they may have shared the same funeral customs as the areas throughout the Atlantic coast and the northwestern area of the peninsula that have not produced any evidence for cemetery areas and whose mortuary rituals have left no recognizable archaeological trace (Ruiz Zapatero and Lorrio 1995).