In the last 30 years, there has being a growing interest in understanding how cultural remains entered the archaeological record (eg, Schiffer Reference Schiffer1987; Tani Reference Tani1995; LaMotta & Schiffer Reference LaMotta and Schiffer1999; Chapman & Gaydarska Reference Chapman and Gaydarska2007; Jiménez Jáimez Reference Jiménez Jáimez2007; Bailey Reference Bailey2007; Lucas Reference Lucas2012; Wolfram Reference Wolfram2013). In fact, this topic is widely regarded as a key subject whose discussion should be addressed in advance of any functional or spatial accounts. However, this is a completely neglected issue in many European archaeological traditions, such as central and eastern European and Mediterranean later prehistory. Assemblage-based formation studies have mainly relied upon faunal and lithic debris so that the contribution of pottery in taphonomic terms remains under-exploited, save in some avant-garde milieux, especially within British archaeology (eg, Bradley & Fulford Reference Bradley and Fulford1980; Sørensen Reference Sørensen1996; Garrow et al. Reference Garrow, Beadsmoore and Knight2005; Brudenell & Cooper Reference Brudenell and Cooper2008; Mercer & Healy Reference Mercer and Healy2008; Lamdin-Whymark Reference Lamdin-Whymark2008; Edwards Reference Edwards2009; Reference Edwards2012; Beadsmoore et al. Reference Beadsmoore, Garrow and Knight2010). Yet even these seminal contributions often focus on the idiosyncrasy of the ceramic assemblages at hand and their procedures are designed for the interpretation of these particular cases studies. The methodological aspects and the criteria for conducting such evaluations or their applicability to further collections are not properly tackled and made explicit.
An analytical protocol is presented to characterise the preservation condition as well as the representation and breakage rates of pottery collections. The variables for conducting such post-excavation operations are carefully discussed here. The ceramic assemblage from a residue-rich depositional context excavated in 1981 has been re-examined. Dated to the late 3rd millennium bc, the El Ventorro site (Madrid, Spain) is well-known internationally for its remains that have stimulated opposing and controversial archaeological readings based on the same ambiguous evidence. Its study has wider implications because of the extraordinary occurrence at the same site of pan-European later prehistoric phenomena such as Bell Beakers and monumental earthworks. Thus, this test-case serves to challenge the formational dynamics of widespread Neolithic locales, such as pit sites and ditched enclosures, as well as the changing roles of Bell Beakers throughout an unexpected range of biographical steps. The assessment of an unusually large ceramic sample – nearly 4000 potsherds, including over 100 Beaker sherds – examined from an integrated, systematic, and explicit hands-on methodology has allowed the evaluation of the fitness of several interpretive hypotheses. This study has ultimately led to an estimation of the potential of such a procedure and its applicability to prospective ceramic collections, even to those unearthed by old archaeological excavations.
LOCATION AND AVAILABLE READINGS
El Ventorro is located in the province of Madrid, in the sedimentary tablelands of the Tagus valley in central Iberia, to the south of Madrid city, on the right-bank floodplain of the Manzanares river (Fig. 1). This setting has been intensively distorted by extensive gravel exploitation as well as urbanisation projects, but many later prehistoric sites are known in its surroundings (eg, Muñoz Reference Muñoz2002). The site was subjected to successive rescue excavations in the early days of such activity in Spain (1962, 1972, 1977, and 1981) (Fig. 1C). These campaigns documented explicit remains of dwelling, subsistence, and craftsmanship activities – large collections of faunal, ceramic, and lithic debris – spatially linked to an outstanding collection of Ciempozuelos-style Bell Beaker pottery, all within several sunken features (Quero & Priego Reference Quero and Priego1975; Priego & Quero Reference Priego and Quero1983; Reference Priego and Quero1992). Particularly remarkable at a pan-European scale was the finding of a metallurgical workshop containing copper smelting by-products and crucibles with Beaker incised decoration – results rapidly presented to an international audience and discussed thereafter (Harrison et al. Reference Harrison, Quero and Priego1975; Harrison Reference Harrison1977, 178–80; Garrido-Pena Reference Garrido-Pena1997, 202; Reference Garrido-Pena2000, 43–4; Harrison & Mederos Reference Harrison and Mederos2001, 121).
Fig. 1 Location of El Ventorro. A. Sites mentioned in the text: 1. Las Pozas (Zamora); 2. Camino de las Yeseras (Madrid); 3. El Ventorro (Madrid). B. El Ventorro in the Manzanares valley. C. Excavation trenches at El Ventorro (after Díaz-del-Río Reference Díaz-del-Río2001, 240, fig. 61)
The above-mentioned fieldwork seasons unearthed 228 m2 in four adjacent sectors, allowing the recording of three elongated gullies dug into the rock – interpreted as ‘pithouses’ – and 23 pits in their surroundings. Initially, the site was envisaged by its excavators as a permanent settlement composed by the aggregation of several households living together in semi-sunken huts made up of wattle-and-daub structures with several annexed multi-purpose underground wells (Priego & Quero Reference Priego and Quero1992). Such a locale presented an uninterrupted occupation throughout the second half of the 3rd millennium bc, according to the few available radiocarbon determinations. This allegedly prolonged occupation led to two discernible superimposed phases, coincident with the internal sub-division of the Iberian Chalcolithic in the 1980s: a Beaker phase in the uppermost layer and a pre-Beaker phase beneath that (Priego & Quero Reference Priego and Quero1992, 357–64).
We shall focus on the third excavation campaign, carried out in 1981 (Fig. 2). The excavated features comprised five pits – Pits 008–012 – and an elongated gully – 16 m long, with 44 m2 excavated –, with an irregular section, 4 m in width and 1.2 m deep, the so-called ‘Pithouse 013’ (Priego & Quero Reference Priego and Quero1992, 83–125; Díaz-del-Río Reference Díaz-del-Río2001, 243–5). This cut feature was filled with several thin ashy and sandy layers, interpreted as a slow build-up of successive occupation floors pertaining to two recognisable dwellings – the deeper pre-Beaker and the upper Beaker-phase occupation – partially superimposed and substituted in a span of few years without stratigraphic hiatus (Priego & Quero Reference Priego and Quero1992, 104–5). Within the ‘uppermost dwelling’, some structures or arrangements of patterned remains were identified, such as a cache of potsherds regarded as a ‘soil’Footnote 1 , a hearth, and several activity areas such as a metallurgical oven with copper smelting by-products and a flint knapping workshop (Priego & Quero Reference Priego and Quero1992, 103–4; 123–5, lams. xx–xxii). In short, these occupation layers were seen as forming inside thatched huts, which despite being regularly cleaned, preserved in situ structures and activity areas, along with fallen or displaced debris in primary position (Priego & Quero Reference Priego and Quero1992, 83–125). Subsequently several authors have commented on such a hypothetical reconstruction of what has come to be a paradigmatic case of a two-phase Chalcolithic dwelling in central Iberia (Garrido-Pena Reference Garrido-Pena1997, 190; Reference Garrido-Pena2000, 43–4; Harrison & Mederos Reference Harrison and Mederos2001, 121; Rojo et al. Reference Rojo, Garrido-Pena and García2005, 525).
Fig. 2 Sector excavated in 1981 at El Ventorro (after Priego & Quero Reference Priego and Quero1992, 85, fig. 32)
Díaz-del-Río (Reference Díaz-del-Río2001, 237–50; 377–8) conducted a critical reassessment of the published data and concluded by offering an alternative account. His chief objections were: (A) the sequence of two Chalcolithic phases for ‘Pithouse 013’ is invalidated by the arbitrary criteria used to support it, as this was a non-stratigraphic excavation, that followed fixed-depth layers of 10 cm which revealed clearly sloping strata in section (Priego & Quero Reference Priego and Quero1992, 101, fig. 39; 103, fig. 41). Moreover, phase discrimination was based only upon the presence/absence of Bell Beaker and metallurgical elements (Díaz-del-Río Reference Díaz-del-Río2001, 244–5); (B) there are no true in situ structures or diagnostic elements of the domestic realm and those appearing – irregular post-holes – might be regarded as an open-air windscreen, indicative of outdoor activities (Díaz-del-Río Reference Díaz-del-Río2001, 246–7); and (C) the extraordinary amount of remains yielded by ‘Pithouse 013’, containing, as it did, the highest accumulations of flint flakes (n=2792), faunal debris (n=3283), granite querns (n=24), and potsherds (n=33,595, including 106 Beaker sherds) per sq. m in Iberia and beyond (Díaz-del-Río Reference Díaz-del-Río2001, 246; Reference Díaz-del-Río2006, 73). (D) In view of these extraordinary quantities of items within ‘Pithouse 013’, he posited its interpretation as a massive midden resulting from successive dumps of large volumes of refuse disposed of during festive social gatherings (Díaz-del-Río Reference Díaz-del-Río2001, 246–50; Reference Díaz-del-Río2006, 73).
ANALYTICAL METHODS
Our main goal has been to assess the various explanations by examining a sample of the material evidence. Thus an analytical protocol has been designed, inspired by recent attempts to deal with the depositional histories of ceramics (Garrow et al. Reference Garrow, Beadsmoore and Knight2005; Brudenell & Cooper Reference Brudenell and Cooper2008; Edwards Reference Edwards2009; Reference Edwards2012; Beadsmoore et al. Reference Beadsmoore, Garrow and Knight2010; Wolfram Reference Wolfram2013). This has been directed to shed new light upon overarching issues such as the composition, origin, timing, and degree of intentionality of such deposits; the representation, dispersion or loss of parts of the same vessels, or the trajectories undergone by sherds after their break and the time elapsed before being deposited into the cut features. The huge number of potsherds found in 1981 – over 36,500 items – made it necessary to choose a representative ceramic sample which could be studied using intensive methods of analysis and recording. Thus, our efforts focused exclusively on feature sherds – decorated sherds, rims, carinations, bases, and handles – recovered within Pits 008–012 and in squares B-1, C-1, D-1, and D-2 for the area of the ‘Pithouse 013’ (Fig. 2). To explore the results of our assessment, the simple sherd count proved an unreliable indicator of relative frequency, being a variable too dependent on the mechanical properties of the ceramics and their patterns of breakage. The combination of attributes is a much more robust method. Thus, we chose the surface of the sherds expressed in square centimetres and their weight in grammes as the two attributes that best defined their size (eg, Solheim Reference Solheim1960; Chase Reference Chase1985, 218; Byrd & Owens Jr Reference Byrd and Owens1997, 315–7).
A scheme using four categories with predefined thresholds was chosen (Table 1). A cardboard template with cut openings allowed us to sort out the ceramics easily and accurately. In response to different patterns of breaking the pieces, two alternative formats for each size were contemplated: a square (Table 1, cell A) and another rectangular cell (Table 1, cell B), both representing the same surface area. The weight of every sherd was measured with a small electronic scale.
Table 1 CERAMIC SIZE CATEGORIES EXPRESSED IN AREAL UNITS (cm 2 )
Secondly, the preservation condition of ceramics is a very informative variable on the vicissitudes experienced before entering the archaeological record (eg, Bradley & Fulford Reference Bradley and Fulford1980; Schiffer & Skibo Reference Schiffer and Skibo1989; Sørensen Reference Sørensen1996, 67–70; Buko Reference Buko1998, 399–403; Edwards Reference Edwards2009, 147–59; 2012, 86–9; Beadsmoore et al. Reference Beadsmoore, Garrow and Knight2010, 125–9; Wolfram Reference Wolfram2013, 82–6, tab. 1; Sánchez-Polo & Blanco-González Reference Sánchez-Polo and Blanco-González2014: 14–16, tab. 2). The basic principles of this approach are that the current state of the fragments provides information about attritional processes (Schiffer & Skibo Reference Schiffer and Skibo1989, 101) and that it is possible to differentiate between pre- and post-depositional alterations. The resistance to such changes depends on the hardness of the fabrics and the nature and intensity of the abrasion processes (Sørensen Reference Sørensen1996, 67; Buko Reference Buko1998, 402). There can always be exceptions but ceramic collections with homogeneous mechanical properties and made under similar firing conditions are subject to comparative evaluation. There is no a widely tested and agreed method for measuring erosion on potsherds. Sørensen (Reference Sørensen1996, 67) proposed three degrees of sherd abrasion – low, medium, and high (Sørensen Reference Sørensen1996, 65, fig. 41). Buko (Reference Buko1998, 400, fig. 13) working on wheel-thrown medieval ceramics, and Edwards (Reference Edwards2009, 148; 2012, 88, fig. 7.11), working on hand-made pottery, proposed an ordinal scale for all sherds. It is this ordinal scale that is used here (Table 2). The unit of analysis has been the individual potsherd, and the scheme includes four classes defined by the condition of surfaces (unpatinated or dull) and the degree of wear on the edges and corners (sharp or rounded and blunt) of each piece (Fig. 3). The variables of area, weight and erosion of all pieces were recorded in a database.
Fig. 3 Examples of Beaker sherds for every abrasion category: 1A & 1B) Grade 1, ‘freshly’ broken; 2A & 2B) Grade 2, local abrasion; 3A & 3B) Grade 3, slight but generalised abrasion and 4A & 4B) Grade 4, highly abraded. (Scales in cm)
Table 2 TYPES FOR ASSESSING THE ABRASION OF SHERDS
Finally an exhaustive and massive re-fitting operation was carried out in order to identify prospective connections between sherds pertaining to the same vessels (cf. Sørensen Reference Sørensen1996; Garrow et al. Reference Garrow, Beadsmoore and Knight2005; Mercer & Healy Reference Mercer and Healy2008, 753–5; Beadsmoore et al. Reference Beadsmoore, Garrow and Knight2010; Edwards Reference Edwards2009; Reference Edwards2012). This procedure was repeated systematically looking for cross-mendings between sherds within the same context – vertical or intra-feature matches – and between pieces from different contexts – horizontal or inter-feature refits. The cases of uncertain or non-adjoining sherds, lacking direct matching but probably from a common pot, were also considered as positive results (Bollong Reference Bollong1994, 17–8, tab. 1; Beadsmore et al. Reference Beadsmoore, Garrow and Knight2010, 126). In order to assess the likelihood of these sherd-to-vessel associations, a scoring template was used to express such relevant observations in terms of inter-sherd matching probability (Blanco-González & Chapman Reference Blanco-González and Chapman2014).
All in all, the above tasks demanded the spreading of the whole pottery assemblage across large tables and on the floor of the two rooms kindly provided by the Museum of San Isidro in Madrid (Fig. 4). The ceramics from each context were kept in independent areas, near their labels, and each set was delimited with coloured tape (Fig. 4). The analysis involved three people for 15 days, spending about 300 person-hours of actual work.
THE CERAMIC ASSEMBLAGE
Out of 36,578 potsherds retrieved in 1981 (Priego & Quero Reference Priego and Quero1992, 90–110), 3837 ceramic fragments have been studied, representing 10.5% of the total assemblage from this sector (Table 3). The bulk of the sample – 3296 sherds – was found within ‘Pithouse 013’ and represent 10% of the ceramics from this depositional context, whereas between 9% and 20.5% of the ceramic contents from the annexed pits were analysed (Table 3). The majority of studied potsherds (84%) are medium or large in size (size groups 2 and 3) (Table 1), and 13% of them are small (group 1) with only 3% of cases over 66.5 cm2 (group 4) (Table 4). Most of the larger sherds – with a relatively short delay between their breakage and deposition – were recovered in ‘Pithouse 013’, but smaller proportions also appeared in the pits (Fig. 5). Regarding their preservation condition (Table 2), ‘freshly’ broken pottery with little or no sign of weathering (abrasion grade 1) dominates with 60% of the sample, while a third of the total sample exhibits partial and local attritional alterations (grade 2) and the most intensely eroded (grades 3 and 4) represent only 7% (Table 4). The pits contained mostly well-preserved potsherds (grade 1), with a weak presence of items classified in grade 2 and almost total absence of the most worn cases (grades 3 and 4) (Table 4 and Fig. 6). Ceramics with partial attritional marks (grade 2) and more intensely eroded pieces (grades 3 and 4) are best represented in ‘Pithouse 013’ (Fig. 6). A Kruskal-Wallis H-test helped us to assess the differences between the values of size (chi square=5.0, p value=0.416) and erosion (chi square=3.143, p value=0.370) between the studied ceramic samples from each feature. Such contrasts are not statistically significant, ie, these ceramic subsets are not dissimilar enough to say that they come from different populations. The important conclusion is that they were subjected to similar processes of fragmentation and attrition.
Table 3 FREQUENCY AND PERCENTAGE OF CERAMICS
key: st./feature: percentage of sherds studied by feature; st./1981: the proportion of sherds studied in the whole 1981 assemblage
Table 4 NUMBER OF STUDIED SHERDS: SIZE (TYPES 1–4) AND ABRASION (GRADES 1–4) PER FEATURES
It has been argued that the longer the history of disturbance and alteration, the smaller the potsherds (Bradley & Fulford Reference Bradley and Fulford1980, 86; Buko Reference Buko1998, 402). Thus, the possible association between the size of sherds and their abrasion (Table 5) has been evaluated by the Kendall-Tau B test (Edwards Reference Edwards2009, 355–8; 2012, 88–9), offering a negative result close to 0 (τ=–0.032 and p value =0.033). This means that the two variables are not related, namely that the smaller fragments are not the worst preserved, and hence both post-breakage and pre-depositional dynamics were relatively independent.
Table 5 RELATIONSHIP BETWEEN THE VARIABLES OF SIZE AND ABRASION
The re-fitting operation (Table 6) has yielded 310 sherd-to-vessel associations comprising 730 potsherds, involving 2–11 sherds per refit; 39 of these cases are physically matching sherds, and the remainder are possible matches lacking a direct refit. According to our scoring template for non-adjoining sherds (Blanco-González & Chapman Reference Blanco-González and Chapman2014), these cases feature medium–high probability re-fits (75–90%). Regarding the type of refit, 96% of these connections are intra-feature refits, linking sherds within the same depositional context – the vast majority within ‘Pithouse 013’ (Table 6, Fig. 2). Some inter-feature or horizontal connections have also been identified between Pits 010, 011, 012, and ‘Pithouse 013’, although no refits have been traced between pits (Table 6; Fig. 2). Importantly 3107 items, representing 81% of the ceramic sample, are ‘orphan sherds’ without any local refit (Schiffer Reference Schiffer1987, 298–302; Bollong Reference Bollong1994, 18, tab. 1; Chapman & Gaydarska Reference Chapman and Gaydarska2007, 81).
Table 6 RE-FITTING OUTCOMES
The number of physical or secure refits are in bold and underlined, the remainder are possible non-adjoining refits
Certain materials deserve further comments because of the clues they provide for characterising the depositional dynamics in operation in this case study. Thus, most of the taphonomic alterations were acquired when detached sherds were undergoing divergent post-breakage trajectories before their eventual deposition altogether within a common cut feature. One case in point concerns the different colours of three conjoining rim sherds from a bowl found in ‘Pithouse 013’ (Fig. 5A). A second concerns a pair of physically matching rim sherds with important lacunae between their breaks – proof of intensive attrition – and again exhibiting different colors (Fig. 5B). A third example is two rim sherds of a bowl with contrasting abrasion effects (grades 2 and 3) (Fig. 5E). The duration of this delay between the discard of ceramics and their deposition in the sunken features might have been prolonged. This occurrence is well illustrated by an initially large slab of a pot rim showing local traces of weathering on its inner surface (grade 2) and subsequent secondary shattering into four smaller sherds (Fig. 5F). A similarly long-lasting pre-depositional sequence can be tracked from two sherds from a hemispherical bowl (Fig. 5D), showing a large external chip prior to its breakage into two sherds, the smaller of them then ground and re-utilised. There are several examples of re-used sherds with intensively polished breaks (Fig. 5C), which inform us about the regular re-cycling of ceramic debris by those prehistoric communities.
The characteristics of the Bell Beaker assemblage is of the utmost importance, with over 180 potsherds found in the three last fieldwork seasons (1972, 1977, and 1981) representing 2.5% of the total ceramics found on this site (Priego & Quero Reference Priego and Quero1992, 231). The 111 Beaker fragments from Pit 011 and ‘Pithouse 013’ feature a remarkable range of taphonomic variability (Table 3): 66% of them are of small and medium size (groups 1 and 2), while the remaining 34% is larger than 12.5 cm2 (groups 3 and 4). Regarding their preservation condition, 74% are well preserved (grade 1) and a significant set, around 23%, lack a ‘freshly’ broken aspect (grade 2) (Fig. 3). We know that both erosion and size are uncorrelated variables and are to be considered as autonomous disturbance dynamics. The scores for the size and abrasion of the Beaker pottery subset have been compared to the rest of the sample by means of the chi square test to assess if their differences were statistically significant. The result (chi square =56.0, p value =0.229) confirms that Beaker fragments did not experience distinctive treatment or idiosyncratic conditions. All ceramic waste, whatever its decoration, shared similar post-breakage cycles and alterations before its deposition. Finally, a dozen very worn sherds, two of them from Pit 011 and the remainder from ‘Pithouse 013’, have been dated to the Neolithic because of their diagnostic form, such as a characteristic type of elongated-section handle (Fig. 6A), or decoration, including different impressed motifs (Fig. 6B & D), finger and nail impressions (Fig. 6C), and impressed applied protuberances (Fig. 6E).
Fig. 4 Assessment of the selected assemblage from El Ventorro 1981 in Museum of San Isidro (Madrid)
Fig. 5 Examples of re-fits and sherds showing different taphomonic marks. (Scales in cm)
Fig. 6 Heavily worn sherds of probable Neolithic date. (Scales in cm)
DISCUSSION
Deposition in domestic structures and their subsidiary underground facilities is a relatively frequent later prehistoric scenario (eg, Domboróczki Reference Domboróczki2009; Bergin Reference Bergin2011; Wolfram Reference Wolfram2013) and the excavators at El Ventorro posited an interpretation centred on this key idea (Priego & Quero Reference Priego and Quero1992). However, only recently, scholars have started paying adequate attention to the specific conditions to be met to support such a reading. All too often, such requirements are not fully considered and, in their absence, a series of fallacious principles are assumed, namely reflectionist prejudices (Chapman & Gaydarska Reference Chapman and Gaydarska2007, 71–3) based on Schiffer’s ‘Pompeii premise’ (Schiffer Reference Schiffer1985; Jiménez Jáimez Reference Jiménez Jáimez2007; Lucas Reference Lucas2012, 102–4). This distorted misinterpretation has been widely applied to prehistoric European pit sites (Chapman Reference Chapman2000). According to this line of argument, the cut features, their fillings, and contents reflected pristine areas of activity with meaningful arrangements and associations faithfully frozen in time.
A brief overview on the local contexts of our case study might be useful to contextualise such statements. There are numerous examples of huts in the central Iberian Chalcolithic. They are highly variable structures in size (5–50 m2) and construction techniques, with semi-sunken bottoms, shallow foundation trenches, stone foundations, and frequently roofing of perishable materials (eg, Díaz-del-Río Reference Díaz-del-Río2001, 220–7; García Barrios Reference García Barrios2005; Liesau et al. Reference Liesau, Ríos, Aliaga, Daza, Llorente and Blasco2013, 141). Discussions of their domestic character have been imprisoned by the descriptive language, as the act of naming these structures constrains and orientates their archaeological interpretation (Hodder Reference Hodder1999, 94–5). Actual semi-sunken dwellings require a series of highly demanding conditions (eg, Bergin Reference Bergin2011) which are not always met by many archaeological cases. Thus, their interpretation as sub-soil domestic quarters has been convincingly refuted in other European regions (eg, Chapman Reference Chapman2000), and similar arguments have been used to refute such accounts for Chalcolithic examples from southern Iberia (Jiménez Jáimez Reference Jiménez Jáimez2007). Most of these suggestions are applicable to ‘Pithouse 013’. Its inadequate technological requirements and limited living conditions – its irregular plan, the poor verticality of its walls, or the absence of external post-holes – discredit any ethnographic analogies with pithouses (contra Priego & Quero Reference Priego and Quero1992, 357; 363). There is no clear-cut separation of the internal and external spaces, since the supposed building had no ‘definite plan’ (ibid., 102) and the morphology of the uneven and sloping layers with patches of refuse accumulations are inconsistent with true occupation ‘soils’ (Schiffer Reference Schiffer1987; LaMotta & Schiffer Reference LaMotta and Schiffer1999; Jiménez Jáimez Reference Jiménez Jáimez2007). On the other hand, the identified spatial arrangements inside ‘Pithouse 013’ also raise skepticism about their primary position (Díaz-del-Río Reference Díaz-del-Río2001, 246). The presence of true soils would be incompatible with the identification of re-fitting sherds between layers 3 and 9, ie, linking the two alleged superimposed buildings (Table 6). We have studied the large potsherds lying flat on the ground surface regarded as the entrance of the ‘upper building’ (Priego & Quero Reference Priego and Quero1992, 103–4; 123). Their low fragmentation state (size 3 is predominant) and fresh, unworn surfaces (grades 1 and 2) do not match the expected taphonomic conditions for items subjected to continuous trampling and friction in a transit zone such as an entrance.
The excavators of ‘Pithouse 013’ suggested the idea of ‘a permanent dwelling, whose floor was regularly prepared and swept, throwing the refuse caused by everyday activity in nearby dumps’ (Priego & Quero Reference Priego and Quero1992, 118). Its contents were envisaged as a Pompeii-like circulating inventory of necessaries at a discrete time in the past, abandoned in its primary context of consumption and lacking formal disposal. According to this, two possible formation processes may be adopted to explain the items found within ‘Pithouse 013’: a) as primary ‘loss’ refuse or micro-refuse that escaped the cleaning activities and became trapped in the floor matrix (Schiffer Reference Schiffer1987, 62; LaMotta & Schiffer Reference LaMotta and Schiffer1999, 21; Tani Reference Tani1995, 233–6) or b) as still usable objects from the abandonment phase left behind but not yet discarded – ie, de facto refuse (Schiffer Reference Schiffer1987, 89–96; LaMotta & Schiffer Reference LaMotta and Schiffer1999, 22). However, 86% of the ceramics from ‘Pithouse 013’ are over 6.25 cm2 (sizes 2–4) (Table 4) and clearly cannot be regarded as primary micro-refuse. On the other hand, within the pits the ceramics exhibit an optimal preservation condition – grade 1 is predominant (Table 4) –, and some re-fits have been identified between ‘Pithouse 013’ and the pits (Table 6; Fig. 2), therefore linking their backfilling. These observations would fit the excavators’ interpretive proposal, with rapid disposal of the broken vessels within the adjacent pits. Nonetheless, the number of ‘orphan sherds’ within the pits remains overwhelming (over 90%), indicating that only a minor portion of the discarded refuse ended up in these subsoil features. Indeed, the pits were filled with a very incomplete sample of sherds, that is, they do not contained representative turnover of everyday waste. The occurrence of vertical re-fits within every pit indicates the retrieval and incorporation of materials from the same provenance, most probably vessels broken not long before the filling of the pits.
It is worth mentioning a special subset of ceramics, comprising 1058 potsherds – representing 27.5% of the total sample – which are medium–large size (types 3 and 4), well preserved (abrasion grade 1) fragments (Table 4) deriving from serving bowls and cooking pots. They were retrieved at different depths within all features and 248 of them are involved in intra-feature refits. This pottery refuse may support Díaz-del-Río’s proposal: vessels used in repeated commensality feasts, being subsequently broken and immediately discarded (Díaz-del-Río Reference Díaz-del-Río2001, 249). Nonetheless, only a part of this group had been broken immediately before deposition, and we can question the relatively rapid filling of ‘Pithouse 013’ (ibid., 248–9). Indeed, our analysis has found polished and re-utilised sherds, perhaps used as lids (?) (Fig. 5C & D) that suggest prolonged life-uses after their break (Chapman & Gaydarska Reference Chapman and Gaydarska2007): the retrieval, handling, and recycling of discarded and even already altered – probably thermally damaged – ceramics (Fig. 5D). The re-use of potsherds in many different ways is well documented ethno-archaeologically (Stanislawski Reference Stanislawski1978; Chapman & Gaydarska Reference Chapman and Gaydarska2007, 75). Their grinding down for subsequent use as grog temper is very likely, as shown by the results of the petrographic analyses of Beakers from Camino de las Yeseras (San Fernando de Henares, Madrid) (Ríos et al. Reference Ríos, García, Aliaga and Blanco2011, 338), and this might partially account for the absence of more conjoining ‘freshly’ broken sherds. Moreover, taphonomic traces indicate that 46% of the potsherds exhibit some degree of wear (grades 2–4) (Table 4) and some re-fitting sherds show differences in their condition which are only understandable if we accept a prolonged interval between their initial breakage and final deposition (Fig. 5).
Importantly, these remarks are also applicable to the Bell Beaker assemblage. The 111 Beaker fragments from Pit 011 and ‘Pithouse 013’ belonged to at least 18 beakers and 21 bowls for individual drinking and nine carinated bowls for containing solid foodstuffs (Rojo et al. Reference Rojo, Garrido-Pena, García, Juan-Tresserras and Matamala2006, 258–9). Their taphonomic heterogeneity and highly incomplete representation prevent us from accepting that all of them were contemporary vessels broken in situ. Quero & Priego (1992, 104) already observed that parts from the same Beakers ‘have been glued despite having different degrees of erosion and surfaces of contrasting colours’. In fact, 67% of the Beaker fragments are small or medium-sized (groups 1 and 2) and 23% have been classified in grade 2 of abrasion, indicating that they might have been discarded in some sort of provisional discard contexts (LaMotta & Schiffer Reference LaMotta and Schiffer1999, 21–2) where they were partially worn or burnt (cf. Garrow et al. Reference Garrow, Beadsmoore and Knight2005, 148–51; Brudenell & Cooper Reference Brudenell and Cooper2008, 22–4; Beadsmoore et al. Reference Beadsmoore, Garrow and Knight2010, 125). There is some new evidence supporting such occurrences. For example, an unusual rich assemblage of Beaker potsherds has been retrieved in ‘hut 2’ at Camino de las Yeseras causewayed enclosure, 20 km from El Ventorro (Liesau et al. Reference Liesau, Blasco, Ríos, Vega, Menduiña, Blanco, Baena, Herrera, Petri and Gómez2008). Their worn condition suggests their extraction from elsewhere and their eventual re-deposition in a domestic context (Liesau et al. Reference Liesau, Ríos, Aliaga, Daza, Llorente and Blasco2013, 143; 147). In short, all the studied ceramics underwent similar degradation dynamics: some were piled up outdoors, in heaps or shallow middens, where they experienced weathering without being trampled or broken into new sherds; others had just recently fractured before being interred (Tables 4 and 5). Thus, at this stage of their ‘biographies’, Bell Beaker wares did not receive any special treatment – they were managed and disposed of in the same way as other plain ceramics. This is a novel and relevant finding based on thorough comparisons with a large sample of plain sherds. Regarding the Neolithic materials, their residual occurrence in Chalcolithic contexts is not surprising: the Manzanares riverbank was intensely occupied in previous times and some 4 km upstream are several Neolithic sites on the same right bank. It is possible that the later groups collected Neolithic sherds as ‘heirlooms’ and brought them to El Ventorro.
All in all, the foregoing observations are not fully consistent with the interpretive proposals put forward so far for this site. In view of the evidence presented here, one of the chief ideas that emerge from our analysis is the important bias of the studied assemblage, which does not constitute a coherent and meaningful unit per se. The selected ceramics from the 1981 excavation have little to do with the living or systemic average domestic repertoire of a household at that time. But nor do these items constitute a kind of remains ‘abandoned in situ, without later cleanings nor excessive shuffle distorting their original arrangement’ (Díaz-del-Río Reference Díaz-del-Río2001, 247). Some materials were transferred and there are no secure patternings in primary position. The analysed collection seems a rather arbitrary and heterogeneous aggregate – ie, one whose spatial associations are not significant from a functional point of view (Lucas Reference Lucas2012, 193–8). It was made up of juxtaposed secondary refuse (Bradley & Fulford Reference Bradley and Fulford1980, 90; Tani Reference Tani1995, 237–8), resulting from various activities, with a wide variability in their origins, timing and trajectories of use and disposal.
These findings open up new interpretive possibilities which require careful re-examination in further later prehistoric contexts. In particular, this case study presents striking similarities with other Neolithic enclosures (Thomas Reference Thomas1999, 38–45; Mercer & Healy Reference Mercer and Healy2008; Beadsmoore et al. Reference Beadsmoore, Garrow and Knight2010). Thus, most of the vessels appear in a very incomplete and fragmentary state, with only sporadic re-fitting and the vast majority (81%) of sherds are ‘orphan fragments’, making their in situ breakage unlikely. It is therefore a very partial outcome, the unintended by-product of complex natural and anthropogenic accretion and depletion processes (Schiffer Reference Schiffer1987; LaMotta & Schiffer Reference LaMotta and Schiffer1999). A portion of the items interred in the cut features might have resulted from feasting episodes performed in their vicinity. Some large ceramic slabs might have been thrown into various receptacles mixed with long-lasting discarded refuse. The latter might have been retrieved from transient contexts where the sherds were recycled and subjected to variable post-breakage and pre-depositional life cycles (Garrow et al. Reference Garrow, Beadsmoore and Knight2005, 148–50; Chapman & Gaydarska Reference Chapman and Gaydarska2007, 75–7; Brudenell & Cooper Reference Brudenell and Cooper2008, 30–3; Beadsmoore et al. Reference Beadsmoore, Garrow and Knight2010, 125; 129–30, fig. 12). It is difficult to define precisely the depositional dynamics responsible for the archaeological image presented here, but it may lie somewhere in the middle of a continuous spectrum of variation (Brudenell & Cooper Reference Brudenell and Cooper2008, 30; Lamdin-Whymark Reference Lamdin-Whymark2008, 175; Garrow Reference Garrow2012, 94; Lucas Reference Lucas2012, 123). The characterisation proposed here may be akin to the concept of a ‘cumulative palimpsest’ (Bailey Reference Bailey2007, 204–5; Lucas Reference Lucas2012, 112–23).
Finally, ‘Pithouse 013’ has been regarded as ‘the largest accumulation of refuse documented to date throughout the Iberian Meseta’ (Díaz-del-Río Reference Díaz-del-Río2001, 246). We must add that this image was part of a bigger picture of which our visibility has been constrained by: a) the depositional cycle, since only a minority of sherds has survived or was deposited in the same spot, and b) the scale of the rescue excavations, which revealed only part of what was probably a monumental ditch (Fig. 1C). The remarkable number of vertical re-fittings within this context (Table 6) mainly involving slightly eroded sherds (grade 2) from the upper and lower layers, points to a prolonged process of backfilling in successive depositional episodes and the inclusion of materials curated or abandoned in its surroundings. On the other hand, we can be confident that at least Pits 010, 011, and 012 and feature 013 were open at the same time and were filled using sediments and cultural debris from the same source, as shown by the horizontal re-fittings (Fig. 2). The well-preserved condition of the adjoining ceramics between the pits and ‘Pithouse 013’ allow us to relate their filling with the use, breakage, and partial deposition of several serving vessels. Thus, the refitting operation confirms the coeval closure of the large ditch and several of the shafts, as well as a relatively independent filling of every pit, but employing shared materials with ‘Pithouse 013’.
The striking resemblances between ‘Pithouse 013’ and another residue-rich Chalcolithic ditch excavated at Las Pozas (Zamora) (Fig. 1A) in 1979 and 1987 (Val Recio Reference Val Recio1992) have already been raised (Díaz-del-Río Reference Díaz-del-Río2001, 237; 249). This parallel is even more appropriate now, since the use of remote sensing techniques – infra-red images edited with GIS tools (García García Reference García García2013) – has revealed the true nature of Las Pozas as a double causewayed enclosure. The occurrence of unusual accumulations of Beaker sherds – some of them abraded – at Camino de las Yeseras (Fig. 1A), another nearby ditched enclosure (Liesau et al. Reference Liesau, Blasco, Ríos, Vega, Menduiña, Blanco, Baena, Herrera, Petri and Gómez2008; Reference Liesau, Ríos, Aliaga, Daza, Llorente and Blasco2013), has also been highlighted here (vide supra). As in many regions of Atlantic and Mediterranean Europe, inner Iberia is rich in such monumental features, with over 50 of them known to date (eg, Díaz-del-Río Reference Díaz-del-Río2004; Delibes de Castro et al. Reference Delibes de Castro, Crespo Díez, Fernández Manzano, Herrán Martínez and Rodríguez Marcos2010). Indeed, El Ventorro and more recently excavated examples (Díaz-del-Río Reference Díaz-del-Río2004; Liesau et al. Reference Liesau, Blasco, Ríos, Vega, Menduiña, Blanco, Baena, Herrera, Petri and Gómez2008; Delibes de Castro et al. Reference Delibes de Castro, Crespo Díez, Fernández Manzano, Herrán Martínez and Rodríguez Marcos2010) share features with other ditch-digging traditions elsewhere in Europe (eg, Thomas Reference Thomas1999, 38–45; Darvill & Thomas Reference Darvill and Thomas2001; Varndell & Topping Reference Varndell and Topping2002) such as their location on rounded locally prominent hills in lowland settings (Fig. 1C) and especially their filling with ‘domestic’ debris and deposits that stood out in terms of their quality or quantity.
In short, there is scope for interpreting El Ventorro as a place for gatherings surrounded by at least one ditch segment, only partially excavated and whose sketch plan can be extrapolated (Fig 1C). Its digging and backfilling would have entailed the movement of huge quantities of sediment, involving large numbers of people. The depositional histories tracked in the ceramic analysis of the fill of ‘Pithouse 013’ can be clearly understood from this standpoint as a cumulative aggregate or palimpsest containing time-averaged residues (Lucas Reference Lucas2012, 106–9), created by seasonal, small-scale, and intermittent social gatherings of dispersed groups (Díaz-del-Río Reference Díaz-del-Río2001, 249; Reference Díaz-del-Río2004; Mercer & Healy Reference Mercer and Healy2008, 755; Beadsmoore et al. Reference Beadsmoore, Garrow and Knight2010, 129). During such festive commensal episodes, people would have interacted, consumed foodstuff, and interchanged animals, items, or know-how in different crafts through hands-on training. Especially at El Ventorro, the role played by copper metallurgy in these gatherings was important (Harrison et al. Reference Harrison, Quero and Priego1975; Priego & Quero Reference Priego and Quero1992). The participants also broke a remarkable quantity of vessels, including Bell Beakers, and disposed of them in a range of standardised ways. The performance of such iterative and protracted depositional practices on the same spot may account for the evidence examined here. It represents the unplanned outcome of a prolonged cultural tradition consisting of the digging and closure of ditch segments and pits, taking its meaning through its very repetition (Lucas Reference Lucas2012, 108).
CONCLUDING REMARKS
This article has presented a hands-on experience addressing object-oriented taphonomy as an under-exploited approach whose potential has not yet been realised. Despite the widespread deployment of ceramic re-fitting and the emerging importance of ceramic taphonomy in the Anglo-American milieu, available contributions lack adequate discussion of their methods, which are often implicit or understated. An integral procedure has been exhaustively presented here, and the validity and prospects of our analytical strategy has been tested with a remarkable ceramic assemblage. Despite the limitations of being an old excavated collection, retrieved in the infancy of rescue archaeology, important insights have been gained.
The initial account of El Ventorro suggested that it was a permanent and open settlement ‘without ramparts or ditches’ (Priego & Quero Reference Priego and Quero1992, 357). The combination of Bell Beaker and massive earthworks at El Ventorro made this case study worth of careful re-examination, to spotlight and discuss widespread concerns and interpretive models in later prehistoric Europe. Thus, the analysis of the patterns of rupture, abrasion and representation of ceramics has led to the rejection of the ‘domestic’ interpretation of ‘Pithouse 013’ and the refutation of the primary character of its ceramic contents, which are not representative of everyday work. We propose an alternative scenario: a place bounded by at least one ditch, whose filling required seasonal and intermittent activity, through a cumulative and gradual pace at the same time as the adjoining pits were backfilled, using a heterogeneous and chaotic aggregate of debris. Just as some vessels were used in community feasts, only a small portion was incorporated into the pits and ditch immediately after their breakage. Others were erratic residues accruing over long trajectories, including Bell Beaker and Neolithic sherds as well as the remnants of previous activities not necessarily linked, either spatially or temporally, with such depositional episodes. A better characterisation of the depositional dynamics responsible for the documented evidence has been achieved. Moreover, these observations, gained via the highly detailed examination of a massive and heterogeneous collection of ceramics, have opened up new interpretive avenues dealing with pan-European later prehistoric phenomena such as the role and meaning of Bell Beakers and the formation of Neolithic ditched enclosures.
Acknowledgements
This paper has been prepared in the course of the European Commission-funded research project Past Fragments (Marie Curie Intra-European Fellowship 298285). C. Vega and E. Carmona were paid by this project to help one of us (ABG) with the study of the assemblage. We are very grateful to A. González Alonso, curator in the Museum of San Isidro (Madrid), for the facilitation of our work and for his generous and crucial collaboration.
