Archaeological context

The 4th and 3rd millennia BC in Campania, and more broadly in Southern Italy are characterised by important cultural dynamics preluding the onset of Bronze Age societies. Nevertheless, the chronology and cultural development of Copper Age Campania were for a long time hard to assess, partly due to the lack of radiocarbon dates for the core contexts and partly to the scarcity of settlement evidence. In recent years, both aspects have improved thanks to large-scale rescue excavations [60–64] and radiocarbon dating campaigns [65–67]. Traditionally, the definition of chronological phases and archaeological cultures in this region has been tied to a typo-chronological approach, mostly based on pottery styles [68–72].

In order to give a more objective and comprehensive outline of the Campania Copper Age phases and cultural traditions, in the present study a distinction was used for periods based on the available radiocarbon dates with reference to the main cultural traditions conventionally identified (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. All radiocarbon dates currently available for Copper Age Campania, calibrated with OxCal Intcal 20, at 2σ, colour coded by cultural tradition with indication of the main volcanic events, red lines (right).

A-MS: Agnano Monte Spina eruption 4420 ± 58 BP; 3335–2913 cal BC 2σ [73]; AP: Avellino Pumices Avellino Pumices 3546 ± 17 BP; 1944−1779 cal. BC 2σ [74].

https://doi.org/10.1371/journal.pone.0343599.g001

Based on the radiocarbon dates and volcanic stratigraphy, five main periods can be distinguished in Copper Age Campania:

• Final Neolithic/Initial Copper Age, 3800–3650 BC (based on parallels with radiocarbon dated sites outside Campania, (13: p. 45–46; 73) characterised by scarce and blurred archaeological evidence often found in caves [71]. The material culture, displaying strong Final Neolithic features, is generally limited to pottery while few lithic tools are reported (Fig 2A). Different cultural traditions have been generally recognised for this time period (e.g., Macchia a Mare-Spatarella, Piano Conte) but their characterisation is still far from comprehensive.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Main pottery types attested for Copper Age Campania.

A. Final Neolithic/Initial Copper Age from Paestum and Pontecagnano; B. Early Copper Age from Sala Consilina, phase I; C. Early/Middle Copper Age from Paestum Gaudo cemetery: D. Late Copper Age Laterza from Paestum; Final Copper Age from Atena Lucana. By permission of the Ministry of Culture, Direzione Regionale Musei Nazionali Campania.

https://doi.org/10.1371/journal.pone.0343599.g002

• Early Copper Age, 3650–3300/3100 BC and Middle Copper Age, 3300/3100–2900/2800 BC characterised by the appearance of complex funerary practices and by an extreme scarcity of documented living areas. Two main burial practices are attested in the region: firstly the deposition of cremated human remains in stone structures (identified as Taurasi cultural tradition) attested only in the Early Copper Age, and several large cemeteries with collective rock-cut tombs (identified as Gaudo cultural tradition) spanning from Early to Late Copper Age. They are respectively characterised by a specific material culture especially expressed by pottery and lithic tools (Fig 2B, C). The first metal objects for the region are attested, mostly daggers or ornaments [75–77].
• Late Copper Age, 2900/2800–2400/2300 BC, characterised by the continuation of collective graves cemeteries and by the onset of large villages with adjoining funerary areas mostly with single flat graves and occasionally rock cut collective tombs (final evidence of the Gaudo culture and beginning of Laterza cultural tradition). Only one date from Castel Baronia sets the Laterza evidence already in the Middle Copper Age. Nevertheless, given the wide error of the date performed in 1980s and the concentration of the other Laterza contexts dates, both in Campania and in Lazio, from ca. 2900 BC onwards, the Castel Baronia date will not be considered reliable for the present study. Few Bell Beaker influences are evident in the ceramic style (Fig 2D), [78].
• Final Copper Age, 2400/2300–2200 BC, characterised by few pieces of evidence emerging form recent excavation and not yet exhaustively published [78–80] in some cases with peculiar Cetina style influences (Fig 2E) [28,81], generally more common on the Adriatic shore.

Regardless of distinctions between facies, cultures or groups, it is clear that the evidence for Copper Age Campania still does not allow clear distinctions to be drawn between different cultures. In order to avoid complex terminologies, and blurred identifications of more-or-less homogeneous cultural groups that are chronologically overlapping, it seems more appropriate at the present stage of research to talk about various, partially contemporaneous and spatially close cultural traditions with both different and shared practices. Therefore, although the main conventional distinctions between cultural traditions were used in this study (Initial Copper Age, Taurasi, Gaudo, Laterza and Cetina influence), these are used critically and tested by comparing contextual and ceramic data, and are ultimately reassessed in the discussion of the results.

Four main Copper Age sites were selected for this study, they are located in the southern sector of Campania region, in two alluvial plains bordered by carbonate mountains, Paestum and Pontecagnano respectively at the southern and northern shore of the Sele plain, the sites of Sala Consilina and Atena Lucana further south in the Vallo di Diano (Fig 3). These were selected on the basis of their well-defined stratigraphy and chronology to provide a wide picture of technological change and continuity during the Copper Age. Three of the sites are multiphase, covering the entire Copper Age and all the main cultural groups traditionally recognised in the area as detailed in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Chronology of the sites and contexts analysed.

https://doi.org/10.1371/journal.pone.0343599.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Distribution of the main sites (shaded dots) currently known for Copper Age Campania.

Pinpointed sites analysed in the present study: 1. Paestum: 2. Pontecagnano; 3. Sala Consilina; 4. Atena Lucana. The map was created using TINITALY DEM91 available at https://doi.org/10.13127/TINITALY/1.0 under CC BY 4.0 license, Europe outline was constructed using “Natural Earth. Free vector and raster map data @ naturalearthdata.com” available at https://www.naturalearthdata.com/downloads/10m-raster-data/. Both maps were modified using the Free and Open Source GIMP 3.0 available at https://www.gimp.org/.

https://doi.org/10.1371/journal.pone.0343599.g003

Detail of the geological setting is provided in the S1 Text.

The site of Paestum has a long history of prehistoric excavations started during the Second World War until the end the 20^th century [77,85–89,91]. Its occupation started in the Middle Palaeolithic to culminate in the foundation of the Greek-Roman city. In prehistoric times, after a Late Neolithic phase (ca. 4500−3800 BC, Diana culture), it has its peak during the Copper Age. An Initial Copper Age phase, probably dated to the first half of the 4th millennium BC, is attested both in the Agorà and in the area of the Temple of Cerere [83,90]. After this minor occupation, a large rock-cut cemetery of over 50 collective graves was established slightly north of the plateau occupied by the Greek-roman city, in Gaudo locality (eponymous of Gaudo culture), and used for almost a millennium, from the mid-4th to the mid-3rd millennium BC [66], (for radiocarbon dates see Fig 1). Almost 2 km south, probably during the 3rd millennium, a further mortuary site was established near the (historic) Temple of Cerere with mortuary practices and grave goods very different to those of the highly coded rock-cut Gaudo cemetery and generally associated with Laterza cultural tradition for its material record. Settlement and mortuary evidence of the same phase are attested about 200m south west, in the area of the Greek Agorà, highlighting the extension of this occupation. The lack of radiocarbon dates for this phase associated with Laterza cultural tradition prevents clarification of its chronological relationship to the Gaudo cemetery; nevertheless, the dates available from other similar sites in Campania and more broadly Southern Italy suggest at least partial contemporaneity (see Fig 1 and [13]).

The site of Pontecagnano, about 26 km north of Paestum, has a long archaeological history, peaking in the Iron Age with the Villanovan occupation that led to the establishment of one of the most important Etruscan sites of southern Italy. Large-scale excavations in the area, mainly in response of modern commercial expansion, have allowed a thorough investigation of the site and its past environment. In total, it has yielded over 10,000 burials from prehistoric and historic periods. For the Copper Age, the main evidence is funerary, represented by a small rock-cut cemetery of ten collective tombs associated with Gaudo cultural tradition [77]. In recent years, a second site, about 1.5 km south-east, centred on an Initial Copper Age hut, was discovered during a large-scale preliminary excavation for the widening of a highway [82].

The site of Sala Consilina, locality Capo La Piazza, is situated on the edge of the Maddalena Mountains in the Vallo di Diano depression. A rescue excavation was carried out in 2006 and 2007 [65,71]. The investigation brought to light an extensive archaeological site, with two distinct phases of occupation, both referable to the Copper Age. The first phase of occupation of the site yielded cremation burials in urns both inside and outside two structures with dry-stone foundation walls. Due to the state of preservation of the remains the context could not be radiocarbon dated [65] but cremation burials are attested in Campania mostly between 3650 and 3100 BC (calibrated radiocarbon dates in Fig 1, sites of Taurasi and Casalbore). The second phase of occupation of the site is exclusively represented by five trench graves near the stone structures with inhumations and a scarce grave good composed only by few ceramic objects. Due to the style of the artefacts and burial customs these tombs have been assigned typologically to a later phase of occupation, corresponding to Laterza cultural tradition. Four of the five inhumations were radiocarbon dated to the first half of the 3rd millennium BC [65], confirming this hypothesis (radiocarbon dates in Fig 1, Sala Consilina tombs 1073, 1074 and 1076).

In the northern part of the Vallo di Diano, at Atena Lucana in the locality of Fosso Aimone, a settlement was explored in the years 2005−2007 during work to widen the Salerno-Reggio Calabria motorway. The site is located at the foot of the valley slope, about 8 km from Sala Consilina [65]. The archaeological record is still under investigation therefore only preliminary information are currently available. The occupation is represented by a large paleo-surface with two rectangular structures with one apse, probably identifiable as houses, and a possible production zone. Charcoal from the firing structure was radiocarbon dated to 2568−2067 cal. BC, 2σ [65,72], (Fig 1), setting the occupation of the site between the Late and Final phases of the Copper Age and the beginning of the Early Bronze Age (generally set around 2200 BC). The ceramic assemblage also points to a date in the last centuries of the 3rd millennium BC due to the presence of decorations and vessel forms of the Cetina style [92].

Form and macroscopic analysis

Macroscopic observations were carried out and recorded, on the whole or substantial amounts of the ceramic assemblage from each of the three main sites analysed. The morphology of the vessels was carefully taken into consideration and types were organised according to the following main use-related properties which can determine the design of the vessel [96]: a) The mouth opening and its dimensional relationship with the maximum diameter of the vessel (i.e., restricted vs. unrestricted opening); b) The presence and shape of the handles, i.e., one or two handles, and their grip. All features that reflect different steps of the productive process were recorded, including forming techniques, surface treatment, decoration and firing. These stages are particularly relevant since they are the outcome of learned moves and gestures carried out by potters with a ‘practical consciousness’ [97] which is often considered to be one of the most conservative aspects of human behaviour [98–102]. Therefore, for the present analysis it was particularly relevant to record these features and verify their occurrence between different sites, phases and cultural entities. The technological features observed macroscopically, by following well developed protocols [22,103,104], are: shaping techniques, surface treatments and decorative techniques.

Petrography in thin section

The samples were analysed in 30µm sections with a Leitz Laborlux 12POL polarising microscope with plane and cross-polarising light. Stitched photomicrographs of about 13 × 3 mm were taken for each thin section with a Leica microscope camera in plane and cross-polarising light. Petrofabrics were described after the method defined by Whitbread [105,106] and modified by Quinn [107,108] and distinguished based on observations of the texture and composition of the samples under a polarising microscope.

X-ray diffraction (XRD)

XRD technique was used to determine the mineral composition of the sample, especially clay minerals which are too small to be recognised by optical microscopy, and to estimate firing temperature ranges [109,110]. The samples were powdered using a planetary ball mill to between 1–10μm grain size. Approximately 2g powder were analysed for each sample. The analysis was undertaken using a Panalytical Aeris XRD with a CuKα1 emitter. Measurements were taken from 5–70° 2θ at a step size of approximately 0.0054° at 39.5 seconds per step. The powdered sample was placed in a circular sample holder with a diameter of 32 mm and a depth of 3 mm. A nickel-beta filter was used on the incident side, along with 0.04 rad soller-slits inserted on the both the incident and detector side of the beam. The analytical configuration also included ¼° divergence slits, a 20 mm beam mask, and a beam knife in the high position, and a 9 mm anti-scatter slit. The total time for the analysis of each sample was 22m 32s. The analysis of the results was completed using the High Score Plus proprietary software package by Panalytical which allows to automatically, and manually, compare the peaks obtained in the diffractogram with a library of known diffractograms of thousands of minerals (in particular the International Center for Diffraction Data (ICDD) Database, and the Crystallography Open Database (COD)). The automatic identification was manually refined also based on the observations made in thin sections.

X-ray fluorescence (WD-XRF)

Major elements were analysed in fused beads, which allowed to overcome the matrix effect, while minor and trace elements in pressed pellets, ensuring less dilution of key trace elements [111]. The samples were powdered using a planetary ball mill with an agate mortar to 110 μg grain size. Preparation of the beads involved fusing 0.600g of the powdered sample with a lithium tetra-borate flux (6.000g) at

1000–1200°C. The mixing ratio of sample and flux in weight was sample 1:flux 10. The error in measurement was kept below ±0.5 mg. Fusion was carried out using a LaNeo Claisse Bead fuser inside a platinum crucible and mould (95% platinum and 5% gold). The analysis of the beads was carried out in a Panalytical Zetium WD-XRF spectrometer with Wroxi calibration standards. Pressed pellets were prepared using 12g of sample mixed in a ball mill with 3g of wax binder (mainly Ultrawax) with a ratio of sample 4:binder 1. When the material available was slightly less than 12g, the analysis was still carried out keeping the ratio constant and taking the lower weight into account during measurement. The 32 mm diameter pellets were pressed in a Herzog press up to a pressure of 200 bar. The analysis was carried out in the Panalytical Zetium WD-XRF spectrometer with Protrace calibration standards.

Statistical treatment of geochemical data

Chemical data were evaluated using different statistical methods. A first approach involved the univariate and bivariate analysis of each element measured, without any transformation. This allowed the investigation of the structure of the data, eventual departure from normality, long tail for variables and outliers [112]. Principal Component Analysis (PCA) was used to explore relationships among the different pottery samples. The data were first log10 transformed to avoid scale effects on the results [111,112]. Negative numbers in the data were avoided when possible, while zeroes were replaced with half the detection limit for the given element [111]. R-studio statistical software was used with the ‘prcomp’ function for PCA calculation, ggbiplot and factoextra packages for plotting of the graphs. The PCA on the correlation matrix was performed using the concentration of major (Al2O3, CaO, Fe2O3, K2O, MgO, Mn3O4, Na2O, SiO2, TiO2) and minor and trace elements (Ag, As, Bi, Ce, Co, Cr, Cu, Ga, Hf, La, Mo, Nb, Nd, Ni, Pb, Rb, Sn, Sr, Th, Tl, U, V, W, Y, Zn, Zr). P2O5 and Ba, Br, Hg, I, Yb, though measured, were not included in the analysis of the data as these could have been affected by post burial contamination processes [113,114]. Elements with mostly negative values or beneath the detection limit in our dataset were also not considered (i.e., Cd, Cs, Ge, Sb, Sc, Se, Sm, Ta, Te). The accuracy and precision [108] of the XRF analysis were determined by running certified reference materials multiple times during each batch of analyses: for fused beads ‘NCS DC87103 Soil’ and ‘GBW 07109 Rock’; for pressed pellets ‘NCS DC87103 Soil’. Elements with an abundance below the limits of detection of the machine were disregarded before submitting the geochemical dataset to multivariate statistical analysis.

Additional information regarding the ethical, cultural, and scientific considerations specific to inclusivity in global research is included in the Supporting Information S2 Text.

Vessel types and macroscopic analysis

The four sites are characterised by a ceramic repertoire that presents substantial changes in the occurrence and frequency of vessel types depending on the context and chronological phase and/or cultural tradition. The revised vessels types are detailed in S1 Text. As highlighted in Fig 4, alongside vessel types of longue durée such as large bowls, small bowls, jars, small jars and cups, there are types with a more restricted occurrence.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Occurrence of vessel types in the four sites investigated.

Early to Final Copper Age, earlier and later periods are not taken into consideration.

https://doi.org/10.1371/journal.pone.0343599.g004

The long-term types are also the most common in the Final Neolithic and continue right through to the Early Bronze Age, albeit with some morphological changes. Closed vessels (askoi, jugs, small jugs and flask vases) and pyxides are exclusive of Early-Middle/Late Copper Age collective graves cemeteries (Paestum and Pontecagnano). Jugs are also attested in the Early Copper Age cremation mortuary context (Sala Consilina, phase I). Hanging jars appear in cremation burials (Sala Consilina Phase I), and in other Initial Copper Age contexts in Campania, such as Casalbore [115] and Grotta dell’Ausino, and continue in the Gaudo repertoire in a smaller version and with a different morphology. Jars and pyxides with vertical handles suitable for hanging or carrying by cords are completely abandoned in Late Copper Age Laterza contexts, as well as closed necked vessels suitable for pouring. The only vessel type which has a later occurrence is the shallow bowl (patera) generally found both in funerary contexts (e.g., the cemetery near the Temple of Cerere, Paestum) and in settlement contexts such as Foglianise, loc. S. Maria a’ Peccerella [116]. This significant change in vessel types suggests an important shift in the material culture, possibly reflecting different habits, both in the daily use of ceramics, and in the ritual sphere (especially given the presence of pouring vessels, mostly in graves) [84].

The main technological macroscopic observations for each context, grouped by period and cultural tradition are summarised in Table 3. For the site of Atena Lucana, macroscopic analyses were carried out only on the diagnostic ceramics sampled for archaeometric analysis, which included the vessels of a style clearly attributable to Cetina influences.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Summary table of the main technological macroscopic observations undertaken on the ceramic assemblage analysed, organised by chronology/cultural tradition.

https://doi.org/10.1371/journal.pone.0343599.t003

Shaping techniques are quite consistent across time and space. The coiling method is attested in all the contexts, with flattened coils ranging between 3 cm and 5 cm in thickness and shaped by pinching, given the alternating U-shaped or rounded breaks (Fig 5.1 and 2; [104]: 55). Near the rims, coils sometimes show a smaller dimension, about 2–3 cm, probably due to a further addition to the vessel walls to build the rim (Fig 5.3) The use of moulds is attested for the first time in the Early Copper Age cremation context of Sala Consilina phase I for the shaping of the characteristic large truncated cone bowls (Fig 5.5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Technological traces related to vessel shaping detected on the analysed samples.

1.Paestum Agorà I; 2, 8 and 7 Paestum Gaudo cemetery; 6 Paestum Cerere; 12 Paestum Agorà II; 4, 10 Pontecagnano Anas; 3, 9, 13-15 Pontecagnano Gaudo cemetery; 5, 11 Sala Consilina phase I. Images out of scale. By permission of the Italian Ministry of Culture. Direzione Regionale Musei Nazionali Campania.

https://doi.org/10.1371/journal.pone.0343599.g005

In the Initial Copper Age context of Pontecagnano (Anas excavation) the same shape was always fashioned by coiling as displayed by the breakage pattern (Fig 5.4). From the Early Copper Age on, the use of moulds is attested in all the contexts analysed mostly for complex shapes such as jugs and open vessels such as bowls (Fig 5.6, 7 and 9). Complex types such as jugs, askoi and cups, typical of Gaudo repertoire were manufactured by modelling the lower and upper parts separately. This process is hinted at by the recurrent horizontal crack at the maximum width of the body, corresponding to the point where the two halves were joined (Fig 5.7) mostly documented in jugs and cups. The two halves of these complex vessels might also have been produced using different forming processes and at different times. The bases are generally rounded, and sometimes slightly flattened or concave, and they are highly homogeneous, suggesting the use of a mould (Fig 5.9), which would also help in supporting the ceramic structure during manufacture when the upper part was added. The use of moulds is documented in several ethnographic cases [104,117] and has also been detected in the Rome area in contemporary Gaudo or slightly earlier contexts [22,29]. Furthermore, the common horizontal fractures occurring along almost the whole circumference points to a multistage forming process, with a drying phase aimed at avoiding the collapse of the vessel under its own weight [104]. Nevertheless this shaping method was not exclusive, since in both the site of Paestum and Pontecagnano some of the jugs and askoi show traces of coils or slabs in the lower portion of the body (Fig 5.8), especially in the case of smaller sizes. In the case of the askoi the irregular shape of their bodies might be the cause of this different fashioning choice.

Handles were applied either by simply spreading onto the outer surface (Fig 5.12) or by inserting the protruding end of a pre-shaped piece into a perforation in the vessel wall (Fig 5.15). In vessels from rock-cut burials (Paestum and Pontecagnano, Gaudo cultural tradition) the handles are exclusively attached by piercing the vessel wall showing a certain degree of homogeneity in shaping tradition among the two contexts and across time. A peculiar case is represented by vertical or horizontal tunnel handles, typical of the first half of the 4^th millennium BC, and attested in the Initial Copper Age site of Pontecagnano (house, Anas excavations) and in the Early Copper Age site of Sala Consilina (phase I, cremation burials). In the first case the vessel wall was modelled, externally pierced and clay affixed to the inside wall to create a small, internal tunnel (Fig 5.10). In the second case, tunnel handles on large bowls or jars are fixed to the vessel wall by spreading onto the outer surface (Fig 5.11).

The attention to surface treatment is consistent throughout the Copper Age (Table 3), with a peak in the practice of burnishing and smoothing accurately in the highly finished vessels of cremated burials and collective graves (Taurasi and Gaudo, Early-Late Copper Age), while it decreases in the Late Copper Age Laterza contexts characterised also by a higher percentage of scaled surfaces. Decorated vessels, generally with impressions or incisions, are most common in Early/Middle and Late Copper Age, with a higher variability in Laterza contexts, and the introduction of new techniques and probably tools in the final Copper Age with the Bell Beaker and Cetina style. Firing treatments are mostly irregular throughout the Copper Age but with a clear increase in reducing temperatures in the Early and Middle Copper Age especially compared to Initial Copper Age and Late/Final Neolithic productions.

Petrography in thin section

Thin section petrography was carried out on 223 samples from the four sites and allowed the definition of 7 main ceramic fabrics based on compositional and textural criteria (summarised in Table 4), and 11 samples each characterised by a different composition, identified as loners (Fig 6). Internal variability within the fabric is referred to as “variant”. The fabric characterisation was based on the first or second most abundant inclusions (natural or temper), since for most of the samples the main mineral components were quartz and feldspars. A second criterion used in the definition of fabrics and subfabrics was the shape and sorting of the inclusions (rounded to angular, poorly to well sorted, unimodal to bimodal sorting). Textural analysis, considering aspects such as the size, shape, frequency, and degree of sorting of inclusions, was also used since these aspects might have had important technological implications [96,108]. The extensive description of the fabrics is provided in S1 Text, while the main 7 fabrics are summarised below, with only a brief description of the petrographic loners.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Summary table of the main petrographic fabrics detected. Organised by site and chronology/cultural tradition.

https://doi.org/10.1371/journal.pone.0343599.t004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Microphotographs of selected pottery representative the 7 main fabrics and petrographic loners.

1. Quartz sand; 2. Quartz sand and grog; 3. Grog; 4. Grog bimodal; 5. Quartz metamorphic; 6. Volcanic; 7. Organics. Petrographic loners: 8. Argillaceous rock fragments; 9. Chert; 10. Carbonatic sand; 11. Carbonatic; 12. Carbonatic and organics; 13. Carbonatic and sedimentary; 14. Sedimentary; 15. Mudstones; 16. Fine calcareous; 17. Fine quartz; 18. Bioclasts.

https://doi.org/10.1371/journal.pone.0343599.g006

1. 1. Quartz Sand: 28 samples

This non-calcareous fabric is characterised by moderately well sorted, quartz-feldspathic sandy inclusions with rare grog (Fig 6.1). Sub-rounded to sub-angular quartz is dominant, followed by frequent k-feldspars (sanidine and/or orthoclase) and polycrystalline quartz, and rare microcline. Volcanic rock fragments and igneous rock fragments are also commonly attested. Mineral inclusions include few/very few plagioclase, chert, hornblende, pisoliths, opaques and pyroxenes, rare calcedony and sphene (detected by SEM-EDS), sedimentary and metamorphic rock fragments. The matrix is mainly homogeneous with a low optical activity, and a mostly strial b-fabric. Voids are macro to micro planar and vughs are mostly parallel to margins, often with traces of burnt organics. Subgroups can be detected based on the size and shape of the grains.

1. 2. Quartz sand and grog: 24 samples

This fabric is characterised by a non-calcareous clay matrix with dominant/frequent quartz-feldspathic sandy inclusions and common grog (Fig 6.2). The quartz-feldspathic grains are mainly fine, sand-sized and moderately sorted, with a few coarser inclusions, generally either grog or sedimentary rock fragments (fine sandstones and siltstones). Volcanic rock fragments (trachytic) are also present, while metamorphic rock fragments are rare to absent. Mineral inclusions also comprise a few plagioclases (more common in samples from Pontecagnano), chert, mica/biotite, opaques. The matrix is mainly homogeneous with a low to moderate optical activity and a b-fabric that is variously strial, striated or speckled. Only two samples (PAE.G.VII.2 and PAE.A.III.19) show traces of sintering. Voids are mostly planar and meso vughs, and occasionally vesicles or channels, generally poorly aligned to margins.

1. 3. Grog: 93 samples

This non-calcareous fabric is dominated by sub-rounded quartz and feldspar inclusions and common grog (Fig 6.3). Opaques (generally Fe-Ti oxides, as identified by SEM-EDS) and rock fragments are common. Other mineral inclusions detected are a few sub-rounded volcanic lithics and mica, while sub-rounded sedimentary rocks (both sandstones and siltstones), biotite, chert, clinopyroxenes, plagioclase, pisoliths and calcedony are rare. Two variants were identified based on variations in voids and sorting. The matrix is rich in silt size inclusions and varies from homogeneous to heterogeneous, showing a moderate to low optical activity. The b-fabric is mostly strial at times striated or speckled. In some cases, textural features and clay bodies suggest an uneven clay mixing. Voids are mainly planar, meso and macro vughs with moderate alignment to margins. Burnt organics were detected in all the samples.

1. 4. Grog bimodal: 32 samples

This non-calcareous fabric is dominated by angular quartz-feldspathic inclusions and grog (Fig 6.4). Its composition is similar to grog fabric but differs in its bimodal distribution of the inclusions and scarce fine fraction. In some cases, grog is distinguishable by the sintering of the matrix and the presence of relic surfaces. Angular volcanic rock fragments (trachyte) are common, and a few chert, plagioclase, pyroxenes, opaques, mica and rock fragments are also attested. Other rock fragments are mainly sub-rounded sandstone, mudstone and rare metamorphic rocks. The non-calcareous, optically-active matrix is mostly homogeneous with a strial or striated b-fabric, in some cases showing irregular clay mixing. Silt size inclusions are rarer compared to the previous fabric (Grog) suggesting the use of a different base clay. Only two vessel samples have traces of sintering (PAE.C.1.3 and a sintered core in PAE.C.4.3). Voids are mainly meso and macro elongated and equant vughs, in some cases preserving charred remains inside vughs with burnt edges. In most samples, the large angular grains of feldspars, both k-feldspars and plagioclase, the abundant quartz and small amount of biotite may suggest an artificial temper based on crushed, unweathered, sedimentary and volcanic rocks and grog. Fragments of parent rocks with the same composition can also be detected.

1. 5. Quartz metamorphic: 13 samples

This coarse-grained fabric is characterised by closely spaced, sub-angular to sub-rounded inclusions of mostly coarse, sand-sized (with a few pebble-sized) grains (Fig 6.5). Monocrystalline and polycrystalline quartz dominate, while coarse quartz-feldspathic rock fragments are common. K-feldspars are frequently attested and plagioclase is common; microcline, opaques, volcanic lithics (trachyte), chert, clinopyroxenes and metamorphic rock fragments are occasionally attested. Sedimentary rock fragments and ARF are present. The non-plastic inclusions are poorly aligned to margins with a moderately sorted bimodal grain size distribution, with quartz and rock fragments as coarser grains. The non-calcareous matrix is mostly homogeneous with a moderate optical activity and strial to striated b-fabric, silt-sized mica and quartz grains and in some cases also iron-rich particles and clay pellets. Pores are mainly elongated meso and micro vughs and planar voids, mostly parallel to margins. Voids with blackened edges, probably due to the presence of burnt organics, were detected in only a few samples. A finer variant with a lower percentage of inclusion is attested in Atena Lucana.

1. 6. Volcanic: 12 samples

This slightly bimodal fabric is dominated by coarser inclusions of angular to sub-rounded k-feldspars, while sub-angular clinopyroxenes and volcanic lithics such as trachyte are common (Fig 6.6). Further frequent inclusions are sub-rounded quartz, opaques, plagioclase feldspars, a few ARF, hornblende, muscovite, chert and a very few pisoliths, biotite, and sedimentary rock fragments. A single subfabric was distinguished based on its higher content of silt-sized quartz and feldspars and a few rounded carbonatic inclusions, possibly suggesting a different base clay. The matrix is mostly homogeneous, and in only a few samples slightly heterogeneous, with a moderate to low optical activity and striated b-fabric. Voids are mostly meso and micro vughs, randomly oriented. This fabric is characterised by a high number of volcanic inclusions, often euhedral. The slightly bimodal grain size distribution, with larger, angular k-feldspar, pyroxene and trachyte grains, suggests either the addition of crushed volcanic rocks as temper or―more likely―it resulting from the weathering of pyroclastic flows, common in the area of Pontecagnano [118]. Pisoliths and iron-rich features are particularly common and attributable to the base clay.

1. 7. Organics: 9 samples

This fine fabric is characterised by few inclusions except for frequently occurring, mostly coarse, sand-sized sub-angular ARF/grog, in some cases also identifiable as clay pellets (Fig 6.7). There is very little fine fraction, mostly represented by quartz which is dominant to common, and a few k-feldspars and opaques. Chert, clinopyroxenes, plagioclase and mica are rare. The main feature of this fabric is the dominant presence of meso- and macro-elongated and equant vughs and planar voids, generally moderately aligned to margins. The shape of the voids, their blackened edges and the presence in some cases of charred remains, points towards a fabric rich in organics, which were probably added as temper. A variant could be recognised based on the presence of calcareous inclusions (2 samples). The matrix is non-calcareous and homogeneous to heterogeneous especially in the samples from Paestum, probably due to irregular clay mixing given the presence of incompletely hydrated clay pellets [119]. The matrix is mostly highly optically active with a striated b-fabric.

This fabric displays a clear paste recipe, characterised by the use of a finer clay, probably slightly refined by hand picking of coarser inclusions, with the addition of organic material as temper.

Petrographic loners:

Twelve samples could not be grouped in broader fabrics given their variable composition and are fully described in S1 Text. They are characterised by different inclusions and in some cases clay matrixes. In particular: Argillaceous Rock Fragments (n:1, from Paestum Cerere, Fig 6.8) characterised by the addition of shale or over-consolidated clay not properly crushed and wetted during the preparation of the paste (as verified by SEM-EDS); Chert (n:2, from Atena Lucana, Fig 6.9) probably crushed and added as temper; Carbonatic sand (n:1, from Sala Consilina I, Fig 6.10), suggesting the use of a sandy calcareous clay; Carbonatic (n:1, from Pontecagnano Anas, Fig 6.11), with micritic calcite grains and occasional shells and bioclasts; Carbonatic and organics (n:1, from Atena Lucana, Fig 6.12) in a calcareous clay, used for a possible daub or clay structure exposed to fire; Carbonatic and sedimentary (n:1, from Atena Lucama, Fig 6.13) with occasional spathic calcite; Sedimentary (n:1, from Sala Consilina I, Fig 6.14) dominated by shale/mudstone fragments, which may, in some cases, also be interpreted as clay pellets, in this case an uneven clay mixing (either of two different clays or of the same heterogeneous clay) is suggested by streaks and by the uneven distribution of the inclusions; Mudstones (n:1, from Pontecagnano Anas, Fig 6.15) tempered with coarse, crushed shales/mudstones and chert; Fine calcareous fabric (n:1, from Paestum Agorà I, Fig 6.16) with grog temper and calcareous inclusions, both micritic and spatic; Fine quartz (n:1, from Sala Consilina I, Fig 6.17) with sintered matrix; Bioclasts (n:1, from Atena Lucana, Fig 6.18), dominated by fine, monocrystalline and polycrystalline quartz with rounded chert, radiolarite and silicified bioclasts, with a sintered matrix.

XRD mineralogy and equivalent firing temperatures

X-ray diffraction analysis provided qualitative mineralogical data of the pottery samples. The analysis mostly confirmed petrographic observations revealing a quite homogenous composition across the fabrics identified (details of the analysis for each sample in S2 Dataset). The clay minerals detected for all the samples are mainly illite/muscovite, often smectite and occasionally chlorite/kaolinite. The other main mineralogical phases detected are quartz, k-feldspar, plagioclase and calcite (only in the carbonatic loners: Carbonatic sand, Carbonatic, Carbonatic and organics, Carbonatic and Sedimentary and Fine calcareous). Amphibole (hornblende) was identified in two samples from Pontecagnano characterised by a higher content of volcanic materials, by its principal characteristic peak at about d ~ 8Å.

The results of XRD were also used to estimate the Equivalent Firing Temperature [120] of the pottery samples. Even though illitic, non-calcareous clays do not show great changes in mineralogy when subjected to the firing temperatures typically utilised for pottery (500–900°C) [121], some conclusions can be drawn: the presence of illite/muscovite peaks in most of the cases, and in some samples also of smectite or calcite, suggests temperatures not exceeding 800–850°C, which is when these phases generally decompose [121–124]. The use of temperatures exceeding 700°C is supported also by the presence of haematite in four samples from Paestum characterised by a reddish-brown colour confirming the use of irregularly oxidising atmospheres. A different group of ceramics fired at lower temperatures could be detected mostly in Paestum’s Gaudo collective burials where the presence of a chlorite (002)/kaolinite (001) peak indicates 500°C and 650°C as maximum firing temperatures [110,121,123–126].The use of a higher temperature can be inferred only for a Fine fabric from Sala Consilina phase I which completely lacks illite/muscovite peaks and displays wollastonite, a high temperature mineral suggesting a reaction between carbonates and silicates occurring between 800 and 1100°C [124,127].

No further observation could be drawn on the firing atmosphere probably due to the relatively low concentration of iron in the samples (c. 6% Fe₂O₃) [128] and the predominant use of irregular firing atmospheres as suggested by the patchy colouring of the vessels.

Geochemistry

Given the mostly sedimentary and carbonatic geology of the area, the four sites predictably show a rather similar mineral composition, as confirmed by petrographic and XRD analyses (details of the analysis for each sample in S3 Dataset). XRF data, statistically analysed through PCA [129] together with element ratio and bivariate analyses help to refine the ‘chemical fingerprint’ of each sample and examine variations in the data in order to detect geochemical groupings. In this way, sharper distinctions between raw materials, compositional groups and provenance can be inferred. Log10 normalised data of the main fabrics (with more than 2 samples) were statistically analysed by PCA (Fig 7). Fabric represented in less than 2 samples were not considered in order to avoid a dispersion of the data, given the strong compositional differences of these fabrics constituted by a single sample, in particular the highly calcareous ones.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Statistical treatments by PCA of geochemical data of the samples grouped by the main fabrics.

A: PCA; B: loading plot with contribution; C: scree plot of the first ten components; D: samples grouped by site..

https://doi.org/10.1371/journal.pone.0343599.g007

The loading plot (Fig 7B) shows that differentiation between samples is driven mainly by trace elements such as Nb, Zr, Th, Pb, Cr, with a considerable contribution from Fe2O3. The first ten PCs together account for 78.87% of the total variance of the system, with PC1 and PC2 representing 29.6% and 14.95% respectively (Fig 7C). The results of the PCA (Fig 7A) show a clear association between some sites. The grog and quartz rich fabrics are mostly scattered displaying mostly a similar composition with some outliers. The vast majority of the samples are scattered and broadly compositionally similar, albeit with some outliers, not allowing the determination of specific groups.

Only two fabrics display a moderate clustering and therefore a clearly different composition: the Organics fabric (characterised by few quartz-feldspathic inclusions, and low SiO2 and Na2O) and the Quartz metamorphic one (displaying a high SiO2 concentration). Chemical variations within the samples might suggest different raw material sources, especially clays, or differences in the variability of inclusions due to the silico-clastic environment of the four site. These might relate to differing geological inputs or other formation processes that impact the trace element chemical profile. Though with some overlapping, samples from the same site display a good grouping suggesting the use of different raw materials, or base clays at each site.

To investigate these less evident geochemical differences, the ratios of some elements, especially trace and rare earth elements (REEs; La, Ce, Nd, Sm), were investigated further. Element ratios (rather than absolute values) are useful to mitigate the impact of variation among the samples of volatile organic matter, known post-depositional impacts, and tempering with quartz, calcite/limestone, and other sedimentary rocks and minerals (see [130] for a recent treatment). The ratios of elements (La, Th, U, Cr, Co, Sc), including some REEs, are useful for identifying variations in the original clay source of the fabrics since these are less affected by differences in tempering [131,132].

The plots of La/Sc and Cr/Th show several different concentrations, indicated in Fig 8. Most of the samples from Paestum and Pontecagnano represent the main concentration, with low values for both La/Sc and Cr/Th ratios. Pontecagnano samples (light blue crosses in Fig 8), show strong clustering suggesting a homogeneous composition except for a few outliers with a stronger volcanic composition.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Bivariate plot of La/Sc and Cr/Th ratios.

The samples from are colour coded by site (left) and by main fabric (right).

https://doi.org/10.1371/journal.pone.0343599.g008

Samples from Sala Consilina (yellow crosses in Fig 8 left), of the Quartz metamorphic fabric also show good clustering, which is quite distinct from the Paestum-Pontecagnano concentration, with a higher Cr/Th ratio. The samples from Paestum (pink triangles in Fig 8 left) display a more scattered distribution compared to those from Pontecagnano, with a higher number of outliers and a possible cluster with higher La/Sc ratio. Samples with carbonatic inclusions are clearly compositionally different, also from the sample made of Fine calcareous clay, probably due to the type of inclusions. The main fabrics Grog, Grog bimodal, Quartz sand and Quartz sand with grog (Fig 8 right) display a great degree of scattering, probably due to their high internal variability due to temper derived either by alluvial contexts or by human interaction.

A better discrimination between the samples can be made on site basis, suggesting the use of different clays specific of each site.

The overall composition of the fabrics points to a local provenance for the raw materials used for pottery making at the four sites. The petrographic composition of the inclusions is fully compatible with the geological features of the area, where predominantly siliciclastic-arenaceous and carbonate formations outcrop along with less extensive pyroclastic fall deposits related to the activity of Somma-Vesuvius, and especially of the Phlegraean Fields. For the sites of Paestum and Pontecagnano, it must be noted that the area of the Sele River Plain is characterised by geological complexity. A study of the possible clay sources of the area has been recently undertaken but is not yet fully published. This related mostly to later ceramic productions (affected by subsequent volcanic eruptions, especially from the Somma-Vesuvio complex, e.g., Avellino Pumices and later eruptions; [133,134]). These clay sources show a rather heterogeneous chemical composition due to the different sedimentary formations of the area [134]. In Fig 9, the composition of the ceramics from the sites analysed in the present study is compared to the composition of coarse Greek and Roman pottery and to clay sources from the area of Paestum and Pontecagnano [134].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Comparison of geochemical data with other local Greco-Roman pottery and clays.

Compositional groups of all samples analysed from the present study compared with later coarse Greco-Roman pottery from Paestum and Pontecagnano (blue and pink areas) and clay raw materials from the Sele River Plain [134].

https://doi.org/10.1371/journal.pone.0343599.g009

The clay raw materials analysed by De Bonis are both alluvial and basinal, from the plain and nearby hills. As highlighted by the graphs, only a few clay raw materials correspond to the low CaO values of the samples analysed. These are mostly from the area of Paestum, specifically a marine clay from Capaccio (CAP2) [134] and an alluvial clay from Fonte, about 5 kilometres far from Paestum (FONTE2) [134]. All other clay samples display higher CaO values and can therefore be excluded with regard to Copper Age products.

The attribution of samples with higher CaO to calcareous clays is uncertain due to the presence in these fabrics of carbonatic inclusions, which could also cause the higher CaO content. Only in the case of the Fine calcareous fabric, with a CaO content of about 5%, does the yellowish colour of the clay and the presence of micritic calcite in the matrix suggest the use of a different clay, slightly higher in CaO and lower in Fe₂O₃. Clay samples with similar values come from Capaccio (CAP3, marine clay) [134] and the Heraion area near the site of Paestum on the estuary of the River Sele (HER1, alluvial clay).

It must be noted that non-calcareous clays, which were the most commonly used clays at the Copper Age sites analysed in the present research, are found either in older basinal formations (Upper Jurassic-Lower Miocene) or as alluvial clays, depending on the stratigraphy. Basinal clays formed during Miocene-Pleistocene are mostly calcareous. Furthermore, alluvial clays are often rich in a quartz-feldspathic component, closely corresponding to most of the Copper Age fabrics analysed in Paestum and Pontecagnano.

The high Zr-low Cr value of one clay sample from the area of Pontecagnano is of interest as this roughly corresponds to the main cluster for Pontecagnano Copper Age samples. This clay source (PCGA1) [134] is derived from the weathering of a pyroclastic deposit in the area of Pontecagnano, supporting the hypothesis based on the petrographic analysis of the use of such a raw material source. The coarse Greco-Roman samples analysed by De Bonis [134], displayed in Fig 9, also show a similar scattered composition for trace elements. A different origin might be inferred for the ARF, Mudstones and Sedimentary fabrics, which are also quite common in the neighbouring areas, consisting mostly of sedimentary environments.

Throughout Copper Age it must be noted that the use of calcareous clays is extremely rare at the sites analysed, as are calcareous inclusions, attested mainly in single samples for each site. This suggests that calcareous clays and inclusions were mostly avoided, despite their abundance in the area [134] and the ubiquity of calcareous rocks originating both from the travertine formations of the Sele Plain and the carbonate Apennine Platform [135]. This preference might also be due to the higher technical knowledge and skills needed to monitor the firing process for ceramic made of calcareous clays [136].

Chemical comparison between the Copper Age sites of Sala Consilina and Atena Lucana and local sources are difficult given the lack of similar analysis on raw materials. Nevertheless, based on the petrographic and mineralogical analyses, some inferences can be made. Both sites display a lower impact of volcanic products compared to Paestum and Pontecagnano, which lie closer to the volcanic complexes. A peculiarity of both sites is the presence of fabrics characterised by polycrystalline quartz grains of metamorphic origins. Metamorphic outcrops are not attested in the Campania region, the nearest being in northern Calabria, but residual and detrital gneiss, granitoids and schists are attested in the sedimentary environments, especially in flysch formations outcropping near the Vallo di Diano (Flysch di Gorgoglione and Lagonegregro succession) [135]. The local provenance of these ceramics is also supported by evidence of the use of a similar raw material in the later Full Bronze Age in the cave of Pertosa, on the edge of the Vallo di Diano basin [95,137].

A further peculiarity of the samples from the Vallo di Diano is the common presence of chert and silicified bioclasts/radiolarites. These formations can be attributed to the Rhaetian–Jurassic radiolarites and reddish, greenish and violet silicified argillites typical of the ‘Scisti Silicei’ formation, outcropping in the area of Sala Consilina, 8 km from Atena Lucana [135].

Raw material selection and processing

Local raw materials were exploited throughout the Copper Age, with a long-lasting and consistent preference for iron-rich, non-calcareous clay mixtures of illite and smectite. Raw material procurement appears to have been the most conservative stage of the manufacturing process, as commonly attested elsewhere in prehistoric Europe and the Mediterranean. Copper Age products are distinctive compared to Late Neolithic ones [138] for the introduction of grog tempering and the almost complete abandonment of calcareous clays. Most raw material sources are site specific and can be preliminarily traced to the immediate hinterland of the sites up to a maximum distance of 8 km. This evidence closely matches ethnographic observations made by Arnold [99,139] and by Gosselain and Livingstone Smith [51], who report the use of raw materials generally within a radius of 7 km from the sites. Choices in the raw materials used for ceramic production are often explained by archaeologists either from an ecological perspective [99,140], with variation related to the geological variability of the surrounding environment, or from a functionalist perspective [141–143], based on the raw materials’ properties and suitability for specific functions. As highlighted in the following sections, in Copper Age Campania ideology and symbolic purposes also played an important role in the determination of the sourcing and processing of raw materials, including the abandonment of calcareous clays and the use of grog temper. No significant variation of raw materials could be detected between chronological/cultural contexts at the same site, testifying to a high degree of continuity in the exploitation of sources and a shared knowledge of their locations and properties.

A further aspect to take into consideration in the interpretation of raw materials procurement is the relationship between people and landscape also in connection with other tasks. In the concept of taskscape developed by Ingold [144], tasks, understood as daily activities in the life of ancient people, are considered as mutually interlocking, and therefore ‘just as the landscape is an array of related features, so - by analogy - the taskscape is an array of related activities’ [144]. Therefore, raw material procurement should be seen as a part of the wide variety of tasks ancient people carried out in the surrounding landscape [93]. The sites taken into consideration in the current study are located in a landscape rich in raw materials, where potters had a wide choice but displayed clear preferences towards non-calcareous alluvial and volcanic clays, as well as sandy, quartz rich tempers or artificial ones. Calcareous and non-calcareous clays are widely available in the area from the coast to the nearby hills. This suggests that raw material procurement could have been carried out in the immediate surroundings of the sites, near the main watercourses (Sele river and Picentino river for Paestum and Pontecagnano, the Tanagro river for Sala Consilina and Atena Lucana) probably in connection with other types of tasks. The tendency towards watercourses might be connected to the procurement of water and wood and other sorts of materials from the surrounding wooden area. Among these are possible wooden tools for decorating vessels, pebbles for burnishing, but also plant fibres for basketry, etc. Given the funerary use of most of the contexts analysed, little information is available on grinding stones, and other types of daily tools. Watercourses appear as a fundamental resource and pole of attraction in the landscape for local communities. A deeper understanding of the dynamic relation between Campanian Copper Age people and the surrounding landscape could be further investigated through detailed surveys for raw materials in the area [93] focusing both on sources of clay [134] and temper.

Technological choices, innovations, and pottery-making traditions

When considering the whole dataset, continuities and discontinuities in ceramic production can be detected between the five cultural traditions associated with the contexts analysed, affecting different stages of ceramic chaînes opératoires between the 4th and 3rd millennia BC, as summarised in Fig 10.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 10. Continuity and change in ceramic production during the Copper Age in Campania in relation to chronology and cultural traditions.

Key: + innovation; – abandonment; | continuity.

https://doi.org/10.1371/journal.pone.0343599.g010

The main innovations and breaks in the productive process displayed in Fig 10, such as the abandonment of fine figulina production and the introduction of grog temper in the first half of the 4th millennium BC, the use of moulds in the second half of the same millennium, a stronger control over firing temperatures and conditions regained by the second half of the 3rd millennium BC, concur with broader changes in funerary customs, domestic structures and material productions, generally regarded as the emergence of new cultural traditions. Nevertheless, these changes are often shared by different groups and might reflect broader trends or a wider circulation of knowledge and information. Attention to surface treatment increased during the Copper Age, both through the introduction of textured surfaces such as scaled decoration and through a careful smoothing or burnishing of vessel surfaces to achieve a degree of polish and shine. While simple vessel types are attested throughout the Copper Age, more complex and peculiar shapes are typical of specific contexts (e.g., composite shapes in the collective burials) and in some cases were triggered by external influences, as in the case of the Bell Beaker and Cetina style pottery.

To unlock new lines of interpretation from a social and productive perspective, the identification of pottery traditions besides cultural labels would be extremely interesting. Nevertheless, this task proved to be complex for two main reasons addressed in the following paragraphs: the limitations of the archaeological data and the relative dating of some of the contexts.

In the delineation and interpretation of different cultural traditions and communities of practice in Copper Age Campania, a fundamental aspect to take into consideration is the partiality of the archaeological data in reflecting all the different ‘activities’ carried out by the human groups that constituted a ‘social world’ [145,146]. In fact, for some time periods and cultural traditions in Campania either only the village or the mortuary sphere is currently visible archaeologically, as highlighted in Fig 11.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 11. Visual representation of the cultural domains and communities of practice analysed for Copper Age Campania.

https://doi.org/10.1371/journal.pone.0343599.g011

Therefore, only some of the constituent domains of a polythetic culture [146,147] can be linked and interpreted. However, it cannot be excluded that some of these domains (e.g., settlement organisation, mortuary practices etc.) could have overlapped between what we currently recognise as different cultural traditions. Indeed, overlaps in material culture and mortuary practices can be detected among the different cultural traditions identified, such as the presence of both Gaudo and Laterza style vessels in some rock-cut tombs (e.g., Pontecagnano tomb 1497) [77].

A further element to consider is the relative dating of many of the contexts analysed. The absence of extensive radiocarbon dating in the cemeteries and dwelling contexts, either for lack of resources or lack of datable sample materials, does not allow the absolute dating of ceramic production for each context analysed without taking into consideration the relative chronology based on typo-chronological parallels with dated assemblages. This implies a possible bias when comparing and connecting archaeological sites and contexts.

In order to test the continuities and discontinuities displayed in Fig 10, and answer the related research questions, the only three contexts analysed directly by radiocarbon dating were considered and their ceramic production assessed without taking cultural labels into consideration.

Among the sites considered in the present research only three contexts were directly radiocarbon dated: Tomb IX from the Paestum rock-cut cemetery in the Gaudo locality; four single flat graves from Sala Consilina phase II; and the habitation layers of the village of Atena Lucana. The ceramic chaines opératoires identified at the three sites is summarised in Table 5. From Sala Consilina only two tombs (1073 and 1076) were considered: those yielding the larger ceramic repertoire.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Main stages of the chaîne opératoire attested in the 3 contexts directly radiocarbon dated taken into consideration in the present research.

https://doi.org/10.1371/journal.pone.0343599.t005

These three sites are chronologically subsequent to one another: the collective grave tomb IX was radiocarbon dated using samples taken from two individuals and spans the second half of the 4^th millennium BC (though some yet unpublished radiocarbon dates prove its continuity also into the first half of the 3^rd millennium BC) [66]; the flat graves date to the first half of the 3^rd millennium BC; while the living area at Atena Lucana dates to the second half of the 3^rd millennium BC.

By looking exclusively at the ceramic data some conclusions can be drawn. At all the three sites potters preferred the use of local non calcareous iron rich clays, widely available in the surrounding alluvial area, both in the Sele Plain and in the Vallo di Diano, with slightly different tempering procedures. Forming practices appear quite homogeneous, with the use of moulding for the bases of complex vessels in Paestum and for open shapes in Sala Consilina, while coiling and slab building are attested for the upper body of complex forms and large containers such as jars or large bowls at all sites. Surface treatments are accurate in all three sites with burnishing and smoothing preferred, and scaled treatment attested only at Sala Consilina (II) and Atena Lucana. Looking at the overall technological choices it can be argued that at Paestum potters repeatedly applied a consistent set of technological choices to a broader range of ceramic types (jugs, cups, askoi etc.) while at the other two sites they applied a wider range of paste preparation methods and surface treatments to a more restricted repertoire of shapes. Decoration techniques consist mostly of impressions and incisions with a wider repertoire employed at Paestum and Atena Lucana, where some new types of tools for impressing are attested, while extremely poor at Sala Consilina (II) where only a plastic decoration is attested mimicking an handle. Firing conditions are extremely irregular at both Paestum and Sala Consilina, while at Atena Lucana they are homogeneously either oxidising or reducing.

Based on these data, seven potting chaines operatoires can be recognised at the three sites, as summarised in Fig 12.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 12. Overview of the ceramic chaines operatoires detected in the radiocarbon dated contexts.

Sites of Paestum, tomb IX, Sala Consilina, phase II tombs 1073 and 1076, and Atena Lucana, based on the samples that underwent both macroscopic and archaeometric analyses. Keys: Sc = Scaled; R = Rough; RS = Roughly Smoothed; AS = Accurately Smoothed; B = Burnished; SB = Smoothed/Burnished; REDOX = Reducing; IRR = Irregular; BC = Black Core; OX = Oxidising.

https://doi.org/10.1371/journal.pone.0343599.g012

In the case of Paestum, a consolidated production can be recognised, with a single chaîne opératoire applied despite the wide repertoire of vessel forms. Besides the use of moulds and local non calcareous clays no major correspondences can be inferred with the other two sites which are located in a different area in the Vallo di Diano (about 50 kms far from Paestum). In the other two contexts, potting traditions appear more variable, with a wider range of tempering traditions and surface treatments, which are even broader at Atena Lucana (with 5 different chaines operatoires compared to 2 attested at Sala Consilina). The two sites share one tempering tradition with surfaces more-or-less accurately treated, suggesting some degree of continuity between the two, chronologically subsequent, productions.

Overall, when compared to the chronological/cultural distinctions displayed in Fig 10, the data seem to confirm the general trends detected for all the sites analysed in the present research, further highlighting the consistency and distinctiveness of operating procedures for pottery making in collective graves, while highlighting some continuity between the Late Copper Age site of Sala Consilina (II) and Final Copper Age one of Atena Lucana. Ceramics destined to be used in the complex rituals of the collective graves appear as a distinctive and controlled strand of production, probably associated with a smaller group of highly skilled potters, compared to a more variable production at the sites of Sala Consilina and Atena Lucana. For these sites access to pottery making traditions and practices could have been broader and extended to a larger group of potters. Nevertheless, the increased attention to and control over firing conditions suggests an advancement and specialisation at least in this step of the production process at the site of Atena Lucana.

Given this, it could be that the different steps of pottery making could have been carried out by different people in relation to other tasks.

Ceramic production and long-term cultural changes

Continuities and discontinuities in the local pottery making traditions and demands for production in Copper Age Campania within their productive and social contexts can be better understood in relation to the broader archaeological context and can in turn contribute to clarifying cultural dynamics. As highlighted in Fig 13, during the Copper Age ceramic production undergoes three main typological and technological turning points, traditionally also recognised as shifts between cultures.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 13. Major trends recognised in ceramic production in relation to chronology, cultural tradition and cultural processes at a regional and broader scale.

https://doi.org/10.1371/journal.pone.0343599.g013

These turning points can be roughly set in the first half of the 4th millennium, at 3650 BC, 2900 BC and 2500 BC, and correspond to important cultural processes taking place both in Southern Italy and more broadly across Europe. In the following sections the results of this research are combined with the broader archaeological context in order to propose a narrative of the development of ceramic production in the whole Copper Age in relation to cultural and socio-economic dynamics.

The first half of the 4th millennium: Between tradition and innovation―The dawn of the ‘Age of Shine’

During the first half of the 4th millennium BC potters in the Campania region preferred a ceramic repertoire characterised by simple, broadly circulating multifunctional ceramic types (e.g., jars, bowls, cups) in continuity with Late Neolithic Diana traditions (Fig 13). Regional and micro-styles disappear, replaced by models that are interchangeable and recognised across larger territories, representing the influence of Late Neolithic Diana, Ripoli and Chassey/Lagozza styles [23,148]. This connection is unsurprising since a certain continuity can be detected in sites’ occupations between the Late Neolithic and Initial Copper Age in Campania, but also in Central Italy [149]. This is testified in Campania by the two coastal sites analysed in the present study, Paestum and Pontecagnano, where a scarcely characterised Inizial Copper Age frequentation follows a more intense Middle and Late Neolithic one. This period is incompletely characterised in Italy and beyond and is generally termed the ‘Neolithic Decline’ due to a consistent demographic decrease across Europe [53,150] and partly in the Italian Peninsula [151]. In the latter, this apparent demographic trend might also be affected by a gap in the archaeological record due both to the major research focus on earlier or subsequent phases, as highlighted in targeted study on settlement patterns in Central Italy [149], and by demographic estimations [151]. In particular, in Campania evidence dated to the first half of the 4th millennium BC is characterised by short-lived settlements that leave more ephemeral archaeological evidence and poorly documented occupations in caves [9,60,152].

Several attempts have been made using a traditional typological approach to distinguish between different phases and facies in Campania during the Early Copper Age [68]. Based on the present analysis, the ceramic types of this first phase are seen to be quite homogeneous for most South Italian contexts, with differences largely limited to the decorative motifs. Therefore, the clear distinction of chrono-typological sequences has been particularly difficult in different areas of the Peninsula [58,68,149,153,154].

In order to better distinguish these Initial Copper Age ceramic productions, larger programs of archaeometric and technological analysis, quite rare for this period, would be beneficial in future research. In fact, it is in this period that an aesthetic shift that started in the Late Neolithic (monochrome horizon, [23]: 47; [58]: 171) reaches its completion: the ornately decorated fine wares of the Middle Neolithic are completely replaced by plain, dark, glossy or textured surfaces that dominate the whole Copper Age, for this reason termed by Robb as ‘the Age of Shine’ [58].

This important aesthetic shift is evident in the progressive abandonment by potters of fine, light coloured, calcareous fabrics (figulina pottery), widely documented at Middle-Late Neolithic sites, used for vessels of exceptional quality (e.g., Serra d’Alto ware). Until the 5th millennium BC, figulina ware was used for high quality consumption vessels (e.g., complex cups with smooth surfaces, homogenous firing condition, buff colour, often painted decoration and baroque handles), partly by virtue of its fine appearance. It required more complex manufacturing operations including the depuration of the base clay, and higher levels of expertise in fashioning, as a result of its enhanced plasticity, and control of the firing process, due to its calcareous composition and homogeneous oxidising firing [17]. The social value of these vessels is highlighted by their visual distinctiveness, by the complexity of their production process, by their long-distance exchange [155,156] and by their preferential occurrence in ritual contexts [157–159]. Nevertheless, figulina wares such as Serra d’Alto are attested also in settlement sites in contexts that suggest they were used in everyday life as well, probably on social occasions involving drinking and food consumption (due to the type of vessel shapes, mostly cups). Their role in social display and as valued means of exchange was probably part of their embedded practical functions [160].

From the first half of the 4th millennium BC, this highly valued figulina pottery became obsolete. This process is particularly evident at the site of Paestum where the fine calcareous clay recipe used for Serra d’Alto and Diana figulina wares at the same site [138] was modified in the Initial Copper Age plain bowls and cups from the Agorà through the addition of grog, probably to enhance the workability of the clay. This shift implies the abandonment by potters of specific clay sources for fine wares in favour of iron-rich, non-calcareous ones already in use for coarse wares (e.g., jars, large bowls etc). This important change has several implications and possible concurrent causes. A clear shift in the demand for socially valued goods [161], caused or followed by a change in the use of ceramic vessels can be recognised. As expected, cups and small bowls are still attested and manufactured with finer fabrics, though they display a lower incidence in the ceramic record. This might indicate a change in people’s ritual and social consumption practices or a change in the role that ceramic objects played in these. From a technological point of view, this disruption in the production of calcareous fine wares can be also understood in terms of a loss of potters’ expertise in fashioning and controlling the firing of calcareous ceramic pastes [124,162]. This could either be due to a break in the transmission of information and practices between potters across time, or to a lower time investment in the production of high skilled ceramic vessels, as suggested also by the modification of the original recipe in favour of a greater workability of the clay.

The modification of earlier, more demanding traditions (the figulina ware now tempered with grog, Fine calcareous fabric) and their subsequent abandonment are the first significant changes happening in the first half of the 4th millennium BC, suggesting that potters tended to a simplification of the ceramic production process. At the same time, they developed innovative features, such as the scaled decoration that characterises domestic ceramic production (particularly jars) across the whole Italian Peninsula until the final centuries of the 3rd millennium BC. In this context, ceramic production expresses an inter-regional similarity characterised by simple types lacking particularly highly finished, regionally distinguished, exchanged ceramic vessels.

As previously outlined, the archaeological evidence for this period in Campania exists only in settlement sites, while funerary evidences are completely lacking. The rarity of mortuary and clearly ritual sites, together with the apparent decrease in the symbolic meaning of pottery, suggest that during the first half of the 4th millennium BC important ideological, symbolic and possibly economic changes were ongoing, but not yet defined, resulting in the substantial changes evident in the subsequent Early and Middle Copper Age phases.

3650 BC-2500 BC: The rise and fall of collective graves and symbolic ceramic repertoires

This transitional period is followed in Campania by a restructuring of local communities around 3650 BC, when marked changes can be recognised in the ceramic record and other archaeological evidence, with the onset of large cemeteries and a decrease of settlement sites. Funerary contexts that were rare in the Late Neolithic and Initial Copper Age become predominant, with dedicated tomb and cemetery spaces and complex ritual practices, often involving the deconstruction of individual identity in favour of a non-individualised community of the dead [9,163,164]. This wide range of changes, occurring in different ways across the Italian Peninsula and Europe, is accompanied by a substantial transformation of ceramic production processes and repertoires. This sees the emergence of skilled potters sharing a clear community of practice responding to demands for, and dedicated to the production of high quality, socially valued goods used by mourners in complex and codified funerary rituals. The broadly circulating vessel types of the Final Neolithic/Initial Copper Age were replaced by regional repertoires, characterised by representative vessel types rarely circulating outside their main, mostly funerary, contexts, such as Taurasi and Gaudo in Campania and surrounding regions, but also Remedello and Rinaldone in North and Central Italy [165,166]. In the Campania region, for a 500 year period (about 3650−3100 BC) relatively large cemeteries and collective graves with cremation or inhumation in different structures become the most common features of the archaeological record, attributed to the Taurasi (cremation) and Gaudo (rock-cut inhumation burials) cultural traditions.

Ceramics associated with cremation burials are highly problematic and deserve a separate discussion. At cremation sites, attested in a patchy manner in Southern Italy, ceramics represent different styles (e.g., Early Copper Age style at Sala Consilina and Caivano, but Gaudo-like style at Taurasi, [76] and presently cannot be linked to other forms of archaeological evidence (e.g., settlement sites). Therefore, in order to address their production systems in a more comprehensive way, further technological and archaeometric analyses on all cremation contexts attested in Campania and Southern Italy are an essential development for future research. Taking only the sites analysed for the present research, potters at Sala Consilina (phase I) produced simple vessel shapes in continuity with the Initial Copper Age but with slightly different techniques. For the first time they used moulds for the shaping of large bowls and attached tunnel handles on the outside of the vessel rather than on the inside as attested in Initial Copper Age Pontecagnano. The large hanging jars and bowls typical of the site were accurately produced with smoothed or burnished surfaces and with the almost exclusive use of crushed quartz of metamorphic origin as temper. Following the trend started in Initial Copper Age, potters at Sala Consilina preferred coarse fabrics in order to give large vessel shapes a stronger matrix to avoid collapse during their construction process [22,29,104]. Fine fabrics were extremely rare, and only in one case is the use of a sandy calcareous clay attested for a large bowl fired in reducing conditions (contrarily to the oxidising figulina tradition).

The technological variation between the Initial Copper Age Pontecagnano and the Early Copper Age at Sala Consilina is striking since the continuity in the vessel types is not paralleled by technological continuity in the forming techniques, suggesting that potters adhered to a different potting tradition but produced types still connected to the previous trends. Based on vessels types and style it can be hypothesised that cremations at Sala Consilina are slightly older than the radiocarbon-dated ones at the site of Taurasi. In fact, the ceramic repertoire from Taurasi and other cremation contexts such as Giardinetto in Puglia [167] and Salve [168] on one side share the same decoration techniques and patterns of Sala Consilina (I), on the other display some of the characteristics of Early/Middle Copper Age such as the abundance of jugs and pouring vessels, and the appearance of specialised shapes. Nevertheless, the current state of the research on cremation burials in Southern Italy does not allow a more detailed interpretation of these practices. New excavations and the re-examination of old ones [169] will hopefully bring to a further understanding of this complex tradition and ceramic repertoire.

In contrast, the typological and contextual analysis shows that the ceramic products deposited in collective rock-cut cemeteries (of the Gaudo cultural tradition) are characterised by a strong internal consistency in types and technological choices across different burial sites (e.g., Paestum and Pontecagnano, 25 km apart). Towards the middle of the 4th millennium BC, potters in Campania exerted a strong creative effort for the formalisation of a completely new ceramic repertoire destined to be deposited in rock-cut collective tombs, introducing pouring vessels, different types of hanging jars, pyxides, etc., for a total of 11 new shapes (see Fig 4). This creative process, already visible in some of the earlier cremation burials, is particularly evident in the long standing Gaudo tradition, where this new repertoire becomes codified and continues with a strong homogeneity for almost 1000 years (about 3500−2500 BC, according to radiocarbon dates from the cemetery of Paestum) [66]. In fact, vessel types, forming techniques and decorative schemes show a certain variability and creativity but always within well-defined boundaries. The data gathered in the Campania region for the Gaudo cultural tradition (with parallels in the area of Rome) [22], display an impressive consistency in terms of production, resulting in a codification of all stages of the chaîne opératoire, including the recipes adopted, the types and morphologies of vessels, the careful smoothing/polishing of surfaces, and the normative use of decoration. This strict codification of the ceramic vessels, both in their visible features (e.g., shape, decoration and surface treatment) and invisible features (e.g., fabric and firing temperatures), finds correspondence in the equally formalised ritual practices performed at the Gaudo rock-cut burials (e.g., fragmentation of jars in the entrance shaft, depositions in the chambers, secondary manipulation of the human remains, etc.) [77].

The predominance of collective mortuary rituals, including both inhumation and cremation, emphasised bonds with ancestors and kin-based groups. Together with the codification and relative standardisation of ceramic production for funerary purposes, this suggests the presence of close-knit human groups that invested in the transmission and perpetuation of communal bonds for almost 1000 years.

Roughly around 2900 BC, alongside large rock-cut tomb cemeteries (that endure until about 2500 BC), large settlements with annexed cemeteries, mostly with single trench graves and more varied funerary goods, emerge once again in several plains and inner areas of Campania attributed to Laterza cultural tradition. In these contexts greater typological and technological variability in the ceramic production can be found both within the same site and between different sites, while ceramic assemblages from settlements and burials cannot be distinguished. Potters continued to use the technological innovations introduced during the 4th millennium BC such as grog temper and moulds for shaping especially shallow and small bowls. The ceramic repertoire is characterised by a smaller number of vessel types than the Gaudo culture (jugs, pyxides and hanging jars are no longer attested) and it has greater internal variability. For example, small bowls are particularly common, but may have a hemispherical, sinuous or carinated profile, an everted, raised or inward bending rim, etc. This variability is also reflected in technological choices, such as paste recipes, surface treatments and decorations, which are not standardised in contrast to the Gaudo tradition.

The same paste recipes are broadly attested for vessels used both as grave goods and for domestic activities. The use of grog is still dominant but accompanied by an increasing use of sandy quartz-feldspathic fabrics, both for coarse and fine vessels. The use of sand as temper instead of grog might be indicative of a slight change in the production process. In terms of structural strength, quartz-tempered fabrics are similar to those tempered with grog, although the thermal expansion of the quartz when heated repeatedly makes it less suitable for cooking pots [34,170]. The increasing use of sand in vessels used both in domestic and funerary contexts implies a change in the production process, rendering the breaking and grinding of large quantities of pots no longer necessary. The reduction in the use of grog may also imply a decrease in its symbolic meaning paralleled by a decrease in the importance of ceramic vessels in mortuary rituals.

Indeed, the normative Gaudo ceramic repertoire is replaced in Laterza mortuary contexts by fewer vessel types with greater morphological simplicity, rarely composite and displaying greater variability in single attributes (shape of the rim, carination, etc.) and decorations. Pouring vessels (jugs and askoi) and hanging jars (including pyxides) disappear from both cemeteries and villages, which have similar pottery productions with the main vessel types widely attested in both contexts. Compared to rock-cut cemeteries, the ceramic repertoire involved in the funerary rite differs significantly, suggesting a change in the ritual practices carried out during the funerary rite. Funerary practices themselves appear more varied within different structures. Rock-cut collective burials are less common in Campania, except for the site of Paestum, in favour of single or double graves, mostly in trenches or pits, organised in cemeteries or located inside the settlements. Tombs are always characterised by fewer grave goods, involving one or two pottery vessels and in some cases metal or stone objects, also in this case typologically differing from Gaudo products.

This variation in burial customs and funerary rituals implies an important change in the symbolic world of Campanian human groups during the 3rd millennium BC. Alongside the longstanding tradition of collective rock-cut burials, flat single graves started to emerge and would become exclusive by the second half of the 3rd millennium BC. These changes in the mortuary sphere were accompanied by the onset of large settlements linked to agricultural exploitation probably boosted by the adoption and development of new agricultural technologies such as the plough and animal traction. This is particularly evident in the Campana plain where, preserved by eruption layers, traces of ploughing and carts can be recognised already in this period, and which increase conspicuously in the Early Bronze Age [171], suggesting an intensification of subsistence practices in particular in large plains richly endowed with volcanic soils such as the Campana plain and the area of Rome. Villages occupy the same areas for long periods with a gradual shifting of the houses, probably due to land exploitation or rebuilding, as exemplified by the sites of Gricignano d’Aversa and Acerra where areas previously occupied by houses became dedicated to cemeteries [60,61,64]. Larger settlements allowed for a more conspicuous ceramic production dedicated to the residents’ daily activities, such as large storage vessels and technical ceramics. In fact, in Laterza contexts pottery seems to respond more to the demands of everyday practicality rather than ritual. New categories of functional ceramic objects appear, such as the tuyère, the boiler lids, and portable slabs (mostly attested from the Bronze Age onwards). This suggests a development, both in ceramic production and in domestic activities such as processing of materials, food processing, metalworking and weaving, which required a new set of ceramic objects, also attested at the Final Copper Age site of Atena Lucana. This supports the hypothesis of an increase in the complexity of food processing activities, using boiler lids and strainers. At the same time, the presence of tuyères suggests the existence of other productive activities at Paestum and Atena Lucana possibly linked to metalworking.

The second half of the 3rd millennium BC: Extraterritorial contacts and the structuring of Early Bronze Age societies

The important changes in ceramics and society seen during the 3rd millennium BC in Campania can also be connected to increasing interconnectivity on inter-regional and even longer-distance scales in the Late and Final Copper Age, roughly from 2500 BC onwards, corresponding to the Bell Beaker and Cetina cultural phenomena. Human groups, including potters, in the Campania region were impacted by and engaged with these phenomena, as testified by the Bell Beaker fragments at Paestum and other key sites, and by the Cetina style of ceramics from Atena Lucana.

Compared to the societies of the Full Copper Age, these Late Copper Age communities appear to have been more open and receptive to external stimuli, which were adopted in full or reworked. In this context, it is possible that, instead of ancestral kin-groups, villages and their communication networks served to consolidate the community and determine its status. In this way, material culture and especially ceramics also moved away from their previous codified ritual character, assuming more dynamic mechanisms for the transmission of information, with more space available for creativity and improvisation, albeit within a shared and recognisable aesthetic.

The social dynamics inherent in the formation of villages and cemeteries, in ritual performances, and in developments in ceramic production during the second half of the 3rd millennium BC laid the foundations for subsequent Early Bronze Age communities, defined in Campania as the Palma Campania cultural tradition, and marked by an even stronger contribution from long distance connections [78,80] that also characterised Southern Europe on an even broader scale. The incomplete picture currently available for the final centuries of the 3rd millennium BC in Campania includes long-distance connections along different routes pointing to North-Central Italy, especially for the Bell Beaker and metal objects [80], but also to the Balkan area for the Cetina style and the spread of tumuli in funerary contexts [172]. Ongoing projects focusing on the mobility of people and objects [80,173,174] will help to further clarify this picture―one in which ceramic production was entangled with social and cultural processes.