Here we present extended parts of the very informative and interesting paper titled “Early seventh-millennium AMS dates from domestic seeds in the Initial Neolithic at Franchthi Cave (Argolid, Greece)“, by Catherine Perlès, Anita Quiles and Hélène Valladas (2018).
Abstract When, and by what route, did farming first reach Europe? A terrestrial model might envisage a gradual advance around the northern fringes of the Aegean, reaching Thrace and Macedonia before continuing southwards to Thessaly and the Peloponnese. New dates from Franchthi Cave in southern Greece, reported here, cast doubt on such a model, indicating that cereal cultivation, involving newly introduced crop species, began during the first half of the seventh millennium BC. This is earlier than in northern Greece and several centuries earlier than in Bulgaria, and suggests that farming spread to south-eastern Europe by a number of different routes, including potentially a maritime, island-hopping connection across the Aegean Sea. The results also illustrate the continuing importance of key sites such as Franchthi to our understanding of the European Neolithic transition, and the additional insights that can emerge from the application of new dating projects to these sites.
Introduction Due to its geographical location and early dates for the Neolithic, Greece usually appears in these broad-scale models as the origin of the spread of farming economies in Europe. These models, however, rest on a fragile basis, since there is actually no consensus on the mechanisms, timing and route of penetration of a Neolithic economy in Greece (e.g. Thissen 2000, 2005; Kotsakis 2001; Runnels 2003; Sampson 2005; Kyparissi-Apostolika 2006; Weninger et al. 2006; Seferiades 2007; Perles 2010). Depending on the presumed origin—local, western Anatolian, or multiple origins—and on the postulated mechanisms— indigenous development, spread by terrestrial or maritime routes, reaction to the ‘6200 cal BC climatic event’ (Alley et al. 1997)—dates older than 6200 or 6400 cal BC are either rejected or, on the contrary, readily accepted. In the most extreme statements, dates as far apart as 7000 cal BC (Perlès 2001) and 6200 cal BC (Weninger et al. 2006) have been defended for the earliest occurrence of the Neolithic in Greece. The late date hypothesis aligns the Greek Neolithic with that of the southern Balkans, thus supporting the view that the Neolithic colonisation of Europe originated from western Anatolia and was triggered by the deleterious effects of the ‘6200 cal BC climatic event’. By contrast, according to the early date hypothesis, no climatic event can be invoked to explain the spread of the farming economy. Furthermore, the early date hypothesis supports a model of multiple origins for the European Neolithic since the early Neolithic in western Anatolia, which is clearly related to that of Bulgaria, may not be earlier than the earliest Greek Neolithic.
The earliest Neolithic in Greece: a debated issue This debate, however, is relatively recent. When dates of c. 7000 cal BC for the early Neolithic of Greece were published in the 1960s and early 1970s, their antiquity posed no major problem. The debate between proponents of traditional ‘short’ chronologies and proponents of the new, ‘long’ 14C-based chronology had abated (Renfrew 1973), and the Greek Neolithic was commonly considered a direct outcome of the Near Eastern Neolithic, itself of greater antiquity (e.g. Weinberg 1970; Theocharis 1973). However, among the four sites with 14C dates from the first half of the seventh millennium, Nea Nikomedeia (Macedonia) was later re-dated by AMS and none of the new dates, on carbonised seeds, predated 6400 cal BC. The two oldest determinations from Argissa (Thessaly) were made on bone, had no precise provenance, and could be considered unreliable. Knossos (Crete) and Franchthi (Argolid) were thus the only two remaining sites suggesting that the introduction of a farming economy in Greece dated back to the first half of the seventh millennium (see compilations in Reingruber & Thissen 2005, 2009).
The rejection of the early 14C dates from Franchthi and Knossos only occurred in the late 1990s, after the first settlements in western Anatolia and Thrace—the presumed direct ancestors of the Greek Neolithic—were excavated and dated at the time to the end of the seventh millennium cal BC. Thissen in particular then stated that, since the early Neolithic in western and north-western Turkey did not go back further than c. 6200 cal BC, the Greek Neolithic could be no older*° (Thissen 2000: 161, see also Thissen 2005). He recognised nevertheless that an ‘aceramic’ Neolithic was represented at Knossos and Franchthi, but he considered these episodes as brief and of strictly local significance*°. More recently, Reingruber and Thissen (2009), despite abundant stratigraphic and artefactual evidence to the contrary (e.g. Jacobsen & Farrand 1987; Vitelli 1993; Perlès 2004), rejected the presence of an Early Neolithic at Franchthi and in the Peloponnese*°. As a consequence, the earliest Neolithic in Greece would have been that of Thessaly, around 6500/6400 cal BC, in line with the more recent dates from western Anatolia, while the Peloponnese would have lagged behind until about 6000 cal BC (Reingruber & Thissen 2009: 762).
An even more drastic stand was taken by Weninger and his collaborators who rejected all dates older than 6200 cal BC. They concluded that Macedonia (Nea Nikomedeia) and Thessaly (Sesklo and Achilleion) were first settled during ‘the 6200 cal BC event’ and considered ‘…that the 8200 cal yr BP aridity triggered the spread of early farmers out of Anatolia into Greek Macedonia as well as in the fertile floodplains of Thessaly…’ (Weninger et al. 2006: 417)*°. Franchthi and Theopetra (Thessaly), as ‘cave sites’, had been left out of the discussion.
*°(NovoScriptorium: So, instead of changing our theory in accordance with the actual finds, we choose to hypothesize impossible (and irrational sometimes) alternatives or exclude a priori other alternatives, just because we want a certain theory to be served. This doesn’t really look at all ‘scientific’! This looks like ‘dogma in Science’ really, something which is disappointing and very dangerous, too)
These readings of the dates are clearly linked to the hypothesis that the Greek Neolithic must come from western Anatolia, and, therefore, cannot be older than the latter. In the meantime, however, one of the present authors (CP) had argued that the influence of western Anatolia was limited to Macedonia. Other origins, through repeated episodes of maritime colonisation, had to be sought for the establishment of the Neolithic further south (Perlès 2005, 2010). According to this model, the dates from western Anatolia no longer constituted a terminus ante quem, and the multiple origins that were argued for made it possible to accept a wider range of dates.
Given the importance of the site of Franchthi in the debate and the unreliability of radiocarbon dates obtained more than 30 years ago, new dates on the Initial Neolithic at Franchthi were clearly required. We shall report here on a series of charcoal and seed AMS determinations related to the Mesolithic–Neolithic transition at Franchthi, including two dates on seeds of domesticated wheat.
The context The Upper Mesolithic was excavated in trenches, FAS, FAN, H, H1B, H1A and G1. It is characterised by intensive tuna fishing*¹ (Rose 1995) and a very specific lithic assemblage, which comprises numerous atypical non-geometric microliths (Perlès 1990). It has already been 14C dated by six conventional radiocarbon determinations. With one exception, they give coherent results between c. 8400 and 7400 cal BC (2σ). This phase has not been re-dated. (The outlying sample comes from H1A, at the limit between the Lower and Upper Mesolithic. The date clearly indicates a Lower Mesolithic contamination.)
*¹(NovoScriptorium: From the paper titled “Late Pleistocene/Early Holocene seafaring in the Aegean: new obsidian hydration dates with the SIMS-SS method“, by N. Laskaris, A. Sampson, F. Mavridis, I. Liritzis, 2011, we learn that already from the 13th millenium Melian obsidian finds from Francthi prove sea-trade and maritime capabilities of its inhabitants. If one takes into account that there appears to have been continual habitation-usage of the cave at least for 40,000 years – with reference to “Franchthi Cave revisited: the age of the Aurignacian in south-eastern Europe“, by K. Douka,C. Perlès, H. Valladas, M. Vanhaeren & R.E.M. Hedges, 2011- it is far more reasonable to assume ‘Neolithisation’ in the area as an indegenous evolution rather than a ‘migration product’, without excluding the possibility of adopting the idea from somewhere else. It is also interesting to note that remnants of 200kg tuna were found at Francthi, suggesting that the ships used to catch the big fish were not small or weak but, rather, sufficiently large and capable)
The Final Mesolithic is present in situ in FAS and FAN (Stratum X1) but is suspected to be reworked in H1A and H1B, due to deep post-Neolithic excavations (Vitelli 1993:33). Both tuna vertebrae and non-geometric microliths have virtually disappeared. New transverse arrowheads appear, and the economy relies on a diverse spectrum based on wild fruit, legumes, cereals, a few land snails and the hunting of red deer, hares, turtles and foxes (Munro & Stiner 2012). The Final Mesolithic had not been 14C-dated previously in the reference trenches. Two dates from other trenches (G1 22 and FF1 43A1), around 7500–6700calBC, were potentially compatible with the Final Mesolithic but the attribution could not be fully confirmed on an archaeological basis. Three new charcoal samples from reliable contexts, FAN 169 and FAN 166, were thus submitted for analysis.
In the Initial Neolithic, domestic sheep, goats, wheat (Triticum dicoccum) and barley (Hordeum distichum) appear in small quantities (Payne 1975; Hansen 1991), together with trapezes on pressure-flaked blades (Perles 1990). The Initial Neolithic was labelled as “possibly aceramic Neolithic” in a preliminary report (Jacobsen*² 1969: 352) in view of the very small number and small size of the sherds it contained, but this denomination was not retained subsequently by the Franchthi specialists (see ‘Initial Neolithic’ in Perlès 1990; ‘Interphase 0/1’ in Vitelli 1993). In FAN, FAS and possibly FF1, it corresponds to the well-defined sedimentary stratum called the ‘grey clay’ by the excavators (Farrand 2000: 50–51).
*²(NovoScriptorium: Reports from Thomas Jacobsen, 1967-1976, suggested a Mesolithic Age 9,000-7,000 BC)
According to Farrand (2000: 50), five dates could be attributed to the ‘grey clay’ (Stratum X2). Two of them, however, FF1 43A1 and FAS 146, come from units that Farrand himself did not list as included in Stratum X2. Examination of the sections shows that FAS 146 may have cross-cut the underlying rocky deposit. The exact lithostratigraphic position of FF1 43A1 cannot be established, but both Perles (1990) and Vitelli (1993) consider it as probably Final Mesolithic. A third sample, from trench A, comes from a unit (A 63) with suspected Middle Neolithic contamination (Vitelli 1993: 37). Accordingly, only the dates from FAS 143 and FF1 44B5 could be securely attributed to the ‘grey clay’ stratum and can be considered to date the earliest Neolithic deposits at Franchthi. Both units were devoid of sherds.
We submitted three additional Initial Neolithic charcoal samples for dating (FAN 159, 158 and 151), all included within the ‘grey clay’ stratum. No sherds were found in unit FAN 159, one ‘chip’ was found in unit FAN 158, and 33g of sherds in unit FAN 151 (Vitelli 1993: 226). Of these three Initial Neolithic samples, two (FAN 159 and 151) gave 14C dates so close to Final Mesolithic dates that the charcoal could be suspected to be intrusive, while the third (FAN 158) gave a date younger by several centuries, corresponding to the Early Neolithic. We thus decided to date indisputable carbonised seeds of Triticum dicoccum from the same stratigraphic context. However, domestic seeds are not abundant in the Initial Neolithic (Hansen 1991: 139–44) and we were only allowed to take samples in units with around 10 seeds of wheat. The best potential sample, from FAS 145, could not be located. We thus chose two seeds identified by Hansen as T. dicoccum from unit FAN 163, at the interface of the rocky layer X1 and the ‘grey clay’ (X2), and from the overlying unit FAN 162, at the base of the ‘grey clay’ stratum. The lithics from both units show some mixing of Final Mesolithic and Initial Neolithic (Perles 1990: 91–93), but there is no wild progenitor of wheat in Greece and wheat was introduced as a domestic species in the Neolithic*³. These domestic seeds thus necessarily date the Neolithic, not the Mesolithic.
*³(NovoScriptorium: How can we be so positive that there was ‘no wild progenitor of wheat in Greece’? The correct thing to say would be that ‘until now, we know not of a wild progenitor of wheat in Greece‘)
Discussion The dates obtained on the carbonised seeds indicate without any possible doubt that cereal agriculture was practiced in southern Greece during the first half of the seventh millennium. The dates from seeds are younger by about 200 years than the dates obtained on charcoals from the same lithostratigraphic context, a difference usually attributed to the ‘old wood’ effect. Charcoal dates for the Initial Neolithic are very close to Final Mesolithic ones, which may suggest some Final Mesolithic contamination. Looking at the sections, that is indeed possible for FAS 146 which is included in the rocky layer X1. The other charcoal dates suggest a time gap of around 100 years between the Final Mesolithic and the Initial Neolithic, whereas the dates on seeds suggest an interval of c. 200 years. This short interval is probably compatible with both Farrand’s observation of some weathering of the Final Mesolithic rocky layer X1 (Farrand 2000: 51) and the strong continuity in the chipped stone tool assemblages, marine molluscs and ornaments between the Final Mesolithic and Initial Neolithic (Shackleton 1988; Perlès 1990:116, 2013). It reinforces the suggestion that the Initial Neolithic reflects a phenomenon of acculturation by local (ex-)hunter-gatherers (Perlès 1990). On the other hand, our new dates obviously do not solve the problem of the status, ‘aceramic’ or ‘ceramic’, of the Initial Neolithic. The few sherds found in some units of the ‘grey clay’ stratum in FAN and FAS all belong to dominant Early Neolithic wares. They can be alternatively considered as intrusive, or as evidence for very scarce use of pots (Vitelli 1993: 39). No new data can be added to this ongoing discussion.
The demonstrated presence of Neolithic farmers at or around Franchthi during the first half of the seventh millennium leaves open the possibility of a brief episode of strictly local significance, as suggested early on by Thissen (2000: 191–92). Whether or not the Early Neolithic follows more or less directly at Franchthi cannot be confirmed from the few and unsatisfactory Early Neolithic dates available. The answer to this question requires another dating programme, focusing on the thick Early Neolithic deposits from Paralia and not the shallow and largely disturbed Early Neolithic deposits within the cave (Vitelli 1993: 44).
Conclusion Our results demonstrate that Franchthi was occupied by farmers before 6500 BC, not much later than the earliest occupation of Knossos in Crete, where an oak acorn was recently AMS-dated to 7040–6770 (OxA-9215) (Efstratiou et al. 2004). These results suggest that Crete and southern Greece may have been occupied earlier than Thessaly, where the few dates earlier than 6500 cal BC come from charcoals, not seeds. No Early Neolithic site has recently been excavated or dated in Thessaly, however, and earlier settlements may well await discovery. Greece can still produce unexpected results concerning the Early Neolithic, as shown by the recent series of dates from Dikili Tash and Mavropigi which demonstrate that Macedonia may have been settled as early as 6500 cal BC (Karamitrou-Mentessidi et al. 2013; Lespez et al. 2013).
All these new dates confirm the antiquity of the Neolithic in Greece, which, on
presently available evidence, precedes the earliest Neolithic occupations in Bulgaria by three to five centuries (Higham et al. 2011), in Italy by five to seven centuries (Binder), and in Spain by at least one millennium (Zilhao 2011). This shows clear phenomena of arrhythmia in the spread of the Neolithic in Europe.
Except for an early seventh millennium date at Ulucak (Derin pers. comm. June 2013), the earliest dates from Franchthi are also slightly older than the currently available dates from sites such as Yesilova, Ege Gubre and Araplı in western Anatolia, Hoca Cesme in Turkish Thrace or Ugurlu on the Aegean island of Gokceada, which do not go back beyond c. 6500 BC (Reingruber & Thissen 2005; Cilingiroglu & Cilingiroglu 2007; Saglamtimur 2007; Derin et al. 2009; Erdogu 2011). Radiocarbon dates, therefore, cannot be taken to demonstrate the chronological anteriority of the western Anatolian Neolithic over the Greek one. There is no doubt, on the other hand, that the Early Neolithic in Macedonia and Bulgaria shows strong links with western Anatolia and compatible radiocarbon dates.
The radiocarbon chronology, as available today, thus concurs with archaeological data and supports a model of multiple origins for the introduction of the Neolithic in Europe.
Research-Selection-Comments for NovoScriptorium: Philaretus Homerides