Research Article: Cereal processing at Early Neolithic Göbekli Tepe, southeastern Turkey

Date Published: May 1, 2019

Publisher: Public Library of Science

Author(s): Laura Dietrich, Julia Meister, Oliver Dietrich, Jens Notroff, Janika Kiep, Julia Heeb, André Beuger, Brigitta Schütt, Peter F. Biehl.


We analyze the processing of cereals and its role at Early Neolithic Göbekli Tepe, southeastern Anatolia (10th / 9th millennium BC), a site that has aroused much debate in archaeological discourse. To date, only zooarchaeological evidence has been discussed in regard to the subsistence of its builders. Göbekli Tepe consists of monumental round to oval buildings, erected in an earlier phase, and smaller rectangular buildings, built around them in a partially contemporaneous and later phase. The monumental buildings are best known as they were in the focus of research. They are around 20 m in diameter and have stone pillars that are up to 5.5 m high and often richly decorated. The rectangular buildings are smaller and–in some cases–have up to 2 m high, mostly undecorated, pillars. Especially striking is the number of tools related to food processing, including grinding slabs/bowls, handstones, pestles, and mortars, which have not been studied before. We analyzed more than 7000 artifacts for the present contribution. The high frequency of artifacts is unusual for contemporary sites in the region. Using an integrated approach of formal, experimental, and macro- / microscopical use-wear analyses we show that Neolithic people at Göbekli Tepe have produced standardized and efficient grinding tools, most of which have been used for the processing of cereals. Additional phytolith analysis confirms the massive presence of cereals at the site, filling the gap left by the weakly preserved charred macro-rests. The organization of work and food supply has always been a central question of research into Göbekli Tepe, as the construction and maintenance of the monumental architecture would have necessitated a considerable work force. Contextual analyses of the distribution of the elements of the grinding kit on site highlight a clear link between plant food preparation and the rectangular buildings and indicate clear delimitations of working areas for food production on the terraces the structures lie on, surrounding the circular buildings. There is evidence for extensive plant food processing and archaeozoological data hint at large-scale hunting of gazelle between midsummer and autumn. As no large storage facilities have been identified, we argue for a production of food for immediate use and interpret these seasonal peaks in activity at the site as evidence for the organization of large work feasts.

Partial Text

Cereal food is one of the main components of the modern human diet. Its integration into the subsistence strategy during the late Epipalaeolithic (c. 12500–9600 cal BC) and Pre-Pottery Neolithic (PPN, c. 9600–7000 cal BC) has been recognized as a very long and complex process involving the selection and utilization of plants, strategies of exploitation of plants and land, the development of cultivation, and ways of processing, storing and consuming plants [1–12]. The establishment of agricultural economies at the end of the later part of the Pre-Pottery Neolithic (PPNB, c. 8800–7000 cal BC), comprising the deliberate, large-scale cultivation of domesticated cereals and other domesticated plants [1, 13–16], was predated by a longer period of experimentation and technological modifications that led to the development of a specialized tool kit for plant food processing [17–24]. Typical implements for cereal processing are pounding and grinding tools used in pairs, comprising a static low implement (mortar, grinding slab or grinding bowl) and an active upper tool that is moved across its surface (pestle or handstone) [25]. The different processes for fragmenting cereals include de-hulling, pearling, polishing or grinding to fine flour and are also ethnographically attested [26]. The aim of all these techniques is to enhance the digestibility of cereals, lower their cooking time and raise their dietary energy [27]. Early direct evidence for the processing of cereals to fine flour through grinding was found in the Early Epipaleolithic site of Ohalo II, dated to c. 21000 cal BC [20, 27, 28] and the Early to Late Natufian site of Shubayqa, dated to 12500–9600 cal BC [29]. However, the regular processing of wild cereals through grinding seems to have been established first in the Late Natufian, as suggested by macrobotanical evidence [11, 30] as well as by morphological changes in grinding stones combined with use-wear analyses [17]. The increase of grinding stones with typical use-wear in Late Natufian contexts [17, 18] and the reduction of forms of bed-rock mortars at the end of the Natufian and during the PPNA (c. 8800–7000 cal BC) [31] can be interpreted as indicating increased processing of cereals as food sources and the establishment of grinding as a more effective processing technique. New analyses seem to confirm the important role these features played in the processing of cereals and the production of beer in the Late Natufian [32, 33]. Flat, large grinding stones and handstones became a supra-regional standard during the subsequent Levantine PPN, constituting an integral part of the architecture [21, 22, 30]. This development seems to coincide with the general trend of increasing use and production of cereals [1–12, 34]. However, there was significant regional variability in the establishment of cereals as one of the main food sources [2, 9].

Göbekli Tepe is situated high on the Germuş mountain range at ca. 770 m asl., offering a wide view over the Harran plain to the south. The mound of reddish soil with a height of about 15 m has a diameter of around 300 m and is characterized by several hilltops divided by depressions. It is surrounded by a limestone plateau, which today mostly shows no sediment cover and very scarce vegetation. This must also have been the case during the Neolithic, as numerous quarry sites, cupholes and petroglyphs on the limestone surfaces suggest [77, 79].

In archaeological analyses, the functions of grinding equipment are usually assessed by use-wear analysis on original finds and comparison of traces to experimentally obtained reference collections [78]. There has also, however, been a trend to separately evaluate use-wear and surface transformations of objects and their formal development [21, 22]. These two lines of analysis should be brought together for a consistent interpretation of tool functions.

In a first step, we conducted phytolith analyses on nine soil samples (S4–S8 Tables). Eight of the soil samples were taken from the major N-S profile in the main excavation area (Fig 4). At the moment of sampling, the highest still standing part of this profile, including the whole sequence through building D´s filling sediments and layer II constructions up to the original topsoil, was located in excavation area L9-69, where a deep sounding had been dug to the bedrock [95]. We obtained samples M11-266, M11-267, M11-270 and M11-274 from an agglomeration of terrazzo floors and from open spaces between the rectangular buildings and the monumental building; while sample M11-264 is from the inside of a rectangular layer II building. Samples M11-263 and M11-269 are from building D´s filling (likely from an older building phase, we took the sample between the youngest and an older ring wall, see below); M11-276 is from adjacent stone slabs. As stated above, the younger phases of the monumental buildings are likely contemporary with the rectangular buildings; samples M11-264, M11-266, M11-267, M11-270 and M11-274 stem from contexts possibly dating to this phase; all but M11-264 and M11-267 were taken on remains of terrazzo floors that probably were affected by slope slides into the monumental building. M13-133 is from a different context, a large limestone vessel in a rectangular building at the northwestern hilltop [51] and was mainly included to test phytolith preservation throughout the whole site.

Approximately one third of the grinding equipment is from the uppermost layer I and thus chronologically undiagnostic (S2 Table). Nevertheless, the spatial distribution of the remaining finds (Fig 11) is indicative of the organization of cereal processing activities at Göbekli Tepe. We will start our overview with the partly younger layer II and then contrast the findings with evidence from layer III.

To facilitate research on the spatial distribution of grinding equipment, we divided the built spaces of layer II into seven zones (S2 Table). Finds on and immediately above floor levels of the rectangular buildings or on floor levels of niches in the buildings (zones 3–5) are considered in situ. For the other zones, including the infill (zone 2) and spaces between rooms (zone 7), dynamic and secondary formation processes have to be considered. Zone 6 refers to grinding stones used in secondary contexts as wall stones; zone 1 is the disturbed uppermost part of the buildings´ fillings and the plough horizon.

This integrated scientific archaeological approach has for the first time produced a basis for assessing the role of cereals at Göbekli Tepe. The massive presence of grinding equipment and standardization in the production and use of handstones hint at large-scale cereal processing in layer II. This is supported by use-wear traces and the presence of phytoliths in samples from their surfaces.