Introductory remarks

Mureybet is the key site for the interpretation and definition of the Khiamian on the middle Euphrates. On the basis of the stratigraphy, Jacques Cauvin (1972, 1978) illustrated his “Révolution des symboles“ before the beginnings of regular agriculture. 

Thanks to the detailed final publication of the site (Ibañez 2008), it now is possible to reinterpret the stratigraphic sequence and compare it to the old and new radiocarbon dates (Evin, Storduer 2008; Hours et al. 1994:404-407).

Originally the “Khiamian” was introduced only because of the then-postulated decline of the Natufian in its later phases: because of the perceived „downward trend” in the Natufian and the differences observed of the then-newly-excavated site of El-Khiam, Echegaray called formerly Natufian points “El Khiam points” and the latest Natufian was named the “Khiamian” (Benz 2000:30-37). However, the reality of this “ill defined entity” is still uncertain. Opinions vary from a ‘cries for help’ to ‘save the Khiamian’ (Ronen, Lechevallier 1999) to the suggestion that the ‘Khiamian’ is merely a mixture of late Natufian with PPNA layers (Sultanian) (Nadel 1990:94, 97; Schyle 1996:189; Kuijt 1997). The work of Kuijt (2002-2004:296-302) has recently demonstrated how difficult the separation of late Natufian and early Khiamian can be, both chronologically and typologically: he had to explain aberrant dates and types by appeal to massive dislocation and erosion processes.

Mureybet is one of the rare sites that allow a detailed comparison of the stratigraphy with many radiocarbon dates. Moreover, the site has been dated by different laboratories. Therefore it is not only possible to compare the stratigraphy with the radiocarbon dates, but also to compare the work of different laboratories – with possibly far-reaching consequences for the interpretation of the radiocarbon dates in the future.

In Phase IB, which initially was not differentiated from the Natufian Phase I and called “Epinatufien”, Khiamian points and a building structure, Maison 37, appeared. More than 40% of the other lithic tools were microliths, which had close affinities to those of the Natufian (Cauvin, Abbès 2008:293-300).^[1]

There are no building structures, only clay floors in the lower Natufian layers, Niveau IA. In consequence, there is a slight typological difference in the lithics and in the architecture. This difference increases during Niveaus IIA/B, which are characterized by changing architectural types and the decline of microliths.

The radiocarbon dates

The analysis of the uncalibrated radiocarbon dates shows that all the dates of Phase IB/IIA/IIB except the new dates of Lyon fall within the same time range (Fig. 1).

There are four dates of Phase IIA that, depending on the laboratory, are either late Natufian (Pennsylvania) or around 9900-10000 BP (Lyon). Evin and Stordeur discarded the old dates of Louvain for the Khiamian layers, preferring those of Lyon. They consider the conventional Ly-date for the Natufian Phase IA (Niveau 1-2) too young, although two dates of Monaco are hardly older. They postulate a clear chronological separation between the late Natufian and the early Khiamian and suggest the beginning of the Khiamian should be dated only about 9600 BC.

As this interpretation will have far reaching consequences for the cultural development of the early Neolithic, we will consider in more detail whether this chronological distinction is merely due to laboratory differences or can be used as proof of separate cultural entities.

There are in total 42 samples from Mureybet. Two of the dates (Lv 604: 9730+/- 140 BP for Phase III [ASPRO 1994:405] and P-1216: 10092+/-118 BP [Radiocarbon 11, 1969, 151]) were omitted from the extensive list of new radiocarbon dates from Lyon (Evin, Stordeur 2008:28-30). ^[2]

For the following calibration all dates were included unless omitted on reasonable grounds by Evin and Stordeur, or if their deviation was higher than +/- 150. Dates of such high deviations would result in time ranges of up to 1400 years (e.g. MC 611: 9840+/-200 BP = 10155-8722 BC cal) and would thus be unable to clarify cultural changes.

In the diagram ^[3] (Fig. 1) all dates are arranged by laboratory and stratigraphy. Those with a cross were discarded by Evin and Stordeur. The dates of Pennsylvania (P) could be attributed only to certain phases but not to specific Niveaus. They were thus placed in the middle of the phases; but their actual positions could vary within these phases according to whether they originally had come from younger or older levels. The same holds true for the sample Lv 604, that as was mentioned above, Evin and Stordeur ignored and for which a precise niveau is missing. It was attributed by Hours et al. (1994:405) to Phase III.

To facilitate the inter-laboratory comparison, the second diagram (Fig. 2) shows only those levels for which dates of Lyon, Pennsylvania and Monaco were available.

Several points can be observed:

1)   The conventional dates of Lyon (Ly-xxxxx) are generally younger than all other dates (see Evin, Stordeur 2008:25ff.). The stratigraphic position of Ly 11.623 is not certain. If its attribution to Phase IA is in fact incorrect, the Natufian layers would then be dated only by Monaco. The younger conventional Ly-dates would then only date Phase IB/IIA, which would better correspond to the stratigraphy.
Concerning Phase III, the difference between the Monaco and Lyon dates remains after calibration, especially for Niveau 14 a/b. The only Ly-date that overlaps with the younger MC-dates comes from the younger Niveau 14b (Ly 11630), whereas the younger Lyon dates come from the older Niveau 14a. However, MC 734 only slightly overlaps with the other MC-dates of Niveau 14a. Compared to Ly 11625 of the lower Niveau 13, it seems to be too old and it is questionable whether or not MC 734 should be retained. The result of this discussion might have some significance for the dating of Maison 47.
Two of the dates (Ly 11.624 and Ly 11.629) are indeed much too young and were discarded by Evin and Stordeur (Fig.1).

2)   The samples dated by Oxford (OxA), Groningen (GrA), and Pennsylvania (P) fall into the date ranges of Monaco. But the dates of Pennsylvania cannot be taken into consideration for comparison due to their only approximately-known stratigraphic position. The Pennsylvania dates that were attributed to Phase II point rather to a Natufian context and their stratigraphic position should thus be questioned.

3)   The dates of Louvain are older than all others for all the niveaus. The only exception is Lv 604 for Phase III, which overlaps with the ranges of the P- and MC-dates.

4)   The dates can be subdivided into three groups: Niveaus 1-6; Niveaus 13-17, and Niveaus 19-20. These phases correspond to the following chronological subdivisions: Phase I/II = late Natufian (10150-9250 BC)/Khiamian (9760-9280 BC) ^[4], Phase IIIA/B = Mureybetien (IIIA: 9350-8600 BC; IIIB 9150-8500 BC), and Phase IVA/IVB = PPNB ancien/moyen (IVA: 8750-7950 BC; IVB 8650-7950 BC).

Comments to Points 1-4

Point 1: Interlaboratory Differences

Comparison of the Lyon/Ly-dates of Mureybet with the dates of Jerf el Ahmar could give some information about the validity of these dates. The stratigraphic position for five of the radiocarbon dates of Jerf el Ahmar have been published. The dates are consistent with the stratigraphy. A general deviation of the conventional Ly-dates thus seems improbable.

The dates of Dja’dé also confirm the general reliability of the Lyon dates because the measurements of Lyon overlap and correlate well with the UtC-date for the same level Utc 2369: 9200+/-100.

The Aswad dates likewise show that the Lyon dates should not be considered too young. Although for the older PPNB-phase of Aswad (de Contenson’s Phase Ia, a PPNA horizon, which has been discarded after the new excavations) all three conventional Ly-dates (12782; 12781; 12107) are much younger than the Gif-dates and slightly younger than the Granson (GrA) dates, this difference is not seen for the later phases of the site.^[5]

In conclusion, generally younger Ly-dates can be demonstrated for Mureybet (Ly-11.6xx-11.788), but this trend is not observed at other sites. Perhaps the dates (Ly 12107; 12781; 12782) of Aswad are also too young; but for further conclusions a detailed verification of the stratigraphy and more radiocarbon dates from the different laboratories are necessary.

Point 2: Validity of Pennsylvania and Monaco Dates, and the Dating of Maison 47

As the dates of OxA, P, GrA and Monaco give relatively similar results, they seem to be reliable for Phase III/IV.

Therefore, the omission of P 1220 and MC 734 for dating Maison 47 should be reconsidered. The main argument by Evin and Stordeur (2008) for the exclusion of P 1220 seems to be that a similar building of Jerf el Ahmar was dated to 9445+/-75 BP (Ly 10649: 9110-8610 BC). But the dates of Jerf el Ahmar that are in favour of a younger date for Maison 47, and were used as an argument for the exclusion of the P- and MC-dates, were exclusively from Lyon.
Moreover, the architectural stratigraphy of Jerf el Ahmar itself suggests the reliability of dates P 1220 and MC 734. The communal building EA 30 of the western tell has an precursor on the eastern tell, EA 7, that resembles Maison 47 because it is both ‘semi-souterrain’ and divided into several compartments. Stordeur et al. (2001:42f.) present two further similarities between EA 7 and Maison 47 of Mureybet in contrast to EA 30:

1)   Whereas the roof of EA 30 was supported by subdividing walls, the roofs of EA 7 and Maison 47 were laid on wooden posts.

2)   Around EA 30 and its precursor on the western part of the tell there are mainly rectangular buildings of a type that existed in the eastern part only at the beginning of Niveau 0. In contrast, around EA 7 there are mainly round buildings, or buildings with rounded corners. The same holds true for Mureybet, where no trace of rectangular buildings was found in the layer of Maison 47. However, this could be due to the small excavation area.

As it is probable that EA 7 and Niveau I/Est can be attributed to the chronological Phase 2 of Aswad, dating between 9310 BC and 8910 BC, the date for Maison 47 might therefore be somewhere between 9300 BC and 8900 BC, a time horizon within which the upper margins after calibration of M734 and P1220 would reach.

Despite these chronological considerations, the dates for Maison 47 (M 734 9950 +/-150 BP and P 1220: 9885 +/- 115 BP) seem to be rather high. M 734 is in fact the oldest MC-date of this Niveau. After calibration it overlaps with the other radiocarbon time ranges of Phase IIIA/B but is out of the 1-σ range of the Ly-date for Maison 47.

Comparison with Jericho, Beidha and Ras Shamra shows that the Pennsylvania dates do not tend to be either too young or too old.

As it will not be possible without another independent dating method (e.g. dendrochronology) to check the radiocarbon dates, all the dates for Phase IIIA of Mureybet have been retained. General differences among laboratories’ results could not be proven. If the older date of Maison 47 would be discarded on the basis of the architectural similarities with EA 30 of Jerf el Ahmar alone, no diachronic developments could ever be reconstructed. Additionally, two other buildings E7 and one undesignated, similar but older than E 30 were found in Jerf el Ahmar.

Point 3:

The exclusion of the Louvain radiocarbon dates by Stordeur and Evin is very reasonable, as the Louvain dates have a high deviation and vary significantly from the mean value of the other dates. They were therefore also ignored here. The only exception is the Louvain date 604.

Point 4:

The following conclusions may be suggested from the above observations of the chronological subdivision (s.a.): On the basis of the radiocarbon dates, the latest Natufian and the Khiamian are not yet distinguishable chronologically. The late dates that date the Khiamian were analysed exclusively in Lyon, and these dates are nearly identical to the one that should date the Natufian layer (Ly 11.623).

The rapid change from the late Natufian to the early Khiamian seems very plausible.

The hiatus of the dates between the first and the second set of dates is the result of the missing samples for Niveau 8-12: this means that the late Khiamian layer IIB is not dated at all. Because of the typological similarities and continuities with the Mureybetian, a long hiatus seems to be improbable.

Nor it is possible by radiocarbon dates to support the stratigraphic subdivision of the Phases IIIA and IIIB. The dates scatter in accordance with the above made observations of laboratories’ differences in both phases in a relatively large range between 9200-8550 BC.^[6] The same holds true for the dates of the Phases IVA/B ranging between 8750-7950 BC.

On the basis of these observations and comparisons, the following hypotheses can be suggested:

1) The early Khiamian seems to be closely related to the late Natufian and a chronological separation of both entities on the basis of radiocarbon dates is not yet possible, although the stratigraphy, the typology, and the social changes support such a differentiation (Evin, Stordeur 2008:24-26). However, the problem is complicated by the small size of the excavated trench for phase IA (35m²) (Natufian) and by the fact that Phase IB (early Khiamian) is dated by only one Lyon date. The late appearance of el-Khiam points is therefore not certain and they may have been introduced before 9500 BC. It is thus still in doubt whether the Khiamian starts around 10000 BC or, as suggested by Evin and Stordeur (2008), only around 9600 BC. The close similarities of the botanical remains of the early Phases I/II at Mureybet with the Epipalaeolithic layers of Abu Hurerya would support a beginning of the Khiamian before the onset of Holocene vegetation  (Willcox 2007:26; 32). ^[7] Changes in the archaeozoological data are observable between Phases IA/IB and Phase II (Helmer 1998:84).

Clearly separated from these early Epipalaeolithic phases, the Mureybetian develops around 9200 BC and resembles the findings of Jerf el Ahmar and Abr’3 (Yartah 2004, 2005), Göbekli Tepe (Schmidt 2006), Cayönü (Özdogan 1999) and Körtik Tepe (Özkaya San 2007).

The possible deviation of the Lyon dates makes it desirable to verify this preliminary observation on other sites. If the analyses of Lyon are correct, sites like e.g. Hatula that are dated only by Gif-sur-Yvette (Gif) should be considered cautiously.

With their extensive publication of the radiocarbon dates of Mureybet, Evin and Stordeur established the basis for a detailed discussion of these dates and an interlaboratory comparison. Although it could be shown on the basis of the radiocarbon dates that the distinction between the late Natufian and the early Khiamian is still problematic, it is clear that the early Khiamian is more closely related to the late Natufian than to the Mureybetian.

On the basis of the radiocarbon dates, it is also possible to discuss the time range for the early communal “polyvalente” buildings of Jerf el Ahmar (s.o.) and Muerybet, with the synchronization of the absolute chronological dates of the latter with the early layers of Jerf el Ahmar as a preliminary working hypothesis. The stratigraphicly later communal buildings EA 30, EA 100, and EA 53 of Jerf el Ahmar date to a later horizon between 8800 and 8600 BC. It must be stressed that the above discussed Ly-date 11.626 for Maison 47 of Mureybet (IIIA) would fall within this later horizon.

The chronological horizon in which the early communal buildings of Mureybet and Jerf el Ahmar fall remains uncertain. On one hand, the rapid succession of the communal buildings EA 30 and EA 53 document a fast succession of architectural phases at Jerf el Ahmar. Similarly it could be suggested that EA 7 is only slightly younger than EA 30. The date of Ly 11626 (8800-8640 BC) would then be most probable.

On the other hand, the stratigraphic succession of Mureybet (Stordeur, Ibañez 2008:61-90) and of Jerf el Ahmar (Stordeur, Abbès 2002) make the attribution of EA 7 and its precursor, as well as of Maison 47 to an earlier horizon between 9300-8800 BC more probable. This would also correlate better with nearly all the dates of Niveau 14a and Niveau 14b (Ly 11630: 9120-8730 BC) of Mureybet.