First of all, we would like to express our gratitude to the editorial board who encouraged us to take part in this epoch-making publication project. We also wish to thank many colleagues—including (in alphabetical order by family name) Makoto Arimura, the Council for British Research in the Levant, the late Donald O. Henry, Juan J. Ibáñez, Cheryl Makarewicz, Christoph Purschwitz, and Tobias Richter—for their kind support in collecting relevant information. Special thanks go to Karin Bartl, Andrew Garrard, Hans Georg K. Gebel, Garry O. Rollefson, Dörte Rokitta-Krumnow, and Denis Štefanisko, who readily provided unpublished information and/or drawings. Among others, frank views exchanged with Avraham Gopher contributed much to the refinement of the basic concept of this paper. Our thanks go as well to two anonymous peer reviewers, whose comments were very useful for improving the final draft. Incidentally, this paper is a result of the comprehensive investigation at the Harrat Juhayra sites, which was financially supported by KAKEN Grant-in-Aid [S] n^o 19H05592 from the Japan Society for the Promotion of Science.

Introduction

2If so, the question is the relationship among the three EPPNB cultural entities, which is expected to provide fresh insight into the PPNA/PPNB transition in the southern Levant. Our last paper proposed a few tentative perspectives on the issue (Fujii 2022b), but the discussion there was rather static in the sense that it treated the Jordanian Badia EPPNB as a homogeneous, single-phase culture. This paper approaches the issue from a more dynamic perspective of the settlement pattern and periodization of the newly defined Badia EPPNB. It is our aim to infuse new breath into the long-standing issue of the southern Levantine Neolithic archaeology from the dry land type (thus more sensitive to environmental changes) EPPNB culture. However, every related piece of fieldwork, including our own excavation at Harrat Juhayra 202, remains interrupted due to the prolonged pandemic, and available research data are still limited. We would like to point out in advance that this paper is a preliminary study toward full-scale discussion in the near future.

Jordanian Badia EPPNB Sites

3To date, the following six cases have been identified as Jordanian Badia EPPNB sites on the basis of both ¹⁴C dates and lithic assemblages (Harrat Juhayra 202 and Mushash 163) or the latter only (Abu Hudhud, Jilat 7, Jabal `Ainab 1, and Jebel Queisa; fig. 1). Before proceeding to the main subject, it would be useful to review the current state of field research (table 1).

Harrat Juhayra 202

Location and investigation

4Harrat Juhayra 202, or HJH-202 for short, is a key site of the Jordanian Badia EPPNB, located at the west-central edge of the Transjordanian drylands. In terms of topography, it occupies the southeastern corner of an elongated lava plateau (Harra in Arabic) that stretches eastward from the foot of Jabal Juhayra, a small-scale extinct volcanic hill at the northwestern corner of the al-Jafr Basin, southern Jordan. The site was discovered during the 2001/2002 winter season survey of the Jafr Basin Prehistoric Project headed by the author and tentatively registered as JF-0201 (Fujii and Abe 2008: table 1). The excavation started in 2016 as part of a comprehensive investigation of the Harrat Juhayra area but remains interrupted until now due to the pandemic.

¹⁴C dates and stratigraphy

5A dozen ¹⁴C dates, including yet-to-be published ones, from the floor deposits of structure 1 (mentioned below) and its surrounding loci converge in the first half of the 9th millennium cal. BCE, indicating that the first occupation at the site dates back to the very beginning of the EPPNB (fig. 2). These dates are coeval with, or even slightly earlier than, those from the northern Levantine EPPNB sites, encouraging fundamental reconsideration of the PPNA/PPNB transition issue in the southern Levant. In terms of stratigraphy, the excavated structure is sandwiched between a basalt bedrock layer and subsequent natural sediment, but in the southern half of the site, a pair of Middle Chalcolithic ossuaries overlies the EPPNB layers (Fujii et al. in press a: fig. 3).

Site and structures

6The distribution range of surface finds and curvilinear masonry walls (exposed in looter pits) suggests that the site formed an oval, relatively dense, hamlet-size settlement with an estimated total area of ca. 0.2 ha. To date, the excavation is limited to its northwestern corner (area 1), where a ground type masonry structure (structure 1) with an oval plan ca. 6–8 m in outer diameter was revealed (fig. 3.1). Aside from a small niche-like space at the northwestern corner, this structure was essentially of a single-room type, opening to the southeast. Outdoor features included a curvilinear windbreak wall beside a narrow entrance and a small open-cut cistern behind the northeastern wall. This structure is much larger in scale and higher in construction quality than those partly exposed in the looter pits, possibly suggesting its use as a communal building like structure O75 or O100 at Wadi Feinan 16 (hereafter WF16; Mithen et al. 2011: figs. 3–4).

Small finds

7Small finds from structure 1 are limited to a chipped flint/obsidian assemblage (n = ca. 15,000) and a ground limestone/basalt assemblage (n = 18) only; no other categories of artifacts, including stone vessels and personal adornments, have been attested so far, although this might be partly due to the limited excavation. Whatever the case, the scarcity of artifact variety is among the distinctive characteristics of the Jordanian Badia EPPNB, shared with the other five sites mentioned below.

Mushash 163

Location and investigation

9Mushash 163, another key site of the Jordanian Badia EPPNB, is located ca. 40 km southeast of Amman, in a semi-arid desert steppe near the northwestern edge of the Badia. An early Islamic desert castle, Qasr Mushash, confronts it with Wadi Mushash in between. This site was discovered in 2012 by the Qasr Mushash Survey Project directed by Karin Bartl and was excavated over four seasons between 2014 and 2017 (Lelek Tvetmarken and Bartl 2015; Bartl 2018, in press).

¹⁴C dates and stratigraphy

10Twenty ¹⁴C dates, including yet-to-be published ones, cover a time range of 8900/8800–8200 cal. BCE, suggesting that the settlement was established almost concurrently with, and used a few centuries longer than, HJH-202 (Bartl 2018: 31). In terms of stratigraphy, the EPPNB settlement is reported to be sandwiched between virgin soil and subsequent fill deposits containing Late Neolithic artifacts (Lelek Tvetmarken and Bartl 2015: 34).

Site and structures

11The excavations revealed an oval settlement ca. 0.25 ha in total area, which is estimated to contain several dozen round or oval semi-subterranean structures/features ca. 2–3 m in diameter (fig. 3.13). The excavated structures/features are arranged side by side a short distance apart to form a relatively dense settlement. Although published information is still limited, it appears that the settlement has much in common with HJH-202 in many aspects, including the settlement size and plan.

Small finds

Abu Hudhud

Location and investigation

13This site is located at the west-central edge of the Jordanian Badia, extending over a basalt-strewn steep slope that fringes the south bank of Wadi al-Hasa, a major drainage system in southern Jordan. It is among 22 supposedly PPN sites found by the Wadi al-Hasa Survey Project directed by Burton MacDonald and was provisionally registered as WHS 1008 in its 1982 final field season (Coinman et al. 1988). Unfortunately, the site is said to have been disturbed by subsequent road work (Rollefson, pers. comm.).

¹⁴C dates and stratigraphy

14Since the site was only surface-surveyed, neither ¹⁴C dates nor stratigraphic information is available. Instead, as mentioned below, the exclusive presence of Helwan points in the projectile assemblage provides grounds for dating the site to the EPPNB (Rollefson 1996: 159, 2008: 74).

Site and structures

15This site is relatively large in scale, and the total area of the main flint scatter or that including an adjacent one (WHS 1007) separated by a gully is estimated at ca. 0.5 ha or ca. 1–1.5 ha, respectively. However, considering the likely effect of the post-depositional disturbance on the steep slope, Rollefson (1996: 159) adds that the real site size was probably smaller than the estimates. If so, it would follow that the site had a settlement size similar to that of HJH-202 (ca. 0.2 ha) and Mushash 163 (ca. 0.25 ha). The surface survey confirmed at least four oval or subrectangular stone alignments with dimensions of ca. 3 × 2.5 m near the upper edge of the site. In addition, several stone alignments or concentrations were also found elsewhere within the site, at least one of which is considered to be a permanent house. In view of the high labor cost required for the house construction on the steep slope, the site is regarded as a hamlet-size, semi-permanent settlement probably with an oval plan (Rollefson 1996: 159).

Small finds

16The lithic collection was examined by Gary Rollefson. The original collection (n = 56) included four Helwan points in addition to three crested blades, ten cores, picks, scrapers, lustered elements, transverse burins, borers/drills, and a Helwan-backed bladelet. Although bidirectional blades occurred commonly, no naviform cores were identified (Rollefson 1998: 104). Incidentally, his subsequent revisit added eight lithic artifacts, which contained four Helwan points (fig. 4.1–4.2), two lustered blades, a borer/drill, and a broken basalt pestle. It follows that all eight projectile points collected at the site belong to the Helwan type, which provided grounds for ascribing the site to the EPPNB. This tentative dating has long been questioned due to the absence of any other clear evidence (e.g., Edwards et al. 2004: 51; Sayej 2005: 44), but the discovery of HJH-202 and Mushash 163 encourages its reevaluation.

Jilat 7

Location and investigation

17The site of Jilat 7 lies in the northeastern Badia, occupying the south bank of the Jilat Gorge known for the existence of an ancient Roman dam (e.g., Politis 1993). This is among the five Dhobaian sites discovered by Alec Kirkbride and Gerald Lankester Harding in the late 1930s and was originally called Wadi Dhobai C. The first excavation undertaken in 1937–1938 dealt with two of the five, Wadi Dhobai B and K, and confirmed an M/LPPNB and epi-Palaeolithic flint assemblage, respectively (Waechter et al. 1938). Wadi Dhobai C, also called Jilat 7, was first test sounded in 1984 and then excavated in 1987 and 1988 under the Azraq Basin Project directed by Andrew Garrard (Garrard et al. 1994).

¹⁴C dates and stratigraphy

18The existence of an EPPNB settlement was suggested by a diagnostic lithic assemblage centered on Helwan points, but the discrepancy with four ¹⁴C dates casts doubt on the identification (e.g., Edwards et al. 2004: 50; Edwards 2016: 60). The excavators suggest that the discrepancy was probably caused by the influx of charcoal remains from the overlying LPPNB occupational layer (Garrard et al. 1994: 75). Whatever the case, it is still certain that the projectile assemblage from the lowest level of areas A and C centers on the Helwan type (Garrard et al. 1994: table 1).

Site and structures

19A flint scatter covers an area of ca. 2,250 m², but the central 700 m² (= 0.07 ha) is estimated to be the real settlement size (Garrard et al. 1994: 75). The excavation in areas A and C roughly in the center of the site revealed a few oval or subrectangular stone-built structures ca. 3–4 m in the longer axis, in addition to two small anthropogenic cuts on the bedrock and 11 bedrock mortars (fig. 4.3). As discussed below, the presence of the subrectangular structures deserves attention. Most of the structures were built using the two-rowed upright slab wall technique common at Neolithic desert settlements in the southern Levant (e.g., Goring-Morris 1993; Fujii 2013: 70–78, 108). As with the cases of HJH-202 and Mushash 163, they appear to have formed a relatively dense settlement, but its general plan is unknown.

Small finds

20The projectile assemblage (n = 45) is dominated by Helwan (66.7%), followed by el-Khiam points (22.2%). In addition, the Salibiya, Jordan Valley, and Jericho/Byblos/Amuq types are also attested as minor components, although some of them might represent intrusive objects from the upper layers. The tool kit also includes Hagdud truncations, pick-like bifaces and glossed sickles, but no detailed information on the core and debitage class artifacts is available. Meanwhile, the ground stone assemblage includes bedrock mortars, handstones, pestles, and shaft straighteners as major components (Garrard et al. 1994: table 2). This site yielded faunal and floral remains, which will be referred to in the next section.

Jabal `Ainab 1

Location and investigation

21Jabal `Ainab 1 is located in southeastern Badia, in the middle of a flint-strewn desert steppe (Hamada in Arabic) that extends near the border with Saudi Arabia (Štefanisko 2016). This site was found in 2013 by Hans Georg K. Gebel during a field trip to Qulban Beni Murra (e.g., Gebel 2016) for film shooting. A limited excavation was conducted in summer 2014 by Barbora Kubikova, Christoph Purschwitz and Denis Štefanisko in the framework of the Eastern Jafr Archaeological Project directed by Hans Georg K. Gebel (Gebel, pers. comm.).

¹⁴C dates and stratigraphy

22Neither ¹⁴C dates nor stratigraphic information is available. As with the case of Abu Hudhud and Jilat 7, the dating of the site depends exclusively on the Helwan type-centric projectile assemblage.

Site and structures

23The site is composed of a total of seven houses or small structural complexes (fig. 4.4). However, it is questionable that they formed a sparse settlement with a total area of several hectares, because the distance between any two adjacent houses/complexes is too large (ca. 50–100 m) to constitute a unified settlement. Thus, it seems more likely that they represent an accumulated picture of the relocations of a (or a few) small campsite(s). In this case, the real settlement size at a certain point in the occupational history would be reduced to less than a tenth of the general appearance.

24The excavation focused on structure A at the northeastern corner of the site, which turned out to be an irregular complex centered on a few rectilinear structures lined with a single row of upright slabs (fig. 4.5). The existence of subrectangular or rectangular structures sheds new light on the architectural landscape of the Badia EPPNB. The excavated complex is estimated at ca. 0.02 ha in total area, which can be taken as the minimum value of the real settlement area.

Small finds

25Small finds from structure A are limited to chipped stone artifacts (n = 5,259) only, and no groundstone tools are included. The naviform core-and-blade technology is commonly used for blank production, but opposed-platform blade/flake cores also occur in small numbers (Štefanisko 2016: 37). The projectile assemblage (n = 31) exhibits the oligopoly situation of Helwan points (100% = 21/21 identifiable samples; fig. 4.6–4.21). However, the remaining ten unidentified samples appear to include at least three Salibiya points (fig. 4.22–4.24) and a Hagdud truncation (fig. 4.25), bearing some resemblance to the assemblage of Jilat 7. Unlike HJH-202, Mushash 163 and Jilat 7, neither sickle blades nor axes/adzes are included.

Jebel Queisa

Location and investigation

26Jebel Queisa is a small site located in the southern Badia, occupying the sloping base of a cliff at the northeastern edge of Jebel Queisa (fig. 4.26). The site was first discovered and test sounded in 1979 and briefly excavated in the following year as a part of the University of Tulsa Prehistoric Cultural Ecology Project, Southern Jordan, directed by Donald Henry (Henry 1995: 345–352).

¹⁴C dates and stratigraphy

27No ¹⁴C dates are available; instead, as in the case of Abu Hudhud, Jilat 7 and Jabal `Ainab 1, the occurrence of a Helwan type-centric projectile assemblage provides grounds for the dating to the EPPNB. In terms of stratigraphy, the supposed EPPNB sub-layer (the lower half of layer C) is sandwiched between the underlying limestone bedrock layer and the two overlying Chalcolithic layers (layers A and B) with an almost sterile sub-layer (upper layer C) in between.

Site and structures

28This small, rockshelter-like site is considered to have been used as a short-term exploitation camp, most likely a hunting station (Henry 1995: 351). The estimated site area is only 50 m² (ca. 0.005 ha), and no structural remains is confirmed at least in the excavation sectors.

Small finds

29The lower half of layer C yielded a chipped flint/chalcedony assemblage (n = 3379) and a shell bead only. The lithic assemblage is marked again by the oligopoly of Helwan points (100% = 3/3 projectiles; fig. 4.28–4.29) and the presence of a Hagdud truncation and disc-shaped opposed-platform blade cores (fig. 4.27). Although the sample size is far from sufficient, the combination of Helwan points and the disc-shaped opposed-platform cores is common to HJH-202, suggesting an early Neolithic date for the site (Henry 1995: 349–352).

Short summary

30Although not always in a unified manner, the EPPNB cultural entities in the northern and southern Levant have so far been defined from the following three major standpoints: 1) ¹⁴C dates falling within the first three quarters of the 9^th millennium cal. BCE, in terms of chronology; 2) the predominance of Helwan points and the increasing use of naviform blades, in terms of lithic techno-typology; and 3) the diffusion of rectilinear structures, in terms of architectural landscape (e.g., Bar-Yosef 1981: 564; Cauvin 1989: 177; Rollefson 1989: 168). Originally, the beginning of village farming must also be added to them (Cauvin 1994: 106), but it requires time-consuming analyses and, for this reason, is usually unconsidered at least during the initial stage of investigation (the study of the Jordanian Badia EPPNB is in this stage).

31The identification of EPPNB sites in the Jordanian Badia depends exclusively on whether the six sites mentioned above fulfil these three requirements. However, the early three sites (i.e., Jebel Queisa, Jilat 7 and Abu Hudhud, in descending order of investigation year) were devoid of clear evidence other than Helwan type-centric flint assemblages. To make matters worse, aside from Jilat 7, the sample sizes of the projectile points at the other two sites were too small to have statistical significance. This explains the reason why the widespread acceptance of the Jordanian Badia EPPNB has been delayed for more than 40 years since the first discovery at Jebel Queisa. However, the recent excavations at the three new sites have drastically changed the situation. Among others, Mushash 163 and HJH-202 completely fulfilled the first two requirements, not only confirming the presence of EPPNB sites in the Jordanian Badia but also encouraging the reevaluation of the previous three sites.

33The existence of the Cypro-EPPNB also supports this judgement. Despite the absence of Helwan points and rectilinear structures, several Aceramic Neolithic sites in Cyprus are defined as EPPNB settlements on the basis of other aspects, including the presence of ¹⁴C dates falling within the third quarter of the 9^th millennium cal. BCE, the introduction of the naviform blade technology, and the innovation of subsistence strategy (e.g., Peltenburg et al. 2001; Peltenburg 2004; Clarke and Wasse 2019). That being so, regardless of population influx from the Euphrates EPPNB, at least the two key sites that completely fulfill the first two of the three requirements are qualified to receive similar treatment.

34Another important question is whether the six EPPNB sites were a part of the southern Levantine Corridor EPPNB or constituted a distinct cultural entity different from it, but the large time lag of the upper limit date between the Badia and Corridor EPPNB sites alone is enough to differentiate between the two. So long as the former began several centuries earlier than the latter, it is impossible that the former was an eastern variant of the latter. In addition, the remarkable difference in natural environment, settlement size, and lithic techno-typology also supports the differentiation. There is no doubt that the third EPPNB cultural entity (i.e., the Jordanian Badia EPPNB) existed in the drylands to the east of the southern Levantine Corridor.

Settlement Pattern of the Jordanian Badia EPPNB

35The above review has shown that the Jordanian Badia EPPNB has a few different types of sites with different locations. Although the total number of sites is still limited, the following discussion explores the settlement pattern of the Badia EPPNB. The sites fall into the following three types.

Hamlet-size settlements

36Unlike the other two EPPNB cultural entities in the Levant, no full-scale settlements have so far been attested in the Jordanian Badia EPPNB; instead, it appears that hamlet-size settlements with a total area of ca. 0.2–0.3 ha were the standard. This type of site is best exemplified by HJH-202 and Mushash 163, both of which are equipped with reaping tools and heavy-duty pounding or grinding implements, suggesting the possibility that they functioned as small farming communities. The same is probably true of Jilat 7 and Abu Hudhud. In fact, although the possibility of contamination from upper layers cannot be ruled out, faunal and floral analyses at the former site suggest that cultivation of cereal and pulse, together with the hunting of gazelle and hare, played an important role in the livelihood there (Garrard et al. 1994: tables 7–13, 1996: tables 11.1–11.2).

37Understandably, these hamlets are concentrated on the western edge of the Jordanian Badia with an average annual rainfall of ca. 50–100 mm (e.g., Jordan National Geographic Center 1984: 112), thereby sharing borders with Final PPNA (FPPNA) settlements such as FW16 (Finlayson and Mithen 2007; Finlayson et al. 2014; Smith et al. 2019) and Zahrat adh-Dhra’ 2 (Edwards et al. 2001, 2002, 2004; Sayej 2004) to the contemporary west. The only exception to this is Jilat 7, but this is probably because the favorable water environment in and around the Azraq Basin allowed for the location of such a remote hamlet-size settlement. Although the details of subsistence activities are still unknown, it is evident that these four hamlets formed the core of the Jordanian Badia EPPNB.

Large or fixed campsite

38Jabal ’Ainab 1 falls into a large campsite if the sparse complex cluster was formed through the seasonal aggregation of multiple migratory groups. However, as suggested above, when the cluster was formed through repeated visits by a single or limited number of small groups, the site would be a small, fixed campsite. It is still unknown which interpretation is correct, but it is certain that the site contains several structural complexes and, in this sense, differs in character from a temporary campsite such as Jebel Queisa without any structural remains. The site is located deep in the southeastern Badia, being devoid of farming-related artifacts. Thus, it can probably be regarded as a remote hunting station.

Temporary campsite

39Jebel Queisa clearly falls into this type of site in view of its small site size and the absence of structural remains and farming-related artifacts such as sickle blades and querns. The rockshelter-like site topography also supports this functional identification. This site is located in the southwestern Badia and can be taken as a temporary hunting camp, although the presence of relatively thick cultural deposits (ca. 20–30 cm; Henry 1995: 347) suggests that it was used over a fairly long period.

Short summary

40The above discussion has shown that while most of the hamlets concentrate on the western edge of the Jordanian Badia, the two campsites are dotted in the remote drylands in the southern Badia. This makes sense when we consider the difference in the site setting, especially water-use conditions, between the two areas.

41The question is the relationship between the two distinct site groups, but it is still unknown whether the remote campsites are related to the radial expedition from the hamlets or the orbital migration of autochthonous hunter-gatherers who adopted a new lithic technology (Crassard and Drechsler 2013; Gebel 2020). Nevertheless, it merits due consideration that unlike the southern Levantine Corridor FPPNA settlement of WF16 (Mithen et al. 2016) and the southern Levantine Corridor EPPNB settlement of Motza (Khalaily et al. 2007), no intra-site burial has so far been attested at any Jordanian Badia EPPNB sites. Although the absence of burial is not inseparably linked to high group mobility, this contrast, coupled with the difference in settlement size and architectural landscape, provides a good reason to think that overall, the Jordanian Badia EPPNB sites were less sedentary than the two settlements to the contemporary west. If so, the seasonal expedition hypothesis would be more likely, but further verification is needed to validate it. Whatever the case, it is certain that the three-tiered sites shared a similar lithic assemblage to constitute a unified cultural sphere newly defined as the Jordanian Badia EPPNB.

Periodization of the Jordanian Badia EPPNB

42Some 30 ¹⁴C dates from HJH-202 and Mushash 163 cover several centuries of the early 9^th millennium cal. BCE, suggesting the possibility of periodization of the Jordanian Badia EPPNB (fig. 2). This is even more so when we consider that, as mentioned below, the latter site has also a few ¹⁴C dates falling into the third quarter of the same millennium. The following discussion approaches the issue through the diachronic transition of the projectile (and core) techno­typology at HJH-202.

Diachronic transition of projectile techno-typology at HJH-202

43The excavation in area 1 produced a total of 241 projectile points, which were recovered from the following five successive layers: layers 3c–3a (or the lower, middle and upper floor deposits of structure 1 and their supposedly contemporary deposits around it), and layers 2 and 1 (or the lower and upper fill deposits in and around it). The analysis counted out some 70 samples ambiguous in stratigraphic attribution, treating the remaining 177 cases only; besides, it unified layers 2 and 1 into the composite layer 2/1 to compensate for the deficiency in sample size.

44The analysis of the relative frequency of projectile types indicates that:

1. the Helwan point is almost always predominant throughout the layers, reaching its peak (66.7%) in layer 3a, then drastically reduced to 22.2% in the uppermost composite Layer 2/1;
2. the el-Khiam point is much less common than the Helwan type and, overall, gradually decreases from 23.2% to 11.1% through the layers;
3. the incipient tanged point is also less frequent but, in contrast to the el-Khiam type, gradually increases from 17.9% to 27.8% through the layers;
4. the same is roughly true of the miscellaneous types (fig. 5.1).

45Although different in detail, similar phenomena have also been observed at Mushash 163 (Rokitta-Krumnow 2019: fig. 10; Rokitta-Krumnow in press), suggesting that they were the general trends of the EPPNB projectile assemblages in the Jordanian Badia. This makes sense when we consider that the PPNA, EPPNB and M/LPPNB projectile assemblages in the Levant are marked by the predominance of el-Khiam, Helwan, and large tanged points, respectively (e.g., Bar-Yosef 1981: 564; Cauvin 1987; Gopher 1994; Shea 2013).

46It is needless to say that the diachronic transition of projectile typology has a close relationship with that of the core typology (fig. 5.2). Although the sample size (n = 43) is not always sufficient, the relative frequency analysis of core types suggests that:

1. single-platform blade cores (which produce short, often irregular, unidirectional tool blanks characteristic of el-Khiam points) are relatively common (27.8 and 37.5%, respectively) in the two lowest layers and gradually decrease to 16.7% in the uppermost layer;
2. disc-shaped opposed-platform blade cores (which produce relatively short, occasionally irregular, bidirectional blanks often used for Helwan points) begin with 61.1% in the lowest layer and decline to 35.3% in the uppermost layer;
3. in contrast, naviform blade cores (which produce elongated, regular, bidirectional blanks suitable to incipient tanged points) start with 5.6% and constantly increase through the layers to account for half of all the samples in the uppermost layer.

47The above observations can be summarized as follows: 1) overall, the lithic production at HJH-202 depended heavily on the combination of Helwan points and opposed-platform cores throughout the layers; 2) however, the early stage of the sequence was accompanied by the earlier combination of el-Khiam points and single-platform cores; and 3) meanwhile, its second half witnessed the steady increase of the new combination of incipient tanged points and naviform cores (fig. 5.3).

48It should be added, however, that the three combinations suggested above are not always exclusive to each other, because e.g., unidirectional irregular blades (used primarily for the production of el-Khiam points) are also detachable from opposed-platform cores or even naviform cores as well as single-platform cores. Conversely, it is also not uncommon that e.g., Helwan points are produced using unidirectional blade blanks detached primarily from single-platform cores (Fujii et al. 2019: fig. 7.4). For these reasons, too much emphasis on the supposed combinations should be avoided, but it is still noteworthy that a certain degree of positive correlation is recognized at least between incipient tanged points and naviform cores. In this sense, the diachronic transition of projectile/core techno-typology can be a marker for the periodization of the EPPNB. The following four-phase division of the Jordanian Badia EPPNB is based on this perspective.

Phase 1

Phase 2

50Phase 2 is the zenith of the Jordanian Badia EPPNB, marked by projectile assemblages centered on the combination of Helwan points and opposed platform cores. Although its precise duration is still unknown, this phase is estimated to cover a few centuries after phase 1, namely, the middle of the early 9^th millennium cal. BCE. Layers 3b/3a at HJH-202 and the middle stages at Mushash 163 belong to this phase. Jilat 7 also probably continued to be inhabited during this phase. Though small in sample size, Abu Hudhud yielded a projectile assemblage specializing in the Helwan type and probably falls into this phase. The same is probably true of the two campsites with a similar projectile assemblage. The expansion of such campsites, together with the presence of the four hamlet-size settlements, highlights the fact that the Jordanian Badia EPPNB experienced its peak during this phase.

Phase 3

51The third phase is characterized by the drastic decline of Helwan points and the replacement by incipient tanged points, both of which are inextricably linked to the decline of opposed platform cores and the increasing use of naviform cores, respectively. This phase is probably ascribed to a few centuries in the middle of the 9^th millennium cal. BCE. Layer 2/1 at HJH-202 and the late stage of Mushash 163 fall into this phase. In addition, in view of the continuation into phase 4, Jilat 7 might also include some habitation during this phase, but nothing can be said about Abu Hudhud limited in information. Meanwhile, the two campsites are devoid of incipient tanged points, suggesting that they were abandoned before this phase.

Phase 4

52This phase is the residuary stage of the Jordanian Badia EPPNB, marked by the appearance of large tanged points such as the Byblos/Jericho/Amuq types produced using naviform blades. To date, such a projectile assemblage has been attested at Mushash 163 and Jilat 7 only, the former of which is associated with several ¹⁴C dates falling within the third quarter of the 9^th millennium cal. BCE (Bartl in press). Meanwhile, the projectile assemblage of HJH-202 is lacking in such large points, suggesting the abandonment before this phase. As discussed below, it appears that the contemporary Cisjordanian hilly country witnessed a cultural shift from the surviving southern Levantine Corridor FPPNA to the newly emerged Corridor EPPNB.

PPNA/PPNB Transition in the southern Levant

53This long-standing issue of the Levantine Neolithic archaeology can be traced back to a new perspective advocated during the 1980-1990s that the real Neolithic revolution associated with village farming started with the Middle Euphrates EPPNB (Euphrates EPPNB). Two models were proposed as to the relationship between the northern and southern Levant. The hiatus model claimed that the new lifestyle did not reach the southern Levant until the stage of the MPPNB, and that the PPNA culture or, more precisely, the Sultanian continued until then in the south (e.g., Cauvin 1994; Kuijt 1997; Kuijt and Goring-Morris 2002). Meanwhile, the time-lagged diffusion model argued that the new wave was transmitted into the south several centuries behind, during the final stage of the Euphrates EPPNB (e.g., Gopher 1989a, 1989b, 1994, 1996; Edwards et al. 2004; Edwards and Sayej 2007; Edwards 2016). Due in part to the support in the fields of archaeozoology and archaeobotany (e.g., Ducos 1993; Willcox 1996), this model has been broadly accepted among scholars. However, the previous discussions targeted the two northern and southern Levantine Corridor EPPNB cultures only, and the presence of the Jordanian Badia EPPNB was beyond imagination. The discovery of the two key sites coeval with, or slightly earlier than, the Euphrates EPPNB requires a fundamental paradigm shift in discussing. In this respect, a series of new concepts, including the Interaction Sphere (Bar-Yosef and Belfer-Cohen 1989) and the Supra-regional Concepts or Polycentric Evolution (Rollefson 2001; Gebel 2004; Rollefson and Gebel 2004; Asouti 2006: 106–111) can be said to have foreseen future transformation of the argument.

54The following discussion approaches the PPNA/PPNB transition issue in the entire southern Levant from the fresh viewpoint of the settlement pattern and four-phase periodization of the Badia EPPNB (fig. 6). Understandably, the periodization in the discussion follows that of the Badia EPPNB. What new perspectives come into sight when we reconsider the issue from the Badia side?

Phase 1: Beginning of the PPNA/PPNB transition in the southern Levant

55The starting point of our discussion is the current perspective that the Badia EPPNB emerged at the very beginning of the 9^th millennium cal. BCE and coexisted for several centuries with the southern Levantine Corridor Final PPNA (Corridor FPPNA) to the west and the Middle Euphrates and Damascus EPPNB (Euphrates/Damascus EPPNB) to the north, followed by the newly emerged Corridor EPPNB (fig. 6.1; Fujii 2022b). If so, the PPNA/PPNB transition in the entire southern Levant, including its arid margin, must be discussed separately divided into two stages: the origin of the Badia EPPNB and the formation process of the subsequent Corridor EPPNB.

56As for the first point, our two recent papers proposed two alternative perspectives. One was a revised version of the traditional diffusion model, arguing that the very rapid southward expansion of the Euphrates/Damascus EPPNB along the eastern edge of the Levantine Corridor resulted in the concurrent occurrence of the Badia EPPNB (Fujii et al. 2019). The other suggested the possibility of a polycentric origin of the Levantine EPPNB culture (Fujii 2022b). Although ambivalent at that time, we are now leaning toward the latter view, because increasing research data not only demonstrates the presence of the Badia EPPNB but also highlights its cultural originality.

57First, there is a fact that the ¹⁴C dates from HJH-202 and Mushash 163 are not only several centuries earlier than those from the southern Levantine EPPNB sites but also contemporary with, or even slightly earlier than, those from the Euphrates/Damascus EPPNB sites. In fact, the comparison of ¹⁴C dates between HJH-202 and e.g., the PPNA/PPNB transitional phase of Jurf el-Ahmar argue for the local origin of the Badia EPPNB. If so, the diffusion model, to say nothing of the hiatus model, would make no sense, although the occurrence of obsidian artifacts at HJH-202 (Cambel et al. 2017) and Nahal Lavan 109 (Burian and Freedman 1988) is suggestive of some contact with the northern Levant and beyond during the EPPNB period (Ibáñez et al. 2016).

58Also suggestive for the local origin of the Badia EPPNB is the remarkable difference in projectile techno-typology. As has often been noted (e.g., Gopher 1994: 190; Adachi 2000: 79–84; Tsuneki et al. 2006: 54–55; Aurenche and Kozlowski 2011: 451–452; Fujii et al. 2019: 192–193; Arimura 2020: 109–112; Fujii 2022b), while Helwan points of the Euphrates/Damascus EPPNB are produced, in most cases, using slender blades with two parallel lateral edges, those of the Badia EPPNB are often produced with squatty blades with two converging lateral edges (fig. 7.2). In addition, while the former are associated with a pair of less developed wings and (occasionally two pairs of) lateral notches often arranged at a relatively higher (i.e., nearer to the tip end) position, those of the Badia EPPNB are in contrast to them, marked by well-­developed wings and lateral notches usually arranged at a relatively lower (nearer to the base) position. Even considering the possible techno-typological change in the course of southward diffusion (Gopher 1989a, 1989b, 1994; Edwards 2016: 65–67), the difference is too great to accept the diffusion model as it is. It is noteworthy in this regard that even the projectile assemblages of the Damascus EPPNB, a transfer point of the diffusion model, are different in many aspects from those of the Badia EPPNB. The series of techno-typological differences can be traced back to the stage of the el-Khiam point, indicating that they are deep-rooted in the local lithic traditions (fig. 7.1).

59Besides, the originality of the architectural landscape has already been noted. There is a possibility that the subrectangular structures attested at a few Badia EPPNB sites paved the way for the subsequent development of full-fledged rectangular structures in the southern Levantine Corridor EPPNB. It is noteworthy in this regard that while EPPNB rectilinear structures in the northern Levant are centered on a multi-room type (e.g., Coqueugniot 1998; Stordeur 2000), those in the southern Levant are usually of a single-room type.

60Taken together, there is a good reason to consider that the Badia EPPNB originated in a local context. Suggestive is its settlement pattern; as noted above, the Phase 1 hamlets are concentrated on the western edge of the Jordanian Badia, sharing borders with FPPNA settlements to the contemporary west. This is best exemplified by the relationship between HJH-202 and WF16, both of which coexisted just ca. 30 km apart with the Fjaje Escarpment in between. A similar relationship might also have existed between Abu Hudhud and el-Hemmeh (Makarewicz et al. 2006; Smith et al. 2016) located in the upper and middle reaches of Wadi el-Hasa, respectively. Understandably, there must have been some contact between any two neighboring settlements. This is even more so in the case of HJH-202 and the LPPNA campsite of HJH-205 located immediately beside it (Kasahara 2018; Fujii et al. in press a). In fact, both sites share water-catchment technology using rock-cut cisterns as well as a variety of lithic artifacts, including tranchet axes/adzes, Beit Ta’amir knives, cuphole mortars and pestles, and obsidian flakes. Seen in this light, there is a good chance that the Badia EPPNB at HJH-202 derived from the Badia LPPNA at HJH-205 in a local context and the Corridor LPPNA at WF16 in a broader context. Though limited in number, the occurrence of Salibiya points at Mushash 163 (fig. 3.14), Jilat 7, and Jabal `Ainab 1 (fig. 4.22–4.23) could be understood in the same context.

61In this respect, it also deserves further attention that the Jordanian Badia has an unexpectedly long cultural sequence. Our previous investigations in the Harrat Juhayra area have confirmed a variety of late prehistoric sites ranging from the Late Natufian settlement of Wadi Qusayr 173 (Fujii 2005a: 42–44) also called Wadi Juhayra (Neeley 2004; Neeley and Peterson 2007: 207–208), through HJH-205 (LPPNA) and HJH-202 (EPPNB), to the M/LPPNB to LN rockshelter settlement of Jabal Juhayra (Fujii 2015, 2017; Fujii et al. 2018, 2020) and the LN open sanctuary of Harrat Juhayra 0 (Fujii 2005b). Furthermore, the Middle Chalcolithic settlement/cemetery/sanctuary complex of Harrat Juhayra 2 (Fujii 2022a; Fujii et al. in press b) and the EBA tabular scraper cache site of Wadi Qusayr 173 (Fujii 2011) have also been investigated. Such a long and uninterrupted cultural sequence, along with the possible existence of wetland remnants around Palaeolake Hasa (Clark et al. 1987), also supports the local origin of the Badia EPPNB. The same is true of the Wadi al-Hasa and Azraq drainage basins, where a long cultural sequence from the Palaeolithic to the Late Neolithic and beyond has been attested (Garrard and Price 1975; MacDonald 1988). The Jordanian Badia was by no means a cultural wasteland for prehistoric populations.

Phase 2: Expansion of the Jordanian Badia EPPNB

62It appears that the coexistence of the three PPN cultural entities (i.e., the Euphrates/Damascus EPPNB, Badia EPPNB, and Corridor FPPNA) continued for a few centuries in the early 9^th millennium cal. BCE (fig. 6.2). The series of ¹⁴C dates from HJH-202 and Mushash 163 are concentrated on this time range, suggesting that the settlement life at the western edge of the Jordanian Badia reached their zenith during this phase.

64The penetration into the remote drylands is highly significant in the sense that it might have accelerated the exploitation of flat, cortical flint nodules ubiquitous in the Jordanian Badia (Bender 1968; Wilke et al. 2007). In fact, the naviform cores from HJH-202 use this type of flint nodules (fig. 3.4), suggesting that the penetration triggered a technological shift from disc-shaped opposed-platform cores (compatible with roundish river stones) to naviform cores (with particular affinity for flat nodules).

Phase 3: Westward retreat of the Jordanian Badia EPPNB

65Research evidence suggests that this phase, probably dated to a few centuries in the middle of the 9^th millennium cal. BCE, witnessed the westward retreat of the Badia EPPNB, which had greatly expanded during Phase 2 (fig. 6.3). To date, EPPNB sites with a projectile assemblage marked by incipient tanged points, a chronological indicator of phase 3, are limited to HJH-202 and Mushash 163. The other sites were probably abandoned before this phase, although the hamlet of Jilat 7 possibly continued to be inhabited during this intermediate phase in view of the presence of the subsequent Phase 4-like assemblage.

66The decline of the Badia EPPNB was probably triggered by increasing aridification toward the 10.2 ka event (Weninger et al. 2009). Given this perspective, it makes sense that the subsequent re-infiltration by M/LPPNB agro-pastoral outposts such as Wadi Abu Tulayha was accompanied by advanced water-use technology combining a basin-irrigation barrage and an open-cut cistern (Fujii 2007). Incidentally, the Badia EPPNB is thought not to have been associated with domestic animals such as sheep and goats (Garrard et al. 1994: table 7, 1996: fig. 11.3), which probably caused the crucial difference from the adaptation (i.e., pastoral nomadization) to the subsequent 8.2 ka event (Fujii 2013).

67It is likely that the southward diffusion of the Euphrates/Damascus EPPNB gradually started during this phase. The Phase 3 projectile assemblages at HJH-202 and Mushash 163 include a limited number of slender type el-Khiam and Helwan points common in the Damascus EPPNB represented by Tell Qarassa North (fig. 3.22; Ibáñez et al. 2010: fig. 7), ushering in the full-scale penetration of the northern Levantine EPPNB.

Phase 4: Second stage of the PPNA/PPNB transition in the southern Levant

68It appears that the coexistence of the Badia EPPNB and the Corridor FPPNA came to an end in the third quarter of the 9^th millennium cal. BCE (fig. 6.4). To begin with, the Badia EPPNB almost disappeared. Aside from Mushash 163 and Jilat 7, no EPPNB projectile assemblages including large tanged points such as the Byblos/Jericho/Amuq types have so far been attested in the Jordanian Badia. The semi-hollowing out of the Jordanian Badia can be understood as an extension of the westward retreat during phase 3.

69Meanwhile, in the contemporary west, the newly emerged Corridor EPPNB replaced the surviving Corridor FPPNA. The formation process of the Corridor EPPNB, another focal point of discussion, also allows for two different interpretations. As mentioned above, one is the time-lagged dispersal model, arguing that new cultural components characteristic of the Corridor EPPNB, including naviform cores, Helwan points, and full-fledged rectangular structures, were gradually spread from the north (Gopher 1989a, 1989b, 1994, 1996; Edwards et al. 2004; Edwards and Sayej 2007; Edwards 2016). This model has been broadly supported since the 1990s, taking the place of the simple hiatus model (Cauvin 1994).

70However, the (re-)discovery of the Badia EPPNB has proved that their likely prototypes (i.e., opposed-platform cores, short Helwan points, and subrectangular structures) existed in the preceding local EPPNB. Significant, among other issues, is the reconsideration of the origin of the Helwan point, the core of the diffusion model. While many of the Helwan points of the Corridor EPPNB still retain a squatty profile proper to the Badia EPPNB, slender Helwan points characteristic of the Euphrates/Damascus EPPNB are rather exceptional in the south (fig. 7.2). This discrepancy has long been understood as representing techno-typological changes in the course of southward diffusion of the Euphrates/Damascus EPPNB (e.g., Gopher 1989a, 1989b), but it is also highly likely that the westward retreat of the Badia EPPNB was responsible for it. Future discussion needs to consider the easterly as well as northerly influence.

Concluding Remarks

71The first half of the discussion has confirmed that overall, the six early Neolithic sites in the Jordanian Badia fulfill the threefold definition of the EPPNB culture, and that they constituted the Jordanian Badia EPPNB different from the other two EPPNB cultural entities in terms of the upper limit date, lithic techno-typology, and architectural landscape.

72The (re-)discovery of the Jordanian Badia EPPNB has led to a fundamental paradigm shift in discussing the PPNA/PPNB transition issue in the southern Levant (fig. 8). The second half of this paper approached this issue from the fresh viewpoint of the settlement pattern and periodization of the Badia EPPNB. Research evidence suggested that:

1. the Jordanian Badia EPPNB probably originated in a local context, at the beginning of the 9^th millennium cal. BCE, and coexisted for several centuries with the southern Levantine Corridor FPPNA to the west and the Euphrates/Damascus EPPNB to the north;
2. in the meantime, it once expanded into the remote dry lands beyond the Jordanian Badia and soon retreated westward;
3. the westward retreat of the Jordanian Badia EPPNB, in conjunction with the southward advance of the Euphrates/Damascus EPPNB, resulted in the formation of the southern Levantine Corridor EPPNB in the middle of the same millennium. It is our current perspective that the PPNA/PPNB transition in the entire southern Levant took place twice over, each time in a different context.

73However, this is nothing but a hypothetical perspective based on limited research data before the pandemic and, understandably, requires further verification. We would like to resume our excavation at HJH-202 as soon as possible and continue to deepen our understanding of the long-standing issue of the southern Levantine Neolithic archaeology.

74Postscript. Our recent excavations at Masyoon (28.631845N, 35.379558E, 1424 m in altitude) in northern Hijaz, NW Arabia, have yielded a beehive-shaped structural complex, a Helwan type-centric projectile assemblage, evidence of skull worship, and several ¹⁴C dates falling within the first half of the 9^th millennium cal. BCE (Fujii et al. 2023a, 2023b). All of these highlight the existence of another EPPNB settlement at the western edge of the Jordanian Badia in the broad sense of the term. The next excavation season, scheduled in the early summer months of 2024, is expected to provide further insights into the relationship between the Jordanian Badia EPPNB and the Neolithization in the Arabian Peninsula.