1Michal Birkenfeld’s monograph on PPNB settlement patterns in Lower Galilee is the published version of his 2017 doctoral dissertation at Hebrew University. It reviews all the known PPNB and probable PPNB sites in the region but with particular focus on Kfar Hahoresh. Birkenfelds’s aim is to investigate, at the regional scale, the socio-economic organisation and structure of the PPNB settlement system in the lower Galilee—from the Zevulon/Acco plains to the Jordan Valley and the Bet Ha-Qerem Valley southwards to the Jezreel/Bet Shean corridor—and to determine whether it changed over the course of that period. However, he also devotes considerable attention to intra-site analysis of Kfar Hahoresh’s lithic assemblage and the spatial distributions of artefacts, a portion of the monograph that is interesting but seems somewhat disjointed from the rest.

2The monograph begins with a very brief introduction to the period and the region before turning to methods for landscape-scale analysis, which include site catchment analysis, site territorial analysis, and Geographic Information Systems. A brief outline of the GIS methods occurs in the appendix 1. The GIS’s main purpose was to estimate the extent of exploitation territories around each site in terms of walking distance, but Birkenfeld also used it to find viewsheds from the perspective of each site, although it is not clear what purpose he meant the viewsheds to fulfil. For the artefact analysis at Kfar Hahoresh, Birkenfeld analyses a sample of the lithic assemblage in terms of reduction sequences and uses point densities and “Hot-Spot” analysis to identify statistically significant spatial clusters of artefacts in the excavated portion of the site.

3Chapter 3 briefly summarises the Lower Galilee’s geographical circumstances and reviews the evidence for PPNB sites in the region, with sites assigned to four categories defined by their respective degrees of archaeological investigation and data availability. The category I includes the excavated and relatively thoroughly investigated sites, Kfar HaHoresh, Yiftahel, and Munhata. The category II includes five sites that have received somewhat less investigation, and categories III and IV still less, with category IV representing sites known only from survey. Birkenfeld provides a brief summary of each site while table 3.1 shows which Neolithic and Early Chalcolithic (Wadi Rabah) periods are represented at each, and table 3.2 provides a fairly terse summary of each site’s environmental characteristics (elevation, slope, aspect, geology, soil type and vegetation). This chapter does not provide any spatial analysis of these sites, which we await until chapter 6.

4Chapter 4 narrows the focus to Kfar HaHoresh, its stratigraphy, architecture, burials, lithic assemblage, and spatial analysis of the lithics. The very large sample of lithics was purposively selected from all available lithics on the basis of secure context, and received basic analysis mainly in terms of tool frequencies by stratigraphic phase (table 4.5). The most interesting aspect of this chapter is the spatial analysis, which consists mainly of “Hot Spot” maps of the distributions of individual lithic classes, such as cores, flakes and blades, and core-trimming elements, and simple artefact densities for the rarer sickle elements, burins, retouched blades, and projectile points. This allows Birkenfeld to identify some patterns, including some inter-phase continuities in the locations of artefact clusters in middens and open areas and a tendency for cores to be spatially separated from the bulk of refuse, perhaps indicating caching of still-useful raw material. Disposal of finished tools shows much the same pattern, other than a tendency for bifacial tools to occur in burials and apparently intentional deposit of caches of blade blanks, finished tools, or even knapping products in isolated cases. This analysis could have benefited from the use of methods that consider interactions among artefact types, such as hierarchical clustering or principle components analysis, instead of considering these types individually.

5Chapter 5 turns to other sites in Lower Galilee, with some focus on the better evidence from Yiftahel, Munhata, Mishmar Ha-Emeq, Tel ‘Ali, and Ahihud. In all of these sites, we have some reasonable evidence for architecture, most of it presumably residential, and some of it conforming to the “megaron” or “pier house” (Banning and Byrd 1988) model, as well as evidence for burials, bone, chipped and ground-stone tools, faunal and plant remains, and “special finds.” The attempt to identify associations among these sites had only limited success, most notably with some evidence that sites in the Jordan Valley differed economically from sites elsewhere in the region, relying more on cereals and perhaps foddered livestock, and may have had stronger ties to sites to the east and north than to Galilean sites. However, the fact that most of the obsidian in the study, which was sourced in Central Anatolia, came from Ahihud, and to a lesser extent Kfar HaHoresh and Yiftahel, would seem inconsistent with this last inference. Perhaps we should not expect interregional networks to be the same for all kinds of exchanges.

6Finally, in chapter 6, we find the spatial analyses that the introduction describes as the main objective of the monograph. Birkenfeld explores tendencies in the sites’ geographic locations, showing that they tend to be on nearly flat land or gentle, north-facing slopes, typically at the juncture between valley floors and foothills, and within 400 m of a stream. He uses GIS to calculate the half-hour, one-hour, and two-hour territories of each site, but indicates that proximity to the modern Israel/Jordan border truncates these estimates for some of the easternmost sites (p. 177). This struck me as rather odd for two reasons. First, even less detailed topographic information from the Jordanian side of the Jordan Valley would have been better than just ignoring this information. Second, the more serious problem here, especially in the Jordan Valley, is that a Digital Elevation Model (DEM) based on modern topography could seriously misrepresent the territories of these sites as they would have been almost 10 millennia ago. The bias may be negligible for most of the sites, which would likely only have experienced localised incision of valleys in the interim, but much of the Jordan Valley could have been somewhat impassable marsh and swamp in this period.

7Birkenfeld presents data on, among other things, the agricultural potential of these territories based on land classifications in a survey of the 1950s. Lands suitable for cultivation belong to classes I to III in this system, and the vast majority of sites have such lands available in more than 40% of their half-hour, one-hour and two-hour territories. What is missing here is whether this differs statistically from a random distribution of territories. It is possible that this does not indicate any decision-making influenced by the availability of arable land, all the more when we consider that the distribution of land suitable for modern agriculture could be rather different than the distribution of land suitable for Neolithic exploitation after millennia of deforestation, valley incision, and pastoral overexploitation. The chapter also considers water exploitation but estimates of access to springs, as Birkenfeld admits, are probably heavily biased by the loss of many springs in recent times due to falling water tables.

8Chapter 6 ends with a consideration of the sites’ viewsheds. It is not really clear what this was meant to accomplish. Sites vary in the size of their viewsheds, and just about the only take-away is that there was very little intervisibility among sites. In any case, viewsheds are even more vulnerable than least-cost paths and exploitation territories to changes that have occurred since the Neolithic, when there could have been considerably more forest cover than today that would have obscured some views. This section would have been more compelling if guided by some theory as to why Neolithic villagers may have favoured, or avoided, broad panoramas or high visibility. Perhaps the former could have helped them monitor movements of migratory game?

9Chapter 7 ties things together and adds some context to present a number of conclusions. One is that the sites may have varied in function, Kfar HaHoresh most notably diverging from more general patterns of site location and exploitation territory. Although other explanations are possible, this could be consistent with excavators’ hypothesis that Kfar HaHoresh was a regional mortuary centre. Some of the other sites vary in their role in lithic production and distribution, ranging from flint extraction sites at Q1 and Giv’at Rabi East to possible concentration of lithic production at Yiftahel. Sites also vary in size but, as Birkenfeld himself notes (p. 209), we should probably not make too much of this, given the problems inherent in estimating the extent of sites occupied over several PPNB phases, with possible “horizontal stratigraphy,” not to mention extensive subsequent Pottery Neolithic occupation, as at Sha‘ar Hagolan. If anything, I am struck by how small most of the sites appear to be, most apparently no more than 0.5 ha.

10With regards to subsistence economy, Birkenfeld provides what we might better describe as plausible hypotheses than conclusions, since they depend more on previous work than on the analyses presented in previous chapters. He sees the half-hour territory as a place for relatively intensive garden cultivation, mainly of pulses, such as lentils and horse beans, with the next zone, up to an hour away, devoted to hunting of wild boar, cattle and gazelles and, perhaps, herding of morphologically wild goats. One might add that harvest of wild cereals could have taken place in this zone as well. The outer zone, up to two hours away, could have had Mediterranean forests and been used for less frequent hunting of deer.

11Turning to regional interactions, he notes that most of the sites are within the half-hour radius of another site, that there was apparently exchange of flint blanks and finished tools over longer distances, and that the Jordan Valley sites participated mainly in a network with eastern and northern connections, rather than in an Esdraelon-to-Acco network. However, local variants of small “votive” axes are an exception to this pattern, while Munhata, at least, participated in the network for some variants of Jericho points also found at Kfar HaHoresh and Mishmar Ha’emeq. Birkenfeld also finds evidence for networks of ground stone distribution that, as we might expect given their relative weight, operated over somewhat shorter distances. He suggests that grain and pulses were commodities traded for these other goods, but this is arguably unlikely over long distances given the low value/mass ratio of crops and therefore prohibitive transport costs where there was no animal or maritime transport.

12Finally, he turns to change over time, and the region’s relationships to the Southern Levant more generally. Settlement went from relatively ephemeral EPPNB to substantial MPPNB sites and then widespread abandonment early in LPPNB followed by limited resettlement in Final PPNB (or PPNC, e.g., Tel ‘Ali and Horvat ‘Uza). Some of these, notably Sha‘ar Hagolan, continued into the Pottery Neolithic. Several lines of evidence point to the region’s outside connections, including import of purple/pink flint from Transjordan, asphalt and malachite from the Dead Sea region, and obsidian from Central Anatolia, while variation between regions in settlement pattern and the adoption of domesticates is also evident.

13Overall, this is a worthwhile work that brings together information that many students of the Levantine Neolithic will find very useful. Its main weakness is in its failure, I believe, to make convincing arguments for the choices that led to settlement locations and for the distribution of resources within their territories. The former suffers from failure to demonstrate that the patterns are statistically different from what we would find in a random distribution of locations, while the latter suffers from what seem like rather unrealistic territories. Possibly because of the coarseness of the DEM (10 m pixels), these territories appear almost circular, which is not surprising where there is little relief but is unexpected in highly dissected highlands and sites on the margins of the Jordan Valley, where we would expect least-cost paths to follow valleys and ridges, not radiate out over mountains and plains indiscriminately.