In several previous papers (Clark, 2016, 2017, 2019), I reported on the intrasite spatial analysis of seven Middle Paleolithic (MP) open-air sites in France. I developed two interconnected methods that I used to directly compare the spatial patterning across a range of site sizes and densities. Lithic artifacts were the only material remaining in these sites, and the dominant spatial patterning was determined by the spatial positioning of core reduction. These methods tracked lithic knapping within these sites and the movement of artifacts from where they were knapped. That study ended in several main findings. First, at these MP sites, differences in site structure could largely be attributed to differences in the length and number of occupations. Second, a large amount of knapping occurred with many usable products left unutilized. Third, lithics selected for use could be identified using these methods, and the lithics selected were often unmodified flakes, in addition to retouched pieces.

In this paper, I present results on the expansion of this study to include two Upper Paleolithic (UP) open-air sites analyzed using the same methods. The analysis of the modern human sites displays two major deviations from the patterning exhibited in the Neanderthal sites. First, unmodified flakes no longer make up a notable component of the selected pieces. There is a much stronger preference toward retouched or bifacially worked pieces. Second, differences in spatial patterning can no longer be largely explained by occupation dynamics. Instead, it appears that site structuring is a product of culturally derived structuring processes. This is the case especially for the more recently occupied UP site, Landry.

An organized use of space is often cited as one potential characteristic indicative of ‘behavioral modernity’ (Wadley, 2001; Bar-Yosef, 2002; Mellars, 2005). The assessment of spatial organization is complicated by several factors, however. Most simply, we might ask what is spatial organization and how would it manifest archaeologically, thousands or hundreds-of-thousands of years after the original ‘organization’ took place? We could have a relatively clear idea of how spatial organization is exhibited in the present, but it is much more complicated to assess its presence left only with the spatial position of lithic artifacts, for example. Moreover, most archaeologists are well aware that the degree of spatial organization might depend not only on the cognitive modernity of a given population but also on the duration of occupation. Even during short stays, groups may adhere to organized routines of site use, but the spatial signatures will often be too ephemeral to detect archaeologically. In this section, I will briefly summarize how various scholars define spatial organization in an archaeological context and how these definitions have been used to characterize the spatial patterning within MP and UP sites.

In his book, “In Pursuit of the Past”, Lewis Binford (1983) argued that modern humans exhibit an organized use of space, with different areas reserved for sleeping, cooking, tool manufacture, and so on. He questioned whether earlier hominins, specifically Neanderthals, would follow such a pattern. In addition, Binford linked the spatial structure of activities to mobility strategies, contending that a foraging strategy would produce ‘intensive’ activity areas, with all activities focused in a limited spatial area, usually around a hearth, whereas a collector strategy would produce ‘extensive’ activity areas, with task-specific activity areas dispersed over a wider area. Moreover, Binford argued that Neanderthals would follow a foraging strategy and modern humans, a collector strategy. This was linked to cognition; a collector strategy required greater planning and forethought, whereas a foraging strategy was based on satisfying immediate needs.

Many scholars drew upon Binford's ideas and sought to test his hypothesis. For example, Simek (1987) compared the spatial organization of Neanderthals and modern humans through a k-means analysis of three layers of Flageolet I (UP) and one layer of Abri Vaufrey (MP). He found the spatial structure of Abri Vaufrey to be homogenous whereas the spatial structure at Flageolet to be heterogenous and more dispersed. In Simek's view, the homogenous distribution was a result of repeated, intensive use of space, indicative of a foraging strategy, and the heterogenous distribution, with some separation of activities, was linked to a collector strategy. For Simek, therefore, it came down to mobility strategies and site function, rather than cognition or behavioral modernity, although those markers were implicit. Hietala and Marks (1981) also sought to link spatial organization with mobility strategy. They compared level 1 to level 4 at Boker Tachtit and found that in level 1, discrete activity areas could be observed, whereas the structure of level 4 was more ‘generalized’. Level 1 was the result of a longer duration of occupation, and level 4 was occupied during a high-mobility strategy. More recently, Riel-Salvatore et al. (2013) also sought to link spatial patterning to mobility strategies, this time within MP layers only, at Riparo Bombrini. Because they did detect some differences in spatial patterning linked to duration of occupation, Riel-Salvatore et al. argued that Neanderthals were capable of organizing their space.

This contrast between an ‘intensive’ vs. an ‘extensive’ use of space and its link to planning and therefore modernity is the fodder that sparked the pursuit of ‘activity areas’ in spatial analysis studies (Clark, 2017). The identification of activity areas would imply an ‘extensive’ use of space, thus a collector strategy of mobility, and consequentially, an ability to plan and function as a fully modern human. However, ethnographic studies reveal that an ‘extensive’ use of space is not always exhibited in extant hunter–gatherer camps, and furthermore, such a pattern might not be detectable in the archaeological record of deep time (O'Connell, 1987; Fisher and Strickland, 1989; Simms and Heath, 1990; Fisher et al., 1991). Binford himself acknowledged that extant hunter–gatherers do not always follow a ‘collector’ strategy but use different strategies depending on the season or environment (Binford, 1980).

Many archaeologists invoke spatial organization and its linkage to behavioral modernity without implying a link to mobility strategies (e.g., Wadley, 2001). In these cases, researchers postulate that our use of space is sometimes practical, but is also often symbolic, for example, in the way space is gendered in many ethnographic contexts (Binford, 1983: 180; Surovell et al., 2022). In my opinion, this argument, and its relationship to the advent of symbolic behavior, is more clearly linked to the emergence of behavioral modernity. The increasing evidence for symbolic behavior is one of the ‘hallmarks’ that has withstood many decades of research and, if anything, only seems to be strengthening over time (evidence for Neanderthal symbolism does not undermine this trend, only reinforces it). Further on in this paper, I argue that the emergence of cultural norms dictating how space is used could be another indication of increasing ‘modernity’ because cultural norms are expected to be more prevalent as culture accumulates. However, the question here is how one is to identify these changes in the use of space.

A major difference in how spatial organization is identified, and used as evidence in this debate, is whether one focuses on artifact patterning or features (or ‘latent’ verses ‘apparent’ structuration; e.g., Anderson et al., 2018). An analysis of features is more straightforward. Hearths, of course, are the most commonly found feature in Paleolithic encampments. Archaeologists assess whether they were constructed in the same place, whether they were ‘built’ by positioning stones or digging into the substrate, and how they are positioned relative to one another and the debris surrounding them (Jaubert and Delagnes, 2007; Stiner et al., 2011; Aldeias et al., 2012; Goldberg et al., 2012; Vallverdú et al., 2012; Spagnolo et al., 2018; Mallol et al., 2019; Murphree and Aldeias, 2022; Clark et al., 2022). These studies are often descriptive in nature and are largely determined by field observations, as well as specialized analyses, such as micromorphology. Sometimes features other than hearths can be identified in Paleolithic sites. These include structures, such as windbreaks, huts, or drying racks, identified via post holes, the positioning of construction materials (i.e., large stones), or phantom ‘walls’ indicated by voids in the distribution of artifacts (Stapert, 1990; Bourguignon et al., 2002; Jaubert and Delagnes, 2007; Kuhn et al., 2009; Wadley and Langejans, 2014; Gingerich, 2022). The frequency of evidence for such features rises sharply throughout the Paleolithic (Clark and Ranlett, 2022; Clark et al., 2022), but in many cases, one must rely on the accuracy of field observations and direct quantitative comparisons are rarely possible.

The positioning of artifacts has the potential to be more objective since the coordinated points of artifacts can be directly compared to one another. However, this requires access to both the coordinated points and the artifact attributes at many sites, which is unfortunately rarely possible, and so most studies analyze the spatial patterning of artifacts at only one or two sites. The analyses used are diverse, and they are often tailored to a particular site, so direct comparison of these studies is seldom possible. Some analyses have produced robust results (Surovell and Waguespack, 2007; Waguespack and Surovell, 2014; Surovell, 2022), but the patterning of artifacts becomes more difficult to interpret with increasing time depth. Frequently, the sole goal is to identify patterning of some sort. It is often difficult to determine, however, what counts as ‘organized’ and what criteria should be used to describe such organization as ‘modern’. These assessments are left to the individual researcher and their interpretation of whether any patterning constitutes as ‘modern’ can differ considerably and has the potential to be affected by underlying predispositions.

When one compares the presence and elaboration of structures over large swaths of time, trends do emerge (Clark et al., 2022). Such trends are harder to detect in the patterning of artifacts, however. This study is meant to fill that void, using the same methodology at nine MP and UP sites so that their patterning can be directly compared. There are certainly other ways to detect spatial patterning than the methods used here; for example, one which focuses on hearth-centered activity areas (see Stapert, 1989, 1990). The point, however, is to be consistent in our application of spatial analyses to a larger sample of sites. Indeed, I was able to detect differences in patterning between the MP and UP sites that would never have emerged had I not such a large database of sites.

There are several limitations to this dataset that must be discussed upfront. One is the obvious low sample size of UP sites (n = 2) compared to MP sites (n = 7). Unfortunately, I was unable to attain access to a larger number of UP sites, and so I decided to proceed with the current sample size, which nevertheless, is the largest direct comparison of MP and UP site structures to date. It is unfortunate too that the two UP sites, Landry and Garris II, have differential access to raw materials, so their use might be somewhat different. All sites in the sample are open-air sites, and they are all located in parts of France where chert is ubiquitous, particularly in riverbeds and drainages. The difference in raw material access is therefore whether a site is directly on top of a raw material source, or merely adjacent to a raw material source, with the exception of Le Prissé where the raw materials were obtained from around 2 km away (Table 1). Therefore, core reduction was still a major activity present at all sites in this sample. Nevertheless, while this analysis has produced some intriguing results, they must be regarded as preliminary and more sites, especially from the UP, must be directly compared before a stronger conclusion can be reached.