Streams as Entanglement of Nature and Culture: European Upper Paleolithic River Systems and Their Role as Features of Spatial Organization

Abstract

Large river valleys have long been seen as important factors to shape the mobility, communication, and exchange of Pleistocene hunter-gatherers. However, rivers have been debated as either natural entities people adapt and react to or as cultural and meaningful entities people experience and interpret in different ways. Here, we attempt to integrate both perspectives. Building on theoretical work from various disciplines, we discuss the relationship between biophysical river properties and sociocultural river semantics and suggest that understanding a river's persona is central to evaluating its role in spatial organization. By reviewing the literature and analyzing European Upper Paleolithic site distribution and raw material transfer patterns in relation to river catchments, we show that the role of prominent rivers varies considerably over time. Both ecological and cultural factors are crucial to explaining these patterns. Whereas the Earlier Upper Paleolithic record displays a general tendency toward conceiving rivers as mobility guidelines, the spatial consolidation process after the colonization of the European mainland is paralleled by a trend of conceptualizing river regimes as frontiers, separating archaeological entities, regional groups, or local networks. The Late Upper Paleolithic Magdalenian, however, is characterized again by a role of rivers as mobility and communication vectors. Tracing changing patterns in the role of certain river regimes through time thus contributes to our growing knowledge of human spatial behavior and helps to improve our understanding of dynamic and mutually informed human-environment interactions in the Paleolithic.

Keywords: Affordances; Ecocultural systems; Focality; Nature-culture entanglement; Pleistocene river systems; Upper Paleolithic spatial organization.

Fig. 1

Coordination scenario discussed by Schelling in his famous book A Strategy of Conflict (1981): y and x mark hypothetical landing positions of two conflicting and heavily armed parties. The framework is defined as an “all or nothing” condition, where contact between the two symmetrical parties will lead to the extermination of both. The objective is therefore to coordinate in space by claiming as much ground as possible while simultaneously avoiding contact and by implication potentially deadly conflict. Schelling shows that the river is the only feature in this scenario which has the ability to “enable” successful coordination, even though it does not divide the available landmass into two identical (or symmetric) parts (redrawn from Schelling : Fig. 7)

Fig. 2

Focality as an emergent property of ecocultural systems rooted in a broad range of factors and their relationships. Natural factors are equally as important as cultural ones, leading to a view which takes into account the spatial quality of a river as nature-culture entanglement. The resulting “personality” is the baseline for a river’s environmental legibility and how people actually read and interpret it. Focality can thus be understood as the disposition of certain river courses to bias human spatial behavior in drawing attention from other landscape features and in anchoring behavior. Vector arrows indicate that the four categories overlap substantially

Fig. 3

Natural river characteristics influence how people experience those fluvial regimes and therefore how to interact with them. The figure exemplifies possible links between key variables defining a river’s fluvial profile and human movement in space: a wide and deep river bed with high flow velocity, b narrow and shallow river bed with low flow velocity, c wide and deep river bed with nonhuman-carrying ice shield, d narrow and shallow river bed with human-carrying ice shield, e river expansion by transgression and periodic flooding events, and f river segmentation by partial shallowization. Black arrows indicate mobility patterns reflecting these profiles

Fig. 4

Chronoclimatic position of the discussed case studies. NGRIP (blue) and NGRIP/HULU (red) curves are used as climate proxy (Weninger and Jöris 2008)

Fig. 5

Schematic map of likely colonization routes to Central and Western Europe during the Early Upper Paleolithic. Large river valleys and the Mediterranean coastline act as primary migration corridors channeling the movement of AMHs. Key sites supporting this view are mapped (for age determinations, see S1, Electronic Supplementary Materials). It should be noted that this scenario opens up the possibility of a bidirectional dispersal to Europe, and to southern Germany in particular: 1 Goat’s Hole (Paviland), 2 Kent’s Cavern, 3 Goyet, 4 Trou Magrite, 5 Grotte Chauvet, 6 La Salpêtrière, 7 Riparo Mochi, 8 Hohle Fels, 9 Geißenklösterle, 10 Grotta di Fumane, 11 Grotta di Cavallo, 12 Franchthi, 13 Keilberg, 14 Willendorf II, 15 Senftenberg, 16 Kozarnika, 17 Temnata, 18 Bacho Kiro, 19 L’Arbreda, 20 El Castillo, 21 Voronezh, 22 Kostenki

Fig. 6

Spatial extension of initial and full Upper Paleolithic entities in relation to important river systems mentioned in the text: a Châtelperronian of southwestern France lying in-between the Ebro formation and the Rhône-Saône river system (Connet ; Pelegrin and Soressi 2007); b Solutrean and Early Epigravettian divided by the course of the Rhône (Mussi ; Banks et al. 2008); c Badegoulian framed by the Ebro, Rhône, Saône, and Loire river systems (Banks et al. 2011); and d Magdalenian and Late Epigravettian separated by the Rhône River (Mussi 2002). Spatial distributions are approximate and claim accuracy only in relation to the critical river courses

Fig. 7

Recorded archaeological occurrences in the Upper Garonne valley from the Pleistocene to the Holocene transition. Upper Paleolithic sites cluster on the edges of the mountainous areas peripheral to the main valley, while the other time slices are also represented in the valley itself. Geomorphology and preservation issues cannot be invoked to explain this pattern because geological strata of the Upper Paleolithic are present but sterile in the valley, and Middle Paleolithic and Epipaleolithic site distribution demonstrates that research history is not the explaining factor. Dot-size indicates the frequency of occurrences. With courtesy of Marc Jarry and Laurent Bruxelles (INRAP, University of Toulouse)

Fig. 8

Distribution of Middle Upper Paleolithic sites yielding Raysse burin technology. The Garonne valley serves as well-defined southern frontier for the presence of Raysse burins in the Gravettian, separating the Raysse facies from the rest of the Gravettian world (modified after Klaric : Fig. 1): 1 Plasenn’al Lomm, 2 Arcy-sur-Cure, Grotte du Trilobite, 3 Arcy-sur-Cure, Grotte du Renne, 4 La Martinière, 5 Les Roches de Pouligny-Saint Pierre, 6 La Picardie, 7 Les Artigaux, 8 Abri Lespaux, 9 Le Fourneau du Diable, 10 Les Jambes, 11 Solvieux, 12 Abri Pataud, 13 La Roque Saint-Christophe, 14 Masnègre ou Masnaigre, 15 La Rochette, 16 Le Flageolet I, 17 Les Morts, 18 Pré-Auberts, 19 Bassaler Nord, 20 Le Raysse, 21 Le Roc de Gavaudun, 22 Les Battuts. Note that the total distribution of the Gravettian exceeds the map by far

Fig. 9

Regional Gravettian groups that center around the female figurine sites of Brassempouy, Laussel, and Balzi Rossi. The Garonne River can be considered the main feature that separates two of these units and organizes their spatial layout (redrawn from Simonet : Fig. 86): 1 Brassempouy, 2 Laussel, 3 Balzi Rossi, 4 Lezia, 5 Le Prissé, 6 Tercis, 7 Pujo-le-Plan, 8 Lespugue, 9 Tarté, 10 Gargas, 11 Gatzarria, 12 Isturitz, 13 Les Battuts, 14 Pech-Merle, 15 Cougnac, 16 Cussac, 17 Pair-non-Pair, 18 Grotte du Vissage, 19 Cro-Magnon, 20 Pataud, 21 Tursac, 22 La Cabre, 23 Le Gratadis, 24 Arene Candide

Fig. 10

Early Upper Paleolithic and Gravettian site distribution on the Iberian Peninsula (modified after Schmidt et al. 2012). The spatiotemporal pattern shows that the Ebro is an important landmark for settlement organization. In the Early Upper Paleolithic (a), the Ebro valley is an important frontier hindering human movement. Only in the later or “evolved” Aurignacian, sites are recorded in the southern part of Iberia. This pattern testifies to the important role of the Mediterranean coastline as a powerful pull factor that channels movement to the south. Note, however, that the status of several sites in southern Iberia as Aurignacian is disputed. In the Gravettian (b), a comparable picture emerges. Sites in the north are still older than in the south, speaking in favor of a scenario in which southern Iberia was successfully settled only during the later phase of the Gravettian. Both maps show only cave sites. AMS dates are used to determine the spatiotemporal pattern (for age determinations, see S2 and S3, Electronic Supplementary Materials). Early Upper Paleolithic sites (a): 1 La Viña, 2 El Cierro, 3 Abrigo de la Güelga, 4 Conde (Forno), 5 Sopeña, 6 Esquilleu, 7 El Pendo, 8 El Ruso, 9 El Mazo de Camargo, 10 Cueva Morín, 11 El Rascaño, 12 Otero, 13 Cobrante, 14 El Polvorín, 15 Venta Laperra, 16 El Castillo, 17 Covalejos, 18 Santimamiñe, 19 Antoliñako Koba, 20 Labeko Koba, 21 Lezetxiki, 22 Coscobilo, 23 Cova Gran, 24 Reclau Viver, 25 L’Arbreda, 26 Mollet I, 27 Abric Romani, 28 Les Mallaetes, 29 Foradada, 30 Beneito, 31 Perneras, 32 Pirulejo, 33 Boquete de Zafarraya, 34 El Bajondillo, 35 Gorham’s Cave, 36 Pego do Diablo, 37 Gruta de Salemas. Middle Upper Paleolithic sites (b): 1 La Viña, 2 Sopeña, 3 Cueva de la Riera, 4 Cueto de la Mina, 5 Llonín, 6 Hornos de la Peña, 7 Altamira, 8 El Castillo, 9 El Pendo, 10 La Garma, 11 Cueva Morín, 12 El Rascaño, 13 El Mirón, 14 Abrigo del Cuco, 15 Bolinkoba, 16 Labeko Koba, 17 Lezetxiki, 18 Santimamiñe, 19 Antoliñako Koba, 20 Ermittia, 21 Aldatxarren, 22 Ekain, 23 Amalda, 24 Aitzbitarte III, 25 Torre, 26 Alkerdi, 27 Zatoya, 28 Roc de Melca, 29 Reclau Viver, 30 Davant Pau, 31 Mollet III, 32 L’Arbreda, 33 Balma de la Griera, 34 Beneito, 35 Parpallo, 36 Les Mallaetes, 37 Barranc Blanc, 38 Les Cendres, 39 Ratlla del Bubo, 40 Las Palomas, 41 Finca Doña Martina, 42 El Palomar, 43 Los Morceguillos, 44 Serrón-La Palica, 45 Nerja, 46 Boquette de Zafarraya, 47 El Bajondillo, 48 Gorham’s Cave, 49 Higueral de Motillas, 50 Higueral de Sierra Valleja, 51 Vale Boi, 52 Gruta de Salemas, 53 Lapa do Anecrial, 54 Casa da Moura, 55 Gruta do Caldeirão, 56 Lagar Velho, 57 Buraca Escura, 58 Buraca Grande, 59 Vale dos Covões (Abrigo 1), 60 Vale das Buracas

Fig. 11

Site distribution from the Early Upper Paleolithic to the Magdalenian and Late Paleolithic in the Ebro catchment (spatial data from Schmidt et al. and Utrilla et al. 2012). The distribution pattern shows that a clear shift in settlement organization occurs during the Magdalenian period. Whereas in the Aurignacian (a) and Gravettian (b), sites are rarely located within the Ebro valley itself, the contrary can be observed in the Magdalenian and Later Paleolithic (c) where sites are even located on the western banks of the river system (for age determinations, see S4, Electronic Supplementary Materials). The shift in settlement preference indicates an important transformation of the Ebro’s role as a feature of spatial organization at the end of the Pleistocene: 1 Berniollo, 2 Urratxa, 3 Atrillon, 4 Anton Koba, 5 Kukuma, 6 Portugain, 7 Atxoste, 8 Legintxiki, 9 Leginpea, 10 Alaiz, 11 Abauntz, 12 Burutxukua, 13 Zatoya, 14 Peña 14, 15 Legunova, 16 Chaves, 17 Forcas I, 18 Forcas I, 19 Alonsé, 20 Cova Gran, 21 Parco, 22 Parco, 23 Guilanyá, 24 Montlleó, 25 Margineda, 26 Molí del Salt, 27 Boix, 28 Colls, 29 Hort de la Boquera, 30 Gato 2, 31 Bolichera, 32 Peña del Diablo 1/2, 33 Vergara, 34 Alejandre

Fig. 12

Site distribution of the Central European Magdalenian (spatial data with courtesy of Andreas Maier; see Maier for details). Sites are highly correlated with first-order rivers indicating an important role of fluvial systems in organizing Magdalenian sociocultural space (for site names and coordinates, see S5, Electronic Supplementary Materials)

Fig. 13

Frequency distribution illustrating the distance between Central European Magdalenian sites and first-order rivers (x-axis: distance to the next first-order river in kilometers, y-axis: number of sites per distance class). The data has been analyzed with the NEAR tool in ArcGIS 10.1 (spatial data with courtesy of Andreas Maier; see Maier for details)

Fig. 14

Relationship between the course of the Rhône-Saône formation, the Rhine River, and the Danube fluvial system and a lithic/chert (sample) and b mollusk/amber (total) displacement vectors in the Central European Magdalenian (arrows point/triangles mark the likely raw material source area). The connection between southwestern Germany and the Main area established by Dreiech-Götzenhain mentioned in the text is highlighted (modified and schematized after Floss and Maier ; compare Table 1 for details)