1. Introduction

Knowledge of the contexts in which our ancestors lived is widely considered to be one of the keys to understanding our evolution (see, for example, Elton 2008; Rosenzweig 1999; Kingston 2007). This is because the hominins, like other animals, do not exist in isolation: instead, they interact with their surroundings, including other organisms, in ways that can influence everything from their anatomy and biology to their behaviour and spatial distribution. Approaches to prehistoric and palaeoanthropological contexts range from the formulation of general theories of human evolutionary environments (as summarised in Potts 1998a; 2007; and Winder 2012b) through work at intermediate – regional or national – scales that analyses the relationships between environmental changes and shifting patterns of human activity, and down to detailed studies of the conditions at individual fossil sites.

As palaeoenvironmental methods and datasets have improved, the emphasis placed on complex and variable habitats has increased. Modern site-specific environmental reconstructions are increasingly likely to identify multiple habitat types and complex conditions in the vicinity of hominin localities (see, for example, Vignaud et al. 2002; Gabunia et al. 2000; WoldeGabriel et al. 2009), and the last few decades have seen the emergence of several new broad hypotheses of human evolution that explicitly focus on variable environments. These include the turnover pulse models (Vrba 1992; 2005; Behrensmeyer et al. 1997), the variability selection hypothesis (Potts 1998b) and the new complex topography hypothesis (Winder et al. 2013).

What few of these ideas and reconstructions consider, however, is the spatial structure of hominin contexts. Most point studies describe a 'mosaic environment' in which several habitats co-exist, but do not explore the ways these might have been distributed in space or, indeed, whether discrete patches (say of forest, grassland and shrubland) are more likely than multi-storey complex habitats or heterogeneous mixes of plants of different categories. And among the broader theories of hominin environments, all – with the single exception of the complex topography hypothesis (Winder et al. 2013; 2014) - focus on temporal, as opposed to spatial, variability. This is very different from the case in more recent archaeological periods, where studies of landscape are regularly conducted and highly valued. In early prehistory landscape studies remain rare despite an increasing body of evidence suggesting that some spatially explicit reconstructions of past environments are possible (Bailey and King 2011; Bailey et al. 2011; Reynolds et al. 2011; Blumenschine et al. 2012b; 2012c; Stollhofen and Stanistreet 2012). This may be because the well-known challenges of taphonomy and spatial averaging are still significant psychological barriers preventing scientists from taking an interest in ancient landscapes. There can be little doubt that these challenges are real, but whether they actually invalidate any or all work on hominin landscapes (defined broadly as 'spatially structured environments') remains unclear.

Whatever the ultimate outcome of the debate over palaeolandscape reconstruction, there remains one line of research to try – the analysis of extant landscapes. Understanding the spatial configurations and variability of current climates, soils, ecosystems and terrains might inform us about patterns in past environments in one of two ways. Firstly, detailed studies of specific modern landscapes might allow us to use analogue methods to reconstruct conditions in the past. And second, if we can demonstrate long-lasting links between spatial patterning and the structure or function of the Earth system (e.g. by clarifying patterns at particular scales that either hold steady under different overall conditions or might be expected to shift in predictable ways), we may learn something about landscapes in the deeper past at the same scales. With extant environments, furthermore, it is possible to study factors such as annual mean rainfall that cannot as yet be effectively reconstructed for the past; to gain a better understanding of the relationships between patterns at different scales and in different regions; and to explore more directly the interactions between landscapes and their inhabitants (see, for example, Winder 2012a; Winder 2014). This approach may also alleviate the risk noted by Kingston (2007, 21) that we might 'overextend and misinterpret the tentative, limited, and scale-specific conclusions' of site-specific palaeoenvironmental studies, and allow us to develop better means of integrating detailed locality data and broader temporal records. At the very least, a decent understanding of extant landscape structures should make it clearer which aspects of hominin palaeolandscapes we can realistically hope to reconstruct from palaeoenvironmental data and how confidently.

To this end, this article does not address existing reconstructions of the environments and landscapes at hominin sites, or how these fit with current hypotheses of human evolutionary environments (current summaries of these issues include those in Potts 1998a; 2007; Winder et al. 2013; Winder et al. in prep; Kingston 2007; Magill et al. 2013). Instead, it explores the spatial structures of extant African environments, to see what they might tell us about the prospects for and pitfalls of palaeolandscape reconstruction. It begins with a study of spatial patterning at different scales (continental, regional and local) and in different parts of the continent (south, east, central and west). The south and east of Africa are the sources of most hominin fossils – a pattern that seems not to be purely taphonomic (Holmes et al. 2005). West Africa is included for comparison and because, while it is not fossil-rich, it is the source of the recent discoveries of Australopithecus bahrelghazali (Brunet et al. 1995) and Sahelanthropus tchadensis (Brunet et al. 2002), while central Africa offers a contrast – an area with no hominin fossils at all. The results of these analyses are then used in conjunction with models of the Earth system, and maps of palaeoconditions, to explore the ways understanding the present might help us reconstruct and understand the past.


Internet Archaeology is an open access journal. Except where otherwise noted, content from this work may be used under the terms of the Creative Commons Attribution 3.0 (CC BY) Unported licence, which permits unrestricted use, distribution, and reproduction in any medium, provided that attribution to the author(s), the title of the work, the Internet Archaeology journal and the relevant URL/DOI are given.

University of York legal statements

File last updated: Wed Mar 4 2015