During the 1950s the dominant paradigm of Behaviorism, which sought to understand human psychology purely in terms of observable behaviors, was supplanted by Cognitivism, which reasserted the reality and importance of internal mental states and processes in explaining overt behavior. This was the “Cognitive Revolution” that led to the establishment of the cognitive sciences, including cognitive neuroscience (Miller 2003). Perhaps reflecting these reactionary origins, the orthodox view in the cognitive sciences has been that cognition consists of the manipulation of abstract, “amodal”, symbols that are distinct from the “modal” systems of perception (e.g. vision, hearing), action (e.g. motor control) and interoception (e.g. body states, emotions) more directly associated with producing behavior (Barsalou 2008). By the 1990s, a further counter-reaction to this Cognitivist orthodoxy had emerged. Although most commonly referred to as “Embodied Cognition” (Wilson 2002), Lawrence Barsalou has suggested the more inclusive term “Grounded Cognition” to describe this emerging paradigm, which includes grounding in perceptual imagery (Barsalou 1999), motor simulation (Decety and Grèzes 2006), and environmental situations (Thelen and Smith 1994) as well as bodily states (Lakoff and Johnson 1980). Grounded Cognition questions the central importance of abstract symbol manipulation in human cognition and replaces it with an emphasis on “grounded” simulation in the brain’s modal systems of perception, action and interoception. For example, semantic knowledge of word meaning appears to be based at least partially on simulation, with color words activating parts of the brain involved in visual color processing, and action words activating motor cortex associated with the relevant body part (e.g. face, arm, leg) (Pulvermüller and Fadiga 2010). There is now substantial evidence that the actions of others (including speech) are understood through internal simulation using one’s own motor system, and that abstract conceptualization and reasoning are supported by the concrete simulation of objects, situations, and actions (review in Barsalou 2008). This move away from the classical Cognitivist view of mental processes as abstract symbol manipulation has encouraged neuroarchaeologists and others to seriously consider: (1) extending the concept of “mind” beyond the conventional boundaries of skin and skull and (2) re-conceptualizing the relationship between perception, action, and cognition.

The central roles posited for simulation and context in Grounded Cognition are consonant with archaeological insights into the importance of artifacts in shaping human experience. Artifacts are not just something people think about, they are something people think with. Malafouris (2004, 2008) in particular has drawn on Andy Clark’s (e.g. Clark 2008; Clark and Chalmers 1998) concept of an Extended Mind to argue that artifacts help to actively constitute cognitive systems, rather than simply influencing internal cognition. For example, a Mycenaean Linear B tablet may be thought of as enabling a redistribution of memory functions outside the brain and indeed as partially instantiating a new “extended” cognitive system that includes reading as one of its behaviors. The incorporation of bodily ornaments, from Paleolithic shell beads to Mycenaean gold rings, into the body image might similarly be seen as extending and transforming the human sense of self. Central to this approach is a questioning of traditional boundaries between mind, body, and environment in the analysis of cognitive systems.

Though largely inspired by work in social science (e.g. Gell 1998; Hutchins 1995) and philosophy (e.g. Merleau-Ponty 1962) this theory of human “Material Engagement” also draws support from neuroscientific evidence that external objects can indeed come to be represented in the brain as literal extensions of the body. Head and Holmes (1911) are commonly credited with originating the concept of a “body schema” localized in the parietal lobe, including the speculation that this internal body representation might be plastic enough to incorporate hand-held objects. This conjecture has now been amply supported by empirical findings. Working with macaque monkeys, Atsushi Iriki (Iriki et al. 1996; Maravita and Iriki 2004) used single neuron recording to show that bimodal (visual/somatosensory) neurons in the parietal cortex actually changed their response patterns when macaques learn to use a rake. For example, “distal type” neurons with somatosensory and visual receptive fields centered on the hand extended their receptive field to encompass the entire length of the tool. Maravita and Iriki (2004, p. 80) suggest that this plasticity may “constitute the neural substrate of use-dependent assimilation of the tool into the body schema”. Although single neuron recording is an invasive procedure that cannot ethically be applied to human subjects, behavioral and neuropsychological evidence strongly supports the existence of similar neural mechanisms in humans. For example, brain-damaged patients suffering from selective spatial awareness difficulties in near (personal) space make errors on a line bisection task when using a stick to point (i.e. an extension of the body) but perform normally using a laser pointer (Berti and Frassinetti 2000).

Iriki’s group has also used tracer techniques to examine anatomical changes in connectivity associated with tool training in macaques (Hihara et al. 2006). They reported the development of new projections from high order visual areas near the temporo-parietal junction to the intraparietal sulcus (see Fig. 7.1), where bimodal neurons are found, and speculate that this physical change in connectivity enables the body schema flexibility of tool-trained macaques. Tracer techniques require killing the subject, and so cannot be applied in humans, however the non-invasive technique of Diffusion Tensor Imaging (DTI, see Sect. Neuroarchaeology Methods) has now provided extensive evidence of experience-dependent changes in human white matter. Examples range from learning to juggle (Scholz et al. 2009) to learning a second language (Schlegel and Rudelson 2012). Voxel Based Morphometry (VBM, Sect. Neuroarchaeology Methods) has similarly documented experience-dependant changes in grey matter density, and there is now widespread acceptance that even adult brains are highly plastic in response to environment and experience.

For Malafouris (2010) this plasticity is yet another reason to consider the brain itself as a cultural artifact, and to focus on the broader organism-environment system as the proper unit of analysis. Archaeology has much to offer neuroscience in this regard, both in taking material culture seriously and in appreciating the full scope of human cultural and technological variation through time and space. Renfrew (2008) has argued that even in the absence of significant biological evolution, the past 60,000 years have witnessed a fundamental, culturally-driven reorganization of the human brain and cognition. Within neuroscience (though perhaps not in the mainstream), Iriki (Iriki and Sakura 2008) has made a somewhat similar proposal that the origins of human tool use, with its intimate relation to body schema representation and sense of self, set up a unique feedback interaction between organism and environment leading to the emergence of consciousness, Theory of Mind, and eventually “scientific and technical civilization”. These ideas about the evolution of organism-environment systems may be further related to the concept of Niche Construction in evolutionary biology (Odling-Smee et al. 2003), which recognizes that organisms alter environments in a multitude of ways that influence the selective pressures acting on their own offspring. For example, termites not only build mounds but are exquisitely adapted to live in them. Over evolutionary and historical time, humans have constructed a complex diversity of cognitive, cultural and technological niches. Neuroarchaeologists would argue that it is no more possible to understand human cognition without reference to this fact than it is to explain the adaptations of beavers without reference to dams.

In line with the name we have chosen for our species, Homo sapiens, accounts of human uniqueness and cognitive evolution often focus on “higher order” capacities for abstract thought, problem solving and symbol use. Grounded Cognition suggests that such higher order capacities are firmly based in the brain’s modal systems for perception and action. This in turn suggests that much of the story of human cognitive evolution may actually concern changing capacities for concrete sensation and action in the world, as opposed to abstract internal representation. This perspective has been most fully developed in the work of Blandine Bril and colleagues on the kinematic organization of stone knapping behaviors.

Bril’s work on stone knapping is empirical and experimental (see below), but in this section we will focus on the broader theoretical framework (e.g. Bril and Roux 2005a), which has its foundations in Ecological Psychology (Gibson 1986; Reed 1996) and the Dynamical Systems approach to cognition and action (Bernstein 1996; Thelen and Smith 1994). Broadly speaking, Ecological Psychology contends that behavior and cognition can only be properly understood as properties of integrated organism-environment systems. For Gibson (1986) this includes the concept of “direct perception”, which rejects the conventional view that perception occurs through the production of internal representations based on limited sensory input (e.g. patterns of light falling on the retina). Gibson instead characterizes perception as an active process of engagement with the environment, so that perceptual information and indeed the very act of perception are not located “in the head” but rather are constituted by the dynamics of the organism-environment system. Gibson’s “anti-representational” stance has remained controversial, but elements of his work have been widely adopted. This includes especially his concept of “affordance”. An affordance is a possibility for action that exists in the relationship between an organism and some aspect of its environment. It is critical to emphasize that an affordance is a relational property—the same vertical wall may afford walking for a spider but not a human. In neuroscience, the concept has been used to understand evidence of motor involvement in object perception by recognizing that perception occurs for action (Fagg and Arbib 1998; Grèzes and Decety 2002) and cannot be easily separated from it. Even the perception of space can be altered by motor planning and activity—for example distances appear greater if there is more resistance to movement (Kirsch et al. 2012). Such findings have led to a questioning of conventional distinctions between sensory and motor systems in the brain, and to the concept of “active perception” (Pulvermüller and Fadiga 2010). As Gibson (1986, p. 223) put it: “We must perceive in order to move, but we must move in order to perceive”.

Bruner (2010) considered the relevance of such sensorimotor integration, and particularly visuospatial integration in frontoparietal neural networks, to classically “cognitive” functions such as working memory and general intelligence. He argues that multisensory integration in upper parietal areas supports the formation of goal-oriented spatial frames for action guiding the allocation of attention and generation of intentions. This interpretation is particularly interesting in light of paleoneurological evidence presented by Bruner that modern human brains are distinguished from other hominin taxa by a differential enlargement of the upper parietal region.

The Dynamical Systems approach is closely related to Ecological Psychology and similarly opposed to mainstream, representational accounts of cognition. Rather than modeling cognition as the manipulation of discrete, amodal symbols, mental processes are seen as arising from continuous variation in coupled perceptual, motor, cognitive and environmental systems that converge on dynamically stable states (attractors) corresponding to adaptive behavior (Barsalou 2008; Thelen and Smith 1994; van Gelder 1998). Like connectionism, this is an “emergentist” perspective which understands cognition as arising from the complex interaction of a large number of simple noncognitive processes (McClelland et al. 2010). As the name implies, the Dynamical Systems approach focuses on patterns of change through time, rather than static cognitive states or “architectures”, and especially on the parameters constraining these patterns. Such influences are often conceptualized geometrically in terms of a “dynamical landscape” with features such as gradients or basins of attraction (van Gelder 1998). It is recognized that human behavior displays a vast number of degrees of freedom (independent dimensions of variation): the joints of the human body, for example, encompass approximately 100 degrees of freedom for movement whereas at the level of individual muscles this number is closer to 1,000 (Bril and Roux 2005a). A central question for the Dynamic Systems approach is thus how this vast complexity is controlled and coordinated. A classic example is provided by Bernstein (1996), who described the arm motions of blacksmiths striking a chisel with a hammer. Surprisingly, Bernstein found that the movement trajectories of individual joints in the arms of these expert craftsmen were highly variable but nevertheless produced a highly consistent trajectory of the hammer head. From this perspective, control emerges from the discovery of effective movement synergies that coordinate action with respect to a goal (or “attractor”) without limiting the degrees of freedom that allow for adaptive flexibility and dynamic stability. Insofar as the discovery of effective synergies changes an organism’s possibilities for action in its environment, it may be considered as a form of experience-dependent affordance perception.

These theoretical considerations, informed by experimental kinematic results, have led Bril and colleagues to suggest that early (i.e. Oldowan) stages of hominin technological evolution were enabled by changing perceptual-motor, rather than cognitive, capacities (Bril and Roux 2005b) and furthermore that the difficulty of discovering the appropriate movement synergies (i.e. the cryptic nature of task affordances) implies that “the recurrence of a learning situation which allows the transmission of the skill, possibly by providing the opportunities for first-hand experience, is likely to have been a prerequisite to the emergence of controlled flaking found in some Early Stone Age sites”. (Nonaka et al. 2010, p. 165). Influenced in part by the work of Bril (e.g. Roux et al. 1995), Stout reached similar conclusions on the basis of functional brain imaging (Stout and Chaminade 2007) and ethnographic (Stout 2002, 2005) evidence. Supporting neuroscientific evidence of human perceptual-motor specializations relevant to tool-making comes from histological studies of occipital visual cortex in humans versus other primates (Preuss et al. 1999) and fMRI experiments showing human parietal cortex responses to 3D form (Vanduffel et al. 2002) and hand-held tools (Peeters et al. 2009, 2013) that are absent in macaques. Similarly, the emphasis placed by Bril and others on the social and environmental facilitation or “scaffolding” of affordance perception/skill acquisition is supported a wide array of research on apprenticeship learning and related topics across disciplines ranging from developmental and comparative psychology to anthropology and education (Fragaszy 2011; Ingold 2001; Lave and Wenger 1991; Rogoff 1990; Tomasello 1999; Vygotsky 1978; Zukow-Goldring and Arbib 2007). The key to understanding the perception-action perspective and its application in neuroarchaeology is to recognize that it does not deny the relevance of “mental”, “cognitive”, or even “representational” levels of analysis (cf. Bril and Roux 2005a) but rather rejects the conventional mind-body dualism which would see cognitive “competence” as separable from motor “performance” (Thelen and Smith 1994). As Nonaka et al. (2010, p. 166) conclude, increasing control in early stone tool-making “is indicative of the emergence of a system, of which learning situation, behavior of an organism, and other biological processes (e.g., genetic activity, neural activity) are components, that stabilizes the transmission of the detection of the constraints and opportunities for action in the environment across generations”.

Interest in the Grounded Cognition perspectives discussed above should not be seen as precluding research focused more directly on the “higher order” capacities for planning, problem solving, and conceptualization that have traditionally been seen as defining uniquely human cognition. The goal of emergentist perspectives is not impose a purely “bottom-up” approach but rather to seek accounts than span levels (McClelland et al. 2010). Functions such as planning that may provide “top-down” control of cognitive processes are generally considered under the headings of “cognitive control” (e.g. Stout 2010) or “executive function” (e.g. Coolidge and Wynn 2001) and are classically associated with prefrontal cortex. Such functions have been of great interest to researchers across cognitive psychology (the study of mental processes, especially using psychometric experiments and formal modeling), cognitive neuropsychology (the study of brain function, especially through work with brain-damaged patients), and cognitive neuroscience (discussed above). This diverse history is reflected in an equally diverse range of partially overlapping terminology and theoretical frameworks that can be quite confusing.

In cognitive archaeology, Frederick Coolidge and Thomas Wynn (e.g. Coolidge and Wynn 2005, 2009; Wynn and Coolidge 2004) have applied the multi-component Working Memory model of Baddeley (e.g. 2003) which derives from cognitive psychology. As discussed by Coolidge et al. (this volume) this model includes a “central executive” responsible for “attention, active inhibition, decision making, planning, sequencing, temporal tagging, and the manipulation, updating, maintenance, and integration of multimodal information”. Working memory capacity, including especially amodal attention regulation, has been linked to “fluid intelligence” (i.e. novel problem solving ability) (e.g. Engle et al. 1999). Coolidge and Wynn (2005) hypothesized that a recent (60,000–130,000 years ago) genetic mutation affecting the central executive of working memory might be responsible for the emergence of fully modern behavior and cognition in the archaeological record (but see Powell et al. (2009) and d’Errico and Stringer (2011) for an alternative account of the emergence of modern behavior). Mithen (1996) made a somewhat similar proposal based on concepts of functional modularity versus “cognitive fluidity” derived from developmental psychology (e.g. Karmiloff-Smith 1992). Mithen proposed that evolution of a novel capacity for fluid integration across multiple “specialized intelligences” explained the emergence of modern behavior. Stout (2010, 2011); Stout et al. (2008) focused on earlier (Lower Paleolithic) time periods using cognitive neuroscience models of hierarchical information processing in prefrontal cortex (e.g. Badre and D’Esposito 2009; Koechlin and Jubault 2006) to suggest a gradual evolution of executive function as well as possible links with the emergence of language (Stout and Chaminade 2012).

Insofar as there is a core concept running across these various theories and applications, it is that executive functions have to do with the selection and organization of adaptive behavior. This is thought to require a variety of constituent capacities loosely falling into categories such as the integration of information across time and input modes or domains, the allocation of attention, and the parsing/production of hierarchical structure (e.g. nesting of behavioral sub-goals). The frontal lobes appear critical to these capacities, although it is increasingly recognized that control processes are distributed across functional circuits integrating frontal and parietal brain regions (Dosenbach et al. 2008). Questions about how to sub-divide or “fractionate” executive function into psychologically real components, and whether these components are functionally localized to different portions of frontal cortex remain controversial and at the forefront of research.

A prevailing view in cognitive neuroscience is that frontal cortex function is organized along a posterior to anterior gradient of increasing abstraction (see Fig. 7.1). As reviewed by Badre and D’Esposito (2009), cognitive “abstraction” has been defined in various ways, including:

Brain imaging experiments designed using these various definitions have consistently documented increasingly anterior frontal involvement as abstraction increases, suggesting that this is indeed a fundamental organizing principle of frontal cortex. However defined, capacities for abstract cognition supported by anterior frontal cortex are clearly important for the planning and execution of complex adaptive behavior, which is commonly impaired in patients with frontal damage (Stuss and Alexander 2007). Research in comparative neuroanatomy has identified evolutionary changes in the organization and interconnectivity of human anterior frontal cortex that may support uniquely human capacities for abstract cognitive control (Teffer and Semendeferi 2012; Reyes and Sherwood, this volume).

The inferior frontal gyrus (IFG) in particular has been linked with human capacities for language, tool use, and social learning, making it of special interest to neuroarchaeologists. IFG includes Broca’s area and has long been associated with speech production and syntactic processing (Broca 2006 [1861]; Hagoort 2005). More recently, IFG has been shown to participate in a range of non-linguistic behaviors from object manipulation to sequence prediction, visual search, arithmetic and music (Fadiga et al. 2009; Fink et al. 2006; Schubotz and von Cramon 2003). It has been proposed that this superficial behavioral diversity stems from an underlying computational role of IFG in the processing of hierarchically structured information (i.e. policy abstraction) (Koechlin and Jubault 2006), leading to speculation that this function may have evolved first in the context of manual praxis before being co-opted to support other behaviors such as language (Pulvermüller and Fadiga 2010). Neuroarchaeological studies of stone tool-making (see below) have reported IFG activation consistent with this idea.

Michael Tomasello (Tennie et al. 2009; Tomasello 1999; Tomasello et al. 2005) is a leading advocate of the influential hypothesis that human cognitive and behavioral uniqueness is largely the product of cumulative cultural evolution rather than (neuro)biological evolution. On this view, an underlying change in human social cognition enabled the high-fidelity cultural learning which subsequently did “the actual work in creating many, if not all, of the most distinctive and important cognitive products and processes of the species Homo sapiens”. (Tomasello 1999, p. 11). Importantly, this cultural mechanism of cognitive change implies a potentially faster (“historical”) rate of change compared to biological evolution. It is thus appealing to conjecture that a single change in social cognition might explain the apparent increase in rates of cultural change that some archaeologists see as marking the initial emergence of “modern human behavior” in the African Middle Stone Age. As mentioned in the introduction, for example, V.S. Ramachandran has suggested that developments in the mirror neuron system of action understanding may have enabled enhanced imitation learning leading to a cultural “big bang” in human evolution.

So-called “mirror neurons” were first described in the in the inferior frontal and parietal cortex of macaque monkeys (Rizzolatti and Craighero 2004). These are neurons that respond both to observed actions and the self-performance of a similar action. Neurons with similar properties are thought to exist in humans (Kilner et al. 2009) and to provide a mechanism of action understanding through internal simulation or “motor resonance”. It has further been proposed that this basic mechanism of action understanding is the foundation for more sophisticated forms of social cognition, including the understanding of intentions (Wolpert, et al. 2003) and Theory of Mind (Gallese et al. 2009). The location of mirror neurons in human Broca’s area and its macaque homolog has also suggested potential links to language evolution (Rizzolatti and Arbib 1998). Michael Arbib’s Mirror System Hypothesis proposes that a primitive anthropoid action understanding system underwent successive evolutionary modifications to support imitation, pantomime, manual “protosign” and ultimately vocal language, thus providing a neural underpinning for earlier “gestural hypotheses” (Hewes 1973) of language origins. It is important to note that macaque monkeys have mirror neurons but do not imitate, so any account of human imitation and cumulative culture invoking the mirror neurons must include evolutionary changes to this system. For example, human mirror neurons respond to pantomimed grasping movements lacking an object, whereas monkey mirror neurons do not (Gallese et al. 2004). These functional differences may be related to the pattern of increasing white matter connectivity of “core” fronto-parietal mirror system regions and related temporal-parietal regions seen across (non-imitating) macaques, (infrequently imitating) chimpanzees, and humans (Hecht et al. 2012).

Although some have linked mirror neurons and motor resonance to the ontogentic development of Theory of Mind, others are less convinced that simulation-based mechanisms can fully account for human understanding of other minds (Saxe 2005). Indeed, the very term “Theory of Mind” implies an actual theory or a set of beliefs about how minds work (Gopnik and Wellman 1992). The debate between “Simulation Theory” and “Theory Theory” is complex and unresolved, but for current purposes does draw attention to brain regions outside the mirror system that are widely thought to be involved in the representation and attribution of mental states: the temporo-parietal junction (TPJ) and the medial prefrontal cortex (Fig. 7.1). As reviewed by Frith and Frith (2006), TPJ appears to be involved in visual perspective taking while medial prefrontal cortex is specifically involved in “mentalizing”—thinking about the mental states of oneself and of others.