Attention schema theory,an interdisciplinary turn? Cognition,culture and institutions

Reductionism versus emergentism and the question of distinct levels of explanation

The epistemic cultures of the social and behavioral sciences, including basic methodological and ontological assumptions such as what counts as empirical evidence and indeed even knowledge, vary markedly across disciplines. Some behavioral scientists advocate methodological naturalism, meaning that humans are ultimately viewed as natural objects that can be studied and explained by means of natural science methods and causal models. In this view, consciousness cannot be viewed as ontologically separate from the level of interacting neurons, suggesting that the former is largely an epiphenomenon. Social scientists reject the reductionist idea that the enormous complexity of the social world can ultimately be understood as matters of genomic and neural dispositions, in favor of holistic approaches that recognize distinct levels of explanation (Bloch, 2012; Lloyd, 2011; Reyna, 2006). The social sciences are thus founded on an emergentist ontology according to which both consciousness and the institutions underpinned by shared intentionality are emergent properties that cannot be reduced to the sum of their constituent parts.

The ontological clash between reductionism and emergentism/holism is manifested in several more specific debates about the puzzling relations between a supposedly immaterial mind and the material brain, such as those concerning the mind-body problem and domain generality/specificity. The latter discussion concerns whether human minds are best understood as a collection of special-purpose processors that evolved to solve specific problems (massive modularity), or whether the mind is best understood as a ‘general-purpose computer’ (Brown and Richerson, 2014). A key question in this regard is whether cognitive functions can be reduced to activities in specific areas of the brain, or whether cognition is an emergent property irreducible to particular brain functions (Lloyd, 2011). The problem harks back to the reductionist position and the question of whether there is such a thing as mental properties distinct from physical properties, and if so, how such immaterial phenomena can exert an influence on the material world. Some philosophers and scientists dismiss dualism, the notion that the mental is distinct from both the brain and the body, in favor of reductionism. Others, such as Searle (2007), advocate property dualism, that is, the idea that consciousness as a subjective, qualitative state is an emergent, systemic feature generated by physical processes in the brain, but that exists at a higher level than that of neurons. This basic and seemingly intractable problem is related to a set of other issues that divide philosophers, behavioral scientists and social scientists.

Cognitive predispositions and the role of culture

A central question dividing social and behavioral scientists concerns the extent to which human cognition is conditioned by the evolved neural architecture of the human brain. While social scientists have tended to downplay the role of the latter, evolutionary scholars highlight the significance of evolved, domain-specific information processing mechanisms that ultimately limit and condition human cognitive and interactive capacities. Domain specific brain structures apply to particular situations and aspects of existence, such as thinking about social relations, objects, causality, space, numbers, or language. Some cognitive and evolutionary psychologists hold that the modular learning machinery of the brain prepares humans not only for learning languages but also for other aspects of sociality, culture, and knowledge, such as mathematics and psychology (Boyer, 2009; Brown and Richerson, 2014; Tomasello, 2000). Researchers are moreover divided on the question of the extent to which human cognition and behavior depend on automated forms of cognitive processing, and the extent to which human cognition involves more general-purpose forms of conscious reasoning. At one extreme end, human behavioral ecologists view human improvisation intelligence as an evolved tool enabling adaptive responses to specific ecological niches. In this view, environmental stimuli evoke various domain specific responses leaving little if any room for more general-purpose forms of conscious reasoning (Brown and Richerson, 2014). So-called dual processing researchers emphasize both the role of evolutionarily adaptive, automatic responses and general-purpose, conscious-processing forms through which humans can address more complex problems. As Greene (2013) puts it, being wired for tribalism (that is, in-group cooperation and out-group hostility) does not mean being hardwired for tribalism as the human brain has various more or less modifiable circuits that compete for control of behavior.

Social scientists have for long been reluctant to recognize the role of any evolved, cognitive tendency shaped by the neural architecture of the brain beyond the most basic level generating awareness and theory of mind. The main emphasis has been placed on the uniquely human capacity to use language, engage in self-conscious reasoning and on the characteristics of social relationships, rather than on natural kind categories rooted in innate neurological or biological features of individuals. For example, while evolutionary scientists tend to view emotions such as fear and anger as natural kinds that can be meaningfully distinguished from each other, social constructivists see emotions as social facts since any bodily change or form of physical arousal must be interpreted by an individual who uses linguistic concepts, labels, and categories, shared situation definitions, and available situational cues to make sense of their experiences (Lindquist et al., 2013). Social categories are believed to have a profound influence on individuals as they may serve to reinforce specific role performances through social ‘looping effects’, a form of social feedback (Hacking, 1995). Moreover, as humans are culture-bearing and meaning-making creatures, it follows that there is no such thing as a generic human nature that can be studied in isolation from specific historical and cultural contexts. Anthropologists, most notably Geertz, criticized sociobiology on the grounds that it is impossible to conceptualize a human essence devoid of culture, as it would lack any recognizable human qualities. Geertz, like Mead before him, emphasized that the human mind is reliant on cultural content extrinsic to single individual brains, that is, shared understandings and representations, in order to operate (Bloch, 2012; Reyna, 2006).

Cultural transmission processes

These fundamental differences in how scientists view the determinants of cognition shape their respective understandings of processes of cultural transmission, that is, processes through which acquired characteristics are passed on between individuals, groups and generations. Social scientists tend to view transmission as open-ended processes whereby culture is passed on via communication, socialization, and imitation. Such processes underlie massive cultural variation as they are largely unbounded by brain structures. At the other end of the spectrum, some evolutionary researchers maintain that the neural architecture of the brain determines the transmission process as well as the cultural content of the mind (Brown and Richerson, 2014). In between these positions researchers disagree on how and to what degree transmission is conditioned by cultural influences and biological factors respectively. For instance, Boyer (2009) maintains that modular cognitive predispositions generate a certain degree of regularity in that certain forms of cultural representations and inferences are more likely to occur, and are thus more easily transmitted than are others. Researchers in the area of gene–culture co-evolution examine how social institutions and genes impact each other. Some researchers in this vein argue that the transmission of human behavior and culture are best explored within a general Darwinian paradigm, meaning that Darwinian theories are extended from the biological domain to that of human history and society and that the transmission of culture and institutions is understood in terms of processes analogous to those characterizing biological evolution (Paul, 2015; Sloan Wilson et al., 2014). Adherents to dual inheritance theory focus on the influence of culture on the reproductive fitness of individuals. In these cases, cultural systems seem to operate independently of biology; the former are deemed as maladaptive possible outcomes of ‘runaway processes’. The latter refers to processes whereby cultural symbols and practices that originally served adaptive functions are gradually transformed into arbitrary ones. In Paul’s (2015) recent version of dual inheritance theory, cultural symbol systems serve to contain ‘the disruptive passions associated with the genetic agenda’, thus facilitating ‘organized harmonious social life’ (p. 309).

Having described the main dividing lines concerning levels of analysis and the roles of brain structures and culture for cognition and processes of cultural transmission, the paper addresses some major shortcomings and limitations associated with both naturalist/evolutionary and culturalist approaches, with specific focus on the extreme positions on each end.

Disciplinary reductionisms, limitations and shortcomings

The first problems considered pertain to the general Darwinist position taken by some evolutionary researchers on the questions of levels of explanation, and mechanisms underlying cultural transmission. This position ignores some specific features of the social world that have kept most social scientists from attempting to study it from a strictly Darwinian perspective. To begin with, beliefs and practices are rapidly changeable. Moreover, social features often spread intentionally and through intellectual interaction, meaning that mental selection processes precede actual selection as humans explore choices through calculations and deliberation (Nelson, 2006). Furthermore, unlike genetic adaptations, institutional phenomena are collective properties based on shared understandings, and they cannot be understood as aggregations of individual attributes (Searle, 2006). In addition, in the social domain multiple socially constructed adaptation criteria are at work at different levels of aggregation. In many areas of the social realm, cultural features and institutions are selected and tested on the basis of social yardsticks, values, beliefs and interests held by influential individuals and groups in specific contexts. Hence, adaptation cannot be accounted for by reference to one single principle, such as reproductive fitness. Likewise, there is no single criterion for maladaptation since multiple adaptation criteria are at work (e.g. individual and corporate profit-maximization, or the promotion of environmental sustainability) at different levels (e.g. individuals, families, corporations or states) (Nelson, 2006). What is adaptive at one level or in one domain may be maladaptive at another level or domain (e.g. maximizing offspring numbers may be maladaptive in terms of maximizing profits). In addition, once a trait has evolved, it can be used for multiple purposes in many different contexts. Thus evolutionary adaptive functions may be coopted and given new and different functions (Graziano, 2013). Finally, the transmission of cultural features is often unrelated to issues concerning survival and reproduction. Institutional scholars of all hues point at a range of different reasons for institutional reproduction (transmission), ranging from elite power, path-dependent legitimation processes, and objectification to various lock-in and self-reinforcing patterns that are completely unrelated to any adaptive function, biological or societal (Mahoney, 2000: 525). Taken together, this suggests that cultural and above all institutional processes of transmission have fundamentally different characteristics from those characterizing biological evolution. It follows that an interdisciplinary approach has to acknowledge distinct levels of explanation, and their complex forms of interaction.

Switching to the other end of the spectrum, cultural reductionism also has serious limitations. To begin with, the concept of culture is essentially contested, and in the last few decades much energy has been devoted to deconstructing Boasian, Kroeberian, Parsonian, Geertzian, structuralist and functionalist notions of culture as relatively tightly integrated networks of meanings, beliefs, attitudes, symbols, rituals, and practice. Systemic notions of culture have been criticized for ignoring psychological, cognitive and neurological factors, intrinsic sources of individual motivation, unconscious forms of cognitive processing, as well as idiosyncratic personal experiences, and the ways in which such factors affect the selection, transformation and recombination of cultural content inside individual minds. Lost in systemic notions are forms of culture at the juncture of the intersubjective or institutional and the cognitive-individual levels. It is therefore not clear how widely shared cultural material must be in order to count as culture rather than as idiosyncratic personal constructs. Cultural theorists have since sought to address such problems by reconceptualizing culture as a complex arena with multiple conflicting voices on the part of individuals with particular life histories, interests and strategies (Bloch, 2012). Among more recent approaches to culture, practice theory and toolkit theory have been particularly influential. The toolkit perspective conceptualizes culture as a ‘tool kit’ of themes and repertoires from which people construct ‘strategies of action’ (Swidler, 1986: 273). In this view, individuals have at their disposal not the entire characteristics of larger cultural and institutional systems, but a fragmented set of skills, cues, habits and heuristics that enable them to select appropriate action strategies in different institutional contexts. Practice theory is compatible with toolkit theory, but while the latter stresses linguistically mediated interaction, and thus explicit cultural themes at the level of semantic consciousness, the former places emphasis on fast, automatic thought processes and implicit habits, practices, skills, dispositions and tacit classificatory schemes (Lizardo and Strand, 2010). In combination, toolkit and practice theory make up a dual-processing approach to culture that may seem to resolve some of the problems associated with more systemic notions of culture. The presumption is that explicit cultural themes shape conscious forms of processing while automated, unconscious processing is conditioned by the acquired ‘habitus’ (Vaisey, 2009). While this dual processing approach takes into account the importance of unconscious processing, it still overlooks the role of psychological and neurological factors and their impact on individual motivation. As Vaisey (2009) acknowledges, the habitus may not be entirely acquired but also shaped by psychological dispositions. In response to such imprecision, Bloch (2012: 150) criticizes practice theories for making cognitive claims that ‘remain totally vague’ since they have not seriously engaged with the cognitive sciences. A full theory of culture cannot do without a more elaborated understanding of cognition. Consequently, social scientists must jettison reductionist models positing that cognition is entirely conditioned by socialization, culture, roles and subject positions, and the notion that culture is not influenced or limited at all by body and brain structures.

Finally, an overemphasis on specific brain modules for cognition ignores the role of general mental phenomena and the fact that there are documented cultural differences with regard to cognition and attention (Masuda et al., 2008). Moreover, in light of new evidence for a higher form of neuroplasticity (that is, changes in connectivity patterns between brain cells) the notion of strict local specialization of brain functions has been called into question. Hebbian plasticity that strengthens synaptic connections between cells appears to be complemented by neural reuse, a form of plasticity that operates at a higher level of brain organization. Through neural reuse, coalitions between entire brain regions are established to support of various complex forms of cognition and tasks. Research on neural reuse indicates that overall patterns of connectivity shape brain function (Anderson, 2015). Taken together, this suggests that both brain functioning and subjective consciousness amount to more than the sum of a set of modules. Subjective consciousness is a unified experience that cannot be reduced to a function of evolved modules, cultural scripts, biographical narratives or social roles (Beatty, 2014). Obviously this should not prevent researchers from distinguishing the specific components of the brain that process language and compute visual experiences, as Searle (2007) points out. Dual inheritance theories viewing genetic aspects and the social world as separate systems face a similar problem as they too disregard how social factors and factors intrinsic to the structure of the brain and body interact in a unified manner (Bloch, 2012). The systemic, unified and subjective quality of human consciousness must be recognized as a fact, meaning that a unified approach to cognition must take into account the complex interaction among intrinsic brain processes and extrinsic influences, such as cultural classification schemas, on cognition.

What is required are approaches that can account for the simultaneous influence of an entire range of interconnected social, psychological, biographical, neural and genetic variables in a parsimonious way. In the following I suggest that attention schema theory (AST), developed by Graziano (2013), is a useful addition to the existing body of integrative anthropological approaches in that it addresses several basic problems concerning consciousness: It is founded on an evolutionary framework; it accommodates the simultaneous influences of the neural architecture of the brain and social and situational influences; it provides a tentative solution to the challenges concerning domain specificity and domain generality; it provides for a limited scope for human agency; it is indeterminist and therefore compatible with the notion of relatively open-ended human developments; and it is open to the existence of distinct levels of explanations and their complex interactions. Finally, attention is located at a point of intersection between intrinsic and extrinsic processes and influences, and it is therefore a particularly useful heuristic for interdisciplinary reconciliation.

Introducing attention schema theory

The main purpose of AST is to account for the brain basis of awareness as a subjective experience, and how we can be subjectively aware of both external content and internal contents such as emotions and abstract thoughts. The theory is that nervous systems evolved an ability to augment the most urgent of incoming signals, and that this selective signal augmentation eventually developed into consciousness, or rather, awareness in the form of an attention schema, or model of attention. Through the process of attention, which includes competition, selection and enhancement of information in the brain, the latter apprehends a signal, directs its cognitive machinery onto it, processes it, and controls behavior on the basis of the information. AST accounts for consciousness as a result of interaction among brain-wide networks and the integration of different kinds of information. Attention binds together prevailing sensory data, be they visual, emotional, cultural, or action oriented, stemming from interconnected brain modules and networks, and/or to varying degrees produced, enhanced or suppressed by specific social, cultural or situational aspects, into a single, unified representation, or a single coherent state (Graziano, 2013).

While the brain produces attention, awareness is known by the conscious mind, a specific informational model constructed on neuronal circuitry that functions as a working description of what it means for a brain to apprehend, focus and process information. Just as the brain generates a model of the body that can be used to control the actual body, Graziano suggests that the brain produces an attention schema, a simplified model of attention that can be used to direct attention. Visual representations or representations of the structure of the body do not contain reports about awareness itself. This is what the attention schema does. It generates a simplified notion about the way the brain attends, a model of the process of attention itself, and this leads it to conclude that it is aware. Thus, the attention schema provides the key to understanding the conscious mind in the sense of awareness of the fact that we attend. Just as the body schema and the body are not synonymous, attention and awareness cannot be equated although they correspond with each other. Like the body schema, which only includes information required for the control of bodily movements, the attention schema does not include a detailed neuroscientific account of the interacting neurons generating attention. The attention schema only includes representations that are necessary for it to track attention, while leaving out many details of the actual phenomena of attention. In this sense the attention schema works as a hub that only recognizes specific kinds of information represented in different brain regions as being within the bounds of consciousness, while other forms of information are integrated in the brain without ever reaching consciousness. AST is agnostic as to whether the schema itself is produced in one brain region or whether it is a product of a complex interaction among several brain regions (Graziano, 2013; Webb and Graziano, 2015).

The attention schema can allocate attention across a range of different types of information, from senses like vision and hearing to internal thoughts, emotions and abstract ideas. Like striking auditory or visual stimuli, salient memories, emotions or abstract thoughts can come to dominate brain processes. The ways in which memories, goals, or emotional states are represented can be biased in such a way that they are more deeply processed and thereby more or less likely to guide actions. However, as noted, awareness can also direct attention in a goal-directed, top-down manner towards specific internal or external content, such that attention on one particular stimulus inhibits awareness of other stimuli. Awareness can function as a driver of behavior since we are more likely to react to aspects we attend to. While it is far from the only factor influencing attention, awareness is a part of the mechanisms of attention control, in that it can act on the brain, direct attention and thereby shape brain processing and behavior (Webb and Graziano, 2015).

According to AST, developing awareness was evolutionarily advantageous since it can be used to guide one’s own behavior and predict the behavior of others. In this sense, the theory provides a limited scope for individual agency (although, as we shall see, this is not equated with a naïve notion of unbounded free will), while indeterminist and open-ended in that it does not view awareness as limited to its evolved function of generating a predictive model of attention. It can be put to a wide range of other uses. One such use has been the promotion of social cohesion, as awareness generates models of the consciousness, or attentional focus, of others, making up the basis for empathy, joint attention, collective intentionality, social alignment, cooperation and community cohesion (Graziano, 2013). Thus, the theory is compatible with the notion of observer dependent institutional facts (in contrast to natural facts existing independently of human observers), making up a distinct, social level of analysis (Searle, 2006).

In the following, the paper will briefly outline how AST, in combination with a neurohermeneutic approach, can be used to analyze the simultaneous interaction between forms of culture and institutions that exist independent of individual experiences with cognitive processes mirroring the workings of the neural structure of the brain.

Bringing it all together

AST serves to clarify how intrinsic brain processes continuously interact with the social and physical environment in producing and shaping specific attentional states. Moreover, the theory can be used to account for how identities and roles, that is, being aware in a specific capacity, contribute to directing attention and thereby influencing bases for action and collective action. In this way, it can be used to clarify the multidirectional and interactive linkages among individual cognition and wider institutional structures of roles and social identities.

Social identities, roles, symbols and concepts, biographical experiences and characteristics of specific relational contexts direct attention to certain kinds of stimuli, details, or information while suppressing others. At the same time, intrinsic brain processes make people differentially inclined to attend to different types of information. All the above aspects can be considered simultaneously as they are unified into a single representation, attention. They act in concert to augment or suppress information, direct attention, and produce or reinforce identities that in turn direct attention and ensuing intentions in specific directions. This suggests that it is very difficult to distinguish among natural and social facts, or biological and cultural aspects, in any real world situation as they are inextricably unified into a human attentional state.

Multiple simultaneous processes at different levels, more or less central to attention, are at work almost simultaneously in generating an attentional basis for the creation of meaning and intentions. Reyna (2006) distinguishes a hierarchy of four levels ranging from automated reflexes and implicit associations that occur beyond attention, to sensation within the attention spotlight, to conscious classification and interpretation (bestowing meaning on an object or subject of attention), and finally the level of desire when representations of how to act are formulated. This schema is useful in charting the interaction between slow and fast, unconscious and conscious forms of cognitive processing, and their respective relationship to attention, identity, and action. Social influences may be most clearly present at the levels of sense-making and goal-formulation but they arguably pervade all four levels, albeit in different ways as different types of social interpenetration work through different types of memory, that is, intrinsic storing capacities. Psychologists distinguish between declarative memory, which includes both semantic memory (of words, facts, concepts) and episodic or autobiographical memory (of specific events and experiences), and implicit, procedural memory (of habituated, automated skills that may be drawn on without being brought into the attention spotlight). Social influences occur, not only at the explicit levels available in the declarative memory but also to varying degrees at levels beyond the attention spotlight, even at the physiological level, e.g. through the coordination of heart rate (Bloch, 2012). Both conscious and pre-reflexive attention processes are shaped by the gradual integration of experience generated by social engagements with the world. As Allen and Williams (2011) suggest, the individual connectome, that is, the shape and balance of neural network connections in the brain that underlie differences in both unconscious and conscious cognitive processes, is affected by socio-economic and cultural circumstances, as well as by environmental factors such as local pollutants (Allen and Williams, 2011).

To analyze how social and cultural influences shape attention, it is helpful to apply a ‘shreds-and- patches’ view of culture, based on the neurohermeneutic approach outlined by Reyna (2002). In this view, explicit memories of sensations tend to have cultural associations as they become linked to particular signs. There are also more or less implicit procedural memories associated with procedural signs or cues that script ‘what to do about what is’. Such signs and cues can become interrelated into procedural schemas. The shreds denote signs and procedural schemas that are shared within any given population, or shared by specific population segments, while patches refer to the idiosyncratic memories that single individuals associate with particular signs. Culture is thus only partially shared and continually changing as new signs are associated with existing sensations, or as novel sensations are linked to existing signs (Reyna, 2002). Shared culture is reproduced and diffused through social learning processes whereby people not only come to associate memories, skills and identities to specific cues or signs, but also learn by understanding others as intentional agents, in the sense of adopting their perspectives of how to understand a specific cue or sign such that they become markers of a shared understanding of a situation. Such an understanding includes not just a certain attentional focus but also a specific way of being aware that is represented by the symbol or the cue (Tomasello, 2000). By understanding the attentional states that others associate with specific signs and cues, people can engage in what Tomasello (2000) calls joint attentional interactions.

Human meaning is generated as the shreds and patches of stored associations, e.g. categories, concepts, and narrative structures stored in various memory networks, are activated and connected to classify, map and make sense of new experiences (Reyna, 2002). Such processes may be more or less conscious. They may occur within the blink of an eye and are then reliant on fast, implicit culturally and socially influenced associations over which the attention schema has little or no influence, or they can be slower and involve a conscious mapping of experiences through the active retrieval of episodic memories, representations, categories, etc., that are brought into the attention spotlight (Reyna, 2006).

By accounting for the basis of our awareness of attention as a subjective experience, AST sheds additional light on the notions of self and identity, that is, conceptions or images of oneself. Signs or cues activate the attention schema, the awareness of attending in the capacity of a certain role or identity, as well as the attentional focus and the intentions associated with it. For instance, occupational roles may contribute to directing attention to specific aspects, and in specific directions, be they real estate prices or public health issues. People are not necessarily explicitly aware of all the attentional states that they activate in response to specific cues and signs, but the awareness of attention provided by the attention schema, in combination with various autobiographical and procedural memories associated with linguistically formulated propositions, enable people to develop explicit knowledge of the capacity in which they are aware of a specific attentional state. In other words, the attention schema contributes to generating the notion of a self with explicit awareness of multiple situational identities and roles associated with different scripts, attentional focuses, and sensory motor skills.

Identities of which individuals are aware can become interactively and reciprocally linked to unconscious, automated forms of processing. For instance, professional identities such as dancer or surgeon are based on automatically run, procedural schemata coordinating learned motor skills that operate without meta-representational consciousness of their function, that is, beyond attention and awareness (Allen and Williams, 2011: 5). Moreover, learned sensory motor skills enabling bodily coordination can be tapped into through joint attention so as to strengthen group identity and cohesion. For example, joint rhythmical coordination of body movements, such as in processes of military drill or any ritual promoting a ‘collective way of looking at the same thing’, appears to promote group cohesion, sometimes to such an extent that individuals experience becoming one with the surrounding group. McNeill has hypothesized that processes of muscular bonding kept together in time through rituals, dance and song became important for human evolution as such processes enabled human groups to enhance their cohesion, increase their size, and thus ensure survival (McNeill, 1995: 5).

Entangling the intrinsic and extrinsic

AST in combination with a neurohermeneutic approach to mental states and culture, as outlined above, can serve to reconcile some of the conflicts between evolutionary and constructionist approaches. This combined approach facilitates analyses of how social and cultural influences shape cognition and emotional responses, and analyses of the possible overlaps and radical differences between humans and other primates. Humans share with other primates the ability to have mental states and track attention. Primates, however, lack the human linguistic capacity through which mental experiences can be made into objects of self-reflection and attention on the basis of representation and signs. If a sensation reaches the human attention spotlight, the encultured human mind will rapidly retrieve and activate memories, representations and concepts to make sense of it. The awareness of attention in combination with the ability to engage in slow, self-conscious reflection using cultural signs and linguistically formulated beliefs, combined with the uniquely human ability to direct attention to more or less abstract propositions, enable people to create a psychological distance from their actions. Taken together this generates a distinct form of human self-awareness linked to various roles, scripts and narratives. This form of awareness arguably differs in kind from that of any other primate (Allen and Williams, 2011). Similarities between primates and humans may be found at the levels of basic automated reflexes and immediate sensations, however. Unlike more complex cognitive processes, such low-level appraisals do not involve linguistically formulated representations and content sentences involving beliefs, desires and propositions, and they therefore can be viewed as distinct from more complex thought processes. Griffiths (2013) maintains that in low-level emotion appraisals, such as reflexively evading a spider out of fear, the information flow in the brain becomes blocked such that examination by other cognitive sub-systems is prevented. Basic emotional responses of this kind are intrinsically ‘action directing’ and occur before any conscious interpretation or formulation of beliefs and desires (Griffiths, 2013).

From the viewpoint of this paper, human attention processes beyond the levels of automated reflexes and some immediate sensations are inevitably pervaded by cultural, social and environmental influences as attention involves the activation of various types of memories of experiences and associated signs and representations, retrieved from multiple brain regions. As Satel and Lillienfield (2013: 16) put it, ‘there is a babel of crosstalk among numerous brain regions as they are strung together in specialized neural circuits that work in parallel to process thoughts and feelings’. It is possible that evolved emotional responses such as disgust are part of more complex cognitive processes beyond the most immediate level of sensation, but they become entangled with cultural representations and memories retrieved from other brain regions and/or coopted for purposes other than those for which they evolved.

Using AST, we can account for how shared understandings of the attentional focus, intentions and expectations associated with specific social roles interact with both general and more idiosyncratic genetic, neurological, cognitive, and psychological factors, as well as individual biographical experiences. For example, individuals may differ with regard to underlying motivational biases and preferences, and personality traits such as extroversion, conscientiousness and neuroticism. It follows that specific roles and social positions, even entire cultural, organizational and institutional regimes, impose different cognitive loads on individuals with different personal characteristics, requiring more or less emotion management and cognitive dissonance-reduction to neutralize conflicting inner and external signals. The workings of organizations and entire social fields, such as segments of the military, the financial sector or academia, may in part mirror the personal characteristics of influential actors in the field. In social environments dominated by specific norms and/or individuals with particular traits, a cognitive load is imposed on individuals with other characteristics as their need for emotion management and impression management increases. At the same time, traits and individual personalities are not fixed but emerge from interactions and feedback processes among intrinsic cognitive and emotional processes and external contexts. Positive feedback processes can strengthen certain role performances such that individuals can perform a role with increasing ease. Identity shifts, such as when people redefine themselves by acquiring a new social identity, e.g. a religious identity, can serve to suppress certain signals and augment others, leading to changes in both attention and action. Neural connections can also be rewired, albeit to varying degrees depending on the type of neural network and age. Age matters since neural coalitions in the brain reinforce existing patterns of connectivity over time (Anderson, 2015).

Education and varied life experiences generate a larger patchwork or toolkit of skills, thinking styles and action repertoires available to the attention schema for selective activation. This is not to say that humans can direct attention in an unconstrained manner, not even among the ‘tools’ retrievable in declarative and procedural memory networks. Tracking and directing attention, like moving one’s limbs, does not amount to free, unrestricted will. Individuals are dependent on the ways in which the brain produces attention and the attention schema, on how different parts of the brain are connected, the implicit associations it makes, and on the types of signals and information that the brain and body make available for the attention schema. Such processes are influenced by genetic predispositions, and the attention schema cannot reach or influence many unconscious brain processes and implicit associations.

In other words, a good deal of the human meaning creation process, in which shreds and patches of stored signs and associations in various memory networks are retrieved and connected to make sense of new experiences, occurs beyond volitional control. However, provided that humans can track and direct attention, the attention schema can be used to direct attention away from certain stimuli while focusing and enhancing other stimuli, or to actively retrieve and draw on specific memories, skills or abstract thoughts. In this way, the attention schema can direct attention from, say, one type of thought to another, thereby activating a specific brain circuit or network such that signals associated with other circuits are suppressed. Thus, AST can increase our understanding of how social learning, cognitive behavioral therapy, meditation and mindfulness practices can contribute to changing attention and ensuing actions.

Individuals are differentially predisposed to apprehend, select, enhance and direct their cognitive machinery to different forms of information. Yet we can learn to direct attention to new types of thoughts, stimuli or entire thinking styles, meaning that neither genomes nor connectomes can be viewed as conclusive determinants of human action. However, notions of genetic determinism can contribute to social looping effects if a label or diagnosis becomes integral to a personal or social identity that makes the attention schema even more attentive to specific signals, thus reinforcing pre-existing tendencies instead of suppressing them. Similarly, research on stereotype threats in psychology has shown that the activation of negative stereotypes can contribute to weaken the performance of stigmatized individuals on various tasks (Schmader et al., 2008).

The architecture of the brain sets some general external parameters for the types of connections, cognitive styles, and forms of sociality that humans can activate and draw on. Thus, there are limits to the variation observed in human social life. Fiske (1992) has suggested that people generate, coordinate, evaluate and make sense of social relations using a limited set of elementary models that have appeared in specific combinations in all known societies across the world (with the exception of market relations): communal sharing, equality matching or in-kind reciprocity, authority ranking and market pricing relations. Similarly, Shweder and Haidt have argued that moral discourses across all cultures are founded on three clusters of moral concerns: the ethics of autonomy (concerns about freedom, rights, harm and justice), community (concerns about duty and the collective enterprise), and divinity (concerns about purity, sanctity and the realization of one’s spiritual nature) (Shweder and Haidt, 1993). Since even a limited repertoire of basic emotions and forms of sociality can be strung together in an enormous number of different combinations, such elementary forms are probably best seen as ideal types rather than ‘real’ or pure types. Thus, even though there are limits to human emotional, social and cultural variation, the scope for variation is enormous.

As individuals have different underlying motivational biases and are differentially attentive to different stimuli and information, they generate idiosyncratic patches, toolkits and worldviews, even if they are exposed to similar forms of social facts and representations. Exposure to religious narratives and practices, for example, affects individuals differently as they select and focus attention on particular themes on the basis of their individual personalities and biographical experiences. It follows that religions and other social facts do not have a one-sided impact on individual cognition and behavior. At the same time, it is imperative to recognize the distinctive character of the institutional level of analysis.

Disentangling the institutional dimension

The relationship between individual, cultural and institutional characteristics and change processes has been subject to frequent misunderstandings and serious distortions. A common fallacy is to explain outcomes at the institutional level by reference to aggregates of individual beliefs and characteristics. A recent example is Wade (2014), who in his book A Troublesome Inheritance accounts for institutional differences by reference to aggregated genetic differences. One obvious misconception here is that the workings of institutional regimes are seen as mirror reflections of aggregated individual characteristics or aggregates of patches, that is, individuals’ experiences, beliefs, representations and understandings. The study of culture in the sense of patches or toolkits has a link to the individual cognitive level, as the selection of cultural repertories on the part of individuals is influenced by individual motivational biases and preferences, and the ways in which automated unconscious processing systems assemble memories to generate meaning. Institutions, by contrast, make up a specific type of shared understanding in that interrelated individuals invoke them simultaneously, meaning that they are less reliant on individual whims. In line with Searle (2006), institutions are here viewed as observer dependent phenomena that exist within frameworks of constitutive rules enabling various joint attentional interactions. Institutional structures exert a top-down influence, as they can compel individuals to conform to preordained roles, in the sense of having to learn to activate particular attentional focuses and associated repertories of action in certain social domains, such as courtrooms, weddings or various sports activities. Thus, institutional structures have causal forces of their own and impact on human biological systems. For example, Birth (2007) argues that biological desynchronization such as disrupted sleep patterns and stress, and ensuing health effects such as altered cortisol cycles, are part of the physical costs of the global time-space compression and the standardization of work and consumption practices associated with the global consolidation of capitalism.

Institutions with different origins can co-evolve to form specific regimes or societal syntaxes. Such regimes regulate an enormous variety of role performances, enabling large, concerted macro-patterns of various joint attentional interactions. Single individuals do not internalize the structures of entire institutional regimes. A regime constitutes a wider macrostructure in which individuals are provided with cues specifying the attentional focus and behavioral expectations associated with specific domains. People thus navigate within externalized institutional structures by picking up cues that generate the order and regularity characterizing wider societies (Lizardo and Strand, 2010). When a large number of individuals share the same understanding of how to act in different roles associated with different rights and obligations in different social domains (e.g. parent in the family, consumer in the market, official in a formal organization), domain consensus has been reached. The term denotes a general acceptance of how to act in an entire set of social domains (Frödin, 2013). Within institutional structures people employ conscious processing at different levels, in different situations and according to different rationales, such as in market situations, only to habitually and unconsciously switch to other rationales of action that they consciously employ in other domains, such as the family, without consciously reflecting on the fact that they regularly shift among identities, attentional states, and rationales of action. Minor differences in the make-up of institutional regimes can have a hugely disproportional impact on a wider society. For instance, Western historical exceptionalism, meaning the simultaneous development of state sovereignty, capitalism, liberal democracy and the rule of law, can be viewed as a result of the convergence of a set of institutions with different origins, forming a specific syntax of domains (on which domain consensus was reached) in which people regularly engage in joint attentional interactions according to a specific pattern. The historical development of this regime type, i.e. the domain consensus underpinning state sovereignty, the rule of law and capitalist liberal democracy, is likely very specific.

In an influential paper, Henrich, Heine and Norenzayan (2010) criticize the tendency in psychology to ignore cultural variation across human populations, as many psychological studies are based on samples drawn from so-called WEIRD societies, an acronym for Western, Educated, Industrialized, Rich, and Democratic. According to Heine et al., WEIRD people ‘are among the least representative populations one could find for generalizing about humans’ (p. 61). While this argument has some merits, it conflates the individual and institutional levels of analysis. Given the current rates of cultural diffusion and migration, there are arguably pockets of WEIRDos all over the world. The WEIRD argument overstates the homogeneity of Western and non-Western countries, as it does not recognize the distinctive character of the institutional level of analysis. While population segments may adopt WEIRD mindsets across the world, it does not necessarily follow that ‘the weirdest people in the world are a harbinger of the future of the world’, as argued by Rozin (2010). Entire institutional regimes do not spread as easily as other forms of culture, since they require domain consensus. Therefore, it is uncertain whether domain consensus on varieties of a ‘WEIRD’ institutional regime type will be reached across the world any time soon. This suggests that cultural and, above all, institutional diffusion processes have specific features that differ considerably from those of biological evolution. That said, social scientists must recognize that human biological and neural structures set some limits to institutional variation, although, as noted, the scope for variation is enormous.

In line with Fiske (1992), we may hypothesize that regimes make up context specific sets of institutionalized social domains within which people regularly invoke a finite set of forms of sociality, underpinned by a finite set of moral concerns. Such an understanding enables analyses to move beyond crude, dichotomous terms such as notions of modern and traditional societies (or WEIRD and non-WEIRD societies). The interrelated social domains making up entire regimes may be more or less interdependent and display varying degrees of functional complementarity. A central question for interdisciplinary research concerns how individuals and groups respond to the global interweaving of institutions with very different origins. An interesting area for interdisciplinary research concerns whether there is a long-term trend towards increasing cognitive systematicity as individuals abandon inconsistent cognitive and cultural styles, as suggested by Pinker (2011), or whether people to varying degrees compartmentalize seemingly inconsistent notions such that different forms of sociality, morals, and social organization co-exist in seemingly paradoxical and sometimes conflicting ways, possibly generating role strain, cognitive dissonance, and increased cross-situational inconsistency.

Using AST to further bring together the social and the natural sciences, a central aim of this paper has been to highlight connections and raise questions for future interdisciplinary research. Obviously, it remains for such questions to be refined, researched and tested.