Realist social ontology and the embedded individual

All three social theorists considered in this paper work from a realist ontological position in their portrayal of the ‘embedded’ individual. Their common ground regarding what it is that constitutes ‘reality’ is summarised in the following six premises:

Premise 1: There can be no social structure without individuals.

Premise 2: The ‘downwards conflation’ of social structure to individuals is avoided: that is, social structures are not simply explained by the properties of individuals.

Premise 3: Individuals depend on social structures.

Premise 4: Individuals are not reducible to social structures: that is, the ‘upwards conflation’ of agents into structure, leading to theoretical analyses of the form ‘structures cause structures’ with no mediation by an analysis of the actions of and relations between individuals (often referred to as methodological holism and functionalism), is ruled out.

Premise 5: Social structures exist prior to individuals: individuals are ‘born into’ social structures that temporally pre-date them. This distinguishes the realist position taken here from Anthony Giddens's (Reference Giddens1984) ‘structuration’ theory. This emphasised the ‘duality of structure’ which ‘centrally conflated’ agency and structure, compressing them together with unfortunate results for the view of individual actionFootnote ¹ (Archer, Reference Archer1995; Mouzelis, Reference Mouzelis1995).

Premise 6: Social structures have causal powers that (i) are not reducible to individual thoughts or actions and (ii) are only exercised through individual actions. Thus, the potential powers of structures are only realised when the exercise of individual causal powers occurs. Beyond these six premises, there are differences over how to characterise 6(ii).

Archer's conception of agency as the starting point

Although we consider Archer, Hodgson and Davis's approaches to issue 6(ii) above, we shall build on Archer's analysis. To date, she provides the most theoretically sustained empirically grounded conception of systematically different types of human mental operation in a realist framework. In the contrast of Hodgson's ‘emergentist materialist’ approach with Archer's approach of ‘causal dualism’, no evaluation of the alternative approaches to causality implied by these works is provided. Nor is a synthesis between these approaches to causality advocated. While emphasising an approach consistent with causal dualism, the paper will indicate, where necessary, where those favouring an emergentist materialist approach might pursue a different line of development. No claim is thus made as to the superiority of causal dualism as distinct from emergentist materialism. Instead, this paper concerns itself with the question of how Archer's approach can (and why it should) incorporate Hodgson's and Davis's insights into habits and collective intentionality respectively.

The concept of modes of internal conversation

As an ‘actor’ and ‘agent’, an individual may be situated in similar contexts, with similar life chances to other individuals in such contexts. However, for Archer that individual is also a ‘person’ who may have a systematically different mode of internal conversation from (some of) these other persons. Consequently, how this individual relates her concerns to her contexts may then systematically differ from (some of) these otherwise similarly situated individuals.Footnote ⁸ Archer (Reference Archer2003) identified four modes of self-talk among her interviewees as part of her research. These modes are not ‘personality types’ but the brief summary below focuses for clarity on the differences in assumed mode of mental operation across types.

The majority of Archer's interviewees were seen by her to follow three ‘modes’ or ‘types’ of self-talk, each ‘type’ taking an ‘active’ orientation towards their structured situations. Their self-talk is described by Archer as ‘instrumental’, orientated to the formation of ‘projects’ in the light of situations and concerns.

First, ‘communicative reflexives’ have past and present lives generally exhibiting a high degree of continuity and repetition or ‘contextual continuity’ (Archer, Reference Archer2003: 184–185). They distrust their own internal dialogues and their decisions are the joint result of internal conversation and external conversations with trusted ‘similars and familiars’, of whom there are usually a relatively large number (Archer, Reference Archer2003: 167–168; Reference Archer2007: 102). Their ultimate concerns are with specific people and they have relatively little problem ‘dovetailing’ their concerns (Archer, Reference Archer2003: 169): their family and friends take priority. Other commitments are fitted around these. Communicative reflexives thus create and maintain a ‘micro-world’ around them of ‘similars and familiars’. Communicative reflexives are thereby insulated from exposure to objective structural or situational constraints (but also prevented from taking advantage of objective situational opportunities). Their orientation towards the constraints and opportunities offered by objective situations is to ‘evade’ them, with a consequent tendency to be occupationally and socially immobile.

Secondly, by contrast, ‘autonomous reflexives’ have a past experience of a contextual discontinuity (Archer, Reference Archer2003: 212). Left to rely on their own judgement (Archer, Reference Archer2003: 235), they show decisiveness in decision-making (Archer, Reference Archer2003: 210–211). Their ultimate concerns tend to prioritise work achievements or commitments but they work at maintaining this while seeking fairly to divide their (remaining) time to other people (Archer, Reference Archer2003: 213). Autonomous reflexives can be very ambitious, can be ‘loners’ and generally wish to be and see themselves as self-reliant (Archer, Reference Archer2003: 214). Given these characteristics, autonomous reflexives are well aware of and confident about weighing up opportunities and constraints of objective situations they face. Their orientation towards objective structures is thus ‘strategic’: they become skilled at knowing and trading off the costs and benefits of the situations they face in order to pursue their work-related concerns and associated projects (Archer, Reference Archer2003: 253), so tending to be occupationally upwardly mobile.

Third, ‘meta-reflexives’, like autonomous reflexives, also have a past experience of contextual discontinuity (Archer, Reference Archer2003: 257) and also rely on their own judgement. However, meta-reflexives tend to have a strong ideal view of how they should be, as persons: their ultimate concern is to live in a context in which they can express certain ideals they hold to (Archer, Reference Archer2003: 258). Often the most self critical of the three modes, their internal conversation is directed at self-evaluation and self-improvement towards their ideals (Archer, Reference Archer2003: 255–156). Meta-reflexives have a ‘subversive’ orientation towards objective structures and situational constraints: they ‘absorb’ objective costs to uphold their ideals and reject objective benefits inconsistent with their ideals (Archer, Reference Archer2003: 289, 293). However, meta-reflexives are equally critical of their situational contexts. Since meta-reflexives often find there is only so much they can do to ‘improve’ themselves, their awareness of contextual ‘incongruity’ (for not enabling them to express their ideals whatever they personally do to self improve) grows, leading them to lateral mobility and several career changes to improve their situational context (Archer, Reference Archer2003: 278). This persistent self and contextual criticism also implies that ‘dovetailing’ of their concerns is problematic for them.

A minority of interviewees of Archer's Reference Archer2003 work did not fit the above three categories. Archer calls these ‘fractured’ reflexives: they engage in self-talk and emote about their situations (Archer, Reference Archer2003: 302) but are unable to conduct instrumental internal conversations to form a plan or action or project to deal with them. Archer calls these ‘passive’ agents, to whom ‘things happen’ (Archer, Reference Archer2003: 303–304).

The marginalisation of habit

Archer's rich conceptualisation of four reflexive types nevertheless builds in systematic avoidance of any ‘dispositional’ interpretation of habit in the mind and, second, minimises a place for habit as repeated ‘behaviour’. Archer (Reference Archer2007: 29–52, 55–58) carries an extensive critical discussion of associated sociological literatures attempting to extend Pierre Bourdieu's ‘habitus’ concept to encompass reflexivity. Notions of ‘flexible’ or ‘reflexive habitus’ (Sweetman, Reference Sweetman2003: 538) in which conscious reflexive deliberation provoked by continual disruption to established socio-economic patterns itself becomes habitual are for her ‘. . .either vacuous or an oxymoron’ (Archer, Reference Archer2007: 55). Archer prefers to exclude a dispositional role for habit. She notes, in her approach, a ‘seeming paradox’:

Archer ‘resolves’ this ‘seeming paradox’ by ‘being precise about what a subject needs to be aware of in order to be influenced’ (Reference Archer2007: 17). Conscious awareness is assumed, the issue becomes of ‘what kind’ of conscious awareness is needed. Archer notes the young academic who, hailing from Britain, finds participation in international conferences a good experience. She does not need to know with ‘full discursive penetration’ that she is benefiting from Britain's colonial past and its impact on the spread of English language. Instead, ‘all she has to recognise consciously and find good is, . . .the ease and fluency with which she makes interventions. . .’ at the conferences (Reference Archer2007: 18, emphasis added).

The emphasised phrase is crucial. The young academic may have developed a habit of fluently intervening. Archer assumes she must always be discursively aware of her fluent action. But why disallow the possibility that the young academic may be unaware of her habitual action? Why not allow that she may or may not develop a later awareness of it? Alternatively, the young academic may have a belief that it is important to make fluent interventions in international conferences. International conference attendance situations may trigger this belief and she acts on it. Archer's approach implies that the academic must be discursively aware that she has this habitual belief. Why not assume that awareness of this habitual belief is something the young academic may or may not develop?

Archer holds that the individual knows their own mind ‘better than anyone else’ (Archer, Reference Archer2003: 21). Their internal conversation is for her a ‘genuinely interior. . .ontologically subjective and. . .causally efficacious’ (op cit: Reference Archer2003: 16) domain of interior mental privacy that ‘. . .is immune from (i.e. not determined by) exterior causes’ (Fleetwood, Reference Fleetwood2008: 197). To concede that an individual without any discursive awareness frequently performs habitual actions and acts on habitual beliefs makes it possible that third parties can often know more of that individual's own mind than that individual.Footnote ⁹

Yet Archer accepts that in certain circumstances, third parties can tell us more about ourselves than we know.Footnote ¹⁰ Nor does Archer's position require us to assume a priori the absence of habit-as-disposition. It can require us to theorise the process by which individuals could become aware of their habitsFootnote ¹¹ thereby preventing themselves being potentially manipulated by third parties (see Section 6). Archer's position is thus not necessarily undermined by allowing a dispositional role for habit in the mind.

In the second place, Archer's larger reflections on individual action in a changing world minimise any remaining importance of habit even as repeated behaviour. She argues that

Archer further claims that ‘. . . “contextual discontinuity” intensifies throughout modernity, and especially during the transition to nascent globalisation, with the consequence that routine or habitual action becomes decreasingly appropriate for all’ (Reference Archer2007: 61).

However, given the increasing pace of life, the world may also be becoming increasingly complex (Hodgson, Reference Hodgson1999: 181), placing greater cognitive demands on individuals. If habits can be automatically created and situationally triggered, this can explain how individuals could become ‘more reflexive’. Greater individual reflexivity may require the reduction of demands on individual cognitive capacities brought about through the existence of (and further development of) institutional patterns that enable habit formation. Moreover, increasing (Knightian) uncertainty about changing socio-economic reality at times of crisis can increase an individual's ‘need’ for habit to combat increased anxiety and stress or ‘ontological insecurity’ (Davis, Reference Davis2003: 122–123; Lawson, Reference Lawson1997: 180).Footnote ¹²

Davis and the nature of collective intentionality

Davis considers that the groupFootnote ¹³ is a neglected agent in economics and it is important to ‘. . .investigate how individuals’ shared intentions facilitate groups acting as agents’ (Davis, Reference Davis2003: 133). Although only individuals form intentions (and, therefore, to avoid Methodological Holism it cannot be held that ‘The group itself believes and intends to do X’), individuals can have intentions about groups. To have a collective intention is not simply another way of describing some kind of societal constraint (Reference Archer2003: 139). A collective intention is voluntarily held, but involves the individual binding themselves to it as a result of the group participation that is involved: ‘. . .expressing a shared intention commits an individual to an intention held by others. . . The individual is bound. . .but voluntarily so’ (2003: 131).

Davis makes the case for collective intentions being neither reducible to another form of instrumental rationality (‘I adopt [the beliefs of members of that group] because it is in my interest to do so’) nor to a form of ‘we-preference’ (‘I prefer to adopt the beliefs of members of that group’) (Davis, Reference Davis2003: 139–142).

A collective or ‘we-intention’ can be said to satisfy two conditions: first, ‘an individual expressing a we-intention believes that this intention is held by other group members’ (Davis, Reference Davis2003: 134) and second, ‘. . .an individual expressing this we-intention believes that it is mutually held by members of the group’ (in other words, person ‘h’ believes that others in group G believe ‘h’ to believe what they themselves believe) (2003: 134).

Collective intentionality, reflexivity and the ‘relative autonomy’ of individuals

Davis's approach to reflexivity is to draw attention to what is involved with group participation. Davis links collective intentionality to individual reflexivity by assuming that

Thus, if individual ‘h’ has an incorrect understanding of what other group members believe or if ‘h’ erroneously thinks that others in that relevant group believe ‘h’ has the same views as they do, then the other members of the group could say to ‘h’: ‘you say we will do this but we have no intention to do it’. This ‘feedback’ then ‘. . .compels individuals in error to re-evaluate what they say and believe’ (2003: 138). To the extent that this ‘re-evaluation’ ‘takes the form of their asking themselves why their own views are at odds with those of others’ (2003: 138) then reflexive behaviour is entailed here.

Davis notes that with continual changes in group association in complex societies, individuals can tend to lose (what he terms) their ‘relative autonomy’: ‘. . .they cease to function as independent distinguishable individuals and instead become enmeshed in ever changing social relationships’ (Davis, Reference Davis2003: 156). Individuals, Davis suggests, often are ‘oversocialised’ (Granovetter, Reference Granovetter1985):

With an ever-more complex social world, ‘oversocialisation’ becomes more likely. Davis does not, however, think that this argument writes the individual out of socio-economic analysis. He distinguishes between ‘capacities’ and ‘capabilities’. A ‘capacity’ is ‘a power that may or may not be exercised’ (2003: 158). A ‘capability’Footnote ¹⁴ is ‘a power that has been developed and cultivated so as to be susceptible to regular exercise’ (2003: 158). Capacities and capabilities are related. ‘One. . .develops certain capabilities when one has the capacities for doing so’ (2003: 158). Davis argues that the potential relative autonomy of the embedded individual depends on the individual having the capability to freely participate in and move across group affiliations, which in turn depends on developing a capacity to self-impose group requirements on themselves. Davis then notes that if embedded reflexive individuals are oversocialised, then they ‘are unable to regularly exercise this capacity [to self-impose group requirements on themselves]’ and so ‘accordingly do not develop the capability for [freely exploring the various opportunities for participating in and moving across various social group affiliations]’ (2003: 158). This ‘individuating capacity’ is ‘inherently possess[ed]’ by people (2003: 158), argues Davis, yet he considers that ‘many, perhaps most’ (2003: 159) individuals fail to develop this capacity. Nevertheless, Davis (Reference Davis2003: 160–163) argues that the socially embedded reflexive individual he portrays is not simply meant to describe reality but also to serve as a policy ideal: ‘Reasoning dynamically, heterodox thinking judges that, if most individuals are not self-determining, that some are creates an agenda for replacing a pure ideal with one that can be made real’ (op cit: 163).

The achievement of relative autonomy: habits and ‘inner conflation’

Archer assumes most individuals, through instrumental internal conversation, consciously choose projects on the basis of their ‘concerns’. Given Davis's argument above, such instrumental self-talk and project choice must also include the individual's choice of groups to belong to and choice of the group obligations to bind themselves to. In this light, Archer's approach amounts to the assumption that ‘relative autonomy’ from groups follows from being able to conduct an instrumental internal conversation. Thus, communicative, autonomous and meta-reflexives are automatically relatively autonomous. Davis by contrast assumes that many, perhaps most, individuals lack relative autonomy and are ‘oversocialised’, so directing attention to how it is individuals can move from being oversocialised to being ‘non-oversocialised’.

I now pursue this difference between Davis and Archer further. Suppose, following Hodgson, that in Archer's approach the mind now has a disposition to generate habitual beliefs. I suppose that (at least some) habitual beliefs are automatically acquired by individuals through group participation.Footnote ¹⁵ I suggest that the movement away from oversocialisation involves individuals becoming aware of influences upon them from their social group participations and obligations that they were (previously) not discursively aware of. We can then infer that individuals are non-oversocialised when they are able to notice and enquire into their own automatic thought patterns (i.e. habitual beliefs) and can ‘integrate’ them into concerns. ‘Integration’ involves two stages. First, a discursive awareness of habitual beliefs arises: they become recognised, by the individual, as ‘concerns’. Second, these new concerns are then (i) added to existing concerns or (ii) used to replace some existing concerns or (iii) are discarded as inconsistent with existing concerns.Footnote ¹⁶ In short, I suggest, persistent vigilance about enquiring into one's habitual beliefs (sustained integration over time) can generate relative autonomy from group influence, thereby achieving non-oversocialisation.

Now Davis sees the development of this capacity for relative autonomy as an issue for all individuals: so whether and how individuals can achieve the above ‘integration’ of habitual beliefs into conscious concerns matters for theory and policy. By contrast, Archer's ‘active agents’ already have relative autonomy, so the issue of ‘integration’ does not arise for them. Archer theorises as if all habitual beliefs are always, instantly and unproblematically integrated into conscious concerns by active agents. The possible reality of habitual belief formation for most people (i.e. ‘active’ agents) is ignored. There is an ‘inner conflation’ of habitual beliefs with conscious concerns in Archer's approach to active agents.

What inner conflation loses for Archer is the more plausible possibility that while most individuals may (as she holds) be ‘active’ agents (they undertake instrumental internal conversations), many, perhaps most of these individuals could (as Davis implies) also fail to experience relative autonomy. Their actions are partly driven by habitual beliefs internalised from their group participations that operate within their instrumental internal conversations, with such beliefs simultaneously unintegrated into their conscious concerns.

Habit as cause of the internal conversation process

Suppose the mind was assumed to have a propensity to form habits automatically given its situations, and that habits determine the operation of any given mode of internal conversation. How might this be understood? In Archer's model of the process of the internal conversation itself,Footnote ¹⁷ the individual engages in a dialogue between their ‘subject-self’ which makes utterances (judgements, statements etc.) and their ‘object-self’ which ‘listens’ and ‘records’ what was uttered,Footnote ¹⁸ presenting it back to the subject-self for further reflection: ‘Confronted by ‘listening’ to the articulated thought, the subject confronts her object’ (Archer, Reference Archer2003: 98).Footnote ¹⁹ There is subsequent further recording of the subject-self's further utterances and so on.

Archer assumes that the subject-self, the object-self and the process of the subject-self/object-self interaction are immune from causation by external structure. Emergentist materialists, however, could assume that the object self/subject self interaction is fully characterised as an instinct–habit process. Deliberation (‘internal conversation’) would be the emergence and operation of ‘higher level’ habits in a layered conception of mind, with mind in principle fully understood as a set of material causes structured by habit and instinct.

Archer is committed to causal dualism, the view that the causal powers of persons are different from those of structures. A causal dualist position must identify a part of the internal conversation process that is immune from causation by material or objective structure.

Habit and an autonomous internal conversation process: a dual concept of mind

Archer depicts the process of the internal conversation as a turn-taking interaction between the object-self and subject-self. To maintain causal dualism therefore we assume that the subject-self is fully autonomous from material causes (i.e. instincts and habits), while accepting that the object-self could be habit structured. The ultimate interiority, subjectivity and causal efficacy of the internal conversation is preserved: the object-self/subject-self turn-taking process is not fully characterised as an instinct–habit structure. Instead, we ask: how might habit affect the mode, rather than the process, of internal conversation?

In order to address this question, we give more structure to the concept of mind. Dropping inner conflation, we explicitly distinguish habitual beliefs beyond immediate discursive awareness from ‘concerns’ available to discursive awareness. We assume that the fully developed mind is a dual structure, with habitual action and habitual beliefs generated in a manner outlined by Hodgson, but with a simultaneously operating internal conversation process operating in a manner outlined by Archer. We assume therefore that individuals continually generate a residue of objective structure (‘internal structure’) within their mind in the form of automatically created habitual beliefs beyond immediate discursive awareness. We assume that these habitual beliefs are created simultaneously with, and on a different level of the mind to, the operation of the internal conversation. We assume that there are relationships between these levels of the mind. Human agency therefore involves not only an interaction between the embodied person and ‘external’ structure but also interactions between ‘internal’ structure (acquired habitual beliefs) and the subjective part of the person's embodied mind (internal conversation).

A hierarchy of capabilities of the mind

To develop the above dual structure conception of the mind, we adopt Davis's ‘capabilities’/‘capacities’ distinction and assume that the mind has four ‘capabilities’ C1–C4:

(C1) The capability to generate actions automatically, triggered by particular situational contexts (the mind's ability to produce habitual actions). (C2) The capability to generate beliefs automatically that can be subsequently triggered by particular situational contexts (the mind's ability to produce automatic judgements as ‘habitual beliefs’). (C3) The capability to engage in ‘instrumental’ self-talk (the mind's ability to form projects or plans of action in the external environment to attempt to realise the individual's principal concerns). (C4) The capability to engage in ‘inner-directed’ self-talk (the mind's ability to treat aspects of the individual's own mental functioning as an object for further enquiry).

Second, we can assume that each ‘capability’ is enabled by a corresponding ‘capacity’ (E) of the mind. So there are four capacities E1–E4 corresponding to capabilities C1–C4.

Third, we assume, following Hodgson, that C1–C4 are a hierarchy of ‘levels’ of the mind. C1 and C2 are the two ‘lower’ levels at which the mind works ‘automatically’ in the Hodgsonian sense. These two lower levels of the mind can be understood as parts of ‘objective external structure’ that have become ‘objective internal structure’ within the embodied mind – brain.

C3 and C4 are the two ‘upper’ levels at which the mind works reflexively as a subjective process of internal conversation, separate from objective structure (both internal and external) in the Archerian sense. The operation of the mind at capability levels (C3 and C4) neither causes nor is wholly caused by the operation of mind at capability levels (C1 and C2). However, there are relationships between the ‘lower’ and ‘upper’ parts of the mind.

Fourth, we make assumptions about the possible nature of these relationships. We first assume a particular relationship between C2 and C3: the self-talk turn-taking process can employ habitual beliefs as ‘gut feelings’ to terminate particular instrumental internal conversations by justifying particular judgements. Whether and the extent to which habitual beliefs are invoked within internal conversation depends on the mind's relative capacity for instrumental internal conversation, denoted E3. Given E3, the greater the number of habitual beliefs an individual has and the wider the range of situations in which these may be triggered, the greater is the individual's automatic employment of habitual beliefs in instrumental internal conversation.

We second assume a relationship between C2 and C4: the self-talk turn-taking process can be directed at making more explicit certain patterns – the habitual beliefs – of the individual's mind.Footnote ²⁰ Which habitual beliefs are scrutinised and how they are scrutinised depends on the internal conversation process C4 itself. However, whether/the extent to which they are scrutinised depends on the capacity E4 of the individual's mind for ‘inner directed’ internal conversation.

Habits, modes of reflexivity and relative autonomy: six hypotheses

These assumptions about the relationship between habitual beliefs and internal conversation offer directions for future research that Archer's avoidance of a dispositional role for habit in the mind (and consequent inner conflation) misses.

First, Archer suggests that communicative reflexives, unlike autonomous and meta-reflexives, have much less faith in their ability to reach decisions by their own self-talk alone. The concept of ‘breadth’ of habitual beliefs may deepen Archer's explanation here.Footnote ²¹ I suggest Hypothesis 1: Autonomous reflexives have a ‘breadth’ in their habitual beliefs that communicative reflexives do not have. Autonomous reflexives have access to a range of automatic judgements that they can make about various situations because their habitual beliefs are numerous and wide in scope. Communicative reflexives have not produced for themselves a numerous and wide array of habitual beliefs as ‘gut feelings’ to be used for making judgements within their own self-talk, so they are observed by Archer to rely more on the judgements of others.

I suggest also Hypothesis 2: that this difference of ‘breadth’ could be caused by the greater ‘contextual discontinuity’ Archer claims the autonomous reflexives have experienced, relative to the communicative reflexives. Exposure to a greater variety in objective contexts (an implication of ‘contextual discontinuity’) is likely to produce more situations in which the repertoire of habitual actions is either inadequate or clashes occur. At level C2, the mind may respond to such situations by automatically predisposing the individual to ‘look outward’ for external guidance as to what to do. Features of the objective situation (‘rules’) can then become internalised as ‘beliefs’. To the extent that such situations are repeated, then such beliefs are automatically invoked: they become habitual beliefs. The mind then has an array of dispositions to judge particular situations in particular ways. I am then suggesting that the instrumentally oriented internal conversation – at level C3 – is able to make use of these automatic dispositions to judge certain situations as an instrument to (potentially more rapid) decision-making. The greater availability of such automatic judgements to those individuals subject to more past variety in situational exposure (greater ‘contextual discontinuity’) may serve to foreshorten their internal conversations so helping to explain why decision-making by such ‘autonomous reflexives’ is more rapid.

Second, what is it that causes individuals who are self-reliant in their decision-making to be meta-reflexives rather than autonomous reflexives? Archer sees meta-reflexives as influenced by contextual incongruity (rather than the autonomous reflexives’ contextual discontinuity) ‘. . .in relation to their dreams’ (Archer, Reference Archer2007: 155). ‘At some point’, Archer argues, meta-reflexives ‘. . .encountered an incongruity between what they would become – their aspirant vocation or the preconditions for it – and their social context, which impeded its realisation’ (2007: 155). However, Archer is less clear on the process within the internal conversation that produces this ‘point’ in meta-reflexives’ lives. I suggest that the way habitual beliefs and the internal conversation interact may be different between persons with one of these two dominant modes.

Consider then Hypothesis 3: the internal conversation of the meta-reflexive is relatively more oriented to treating its habitual beliefs as objects to be transformed into explicit ‘concerns’ in a process of self-examination, and less oriented to using habitual beliefs as automatic judgements.Footnote ²² Thus, it may at least partly be because meta-reflexives actively engage in the ‘integration’ of acquired habitual beliefs into concerns (unlike autonomous and communicative reflexives), that meta-reflexives are relatively more likely to notice such ‘incongruous’ moments in their lives, prompting their characteristic lateral mobility. I also suggest the related Hypothesis 4: the ‘strategic’ orientation to external structures demonstrated by autonomous reflexives is at least in part due to their lack of enquiry into their gut feelings.Footnote ²³ I further suggest Hypothesis 5: Davis's desired state of individual ‘relative autonomy’ or non-oversocialisation may rest on more individuals having a developed capability level C4 (as is characteristic of meta-reflexives). Those with developed C4 are then most likely to become aware of habitual beliefs they have non-consciously acquired through their group participations.

Finally, why are Archer's fractured reflexives ‘passive’ agents?Footnote ²⁴ I offer Hypothesis 6: relative to ‘active’ agents, ‘passive’ agents may be more dependent on habitual beliefs and actions (they have access to capacities E1 and E2 and so may have highly developed capabilities C1 and C2) but lack capacity (E3) for instrumental internal conversation. Enquiry into the basis of these capacities is therefore indicated. Social structures and institutions are likely to play a role in enabling or disabling these capacities.Footnote ²⁵

To maintain inner conflation prevents pursuance of these hypotheses. The ‘breadth’ of a person's habitual beliefs, the relationships between their habitual beliefs and their internal conversation and the relative capacities behind these habit capabilities and internal conversation capabilities are disregarded under inner conflation. Yet, we have suggested that to include habit in this way does not have to undermine the ultimate interiority, subjectivity and causal efficacy of the internal conversation, if, as a causal dualist this is what one wishes to maintain.

Some implications for heterodox economics

Can the somewhat abstract arguments above offer more direct guidance to research into heterodox economic analysis of individual behaviour? We very briefly consider three directions.

First, (post-Keynesian) heterodox economists could enrich their analysis of individual resource usage and acquisition by exploring the links between dominant modes of reflexivity and empirically recognisable socio-economic groupings. Is there a greater prevalence of addictive consumption among groupings containing (the more habitual) fractured reflexives? Are there systematic differences in the use of informal relative to formal information channels between groupings respectively intensive in communicative and autonomous reflexives? Is there a clear tendency to spend to both create and to maintain ‘micro-worlds’ by groupings rich in communicative reflexives and what kinds of expenditures are entailed? Might one observe more aspirational and more ‘self-help’-oriented spending but also relatively high savings ratios from groupings strongly featuring autonomous reflexives? Could groupings intensive in meta-reflexives exhibit a relatively stronger resistance to advertising influence and emulative pressures and show a more ‘principled’ ethical sensitivity (maintaining this stance during recessions more than other groupings)?

Second, at the macro level, one might deepen the heterodox behavioural linkage from uncertainty to Keynes's (once more extremely pertinent) concept of liquidity preference. Consider the following hypothesis: Increased uncertainty (as in the current financial crisis) undermines the basic trust of all reflexive types in the ongoing stability of structures, which reduces the confidence of individuals in their own habitual action repertoires (the mind's level one capacity E1 for habitual action is weakened).Footnote ²⁶ As a result, the mind's level one capability to perform habitual actions (C1) is inhibited under increased uncertainty. So inhibited, individuals (and their associated organisations) find it more important to be able to delay actions. The preference to hold/hoard cash – as a means by which (spending-related) actions can be delayed in a contract-using monetary production-based capitalist economy – correspondingly strengthens. Greater uncertainty may therefore result in structurally induced inhibitions on habitual action performance with the observable systemic counterpart being increased liquidity preference.

Finally, Len Doyal and Ian Gough's (Reference Doyal and Gough1991) pioneering welfare analysis identified individual needs for (physical) ‘health’ and ‘critical autonomy’ (broadly the ability to engage critically with socio-economic structures/institutions). The role they identify of various ‘societal preconditions’ for need satisfaction may be clarified by a more explicit view of the mind. Human need satisfaction may also require a balanced development across all four mind capabilities. No reflexive ‘type’ is necessarily both ‘healthy’ and ‘critically autonomous’ as all types can be unbalanced in their mind capabilities. Furthermore, it may be that some or all existing capitalist socio-economic structures may promote (and even require) unbalanced capability development across individuals (perhaps relative overdevelopment of capability C3?). Investigation into the nature of the four ‘capacities’ that enable mind capabilities to develop is urgent. Such capacities may be, as with E1 above (confidence), and as implied by Doyal and Gough (Reference Doyal and Gough1991), culturally/institutionally formed and developed. Heterodox economists might then identify those structures and institutions supporting such capacities since this may enhance need satisfaction.