The extent of cognitivism

Introduction

The ‘epistemological turn’ of ‘the Descartes–Locke–Kant tradition’, to use Richard Rorty’s terms (2009), has undoubtedly been one of the most influential as well as most criticized turns of modern western philosophy. Essential to this turn was that in it the centre stage of philosophical interest was taken by the question of the cognitive mechanisms by which human beings have knowledge. One of the important contemporary offspring of the turn can be named cognitivism, broadly characterized as the view that the engine of human (individual and collective) action is the intentional, dispositional, or other mental capacities of the brain or the mind.¹

Cognitivism is undoubtedly very deeply rooted in our intellectual culture and its many forms in different areas of culture probably have more proponents than critics. One critic, Fernando Vidal (2009: 6), spoke of ‘the cerebral subject’ which denotes the cognitivist view about the essentiality of the brain or the mind to the constitution of human subject, the subject’s identity and capacity to act. ‘[T]he brain is the only part of the body we need in order to be ourselves’, Vidal described the core idea of the cerebral subject conception. ‘Brainhood’, Vidal continued, is a ‘property or quality of being, rather than simply having, a brain’ in the sense in which personhood might be thought to designate a property or quality of being a person. In Vidal’s terminology, ‘[t]he brainhood ideology’, in turn, denotes the wider intellectual-cultural orientation to conceive the human being as a cerebral subject.

Also Vidal identifies the formative historical roots of the cerebral subject in the aforementioned Descartes–Locke–Kant tradition, in particular in its view of human selfhood with consciousness as its only necessary property. This was characterized by Vidal (2009: 7) as the view of the human being ‘as autonomous agent of choice and initiative, and the corresponding emphasis on interiority at the expense of social bonds and contexts’. In the paper from which these passages stem, Vidal’s main focus is on the role and character of the cerebral subject in a variety of areas from general audience and media discourse to medicine. This can be seen as another episode in the philosophically and intellectual-historically very significant drama played around cognitivism.

In an earlier episode of that drama, the modern classic anthropologist Clifford Geertz articulated the core idea of cognitivism as it appeared in his field thus:

… [v]ariously called ethnoscience, componential analysis, or cognitive anthropology (a terminological wavering which reflects a deeper uncertainty), this school of thought holds that culture is composed of psychological structures by means of which individuals or groups of individuals guide their behavior. (ibid.: 11)

Geertz’s account is interesting because it shows how common it is to equate the causal domain of brain neurology with things like psychological structures and the mind more generally. Some may want to claim that these are distinct domains, but most philosophers at least would insist on naturalism and argue that the reality and phenomena investigated in philosophy and other human and of course natural sciences have a fundamentally (non-supernatural) physical existence or essence. Cognitivism, in turn, can be seen as a form of naturalism saying that the brain is the ultimate ontological seat of knowledge, beliefs, or any other aspect of human-mindedness and intentionality that we may want to employ to explain human action. A case in point is the philosopher John Searle. His analysis of the construction of social reality sets off from the conviction that ‘we need to figure out how social reality fits into our overall ontology’ such that we do not end up ‘postulating different ontological realms’ for physical and social facts (Searle, 1996: 5; Searle, 2010: ix. Similar sentiments abound in philosophical literature; e.g. Elder-Vass, 2012; List and Pettit, 2011; Tuomela, 2002, 2007). For Searle, social reality builds upon the familiar, allegedly well-known ontological basis of beliefs and other representational and non-representational states and processes in the heads of individuals albeit such that these are aligned in ways that produce collective recognition and acceptance of the rules, concepts and the like of the society in question.

Where Searle is unashamed in his naturalism and cognitivism, in philosophy and sociology many others have appeared to reject cognitivism, but retained its functional core, or so this article argues. That is, many non-cognitivists continue to describe human social action and sociality by reference to collectively known or mastered concepts, norms and values, a position that arguably is a cognitivist one, for it views human action in terms of its alleged source in some features of the human ontological make-up, most often the mind or the brain. This way, I shall also attempt to show that once the tap of critique of cognitivism in human sciences has been opened, it cannot easily be shut again before substantial areas have been affected – some of which we may have thought will not be affected by such criticism. As such, the chief aim of the article is to explore the logical extent of the critique of cognitivism. The article argues that such a critique implies a more radical philosophical position regarding the role of human mentality in explanations of human action than is commonly acknowledged. I believe that this view, if true, represents a notable development in the debate about cognitivism and thereby it has significance in the history of modern philosophy and the human sciences more broadly. It should be stressed that the intention of the article is to play the devil’s advocate and to explore the extent of the critique of cognitivism. I believe this is a valuable exercise on its own account. In the final section of the article I shall offer some observations about the virtues of a consistently non-cognitivist account, but the development of a full-blown non-cognitivist alternative is beyond the scope of this article.

Latent cognitivism

Searle’s naturalism can be characterized as the methodological conviction that any respectable view of social reality, and of individual and collective action in society, can only be naturalistic (Sellars, 1922). I shall suggest that, while many other authors often explicitly reject cognitivism, their views about individual and collective action, and social reality more generally, function just like in cognitivist theories.

Before articulating that charge in detail by way of examples, I want to note that there exists a body of literature critical of what these authors term ‘latent cognitivism’ in philosophy and social theory (Pleasants, 1996: 236). In the view of these critics, cognitivism, first of all, is the view that ‘treats belief/knowledge as the wellspring of action and, accordingly, it requires that cognitive operations be understood to underpin, or even determine, sequences of action’, wrote the critics of cognitivism Graham Button and Wes Sharrock (1993: 3). Similarly, Rod Watson and Jeff Coulter (2008: 1) characterized cognitivism as ‘the view that people behave on the basis of various states or processes involved in the “workings of the mind”’ with the associated ‘tendency to reduce social phenomena to postulated individual properties or attributes of mind’. (For other similar views, see Shotter, 2006, 1996; Sharrock and Dennis, 2008; Sharrock, 2004; Pleasants, 1999.)

Latent cognitivism, on the other hand, in the view of Nigel Pleasants, is exemplified, for instance, in Anthony Giddens’s casting of tacit knowledge possessed by knowledgeable social actors as the foundation of practices. ² Pleasants argued that Giddens’s theory ‘merely shifts…from the conscious (discursive) to the non-conscious (tacit) cognitive powers of individuals’ (Pleasants, 1996: 241). Thereby, Pleasants concluded, ‘philosophically [Giddens’s] position is closer to Descartes and Kant than to Wittgenstein’ whom Giddens claims to follow (ibid.: 239). Graham Button and Wes Sharrock’s view of latent cognitivism is very similar although their context of discussion is somewhat different. They criticized ‘the constructionist view-point in the sociology of scientific knowledge’ for arguing ‘that reality is constituted by “cognitive” operations such as interpretative practices in discourse’. This effectively ‘continues a Cartesian tradition because it preserves cognitive operations as the basis of reality-construction’ (Button and Sharrock, 1993: 2, 3). Although not a critic of cognitivism as such, Also Stephen Turner has advanced perceptive philosophical criticism of a wide range of notions such as ‘tradition, tacit knowledge, Weltanschauung, paradigm, ideology, framework and presuppositions’ as ‘epistemically elusive’ and functioning in a cognitivist fashion as somehow disposing ‘thought or action in a certain way’ (Turner, 1994: 2, 43; see also Turner, 2010).

The key observation that we can see to be made by all these thinkers is that, in (latent) cognitivism, notions such as knowledge of criteria, as well as those listed by Turner, effectively function as the mental source of individual and collective action. The details may vary, but fundamentally (latent) cognitivism explains social reality by appeal to collectively held beliefs, mastered concepts, internalized norms, values and the like which act as the mental engine, the source, of human (collective) action. This idea forms the basis of my critique that I articulate with examples below. I would like to note again that these aforementioned critics of cognitivism also quite happily pass from the ostensibly mental ontological and causal ‘cognitive operations’ and ‘workings of the mind’ to something ontologically and causally perhaps more vague, such as a world-view, making no deciding difference between these domains. I do not wish to challenge this equation of domains. On the contrary, I believe it is essential to note that in the talk of such things as a world-view as the engine of human action the former effectively, albeit perhaps implicitly, works cognitivistically as the mental source of human action. In a yet more general sense, the critique raised below concerns accounts of human action by reference to its source often understood as something mental in the brain or the mind (a belief, representation or concept, for example), but may include also instincts, which one may perhaps not typically designate as mental processes as such. Hence, cognitivism and naturalism, as understood in this article, are part of the ontologizing tendency to explain human action by appeal to its alleged source in the human physical make-up, i.e. ontology, and some allegedly isolable and nameable entities therein such as instincts, beliefs and dispositions but also possessed concepts, internalized norms and values and the like.

Aligning the sources

Let us now turn to examples of latent cognitivism in philosophical and social-theoretical literature. Noteworthy is that the cases discussed below are Wittgensteinian thinkers. Arguably Wittgenstein is one of the most original critics of cognitivism in the history of modern philosophy. However, as noted by Paul Johnston, ‘Wittgenstein is widely regarded as one of the greatest philosophers of this century [the 20th] and yet the central thrust of his work is emphatically rejected by the current philosophical community’. In Johnston’s view, there has been in particular a failure to appreciate that Wittgenstein proposed ‘a radical new approach to the whole topic of the Inner’ (Johnston, 1993: ix, x). A similar view was recently advanced by Avner Baz, writing that there is a ‘deep hostility and dismissive attitude within wide circles of mainstream analytic philosophy toward Wittgenstein’s work, even if not toward some of his isolated “results”’ (Baz, 2012: xi).

The first Wittgensteinian whose (latent) cognitivism I shall highlight is David Bloor. Bloor and the so-called Strong Programme have of course been very influential in sociology, history and philosophy of science encouraging anything but cognitivist approaches, but I argue that upon a critical inspection of the programme’s philosophical methodology a different story begins to emerge. Perhaps, then, for some the reading of Bloor I provide below might seem unfair and selective, but for the same reason one could also see the reading as highlighting an internal tension within the programme’s views.

Bloor lays out his view by discussing Wittgenstein’s case of a pupil being taught to work out a series of numbers according to a formation rule (Wittgenstein, 1958: § 143). The problem, as alleged by Bloor, is twofold. First, a finite amount of examples cannot possibly convey a general rule that reaches to infinitely many instances, Bloor thought. The second problem is that ‘we cannot possibly conjure up in our mind, or bring into consciousness, all of the cases to which the rule is “meant” to apply’, as Bloor put it (1997: 13). A dummy answer to this challenge construed by Bloor is the meaning-deterministic view that there is an idea in the mind conveyed in teaching that somehow contains all the infinite number of instances in which it applies. Bloor argues here that ‘[w]hen we are confronted with a finite set of example we do not extract from them any general idea, rather, we instinctively pass on to what strikes us as the next step or the next case’ (ibid.: 13–14).

As we shall see shortly, this idea of instinctive passing on is not the whole extent of Bloor’s account, but let us pause for a moment to reflect on Bloor’s offering here. Whatever strengths or weaknesses this idea of instinctive response may have, and whether it is a plausible Wittgenstein interpretation, I think it should strike us immediately that the view is cognitivist in the broader ontological sense of the word: it posits something in the human physical make-up, an instinct, by the guidance of which we allegedly pass on from case to case.

Here it must be noted that a prominent critic of cognitivism, Jeff Coulter, also subscribes to the idea of an instinctive reaction, or at least something very similar to it. Coulter wrote:

For Wittgenstein, our basic concepts of ‘cause’ and ‘effect’ are grounded in our immediate responses, our conduct toward states of affairs which puzzle us, trouble us, etc. There is no inferring, guessing, concluding, and the like. In other words, no ratiocination is involved. (2005: 82–3)

Now, again regardless of whether the idea of instinctive response is plausible or not, is there not at the very least a tension in Coulter recommending an approach of not trying to describe, as he puts it (2005: 91), ‘any sort of underlying “mental apparatus” at all, but of explicating some rules or procedures with which their conduct is in accord’, and that he at the same time put forward a view about instincts as underlying so much of our dealings with the world? For to speak of instincts would seem to be precisely about describing an ‘underlying mental apparatus’ where ‘mental’ is understood in the broad sense in which an instinct designates a nameable item in the human physical-mental apparatus by which we react to external stimulus. Of course, Coulter’s intention may be to limit cognitivism to concern intentional and representational states, interpretations and the like and then dispute that form of cognitivism. Yet, the question arises: why do we deny a role in our approach to these aspects of the mental apparatus but not to instincts? Surely all forms of mental activity must equally be subject to criticism if we want to be consistent in our rejection of cognitivism.

Let us now return to Bloor. Bloor’s exposition continues with his noting that the idea of instinctive passing on can say ‘nothing about whether the move to the next putative instance of a rule or concept is right or wrong’. Correctness requires a standard of comparison which is social in nature and based on consensus. As Bloor argued: ‘following a rule counts as a “right” step…if it is aligned with the steps everyone else, or nearly everyone else, takes’. That is to say, ‘[t]he normative properties of rules, then, do not derive from the instinctive sources of individual activity, but from the alignment of these different sources in the majority of cases’, Bloor concludes (1997: 15–16).

Note again carefully the cognitivistic character of the view proposed here by Bloor: the source of collective action resides in the minds of social actors, albeit such that these sources are aligned. That is to say, cohesive social action flows from all members, or nearly every member, of a community being in mental possession of the same rules, concepts, instincts and what-not, or so Bloor seems to be saying.

The cognitivistic character of the programme’s methodology becomes even more apparent if we look briefly at a thinker whose ideas Bloor designated as ‘[t]he most developed account [of social institutions], and the one that I have found most useful’, namely, ‘“Barry Barnes”: “Social Life as Bootstrapped Induction”’ (Bloor, 1996: note 3). In the said paper of Barnes, the programme’s view is discussed by the example of classification: how does the general term ‘leaf’ come to stand for all leaves? Barnes too uses a dummy position called ‘the pattern-recognition stereotype’ according to which as we observe various particular instances of leaves we acquire ‘a pattern for leaf’ with the aid of which we are able to conduct ‘routinized, habituated, pattern-matching’ procedures (Barnes, 1983: 525). Here Barnes speaks of the human being making a classification as a ‘designation device’, a ‘reification’ representing ‘the property of an individual agent, his means of routinely attaching a specific term to instances’ (ibid.: 527). Now, just like Bloor, Barnes rejects the pattern-recognition stereotype on the grounds that no one designation device alone can account for correct usage. Rather, again as Bloor, Barnes argues that the individual designation devices must be aligned in their judgements such that individual devices label the same instances as of a particular kind. Again, it should be obvious that the account given here is cognitivist: a human community is, perhaps not solely but essentially, a community of designation devices in joint mental possession of rules, concepts, instincts and the like that drive their activities.

Knowing the criteria

Let us now turn to our second example of (latent) cognitivism in Wittgensteinian philosophy, Peter Hacker’s views on concepts and criteria. Hacker, if anyone, might be thought of as a non-cognitivist, but it will be instructive to inspect more closely whether there might be internal tensions in Hacker’s views and learn from this with regard to detecting certain dangers facing Wittgensteinians of lapsing into cognitivism.

To be sure, Hacker does explicitly deny cognitivism saying, for instance, that

…[w]e are inclined to think of counting, calculating, inferring or constructing proofs as primarily cerebral activities. The essential business seems to be carried on in the medium of the mind. This is misconceived. (ibid.: 118)

However, upon a closer inspection, the way Hacker uses the notion of criteria appears (latently) cognitivist, for it functions in a cognitivist manner as the mental source of human action. Crucial is Hacker’s characterization of concept-possession as knowledge of the criteria for its application. Hacker writes, for example, that when ‘we are engaged in concept-formation, [we are] introducing criteria for the application of concepts’ (Baker and Hacker, 2009: 51). The criteria are collectively known in that ‘there are public criteria in behaviour for whether an act accords with a rule or not’, Hacker says (ibid.: 131). According to Hacker, to understand a concept is to be able to use it: ‘the speaker’s correct applications…exhibit the ability in which understanding consists’ (ibid.: 68). The criteria define the concept: ‘[l]aying down such criteria of correctness is necessary to delineate the scope and content of any technique’, he says (ibid.: 144). Therefore, on Hacker’s view, understanding comes out as knowledge of the criteria for correct concept application which is exhibited in use.

While Hacker would surely protest, the notions of knowledge of criteria and display of abilities that in Hacker’s discussion emerge as central notions for conceptualizing human sociality, unwittingly express the aforementioned ontologizing tendency in that human action is viewed in terms of competent social beings’ possession of knowledge of criteria, which is to view human action in terms of its alleged source in some representational or non-representational aspects of the human physical make-up. In other words, at the heart of Hacker’s view of concepts is the subject’s knowledge or understanding of criteria which is, I think, naturally cast as that subject’s mental possessions, possessions that underlie the exhibition of this knowledge and understanding as abilities – or how else are we to understand the grasp of criteria if not as something that the subject has and on the basis of which the subject acts? ³

To see this better, consider whether Hacker’s views could be adopted by an explicitly cognitivistically orientated thinker to advance a cognitivist view. It seems they easily can, for it is possible that upon hearing Hacker talk about ‘introducing criteria for the application of concepts’ and about such criteria as delineating ‘the scope and content of any technique’ that we display in competent behaviour, the cognitivist might be inspired to set upon finding the ‘neurophysiological realization’ or ‘underlying structures that make the behaviour possible’, as Searle has described his philosophical project (Searle, 1996: 129; see Elder-Vass, 2012, 2010, for a similar causal ontological approach to social ontology). Thus, it seems there is nothing to stop the cognitivist from giving a cognitivistic spin to Hacker’s views on criteria and associated notions. Also, the cognitivist might argue further that Hacker’s view functions as an account of human sociality precisely because of its latent cognitivism: if the notion of criteria is to do any work in explaining human action, it is because the criteria are thought of as known by knowledgeable actors and as such are the engine of their activities. This seems justified, for characterizing understanding as displayed in the ability to act and respond competently in accordance with criteria would seem to work insofar as the subject actually has a grasp of the criteria. While it may seem that we are here unfairly imposing a cognitivist project upon Hacker, one that he has explicitly denied, the considerations advanced here do seem to me to expose an internal tension in Hacker’s views between explicit rejection of cognitivism and its simultaneous compatibility with Hacker’s views.

A consistently non-cognitivist conception of human action and human sociality, on the other hand, would involve tying them, not inwards to the subject’s grasp of criteria (i.e. what the subject is alleged to hold in mental possession), but outwards to the subject’s accomplishment in the process of human interaction. Something like this is echoed in Lars Hertzberg (2010: 123) arguing against Hacker that ‘to relate to someone as a speaker is not a matter of noting that she lives up to some standard or conforms to some pattern’ but, as I would put it, to engage in ongoing interaction with her. The claim here is not so much that Hacker might actually have such a picture of a speaker’s relation to another, but that the terms Hacker chooses to express his view suggest such a picture. In other words, in the non-cognitivist perspective the focus is on the process of interaction, but in the (latently) cognitivist view the focus ends up upon the deeds of the individual as evidence of her concept possession – the ability exhibited in the deeds is said to define the concept by exhibiting its use. More properly, however, the criteria, one might say, describe a slice of human interaction, not the alleged mental possessions of a competent subject which the subject after all might or might not have.

The non-cognitivist perspective is, in fact, a view that Hacker is well aware of and that he espouses in another paper, and hence my suspicion that there might be an internal tension in Hacker’s views. In the said paper Hacker writes that to obtain an idea of what the concept of belief is we need ‘to examine the needs which the concept of belief satisfies, what purposes it fulfils’ (Hacker, 2004: 219). This he does very well by looking at a number of uses of ‘belief’ and cognates, for instance, to convey a degree of uncertainty about information provided as in ‘I believe the train will arrive at 5:15 – but you’d better double-check’. In such a description there is no issue with what mental possessions the person making the utterance has – that is, what criteria the person grasps – but that the utterance has particular effects in the interaction of that person with another. Yet, to turn the frame around and consider action as evidence of possession of knowledge of criteria, is to espouse (latent) cognitivism. The fundamental issue in such a turning of the frame, just as in Bloor, is that while language is seen as situated and used in connection to human activities, the use is in turn said to build upon collectively held knowledge of criteria. This latter move is the latently cognitivist one that once again draws back in cognitive operations as the foundation of human action.

One might note that here we have, as it were, a form of collectivist cognitivism in which collective action is seen as driven fundamentally by collectively held knowledge of criteria. This can be seen as a philosophical formulation of the immensely popular and widespread view that collectives, cultures and ages are fundamentally defined by their ideas. A particular application of this view is the ideological characterization of the modern human relationship to nature at the root of which is the Christian view of nature, as Lynn Townsend White’s classic paper argued (Townsend White Jr, 1967; see also Dunlap, 2006; Dickson, 2000; Shove, 2010). These views have a notable classic proponent in Max Weber (1930) and his argument from the Protestant ethic and the spirit of capitalism as having played a significant role in the different courses of historical development taken by the West as compared with the East in the early-modern times. Again, the structural functionalism of the seminal American sociologist Talcott Parsons (1954, 1949) made significant use of the idea of collectively shared ‘normative orientations’ as underpinning cohesive collective action. These observations point to the wider significance of the debate about cognitivism that I shall return to at the end of the article.

These mentions also serve to illustrate the deep embeddedness of (latent) cognitivism in our intellectual culture. They illustrate what Fernando Vidal named the brainhood ideology. According to this ideology, it is the brain and its capacities that endow individuals with their identity as well as their capacity to act as persons. In collectivist cognitivism this view mutates into one in which members of a collective and their activities are characterized by the ideas that drive them. It is things like knowledge of criteria understood as mental possessions of individuals that in the views we have reviewed effectively account for the identity of a person as a member of a community, for example. On this account I claim that the aforementioned views are essentially functionally cognitivist.

Mental predicates

The main charge I raised above against a number of Wittgensteinian (or Wittgenstein-inspired) philosophers was that they are latently cognitivist in the sense that in their writings notions such as criteria function just like in an openly cognitivist view, namely, as the mental source of human action. Another way to articulate the charge from latent cognitivism could be to use the words of John Searle. Searle defends his naturalistic cognitivism saying that ‘[b]ecause we do not know how these [shared mental] structures function at a neurophysiological level, we are forced to describe them at a much higher level’ (Searle, 1996: 129). In the argument of this article, collectively possessed knowledge of criteria is such a higher-level description that latently cognitivist views build upon and wherein lies the essence of their latent cognitivism. It is true that, for example, Bloor and Hacker do not speak about mental ontology and neural processes as such. They speak of knowledge of criteria. This is, however, as we have seen, functionally cognitivist in that it is these collective mental possessions (consensus in Bloor) that define groups and that drive and thus ultimately explain their activities. We must, I think, conclude that until such latently cognitivist aspects have been identified and rejected, the philosophical critique of cognitivism cannot be seen to have reached its full extent. This illustrates the meaning of the claim made in the introduction that once the tap of critique of cognitivism has been opened, it cannot easily be shut again before the critique has spread over a large area which the critique could previously not be thought to concern.

At this point one might despair that certainly one must be able to talk about knowledge, beliefs and the like without immediately being branded a cognitivist. This indeed takes us right to the core issue with cognitivism as Watson and Coulter (2008: 12) observed: ‘a major battleground in social science concerns the treatment of the “mental” predicates’ by which they meant mental predicates such as thinking, understanding and motive. They noted, quite correctly, that ‘[w]hen addressing issues concerning mental predicates we are…facing a whole intellectual culture that reproduces these puzzles, dilemmas, fudges, evasions and category-mistakes’ (ibid.: 9) and illustrate this with a wealth of examples (the names implicated here include Émile Durkheim, Erving Goffman, Herbert Blumer, George Herbert Mead, Norbert Elias and others).

One of the themes Watson and Coulter highlight concerns the notion of Einzel-Individuum [sole individual] and the ‘“internal”–“external”’ opposition ‘in terms of which’, Watson and Coulter said, ‘classical sociological orthodoxies are still cast’ (Watson and Coulter, 2008: 10). By contrast, in a truly non-cognitivist position, Watson and Coulter argued, mental predicates should not be treated as referring to the mental possessions of an ‘actor/agent…as the bearer or container of the referents of the “mental” vocabulary’. Instead, we should ‘look at motive avowal and ascription as naturally occasioned, naturally organized and naturally situated practices’ and ‘treat these practices as normative’ observing that ‘[a]vowal and ascription procedures for mental predicates are hedged around by norms’ (ibid.: 12).

Is there not, however, a real danger of lapsing once again back into (latent) cognitivism here? For we should be careful not to understand norms, which according to Watson and Coulter hedge the use of mental predicates, in the same manner as in Hacker and Bloor above, namely, by casting them in the latent cognitivist fashion as knowledge possessed by competent members of a community. Watson and Coulter (2008: 13) pointed out that mental predicates are ‘person-level predicates and not names for “processes of/in the mind”’ and obviously a lot turns on what this person-level is, that is, whether it is cashed out in a non-cognitivist or a latent cognitivist fashion. Settling the issue would involve a lengthier discussion and so I will not pursue these doubts here any further, but note that this may, just may, be another place where the critique of latent cognitivism spreads further out than we have thought.

What is non-cognitivism?

The main argument of this article has been that many non-cognitivist thinkers have unwittingly introduced elements of cognitivism back into their positions by the latently cognitivist view that human sociality is fundamentally grounded in collectively held instincts and knowledge of criteria and norms. The full development of a non-cognitivist position is beyond the scope of this article, but I close with a discussion of the implications of this argument regarding the character of non-cognitivism as compared with cognitivism that hopefully also sheds further light on what non-cognitivism could be and why it matters.

Above, in discussing Hacker’s views on criteria, I argued that the way Hacker frames concept possession as knowledge of criteria displayed in competent conduct suggests a latently cognitivist picture of human knowledge as the essence of human sociality. As a first approximation, I argued that a consistently non-cognitivist view would take the ongoing and developing processes of human interaction as the ‘stuff’ or phenomenon upon which our attention is focused. This is an idea recognizable in a range of literature under such labels as practice theory (Schatzki et al., 2001; Shove et al., 2012), networks (Latour, 2005) and the Wittgensteinian terms ‘language-game’ and ‘form of life’. In Wittgenstein, this point of view was expressed in passages like

…[h]ow could human behaviour be described? Surely only by sketching the actions of a variety of humans, as they are all mixed up together. What determines our judgment, our concepts and reactions, is not what one man is doing now, an individual action, but the whole hurly-burly of human actions, the background against which we see any action. (Wittgenstein, 1981: § 567)

Non-cognitivism is characterized by the attempt to see human life as a set of ongoing, dynamic exchanges (hence the word ‘process’ above’) and to avoid a static view of human sociality, say of a culture or a culture’s relationship to nature (more of which below), as an assemblage of particularly disposed and knowledgeable actors. The difficulty in maintaining the dynamic perspective resides surely in the constant invitation present in our intellectual culture to think statically in terms of identifiable entities and their properties, ⁵ say, in terms of individuals with beliefs, intentions, abilities, capacities and the like.

Wittgenstein’s non-cognitivist approach makes itself felt in that, as Hans-Johann Glock (2010) has argued, in Wittgenstein the question of what a concept is gets ‘operationalized’ and discussed as the question about what it is to possess a concept. This involves looking at the use of concepts as well as explanations of concept meanings. I would agree with this except that in the point of view I advocate, the correctness of someone’s performance in these respects marks out, not that this person is (or is not) in possession of a concept, but that a move (Wittgenstein, 1958: § 22, § 49) has been made within the system of a language-game. ⁶ Conceptualizing speech and conduct as moves made within some system of language-game serves to highlight the systemic and developing character of human interaction and that what happens in its course has significance precisely because the happening affects the course in some way.

It is important to note that the notions of making a move and the developing character of human interaction constitute a categorically very different framework for looking at human interaction as compared with the (latently) cognitivist ‘mastery as knowledge of criteria’ picture. In the latter view, human action and interaction is looked at – or unwittingly comes to be looked at given the kinds of notions used to formulate the position – cognitivistically in terms of what actors know, this knowledge appearing as the source, the driver, of their activities. In the non-cognitivist view, from the start the focus is on the organization and development of systems of human interaction and, importantly, it is explicitly acknowledged that there will in fact be differences in the roles and positions of different actors and thereby in their interests and knowledge of the system in which they operate – a point that seems to me highly significant and to which I shall return below. Also, a move made in a language-game appears in the first line as a movement in the developmental course of interaction – in the hurly-burly of human actions, as Wittgenstein put it above – not as a knowledgeable subject’s exercise of intentional capabilities. In this sense, for example, Wittgenstein would not describe a teaching situation by the mental processes that the pupil allegedly goes through — ‘the reading connection has been set up’ (Wittgenstein, 1958: § 158) – but he laid it down as the familiar developing process of giving examples, trying to repeat them, and so on. Evident again here is Wittgenstein’s non-cognitivism as a perspective focused on the ongoing, dynamic character of human interaction characterized most essentially by the attempt to achieve something in its course.

A further hallmark of non-cognitivism, as I see it, is that to focus on the processes of human action is also to look at it as purposeful in a particular way. By contrast, as a rule, cognitivist approaches seem to ignore purposes and produce questionable statements like ‘[w]hen I say, for example, that I am able to speak English, I am talking about a causal capacity of my brain’ (Searle, 1996: 129). This is questionable because the statement fails to say anything about the significance or purpose for which one might actually say ‘I’m able to speak English’. Certainly it is very rarely, if ever, that someone utters the aforementioned sentence for the purpose of highlighting a causal capacity. One can easily imagine ordinary circumstances, and an ordinary purpose, for which one might utter this sentence – say, I tell this to a group of English-speaking tourists who have been struggling to ask me for directions in my native language. And ‘by that’, as Wittgenstein observed, the situation ‘loses everything that is philosophically astonishing’ about it (Wittgenstein, 1975: § 622).

The general contrast highlighted here between cognitivism and non-cognitivism, one might say, has to do with contrasting the non-cognitivist interest in the organization and development of the (recurring) course of human interaction – a practice, language-game, or form of life as one might also say – with the traditional philosophical interest in (positing) internal states or processes within the subject and then ‘analysing’ in what sense and under what circumstances those states and processes are knowledge or beliefs and as such true, justified and the like. This was recently noted by Avner Baz in his critique of a number of philosophical analyses of ‘to know’ and its cognates. Baz identified the fundamental conviction in philosophical epistemology that ‘to give an account of x – knowledge, truth, or what have you’ – is a matter of establishing when, and under what conditions, does y – ‘a sentence, a statement, a belief, a proposition, or what have you’ – stand in a right type of relationship to what makes it true or what makes it knowledge (Baz, 2012: 56–7). Baz suggests, however, that what typically is at stake in situations where ‘to know’ and cognates are used, is more properly said to be the ‘human needs, interests, and concerns that give those uses their specific point’ (ibid.: 45), and not whether a particular relationship obtains between different kinds of objects (mental, metaphysical, etc.). That is to say, instead of by way of a cognitive theory of knowledge, human interaction is looked at as a developing process aiming at something and avowals of knowledge, for example, as facilitating or bringing about obstacles or changes of direction. This is a challenge going straight to the core of ‘two and a half millennia of Western philosophy and a century or so of analytic philosophy’ and the ‘numerous impasses and puzzlements but no truly satisfying analysis or theory of the envisioned sort’ it has produced, Baz argued (ibid.: 47).

A similar shift in focus is suggested in the critique of the cerebral subject conception by Alain Ehrenberg, in particular of what he calls ‘the “social” brain’ conception in neuropsychology, which Ehrenberg characterized as ‘the idea that not only mental disorders, but even social “behaviours” can essentially be explained in terms of cerebral activity’ (Ehrenberg, 2010: 117). Ehrenberg gives us a hint as to what in his view an alternative non-cognitivist approach would look like as he writes that (in sociology) ‘[w]e are not interested in the actors in the drama as individuals with all sorts of feelings’ but in their ‘role within a social framework’ (ibid.: 128). And indeed, one can see here something parallel to the critique of latent cognitivism raised above in that in sociological theory and philosophy many have indeed been interested in individuals ‘with all sorts of knowledge’ about the social, and furthermore cast shared knowledge (of criteria, i.e. consensus) as underlying social institutions. Ehrenberg has recommended an alternative in which we should be interested in ‘social frameworks’ within which individuals operate and not in whatever individuals happen to know or believe about these frameworks. Echoing a similar, if you like, systemic focus, in the passage quoted at the beginning of this article, Vidal wrote that the cerebral subject conception emphasizes ‘interiority at the expense of social bonds and contexts’ (Vidal, 2009: 7). This is not, it should be stressed, to deny mental processes and to espouse some sort of supernaturalism or ineffability (whatever that might in fact mean), but to contextualize ascriptions and avowals in a particular way. The point can be seen to parallel the view expressed by Ehrenberg (2010: 126) in the context of neuropsychology: ‘[s]ympathy and empathy are not universal affective-cognitive mechanisms’ as looking at these emotions in a cognitivist manner would have it. Instead, they could be considered, for instance, as ‘social features of an egalitarian individualistic order’, Ehrenberg writes. Studying such an order and what empathy and sympathy mean there, gives us an essentially better, deeper and more complex, view of the role of emotions in human life, and the same is true of mental predicates more generally. Perhaps in a summary fashion one could say that, with words like Ehrenberg’s ‘framework’, Vidal’s ‘context’ and Watson and Coulter’s ‘practices’, non-cognitivist thinkers have attempted to prioritize and establish a view of the larger surroundings in which talk of mental predicates, for instance, occurs.

This takes us to the final point against (latently) cognitivist approaches to human action. In its emphasis on collectively held knowledge of criteria as the engine of collective action, latent cognitivism unduly intellectually homogenizes social actors and does not thereby give us a realistic view of the complexities of human life. The argument can be made that social institutions, and human life more generally, are so complex as well as varied, and involve individuals acting in different positions with differing perspectives, interests and knowledge, that it is an gross oversimplification to attempt to explain all that by reference to shared knowledge.

A topical example, given the ongoing concern with human-induced environmental change, is the human perception of nature and the human relationship to nature more generally being more complex than can be accounted for cognitivistically. In such a context, Elizabeth Shove has recently challenged the so-called ABC framework ‘in which “A” stands for attitude, “B” for behaviour, and “C” for choice’ as a tool for conceptualizing and addressing environmental problems. Employing such a framework means that ‘responsibility for responding to climate change is thought to lie with individuals whose behavioural choices will make the difference’, Shove argues (Shove, 2010: 1274). One key problem with this framework is to capture the complexity of attitudes and behaviour, for, as Shove puts it, ‘there is no obvious limit to the number of possible determinants and no method of establishing their history, their dynamic qualities, their interdependence or their precise role in promoting or preventing different behaviours’ (ibid.: 1275; see also Butler, 2010; Hards, 2011). The issue is highly significant, for arguably the human causing of environmental problems cannot be explained cognitivistically by the actions of actors with particular environmentally questionable attitudes and values, but by attending to the environmental effects of the interaction in division of labour of differently placed, often globally dispersed, actors with differing roles, position, interests and knowledge (see Dickson, 2000; Bellamy Foster, 2002; Karlsson, 2012; Arponen, 2013). The contrast emerging here between viewing the human relationship to nature as a matter of attitudes (values, concepts) and viewing it as mediated by the organization of the interaction of differently placed actors mirrors the contrast between cognitivist inward interest in knowledge, values and norms as the engine of human action and the non-cognitivist outward interest in purposes, goals and products of organized human interaction.

At least since Peter Winch’s Wittgenstein reading in his seminal The Idea of a Social Science (Winch, 1980), Wittgensteinian philosophy has, justifiably or not, been charged with relativism. This charge is motivated by the Wittgensteinian idea that human sociality builds upon collectively known and self-justifying (self-referential, performative) criteria, concepts, values and norms which can be radically incommensurable with other conceptual systems. Non-cognitivism can, I believe, begin to challenge this paradigmatic idea of human sociality as founded upon a cognitive agreement in criteria. ⁷ Towards this end, alongside the example of the human relationship to nature discussed above, outside philosophy and sociological theory, for example, Philip Converse’s (2006[1964]; Campbell et al., 1960) classic works in political theory exposed significant differences in people’s interests as well as gaps in our knowledge of political systems within which we nonetheless collectively perform. Again, in the study of national cultures, for instance, Brendan McSweeney (2002: 112; see also McSweeney, 2009; Uzzell, 2010) has questioned the existence of ‘systematically causal national cultures’ saying instead that ‘we need to know more about the richness and diversity of national practices and institutions – rather than merely assuming their “uniformity” and that they have an already known national cultural cause’ (see also Kottak, 1999; Kottak and Kozaitis, 1999). Similarly, Max Weber wrote that ‘[s]uch notions as “the will of the people”, the true will of the people, ceased to exist for me years ago; they are fictions’ (quoted in Turner, 2010: 119; emphasis removed). In sociological theory, the idea of an epistemic agreement as the foundation of human sociality has been referred to by Margaret Archer (1996) as ‘the myth of cultural integration’.

These examples invite us to note that in human sociality, and in some of its particular expressions such as national cultures and the will of the people – but also in understanding and knowledge more generally – there is always a lot more going on than an avowal or attribution of mental possessions to an individual or a group. In the above examples, phenomena such as the human relationship to nature or a national culture, are looked at non-cognitivistically as realized or operationalized as ‘moves’ (utterances and actions) made within the context of ongoing and recurring processes of human interaction. A national culture, for instance, can be a political manoeuvre but also a pattern of action and speech rooted in past and present events and annual celebrations, and it can represent sectional interests, be expressive of a power relation, inequality and the like. Non-cognitivism, with its focus on ongoing, dynamic processes of human interaction as the manner in which a national culture or the human relationship to nature is practically signified or not – namely, as action – is essentially better placed to explore this variety and complexity than (latent) cognitivism.