Introduction

Why do people stick with their unhealthy habits despite adverse consequences? This is a pressing question for both public health research and policy-makers. For example, the overweight and obesity prevalence has been steadily growing in all Western societies (Ng et al. 2014). Smoking continues to be a major public health problem even though its health risks are widely recognised (Reitsma et al. 2017) and many behaviours that are acknowledged being essential for healthy lifestyles have not been universally adopted, such as getting enough exercise or eating sufficient amounts of vegetables (Spring et al. 2012). Since risky behaviours are more prevalent in lower socioeconomic groups, understanding why unhealthy behaviours are so resistant to change is vital to tackle inequalities in health. In this article, we argue that there is a theoretical tradition which has been unexplored in this context even though it is well suited for examining the core questions of health behaviour research. This tradition is pragmatism and its conception of habits, which offers a dynamic and action-oriented understanding of the mechanisms that “recruit” individuals to risky health-related behaviours.

Health-related behaviour is often understood as an issue having to do with the individual and guided by motivations, intentions, self-efficacy and expectations, as it is the case with influential and widely used planned behaviour theories (Ajzen 1985) and the health belief model (Strecher and Rosenstock 1997). In this line of thinking, individuals behave the way they do because their intentions, knowledge, beliefs or motives lead them to do so (Cohn 2014; Nutbeam and Harris 2004). The individualised approach is especially visible in many psychological theories of behaviour change and in interventions and programmes designed on the basis of these theories (Baum 2008; Blue et al. 2016). Behavioural interventions have become increasingly important in public health promotion despite weak evidence for their overall effectiveness in generating long-lasting changes in behaviour and their potential to reduce inequalities in health (Baum and Fisher 2014; Jepson et al. 2010).

Research on social determinants of health often takes a critical stance towards psychological theories and recognises social structures as key contributors to health and health-related behaviours (Mackenbach 2012; Marmot and Wilkinson 2006). In health sociology, concepts and measures related to power, cultural norms, social circumstances, societal hierarchies, and material resources, for instance, are used to refer to structural constraints and modifiers of individual action and related outcomes. A large body of research has shown that education, occupational status, financial resources, living area, gender and ethnicity all affect ill health and life expectancy and the ways in which individuals act upon their health. The better off people are, the more likely they are to lead healthy lives and adopt healthy lifestyles (Marmot et al. 2010; Pampel et al. 2010).

While social structure undoubtedly constrains people’s behaviour, people can also exert agency, as they are able to consider different options and to act in discordance with their structural predispositions and social circumstances (Mollborn et al. 2021). The key question in sociological theory is, thus, how individual behaviour can be simultaneously understood as shaped by social structures and as governed by individual choices. It is not enough to state that both social structures and individual intentions are important in explaining behavioural outcomes. One also needs to understand how and why social structures enable or generate particular kinds of behaviour within the context of people's everyday lives. Sociological theorisation on health inevitably falls short if it fails to confront this issue, thus leading to an insufficient understanding of factors that shape health-related behaviours (Williams 2003).

In this article, we first take a look at sociological theories of health-related behaviour, to which the concepts of lifestyle and, more recently, social practices have been central. Then we move on to discuss the pragmatist concept of habit. The concept of habit has often been used in research on health-related behaviours and behavioural change, and it has proved to be useful in explaining continuities in behaviour (Gardner 2015; Lindbladh and Lyttkens 2002). We argue that previous research has not taken into account the pragmatist understanding of the concept as an important contribution to theorisations of health lifestyles and practices. Pragmatism’s dynamic and action-oriented understanding of habits helps in conceptualizing how practices are formed in interaction with material and social conditions and what the mechanisms are by which practices recruit individuals.

In pragmatism, habits are understood in terms of problem-solving; they are active and creative solutions to practicalities of everyday life and responsive to change, not mere blind routines. We, therefore, focus on the creative and active nature of habit formation, which can be understood as mechanisms by which behavioural patterns emerge. The pragmatist approach not only opens new perspectives in health research but can also give new tools for preventing non-communicable diseases and reducing inequalities in health. Next, we discuss theories of lifestyles and social practices and go on to show how the pragmatist theory of habits anticipated many of these insights (historically speaking) but also developed its own framework for analysing the inherent habituality of action.

The interplay between structure and agency: lifestyles and social practices

Attempts to bridge the gap between social structures and individual action in health sociology often draw from a loose tradition of practice theories. They are all based on the attempt to overcome methodological individualism without leaning too much towards methodological holism (Maller 2015). This means that practice theories try to take into account both individual action (methodological individualism) and the role social structures play in explaining action (methodological holism). From the perspective of health sociology, the fundamental question is how to understand the interplay between individual agency and structural factors in health-related matters, such as smoking, drinking or food consumption. In this respect, two concepts have been central: lifestyles and social practices.

Biomedical or social epidemiological approaches, which dominate health inequality research, typically frame lifestyle as a set of individual, volitional behaviours (Korp 2008). Lifestyle is thus a sum of individual health-related behaviours, such as ways of consuming alcohol or dietary habits. In sociological literature, lifestyle is seen as a collective attribute: lifestyles are shared understandings and ways of operating in the world that have been generated in similar social circumstances (Frohlich et al. 2001). They develop over the life-course (Lawrence et al. 2017; Banwell et al. 2010) and are shaped by social and material conditions (Cockerham 2005). As such, lifestyles are not merely outcomes of choices or personal motives and preferences, but they reflect an individual's position in a wider social structure and are fundamentally shaped by those structures.

Cockerham (2009, p. 159) defines health lifestyles as “collective patterns of health-related behaviour based on choices from options available to people according to their life chances”. In his Health Lifestyle Theory, Cockerham draws from Max Weber’s concept of lifestyles, in which lifestyle-related choices are seen as voluntary but constrained and enabled by life chances that are essentially structural: similar life chances tend to generate similar patterns of voluntary action, thus generating patterns of behaviour (Cockerham 2009). Cockerham (2013) considers life chances as consisting of a variety of structural determinants, such as class circumstances, age and gender, which collectively influence agency and choices. The interaction between choices and chances constitutes dispositions to act and resulting lifestyles may have varying effects on health. Health-related behaviour is shown to be clustered within individuals and by socioeconomic status (De Vries et al. 2008; Portinga 2007), yet health lifestyles are rarely uniformly health-promoting or health-compromising, and there is a considerable amount of variation in health behaviour between individuals with similar socioeconomic characteristics (Mollborn and Lawrence 2018; Pronk et al. 2004).

Cockerham’s approach, like many other approaches to health-related behaviours (Williams 1995; Frohlich et al. 2001; Carpiano 2006; Gatrell et al. 2004; Korp 2008), draws on Pierre Bourdieu’s concept of habitus. Habitus is a set of dispositions that generate class-specific ways of operating in the world (Bourdieu 1984, pp. 101–102). Habitus develops during the socialisation process in interaction with social circumstances and social relations, and it generates tastes, choices and practices that are subjectively meaningful in given contexts. Accordingly, people accommodate their desired way of life in accordance with their assessment of their circumstances and available resources (Cockerham 2005).

From a Bourdieusian perspective, health lifestyles are a product of life conditions and available resources, as well as people’s preferences and tastes, which are formed in class-specific circumstances. People’s dietary patterns, leisure activities and ways of consuming alcohol therefore reflect class relations and distinctions. Bourdieu’s ideas on habitus and practices highlight how people’s day-to-day activities tend to be, to a great extent, routine-like and taken for granted: once established, a habitus governs behaviour, enabling everyday practices to be acted out without conscious deliberation. Thus, Bourdieu’s approach explains why lifestyles are not random by underlining the importance of class-specific social conditions internalised in the habitus.

Bourdieu’s approach has been repeatedly criticised for exaggerating objective social structures at the expense of agency and reflexivity (e.g., Adams 2006; Frohlich et al. 2001; Archer 2005). Critics have claimed that Bourdieu’s concept of habitus does not allow for voluntary action and thus assumes that existing social structures are reproduced almost automatically. While Bourdieu acknowledges the importance of agency, he still prioritises structural determinants of action at the expense of individual choices, preferences and subjective understandings (Jenkins 1992). In more recent discussions, however, the notions of reflexivity and flexibility of habitus have been more central and the idea of an over-controlling habitus has been rejected (Cockerham 2018). Silva (2016) has noted that Bourdieu’s conception of habitus changed over time so that in his later work habitus is more ‘elastic’ compared to his earlier work. In fact, Bourdieu’s later ideas of the role of reflexivity in situations when habitus and field collide are very close to pragmatism (Bourdieu 1990; Bourdieu and Wacquant 1992; Crossley 2001). Yet, Bourdieu gives priority to social class in the process of lifestyle formation. This means that socioeconomic status determines to a great extent what people do (Gronow 2011). The impression that structures determine can be seen as a result of Bourdieu’s emphasis on class-related determinants of action. Regarding the possibility to modify health-related habits, Crammond and Carey (2017) have emphasized that Bourdieu's notion of habitus does not give credit to public health initiatives or to changing conditions for influencing habitus and behaviour.

More recently, the concept of social practices has been suggested as a general conceptual framework for analysing and understanding health-related behaviour. While there is a variety of so-called practice theories and no integrated theory of practice exist, we concentrate on practice theoretical approaches and applications that have been central to the fields of consumption (e.g. Warde 2005; Shove 2012) and health sociology (e.g., Blue et al. 2016; Maller 2015; Meier et al. 2018; Delormier et al. 2009). In these fields, Reckwitz's (2002) influential article is commonly cited as the source for defining social practices as routine-like behaviour which consist of several interrelated elements, such as bodily and mental functions, objects and their use, knowledge, understanding and motivation (ibid., p. 249). According to Shove et al. (2012), practices integrate three elements: materials (objects, goods and infrastructures), competences (understandings, know-how) and meanings (social significance, experiences). Practices can refer to any form of coordinated enactment: preparing breakfast, having a break at work or having after-work drinks. Similar to lifestyles, social practices turn attention away from the individual and their intentions and motives towards the routinised ways people carry out their daily lives (Warde 2005). The idea is to look at people as carriers of practices because practices guide human action according to their own intrinsic logic (Reckwitz 2002). In other words, practices are relatively stable ways of carrying out a set of elements in an integrated manner. It follows, therefore, that they are both performances enacted more or less consistently in daily life, as well as entities that shape the lives of their carriers (Shove et al. 2012).

The social practices approach points out how smoking, drinking and eating should not be seen merely as single behaviours, but rather as parts of collectively shared practices, which intersect with other everyday routines (Mollborn et al. 2021). For example, in understanding drinking behaviour, one cannot separate the act of drinking from other aspects of the drinking situation, such as the kind of alcohol being consumed, how, where and with whom it is done, and for what purposes (Meier et al. 2018; Maller 2015). Drinking, smoking and eating, accordingly, are not single entities but parts of different kinds of practices, performed and coordinated with other activities of daily life (Blue et al. 2016).

As the main aim of practice theoretical approaches is to explain the stability and continuities of behaviour, the approach has difficulties in grasping the role of individual agency in the enactment of practices. According to critics, in some versions of practice theory, the role of individual carriers and the ways in which they make sense and experience practices seems to be more or less neglected (Spaargaren et al. 2016; Miettinen et al. 2012). Consideration of individuals’ sense of doing things is particularly important when studying aspects of human behaviour that can have adverse consequences and are unequally distributed within society. Therefore, we argue that the practice theoretical approach would benefit from more theorization on individual agency and the mechanisms by which individuals adopt and become carriers of practices. For health sociology, the question of how practices change and how people are recruited as carriers of practices is particularly relevant: how can healthy practices be adopted or how can practices be modified to become healthier? We argue that these issues were fruitfully conceptualized by the philosophical tradition of pragmatism with its concept of habits, which takes the individual actor as a premise without losing sight of the force of everyday routines.

Habits as dispositions

In recent decades, pragmatism has become an important source of inspiration for many social theorists (e.g., Joas 1996; Baert 2005; Shilling 2008). For example, Joas (1996) has argued that pragmatists had a unique viewpoint on the creativity of action, whereas for Gross (2009) pragmatism is a key point of departure when discussing social mechanisms. Pragmatism has been previously introduced to health research, for example, in relation to the epistemological problems of different kinds of health knowledge (Cornish and Gillespie 2009) and health services research (Long et al. 2018). Here, we focus on the aspect of pragmatist thought we find most relevant for health sociology, namely, its concept of habits.

Classical pragmatist philosophers were active at the end of the nineteenth and the beginning of the twentieth century. They included the likes of George Herbert Mead, William James, Charles S. Peirce and John Dewey. We mainly draw inspiration from John Dewey for his insights into the notion of habit. However, all classical pragmatists shared a similar understanding of the essential role habits play in explaining action (Kilpinen 2009). Thus, even though classical pragmatists may have differed in their point of emphasis, Dewey’s notion of habits is in many ways representative of the classical pragmatist understanding of habits. In this conceptualisation, habits are acquired dispositions to act in a certain manner, but they do not preclude conscious reflection.

Pragmatism, like the social practices approach, puts emphasis on contextual factors and the environments of action in understanding how habits are formed and maintained. Thus, one can argue that pragmatists were precursors to practice theorists. First and foremost, pragmatists highlighted the interaction between environments, habits and actors, by pointing out that people are constantly in the midst of ongoing action. Pragmatism also has an affinity with behaviourist psychology, which emphasises the role of environmental cues in triggering action. Behaviourists maintain that once an actor is conditioned to a reaction in the presence of a particular stimulus, the reaction automatically manifests itself when the stimulus is repeated. Say, a smoker might decide to give up smoking but the presence of familiar cues (e.g. cigarettes sold at the local grocery store, workmates who smoke) automatically triggers a response that results in a relapse. Classical pragmatists also thought that everything we do is in relation to certain environmental stimuli, but they did not think of the relationship in such mechanical, automatic terms (Mead 1934).

What acts as a stimulus depends on the part the stimulus plays in one’s habits rather than on simple conditioning (Dewey 1896). Thus, people are not simple automata that react to individual stimuli in a piecemeal fashion but rather creatures of habit. This means that individual actions get their meaning by being a part of habits (Kilpinen 2009). What may trigger the smoker’s relapse is not the presence of isolated cues but the habits that they are a part of; having a morning coffee, passing by or going to the local bars and grocery stores, and taking a break at work. Habits make the associated cues familiar and give them meaning.

The term habit, both in sociological literature and in common usage, typically refers to an action that has become routine due to repeated exposure to similar environmental stimuli. In this conception, the behaviour in question may originally have been explicitly goal-directed, but by becoming habituated, it becomes an unconscious, non-reflexive routine. As such, habits interfere with individuals' ability to act consciously. In practice theoretical approaches habit is similarly paralleled with routine-like ways of doing things. According to Southerton (2013), habits can be viewed as “observable performances of stable practices” (Southerton 2013, p. 337), which are essential for practices to remain stable (Maller 2015). In addition, habits are often understood as routines in popular science. According to Duhig (2012), the habit “loop” consists of the association between routines and positive rewards.

Pragmatists tend to see habits somewhat differently—as inner dispositions. This conceptual move means that habits have a “mental” component and habits can exist as tendencies even when not overtly expressed. Habits are thus action dispositions rather than the observable behaviour to which they may give rise to (Cohen 2007). As tendencies, habits include goals of action and not mere overt expressions of action; in other words, they are projective, dynamic and operative as dispositions even when they are not dominating current activities (Dewey 1922, p. 41). Habits make one ready to act in a certain way, but this does not mean that one would always act accordingly (Nelsen 2015). To paraphrase Kilpinen (2009, p. 110), habits enter ongoing action processes in a putative form and we critically review them by means of self-control. In this way, habits are means of action: habits “project themselves” into action (Dewey 1922, p. 25) and do not wait for our conscious call to act but neither are they beyond conscious reflection. According to classical pragmatists, habits thus do not dictate our behaviour. Rather, habits constitute the so-called selective environment of our action. They give rise to embodied responses in the environments in which they have developed but, as dispositions, habits are tendencies to act in a certain manner, not overt routines that would always manifest themselves in behaviour. What distinguishes habits from inborn instincts is their nature as acquired dispositions.

Moreover, habits guide action and make different lines of conduct possible. This is easy to see in the case of skills that require practice; for example, being skilful in the sense that one habitually knows the basic manoeuvres, say, in tennis, does not restrict action but rather makes continuous improvement of the skill in question possible. Simply reading books on tennis does not make anyone a good player of tennis and therefore actual playing is required for habit formation. Furthermore, once habits are acquired as dispositions, not playing tennis for a while does not mean that the habits and related dispositions would immediately disappear.

In the pragmatist understanding, habits are not the opposite of agency but rather the foundation upon which agency and reflexive control of action are built. Purely routine habits do, of course, also exist but they tend to be “unintelligent” in Dewey’s conceptualisation because they lack the guidance of reflective thought. Furthermore, Dewey (1922, p. 17) argued that conduct is always more or less shared and thus social. This also goes for habits, since they incorporate the objective conditions in which they are born. Action is thus already “grouped” in the sense that action takes place in settled systems of interaction (ibid., p. 61). This is where Dewey’s ideas resemble practice theory most because the grouping of action into settled systems of interaction can be interpreted to indicate the kinds of enactments that practice theory is interested in. While repeated action falls within the purview of habits, Dewey (1922) was adamant that habits are dispositions rather than particular actions; the essence of habit is thus an acquired predisposition to particular ways or modes of responding in a given environment. Compared with practice theories, this notion of habits underscores competences (understandings, know-how) and meanings (social significance, experiences).

Because habits are dispositions, they are the basis on which more complicated clusters of habits and, thus, practices, can be built. This means that practices can recruit only those who have the habits that predispose them to the enactments related to a practice.

Habits as practical solutions

In the previous section, we explained that pragmatists did not think of habits as mere routines. To be more precise, Dewey distinguished between different kinds of habits on the grounds of the extent of their reflexivity. Dewey labelled those habits that exhibit reflexivity as intelligent habits. Smoking is an example of what Dewey called “bad habits”: they feel like they have a hold on us and sometimes make us do things against our conscious decisions. Bad habits are conservative repetitions of past actions, and this can lead to an enslavement to old “ruts” (Dewey 1922, p. 55). Habits hold an intimate power over us because habits make our selfhood—“we are the habit”, in Dewey’s (1922, p. 24) words. However, habits need not be deprived of thought and reasonableness. So-called intelligent habits, in which conscious reflection and guidance play a part, were Dewey’s ideal state of affairs. Dewey (ibid., p. 67) thought that what makes habits reasonable is mastering the current conditions of action and not letting old habits blindly dominate. There is thus no inherent opposition between reason and habits per se but between routine-like, unintelligent habits and intelligent habits, which are open to criticism and inquiry (ibid., p. 77).

Many forms of health-related behaviour can be characterised in Dewey’s terms as unintelligent habits. We stick to many habits and rarely reflect on them in our daily lives. However, that there are intelligent and unintelligent habits does not necessarily imply that all healthy habits would be intelligent in the sense of being open to reflection. Further, the unhealthiness of a habit does not in itself make a habit unintelligent in the sense of being an unconscious routine. Rather, all habits are intelligent in that they have an intrinsic relationship with the action environment. They help the actor to operate in a given environment in a functional and meaningful way. For example, smoking can be seen as meaningful in many hierarchical blue-collar work environments, where the way in which work is organised determines, to a great extent, workers’ ability to have control over their working conditions. Smoking can be used as a means to widen the scope of personal autonomy because in many workplaces a cigarette break is considered a legitimate time-out from work (Katainen 2012). Smoking can thus be seen as a solution to a “problem” emerging in a particular environment of action, the lack of personal autonomy. In this sense, it is an intelligent habit that enables workers’ to negotiate the extent of autonomy they have and to modify their working conditions (ibid.). As shared practices, cigarette breaks motivate workers to continue smoking and recruit new smokers, but when smoking becomes a routine, reinforced by nicotine addiction, it does not need to be consciously motivated (see also Sulkunen 2015). In the context of highly routinized moments of daily smoking, reflection on the habit and its adverse consequences to health is often lacking (Katainen 2012). This means that the habit in question is not fully intelligent in Dewey’s terms.

The mechanisms of adopting so-called bad habits can be very similar to adopting any kind of habit if we understand habits as enabling a meaningful relationship with the environments and conditions in which they were formed. This idea also helps us rethink the socioeconomic patterning of health-related lifestyles. We do not have to assume that people in lower socioeconomic positions always passively become vehicles of bad habits due to limited life chances. The pragmatist view on habit presupposes an actor who has an active, meaningful relationship with the environment, that is, an actor with a capacity for agency, as our illustration of habits as a way to increase worker autonomy shows. Unlike practice theory or Bourdieu’s concept of habitus, the pragmatist concept of habit explains habitual action as a solution to practical problems in daily life. For pragmatists, action is always ongoing, and those activities that work and yield positive results in a given context have the potential to become habitual. We thus use habits to actively solve problems in our living environments, adapt to the fluctuating conditions we live in, and also modify these conditions with our habits.

Habits, doubt and change

So far, we have discussed habits as a relationship between the actor and the environment of action. We already hinted at the pragmatist idea that habits can be reflexive, and we now move on to discuss in more detail how and why habits change. According to Shove et al. (2012), practices are formed and cease to exist when links between materials, competences and meanings are established and dissolved. Additionally, practice theorists have suggested that practices may change when they are moved to a different environment or when new technologies and tools are introduced (Warde 2005). Actors may learn new things and perform practices in varying ways as performances are rarely identical (Shove 2012). However, it is insufficient to assert that practice theory assumes an active agent with transformative capacity if the underlying view of agency is passive and practices are the ones with agency to recruit actors. Furthermore, the question remains as to when actors are capable of being transformative and when they are confined to the repetition of practical performances.

The pragmatist understanding of how habits change, and when and how actors exercise their agency, originates in Charles Peirce’s thought. Peirce (1877) argued that we strive to build habits of action and often actively avoid situations that place our habits in doubt because doubt is an uncomfortable feeling. However, habits are nevertheless subject to contingencies and unforeseeable circumstances. Doubt cannot thus be avoided and it manifests itself in the crises of our habits that take place in concrete action situations and processes.

How should one then go about changing habits? This is a central question in all health sociological theory and has significant practical implications. Dewey (1922, p. 20) was a forerunner of many modern views in that he saw that habits rarely change directly by, for example, simply telling people what they should do. This presupposition is well acknowledged in critical health research, which has repeatedly pointed out that there is a gap between guidelines of healthy living and people’s life worlds (e.g., Lindsay 2010). It is usually a better idea to approach habit change indirectly by modifying the conditions in which habits occur. In the case of unwanted habits, conditions “have been formed for producing a bad result, and the bad result will occur as long as those conditions exist” (Dewey 1922, p. 29). Dewey’s emphasis on the role of conditions is well reflected in modern public health promotion, which rely on population-level measures and interventions. Yet, Dewey’s notion of the conditions of habits goes beyond macro-level measures, such as taxation, restrictions and creating health promoting living environments, to cover more detailed aspects of our daily life. According to Dewey, changing the conditions can be done by focusing on “the objects which engage attention and which influence the fulfilment of desires” (ibid., p. 20). Assuming that simply telling someone what they should do will bring about a desired course of action amounts to a superstition because it bypasses the needed means of action, that is, habits (ibid., pp. 27–28).

Interestingly, Dewey’s ideas of behaviour change have many similarities with the approach known as nudging, as both want to modify environmental cues to enable desired behavioural outcomes (Vlaev et al. 2016). According to both of these approaches, behavioural change is often best achieved by focusing on the preconscious level of habitual processes rather than appealing to the conscious mind by informing people of the potential risks associated with, for example, their dietary habits. Despite these similarities, the pragmatist view of habit change cannot be reduced to the idea of modifying people’s “choice architectures”. As Pedwell (2017) has pointed out, advocates of the nudging approach fail to sufficiently analyse how habits are formed in the first place and how they change once nudged. In the nudge theory, habits are analogous to non-reflexive routines, and the change in habitual behaviour occurs due to a change in the immediate environment of action. As a result, nudge advocates conceptualize the environment through a narrower lens than pragmatists and they are less concerned about how broader social, cultural, and political structures influence and shape everyday behaviour (ibid.).

According to pragmatists, changing habits is something that we do on a daily basis, at least to some extent. This does not mean that we would ever completely overhaul our habits. Dewey (1922, p. 38) thought that character consists of the interpenetration of habits, and therefore a continuous modification of habits by other habits is constantly taking place. In addition, habits incorporate some parts of the environments of action, but they can never incorporate all aspects of the contexts of action. What intelligence—or cognition in modern parlance—in general does is that it observes the consequences of action and adjusts habits accordingly. Because habits never incorporate all aspects of the environment of action, there will always be unexpected potential for change when habits are exercised in a different environment (even if just slightly) than the one in which they were formed (ibid., p. 51).

Different or changed contexts of action imply the potential to block the overt manifestation of habits. For example, if workplace smoking policies are changed so that smokers are not allowed to smoke inside, the habit of smoking needs to be reflected upon and the practice of workplace smoking modified. If the employer simultaneously provides aid for quitting smoking, or even better, creates conditions for work which would support workers’ experience of agency and autonomy, some may consider breaking the addiction, at least if colleagues are motivated to do the same thing. Such contextual changes lead to moments of doubt in habit manifestation and thus compel us to reflect on behaviour and, in some cases, to come up with seeds for new habits. The habit of smoking can be seen as a way of dealing with “moments of doubt”. It is a solution to certain problems of action in a given environment, as in the previous example of workplace smoking and autonomy. If the original context for which the habit was a “solution” to changes, it becomes easier to change the habit as well.

Pragmatist thinking thus suggests that here lies one of the keys to reducing unhealthy behaviours. By modifying the environments of habits, it is possible to create moments of doubt that give ground to the formation of new habits. Contrary to nudge theorists, however, pragmatists are not only concerned with promoting change in individual behaviours and its immediate action environment but also in the sociocultural contexts of habit formation by enabling people to create new meaningful capacities and skills (Pedwell 2017). The pragmatists also considered the consequences of moments of doubt on habits. Dewey (ibid., p. 55) argued that habits do not cease to exist in moments of doubt but rather continue to operate as desireful thought. The problem with “bad habits” is that a desire to act in accordance with the habit may lead to solving situations of doubt by changing the environment so as to be able to fulfil the habit rather than changing the bad habit. For example, new smoking regulations intended to decrease smoking may not lead to an actual decrease but rather to a search for ways to circumvent the regulation by smokers.

A crisis of a particular habit thus need not always result in changes in behaviour, as the disposition does not change overnight and may lead to looking for ways to actively change the environment of action back to what it used to be. Furthermore, the crisis (i.e., situation of doubt) may simply be left unresolved. This is what often happens when people are exposed to knowledge of the adverse consequences of their behaviour. There might be a nagging sense that one really should not behave the way one does, but as long as the environmental cues are in place, the habit is not modified, especially if one’s social surroundings reinforce the old habit (e.g., other people also continue smoking at the workplace). It can also happen that one makes minor changes in behaviour, for example, by cutting down instead of quitting smoking—which can in time lead to falling back on earlier smoking patterns. New workplace smoking policies, therefore, often mean that the practice of smoking is modified, and the smokers adopt new places and times for smoking. While old habits often die hard, discordances between habits and their environments can nevertheless trigger reflection and thus have a potential for change.

Discussion

We have argued that the pragmatist understanding of habits is an often-overlooked forerunner of many modern theories of health behaviour. While the health lifestyle theory helps to analyse the factors by which health lifestyles are patterned and points out that both contexts of action and individual choices are important in lifestyle formation, it is less helpful in empirical analyses on the mechanisms by which particular patterns of behaviour emerge in the interplay between choices and chances. The social practices approach further elaborates the relationship between choices and life chances by turning attention away from the structure-agency distinction towards enactments of everyday life and on how people go about their lives by carrying social practices. However, the social practices approach runs the risk that individual action becomes a mere enactment of practices. Thus, the practices are the true agents and people become mere carriers of practices. In this context, the pragmatist notion of habits can be useful in grounding practices within the clusters of habits that people have, thereby enabling them to be recruited by specific practices.

To conclude the paper, we want to stress some of the key pragmatist insights into the theorization of health lifestyles and practices. First, unlike practice theories, pragmatism takes individual actors and their capacity to meaning making and reflexivity as a premise for understanding how habits are formed and maintained. Thus, from the actor's point of view, habits, even “bad” habits, should be understood as functional and meaningful ways of operating in everyday circumstances. Habits are creative solutions to problems confronted in everyday life and reflect individuals’ relationships to the material and social world around them. Action that proves useful and meaningful in a particular context is likely to become habitual. In the context of health inequalities, risky health-related habits can often be seen as a way to strive for agency in circumstances that provide little means for expressing personal autonomy. We suggest that this insight should be at the core of designing any public health or behavioural change interventions tackling health inequalities.

Second, pragmatism suggests that habits should be understood as dispositions; people are recruited by practices only when their dispositions enable this to happen. Often a lot of habituation is required before the predispositions are in place that make recruitment possible. Third, pragmatism provides tools to analyse how moments of doubt enter habitual flows of action. Doubting habits is an inherent part of our action process, but habits are called into question especially by changes in the environments of action that make particular habits problematic. This, then, can lead to the development of new or modified habits as a response to the “crisis” of action. If the social and material environment of action, to which the habit is a response, stays more or less the same, the habit will be difficult to change.

The pragmatist conception of habits, while emphasizing agency and reflexivity, does not ignore the significance of materiality and routines in daily conduct but is able to incorporate these elements of action in a way that benefits empirical analyses of everyday practices. Pragmatism thus suggests a variety of research settings to investigate the mechanisms by which health-related habits are formed. Here, we provide a few examples. On a macro level, it is important to observe how organisational, technological, or legislative changes are manifested in different contexts and how they modify and enable habitual action in different social groups and settings. Structural measures to promote public health are likely to invoke varying effects depending on the contexts of action of different population groups. Although the physical environment may be the same, the environment of action is not the same for everyone. In pragmatist terms, new policies can be understood as modifications of action environments, which potentially create moments of doubt in habitual action. For example, there is considerable evidence that smoke-free workplace policies reduce workers' smoking (Fichtberg and Glantz Stanton 2002), but more research is needed to determine how different socioeconomic groups are affected by these policies. Macro-level policy changes create an excellent opportunity to study how policies give rise to new patterns of health-related behaviour, how policies are implemented in different contexts, and how reactions to policies and their effects vary depending on socioeconomic circumstances.

A micro-level analysis of health-related behaviour, on the other hand, could focus on the triggers of the immediate environment of action—material, social or cognitive—to examine how habits are formed as practical and creative solutions to specific problems and what kinds of factors create situations of doubt and thus include the potential for habit change. Research should analyse how moments of doubt regarding health-related habits emerge in differing socioeconomic contexts, as well as why unhealthy habits can and often do become deeply routinized and resistant to change. Furthermore, it is essential to find out the problems in relation to which particular habits of action have been formed. In both micro- and macro-level analytical approaches, people’s reflexive capacity and the pursuit of a meaningful and functioning relationship with their environments should be at the core of analysis.

Methodologically, we suggest that the pragmatist approach to health behaviour research calls for methods that integrate the observation of action and people’s accounts of and reasoning about their conduct. Ethnography is one research method suited to this task. With participant observation, it is possible to access lived experiences in local settings through which larger policies affect health (Hansen et al. 2013; Lutz 2020) and hard-to-reach population groups (Panter-Brick and Eggerman 2018). So far, ethnographic studies have been rare in health inequality research (e.g., Lutfey and Freese 2005). One way to proceed is provided by Tavory and Timmermans (2013), who have suggested pragmatism as a theoretical-methodological basis for constructing causal claims in ethnography. They propose that a useful starting point for observation could be the process of meaning making: how individuals creatively navigate their conduct when confronting moments of doubt and how they make sense and respond to them in more or less habitual ways. However, surveys can also be used in creative ways to investigate people’s habits, for example, using mobile apps that ask and/or track what people are doing. Other methods besides ethnography are thus needed to test the causal claims made by ethnographers.

Lastly, research is needed on how educational systems predispose people to develop reflective habits. One possible explanation for why knowledge about the adverse consequences of health-related behaviour is correlated with people’s socioeconomic status, and especially their level of education, is that a higher level of education makes one more sensitive to knowledge-related cues for behaviour. This is because higher educational levels tend to bring about the habit of reflecting on the basis of new knowledge. Education is intimately related with a habit of thinking of things in more abstract terms—distancing oneself from the specifics of particular situations and moving towards more abstract thinking. A high level of education also means the absorption of new knowledge has become habitual. Unfortunately, there are no shortcuts to developing such capacity. This is one of the reasons for why merely providing information on health-related issues will affect different population groups differently.