Cycling in an ‘ordinary city’: A practice theory approach to supporting a modal shift

Abstract There has long been evidence of the benefit of a modal shift toward cycling can bring to meeting several pressing urban challenges including ill-health, climate change, and poor air quality. In the wake of COVID-19, policy-makers have identified a modal shift toward cycling as part of the solution to mobility challenges introduced by social distancing measures. However, beyond exemplar areas, cycling has been largely characterized by a stubbornly-low modal share. In this paper, we use the ‘ordinary city’ – in cycling terms – of Liverpool as a case study to understand this. We apply practice theory in doing so, finding the provision of materials for cycling is the key factor in supporting a modal shift. Not only do they provide the means to support the practice of cycling in the city, but they also have a key role in shaping individuals perceptions of, and the skills required to cycle. We then reflect upon the utility of practice theory in understanding the patterns of everyday life, finding it was particularly well suited in understanding the interactions between different factors which influence modal choice. We go on to identify practical challenges in its application within our analysis raising questions around an inconsistent analysis of influential factors including ‘driver behavior’ and ‘political commitment’. We suggest how this might be overcome, through the isolation of such factors within a category of ‘action of others’, this we argue means the findings in this paper have broad relevance to researchers and policy-makers alike.


Introduction
The COVID-19 pandemic has severely disrupted the operation of public transportation networks due to social distancing measures, and there have been suggestions that fears of infection may have an impact on the individual's modal choice (e.g. Douglas et al., 2020). If true, in the short to medium term there are plausible scenarios in which a modal shift toward private motor vehicles occurs. Such a scenario risks increasing inequalities in accessing economic and social opportunities, as well as exacerbating a suite of existing environmental and health challenges. However, encouraging and facilitating a shift toward active travel, for example, cycling, will even in the event a shift toward private motor vehicles help to mitigate such impacts.
Globally a quarter of adults are classed as inactive (Guthold et al., 2018), and 9% of global deaths are attributed to a lack of exercise (Lee et al., 2012). There is a wealth of evidence that regular cycling delivers health benefits, both in terms of lower incidences of cardiovascular disease and cancer (Celis-Morales et al., 2017), and in terms of well-being and quality of life (Anokye et al., 2012). Many urban areas exceed legal limits for air pollution (WHO, 2019), cycling has been identified as a key means to reduce airborne pollutant within the transport sector, resulting in further health benefits (Holgate et al., 2016). The pandemic has provided a pertinent example of how these factors can influence a population's resilience to current, and future health risks, with obesity and poor air quality both being tentatively identified as factors increasing the risk of COVID-19 hospital admissions (Lighter et al., 2020;Wu et al., 2020).
However, despite numerous strategies which acknowledge the positive effects of increased rates of active travel (e.g. DfT, 2017b in England), away from exemplar case studies cycling rates remain stubbornly low. In England for example, and even despite a small rise over the past 15 years, cycling still only had a 2% modal share in 2018 (DfT, 2019). Similarly, whilst many of the UK's urban areas have implemented cycling strategies they have enjoyed mixed success, reflected in wide variations in cycling rates across local authorities (DfT, 2019).
Though short-term increases in cycling following COVID-19 have been observed, for example, in the UK (DfT, 2020) the longer-term modal share indicates that to capitalize upon this and to achieve local policy ambitions for cycling, there is a need to understand what factors are most important in encouraging and sustaining individuals to cycle. In this spirit, this paper examines some of these issues through a case study of Liverpool, UK. Whilst acknowledging the research which suggests that the understanding of local contexts is vital for a successful approach to cycling (Larsen, 2017b, Sheldrick et al., 2017, in contrast to exemplars (e.g. Copenhagen or Amsterdam), we argue Liverpool embodies the shortfall in cycling uptake. As such, we believe this study can have relevance to urban areas in many developed nations.
To achieve this aim, we draw upon the principles of Practice Theory  to consider both the potential barriers and enablers to cycling. The following section introduces this theory, underlining its relevance to urban cycling, before considering the everyday barriers and enablers of urban cycling.

Practice theory
To address the aforementioned urban challenges there is a need to shift travel behavior from private motor vehicles toward other modes (i.e. cycling). In popular policy discourse, the focus of this behavior change is placed at the individual level . Following this, interventions seek to change the conditions in which decisions are made, in a bid to influence an individual to make the preferred choice. For example, an application of the Theory of Planned Behavior advocates marketing campaigns to change attitudes and values, which subsequently drive behavior (Ajzen, 1991). Approaches informed by behavioral economics similarly place the individual at the center, creating 'choice architectures' which incentivize 'good' choices (Thaler & Sunstein, 2008). In contrast practice theory emphasizes the practice itself, instead of the individual as the 'unit of analysis' (Reckwitz, 2002). Reckwitz (2002, p. 249) defines a practice as a 'routinized type of behavior' undertaken by a carrier (e.g. a cyclist). This work builds upon that of earlier authors such as Taylor (1973, p. 27) who sees 'meanings and norms implicit in [ … ] practices not only in the minds of actors but out there in the practice themselves'. Bourdieu (1990, p. 52) sees practices and habitus in a recursive relationship, with the latter 'always oriented toward practical function'. Bourdieu and Wacquant (1992) see an explanation for the regularity and predictability of social life within the habitus. While the theory of structuration sets out the study of social science as 'neither the experience of the individual actor, or the existence of any form of social totality, but social practices ordered across space and time' (Giddens, 1984).
As part of the second generation of practice theorists, Reckwitz (2002) drew together this philosophical work to set out 'ideal type' of practice theory which defines the key distinction from alternative cultural theoriesthat the social is located within the practice (Everts et al., 2011). Schatzki (2001, p. 12) sees the theory as one that seeks to understand relationships between everyday activities and wider society, where the 'practice [is] the source and carrier of meanings, language and normativity'. The practice is therefore seen as the site where agency, structure and materials act recursively to reproduce patterns of activity (Reid & Ellsworth-Krebs, 2019). Reckwitz's (2002) 'ideal type' of practice theory defined practice as the combination and interrelation of three separate elements: materials, competencies and meanings. A practice will then reflect the changing circulation and relationship of each element. So to understand and influence the trajectory of a practice we must understand the nature of elements which comprise it. Researchers have done so examining a range of activities, for example, showering (Hand et al., 2005) and like this study, cycling (Watson, 2013, Larsen, 2017b. Shove et al. (2012, p. 14) further defines these elements. Materials are 'things, technologies, tangible physical entities and the stuff of which objects are made', in the case of cycling this includes the bicycle, roads, storage facilities. Competencies refer to understanding and practical knowledgeincluding 'skills', and 'technique', for example, the ability to ride their bicycle and navigate road-traffic networks. Meanings are considered as the emotional, social and symbolic significance of the practice.
In short, practice theory is a framework which enables the study of routines of practice to provide a better understanding of the social and infrastructural conditions in which these practices can flourish (Shove & Spurling, 2013), thereby supporting behavioral change. However, whilst Keller et al. (2016) sets out its value in informing policymakers, Shove (2014) instead is skeptical of its applicability without a reconfiguration of policy-making processes. The move away from the individual can highlight non-individual-centered interventions which can be more effective in stimulating change (Gill & Gill, 2012).

The materials of cycling
Cycle design guidance focuses upon improving the environment (materials) in which a bicycle is operated, and, in doing so, typically identifies five factors which have the biggest influence on cycling: 'Coherent, Direct, Attractive, Safe and Comfortable' (CROW, 2016). This approach relies upon the logic drawn from a rational decision-making process, whereby a series of factors are traded off against others, which can result in choosing a different mode if they cannot be satisfied (Parkin et al., 2007).
A perception of risk is regularly cited as the primary barrier to cycling, and through a systematic review Lorenc et al. (2008) find this perception is closely linked to the quality and quantity of cycle infrastructure. Furthermore, Reynolds et al. (2009) review of safety literature found clearly marked cycle-specific infrastructure was consistently safer than onroad riding. Buehler and Pucher's (2012) study of cycle lane provision and cycling rates across 90 US cities estimated that a 10% increase in cycle lanes leads to a 3.1% increase in cycle commuting. Buehler (2012) analyzed of data in districts with contrasting cycling rates in Washington DC, estimating that an additional mile of cycle lane per 1,000 population leads to an 11% increase in the chances that an employee cycled to work.
The literature also indicates that the presence (or absence) of materials beyond cycling lanes can be important in shaping cycling in a locality. In their study of Dutch, German and Danish cities where cycling strategy had been credited in sustaining high rates of cycling, Pucher and Buehler (2008) argue that while separated cycle facilities were a cornerstone of safety, alone they could not guarantee a significant increase in cycling rates. Other materials included safe and adequate cycle parking at home, work and in urban realm. Studies in European and American urban areas found such infrastructure reduced crime, and feed into notions of cycling as a safe activity and supported cycle commuting (Aldred & Jungnickel, 2013;Buehler, 2012). Providing effective cycle parking can also tackle 'fly-parking' which Larsen (2017a) found creates negative connotations of cyclists, even in cities with an established 'cycling culture' such as Copenhagen.

The variation, circulation and interrelation of elements
Where certain materials are lacking it can lead many to choose an alternative mode of transport, but for others it leads to an adaption of their cycling behavior. For example, Latham and Wood (2015) describe how, in the absence of dedicated infrastructure cyclists in London 'recolonize' the road network by breaking rules, for example, by running red lights or pavement-riding. Though some cite safety as a motivating factor, others cite gaining a time advantage. Both re-enforce an image of cyclists being a nuisance, for example, Fincham's (2006) work in London suggest that running red lights is the cyclist behavior which most annoys drivers. As well as undermining latent sympathy toward an unfriendly cycling environment, the negative meaning of deviance can also discourage others from taking up cycling, as found in Hull and Cambridge, two cities with historically high cycling rates (Aldred, 2013).
The perceived need to invoke certain behaviors whilst cycling also requires practitioners to possess certain competencies to adapt to the car-dominated urban environment. This is a barrier to recruitment, and a perceived inability to possess, or develop these competencies is likely to be related to an individual's personal characteristics, or through concern on behalf of significant others, for example, children (Pooley et al., 2013). Studies in five UK urban areas with an varied participation in cycling found that individuals who do not cycle may perceive cycling as an activity for 'sporty' peoplealienating those who are either unwilling or unable to embody this meaning (Aldred, 2013;Steinbach et al., 2011).
Building upon this, Pooley et al. (2013) argue that how cyclists are perceived alone is enough to discourage people from cycling. For example, in places with a low cycling modal share, someone who travels regularly by bicycle may be seen as 'odd'. Villamor et al. (2008) suggest materials (e.g. Lycra) worn by cyclists may play a significant role in the characterization of cycling, particularly in localities where cycling holds a small modal share, contrasting with 'typical' cyclists in places with higher modal share, for example, in the Netherlands, who wear 'everyday clothes', reflecting cycling's status as a mainstream means of transport.
De la Bruheze (2000) historical perspective of cycling in Britain supports this analysis, speculating that a threshold was crossed, whereby the shift toward motorized transport meant that cycling struggled to maintain legitimately as a means of transport. Whilst it is clear it was the changing materials of transport were fundamentally facilitating this shift, Shove et al. (2012) highlight that the marginalization of cycling meant the practice became associated with sport and leisure in the UKand therefore something not considered 'normal', thereby re-enforcing, and accelerating the modal shift which was underway. Shove et al. (2012) draw comparisons with Netherlands and Denmark, where cycling rates were initially higher than the UK, and never dropped so low, meaning that cyclists were never marginalized, preserving its status as everyday transport.
Going further, Steinbach et al. (2011) argue certain individuals may associate cycling with being 'poor' or of a 'low social status', whilst other groups (e.g. middle-class professionals) may instead emphasis the utilitarian nature of cycling, which allows them to combine exercise with journeys made in the optimum time. We can began to see how different meanings are attached to a practice by different groups, perhaps contributing toward large variations in rates of cycling between different groups, for example, in London a cyclist is disproportionally likely to be affluent, white and male (Steinbach et al., 2011). Here, Fishman et al. (2012) study of cycle-hire participation in Brisbane (which is characterized by a lower modal share) suggests that when people see those 'like themselves' cycling they are encouraged to try it out. These studies are important in illustrating the way that practitioners themselves can influence the meanings which are attached to cycling, and where class and ethnicities might affect how meanings are attached to a practice: something that appears at odds with the de-centering of the individual through practice theory.

Approaches to cycling research
Overall, Practice Theory is useful in exploring the different environments in which cycling takes place and how this influences the sometimes contradictory meanings which are then attached to the practice In the cases of materials for cycling, we can see the illustrations of the interactions between materials and the meaning attached to cycling. This review has underlined the diversity of cycling research, studies which approach use systematic reviews and statically driven analysis can effectively identify factors which are likely to drive cycling participation. However, such an approach often neglects the inter-related nature of these factors, whilst also failing to capture locally specific factors. Research which focuses upon single (or small number of) localities, instead can command a richer understanding of locally specific factors and, importantly how these factors interact. These might relate to materials, meanings or competencies which then disproportionally attract or detract potential recruitments to cycling. Watson (2013, p. 124) conceptualizes such factors as 'systemic sticking points' which can limit recruitment or promote defection from cycling, and as we argue an approach which focuses on a particular locality is likely to be more successful in accuracy identifying these factors.
Overall, therefore, practice theory is a useful framework to understand and evaluate which elements either support or undermine urban cycling. It gives weight to the material factors commonly identified as important to supporting mass cycling, but also integrates the important role and consequences of personal attitudes, attributes, social and cultural norms (Handy et al., 2014) in the competencies and meanings of a practice. Clearly, much can be learnt from localities considered as 'successful cycling cities' (e.g. Pucher & Buehler, 2008), but as we will argue a focus upon areas where cycling is more marginal can be revealing in understand the plethora of factors which attract, or detract different groups to participate in cycling. Importantly, and per Larsen (2017b), these factors are often interconnected and, in most instances, would not represent a 'silver bullet' in isolation. Now, we will explore these issues within the context of Liverpool, UK. Here, and in a city that is both emblematic, and symptomatic, of broader malaise toward urban cycling, we explore the extent to which a variety of factors influence cycling within the city and how they might be overcome.

Methodology
Liverpool serves as the focus of this study for two reasons, firstly, we see the city as 'ordinary' in cycling terms, it lacks an established 'cycling culture', and cycling occupies a similar modal share to the UK average at just under 2% (DfT, 2019). However, in other ways the city is unusual, since 2010 the Liverpool City Council (LCC) has managed one of the largest budget reductions in England, a 63% decrease (Liverpool Express, 2019), whilst ranking as the fourth most deprived local authority in England (ONS, 2019). It lacks significant cycle infrastructure, lagging behind other UK core cities and leading European exemplars (Nurse & North, 2013), which is reflected by an appalling road safety record, in which 201 cyclists were killed or seriously injured between 2012 and 2016, the highest of any English metropolitan borough (DfT, 2017b). We argue that Liverpool, by virtue of its distance from exemplar status, can offer useful policy lessons to urban areas in a similar position.
That said, LCC is not without a coherent cycling strategy and the Liverpool Cycle Revolution (Liverpool City Council, 2014) represents the most recent articulation of this. The strategy's key target is that 10% of all trips will be made by bicycle by 2025. LCC admits that this is not without challenges, particularly when LCC cannot meet the minimum spending of £10 per head per year as recommended by the 'Get Britain Cycling Report' (APPCG, 2013). Going further, the strategy has limited detail on specific projects, instead placing a greater emphasis on 'developing' plans for cycling, and investment in enforcement and cycling education programs. In planning terms, the Liverpool Draft Local Plan (Liverpool City Council, 2018a) also seeks to maximize accessibility by active travel, most notably with a policy which protects existing and planned schemes as a means to continue developing a comprehensive cycling network. Latterly, in response to COVID-19 the City Council has planned to implement a series of 'pop-up' cycle lanes (Liverpool Express, 2020).

Site selection
To explore some of the barriers and enablers to cycling further, we focus on the single ward of Princes Park. Selected because of its ability to capture many of the issues discussed previously, the site affords insights on two main fronts. Firstly, the wide leafy boulevard of Princes Avenue is a key cycle commuter thoroughfare to the Liverpool city center, connecting the urban core to the South of the city. Despite this, there currently is limited cycle infrastructure on Princes Avenue itself, although a major upgrade is planned which includes a segregated cycleway. 1 Secondly, our site selection reflects our desire to examine an area where there is significant potential to increase cycling participation. Therefore our approach deliberately targeted both those living within the Princes Park ward and those who live and work beyond the area. The Princes Park ward is one of the most deprived in both Liverpool, and England as a whole (Liverpool City Council, 2018b), whilst 2011 Census data indicating that 66% of households within the ward have no private vehicle access. Yet despite its proximity to the city center, of those in employment, just 3.4% cycle to work as their main mode of travel (overall participation in cycling is likely to be higher, though no such data exists at ward level), with 26% currently walking to work (ONS, 2013). While this somewhat limits the scope for environmental benefits, a modal shift toward cycling may facilitate enhanced access to economic and social opportunities (Raj e & Saffrey, 2016). Combined, therefore, Princes Avenue, along with the surrounding streets contained within the Princes Park ward, provides a platform from which to explore the barriers and enablers to cycling in Liverpool.

Study design
Participants were recruited within the Princes Park ward as they traversed the area either by bicycle or walking, with data collection taking place across eight different sessions, split throughout the day across weekdays and a weekend. In total there were 95 interview participants: 55 cyclists and 40 non-cyclists, who were made up of a mixture of those living within Princes Park and those working within or traveling through the area.
The views of these interviewees were supplemented by detailed interviews with key cycling policy-makers and stakeholderseach focusing on the themes discussed thus far, and the broader efforts made by the city in overcoming them. Those interviewees included a Transport Planner from LCC, members of a local cycle advocacy group, and a local social enterprise promoting cycling.
The study collected data via a mixed methods approach, incorporating a survey and follow-up interviews, the survey was primarily a recruiting tool whilst also producing quantitative data. An initial question established a differentiation between cyclists and non-cyclists, the former being defined by cycling at least three times a week (DfT, 2017c, Cycling UK, 2017a. This allowed the researchers to explore the differing experiences and perception of cycling in the city (i.e. their choice to cycle or not). The survey also informed the interview approach, which provided an opportunity to discuss the specific issues raised in greater detail, allowing the researchers to gather a richer account of cycling and mobility decision-making.
The interviews explored the everyday experiences and perceptions of cycling such as the infrastructure, benefits and challenges of cycling. Some interviews had a greater focus on the absence of infrastructures, where we invited participants to elaborate upon their expectations of infrastructure, and to propose interventions to support cycling. Another focus was upon the competencies of cycling, and how other forms of mobility (e.g. walking and driving) impacted their experiences of cycling.
As previously discussed Practice Theory attempts to decenter the individual, however it is inevitable that different individuals possess contrasting experiences and perceptions of the social world, and these differences are captured through our discussions. We transcribed, coded and analyzed this data, placing each code within the materials, meanings and competencies framework. In some cases this posed a significant challenge, which we discuss further in our analysis.

Study findings
Our findings are presented across four sections, each representing the broad themes which emerged from our work. First is the centrality of materials in defining a practice reflecting a tendency of both groups of participants to link the (lack of) materials with the meanings they attached to cycling. As we explore in the second section the competencies of participants allowed them to overcome the deficiencies in materials, and the tensions between different mobility practicesthe focus of the third section. In our analysis, we reflect upon our experiences of applying practice theory in the context of cycling research, deriving several conclusions for policy-makers to support further recruitment toward cycling and suggestions for developing the utility of the frameworks for future research.
As part of the survey participants were asked to rate their perceptions of the cycling experience in the city. The mean scores (with five representing the maximum) of both cyclists (2.9) and non-cyclists (3.0) revealed that in broad terms both group's perceptions do alignthat is, neither presented an overly-pessimistic nor overly-optimistic view. Reflecting the discussion above they acknowledge that Liverpool is something of an ordinary city in cycling terms neither utopia nor death trap.

Materials define meanings
Where cycling in Liverpool is perceived as poor there was a broad alignment with previous literature (e.g. Lorenc et al., 2008), in that perceptions of cycling as a dangerous activity was the greatest deterrent for many non-cyclists. Here, the provision of materials for cycling was inextricably linked to the perceptions of danger and safety.
In accordance with studies suggesting that a lack of cycle lanes inhibits cycling rates, both cyclists and non-cyclists bemoaned a lack of safe cycling infrastructurewith nearly two-thirds (60) mentioning cycle lanes. In particular, 27 non-cyclists cited the lack of cycle lanes as the major preventative factor in taking up cycling, whilst others said if they did cycle they would only consider pavement-riding. Cyclists frequently raised the dangers of particular roundabouts and roads, typically linking this to an absence of dedicated cycling infrastructure. The discussions regarding the roads were, therefore, overwhelmingly negative. The most frequent complaint from cyclists regarded the poor road surface (i.e. 'potholes'), particularly on side streets which participants felt could otherwise be attractive for cycling. This criticism extended to the limited number of segregated cycle lanes that did exist, with cyclists complaining about the poor design of dedicated lanesparticularly where their design led to potentially dangerous reentry onto the main carriageway. However, there was praise for dedicated infrastructure which did exist, including advanced green lights for bicycles. These were seen to make cyclists feel saferillustrating how dedicated infrastructure can have a positive impact on the meanings attached to cycling.
When cyclists discussed potential changes, many advocated the implementation of segregated cycle lanes, emphasizing their role in creating a safer cycling environment. Many added that any money spent providing cycle lanes must also be accompanied with enhanced maintenance and enforcement programs, perhaps reflecting the condition of existing infrastructure. Conversely, one cyclist dissented from the prevailing view, arguing that implementing lanes meant accepting that cyclists do not belong on roads. Whilst this was a minority view, the provision of materials may have the unintended effect of validating negative meanings of attracted to cycling, that is, cycle lanes are needed due to dangers of cycling.
Whilst discussions largely focused upon aspects of cycling in motion, it was clear that materials supporting the bicycle 'at-rest' also had the potential to shape meanings. A lack of secure cycle-storage facilities was regularly cited as a factor which discouraged regular cycling, supporting Aldred and Jungnickel (2013), this was often linked to living in a flat or shared accommodation with no significant storage space. In public spaces, particularly within the city center, there was seen to be a lack of 'official' cycle parking, forcing cyclists to resort to 'fly-parking'. These material deficiencies result in participants associating the parking of cycles with danger and theft, with nine participants recounting their own experiences of theftsuggesting perceptions of risk were not unfounded.
Six cyclists discussed the need for greater access to showering and changing facilities at work, with one respondent describing efforts by his employers as 'tokenistic'something which, perhaps, reflects the broader view of the provision of materials for cycling in the city.
Notwithstanding the evidence of a stressful and problematic experiences of cycling in the city (typically attributed to the material deficiencies described above), it was clear that almost all participants who cycled derived significant enjoyment from it, especially during summer months. Four cyclists felt that it helped them reach a positive mental state especially when part of their daily routinewith one respondent likening their cycle commute to meditation, saying it put them in a 'good headspace' before work. However, the broader evidence does support this beneficial characterization of the practice (Pooley et al., 2013, Anokye et al., 2012, and that cycling, alongside walking, is seen as the most enjoyable mode of transport (DfT, 2017a). It is clear that there is a contradiction: even though cycling may be perceived as 'bad', it is still good.

Competencies and material deficiencies
While it was clear that materials had a significant impact upon the meanings, and subsequent recruitment to the practice of cycling in the city, there was a sense that this may be overcome through increased confidence (e.g. through experience and appropriate skills training). For example, 43% of cyclists saw Liverpool as a dangerous cycling environment, but qualified this by describing how their own competencies enabled them to overcome this. For example, statements referring to material deficiencies were often followed up with references to competencies, such as cyclists suggesting they had the confidence and skills to overcome the danger from narrow roads by cycling in the middle of the carriageway (see Cumming, 2012). However, a repeatedly-made point from non-cyclists suggested at a vicious cycle within the current situation, questioning how they could be expected to develop the necessary skills and confidence on busy roads with no cycling infrastructure. In response, participants frequently evoked a strategy of 'rule-bending' (Latham & Wood, 2015) to deal with the combination of a lack of materials and competencies, such as pavement-riding, even when they felt unease in doing so.
Corresponding with Aldred's (2013) work, there was also a perception that cycling was something for 'sporty' people, with some non-cyclists suggesting that they felt too unfit to cycle, or that ill-health now prevented them from cycling. In both instances there are inherent Catch-22s at playnot least where people are dissuaded from cycling because they perceive it as 'not for them', yet a key way to instigate change is if more people like 'them' take part.
There were some indications from participants regarding how these issues might be combatted. Supporting work by Akar and Clifton (2009) workplace and public services for cycling were seen by two cyclists as an important space for building competencies and thereby enabling recruitment to cycling. Primarily this focused on gaps in repair skills, but also was seen as an opportunity to discuss and develop competencies such as route advice and riding techniques, which might then aid less-confident cyclists to overcome material deficiencies. Beyond this, one interviewee commented that their street-based service means 'people can see us and join in: visibility is very important', suggesting a dual-function of cycling education and promotion.

Tensions between mobility practices
A key theme present throughout the research was the tensions between different mobility practices, primarily between auto-mobility and velo-mobility. These factor were problematic to clear categorized within practice theories framework, which we explore in detail in Section 5.1. This often stemmed from the aforementioned lack of materials for cyclists (e.g. dedicated cycle lanes), which compels practitioners to share materials (e.g. road space) with other road users. Here, the dominance of auto-mobility in the urban environment was clear, firstly through high volumes of road traffic, particularly at 'rush-hour' and in the city centerboth key times and spaces for cycle commuters. Secondly, parked vehicles were seen to introduce additional hazards and compromised efforts to support cycling, for example, parked vehicles blocking cycle lanes. Both points were seen to increase perceptions of danger, therefore influencing a decision not to cycle.
However, whilst the issues are connected, both cyclists and non-cyclists instead more frequently cited the behavior of drivers as a greater factor in creating a dangerous cycling environment. This varied in severity, from a lack of signaling to an incident of assault. In many cases, it was clear that this perception of danger emerged as a cumulative effect of many minor incidents. This aligns with the work of McKenna and Whatling (2007) who described how drivers can create a hostile atmosphere for cyclists on the road through intimidation and 'near-miss' incidents. This might be interpreted as an indication that certain practitioners of auto-mobility lack the necessary competencies, to drive in the presence of cyclists, or perhaps to drive more fundamentally. Yet there are also more extreme examples of aggression toward cyclistsillustrated by an incident where one female participant was assaulted, whilst another was told 'get off the road you fat b ÃÃÃÃ ' by a driver. In both cases the participants explicitly linked these incidents to their gender, aligning with the work of Heesch et al. (2012), who suggested women were more likely to cite aggression from motorists as a constraint to cycling. Whilst some practitioners are able or willing to use competencies to overcome the inherent tensions between motilities in urban areas, others are less able to do so. For example, Steinbach et al. (2011), concluded that white males have greater confidence to ride 'assertively' to defend against such aggression.
The sharing of materials also may threaten some of the meanings which some participants attached to driving (when they referred to their experiences of interacting with cyclists whilst driving), which may help explain some of the hostility described above. The presence of cyclists (or indeed the materials of cycling) can be seen as a threat to the qualities of driving, for example, cycle lanes replacing the convenience of on-street parking or a cyclist blocking an 'overtaking' maneuver, leading to some practitioners of driving to respond with hostility.
There can be a tendency to classify the practices of walking and cycling within a user group of 'active travel' there were also tensions between these practices that, to some extent, mirror the aforementioned tensions with auto-mobility. In this case 'deviant' cyclists threaten some of the cited qualities of walking, for example, safety and relaxation, which are partially derived from separation from other motilities. A lack of awareness, excessive speed and pavement-riding were all raised by participants who were critical of cycling. Respondents were particularly critical of those relying on cycling as an economic function (i.e. couriers) with one participant saying '[Couriers] seem to behave especially badly', alongside those cycling for sportoften conceptualized as being 'Lycra-clad'. Whilst participants did express sympathy toward pavement-riding (i.e. to avoid the dangers on roads) there remained anger at cyclists for doing so. This ire also singled out cyclists riding through parks ironically, often part of the national cycle routes, and one of the few places where off-road cycling infrastructure was provided.
This highlights a critical issue for a modal shift toward cycling in urban areas: in the absence of dedicated infrastructure, where should potential recruits build up the necessary competencies? While some are willing to 'rulebend' and use pavements, these comments highlight that others are not willing to do so, and may even view seemingly appropriate open spaces such as parks as inappropriate for cycling. Logically, it would seem that these individuals are unlikely to even experiment with cycling in the absence of dedicated spaces for doing so.
We can also see that experienced (i.e. 'Lycra-clad') and inexperienced (i.e. 'pavement-riders') cyclists alike can themselves contribute to the negative meanings attached to cycling. This reinforcing of negative stereotypes, even those stemming from entirely lawful actions, can, perhaps, represent a vicious cycle resulting in negative meanings being perpetuated, thus creating a further barrier toward the recruitment of cyclists.

Local institutional and political commitment
In our discussion of practice theory, we acknowledge that the three elements do not exist in isolation, but rather overlap and interact to shape a practice. However, through our study, we found that there was one issueconsistently raised by participantswhich did not readily map onto any individual element of practice theory. Rather, we observed that this element could be the creator of materials, maker of competencies and generator of meanings. Or not. This was the commitment of the local authority in encouraging and facilitating cycling in the city.
As discussed above, both Liverpool City Council and the city region of which it is a part are ostensibly committed to cyclingevidenced through cycling policies which explicitly include commitments to supporting cycling (Liverpool City Council, 2014, Liverpool City Region, 2018. The strategies emphasize measures on bicycle education and enforcement programs, but often lack details on investment and implementation of materials (e.g. dedicated cycle lanes). Whilst there is evidence in this study that such an approach may support recruitment to cycling, it does fail to tackle the more substantial issues around material deficiencies, the tension between competing mobility practices and the subsequent influence upon meanings and competencies attached to bicycling.
Given the backdrop of financial austerity (Sturzaker & Nurse, 2020) in a challenging socio-economic context, we inferred that this bias in spending reflects the relatively higher costs of provisioning materials in comparison to other measures. Yet, as one of our interviewees argues, whilst it can be difficult to get funding for cycling infrastructure, there are opportunities, for example, through the city-regional Combined Authority Growth Fund, and the national Transforming Cities Fund. Instead, the same interviewee perceived the greater barrier to be technicalfor example, lack of road space to implement cycle infrastructure.
Whilst the interviewee framed these issues as technical matters, they suggested that the local political leadership on the promotion of cycling did also create additional barriers to investment, and that decision to prioritize certain road users, for example, private motor vehicles over cyclists was in reality a political judgment. Elsewhere, evidence does support this, for example Steve Rotheram, the current Metro Mayor did include active travel in his manifesto (Rotheram, 2016), but failed to respond to specific pledges set by Cycling UK during the election campaign (Cycling UK, 2017b), even when other candidates did. Furthermore, though Rotheram did appoint a city-region Cycling Commissioner, he was amongst the last of the city region mayors to do so. Similarly, Joe Anderson, the City Mayor, did not mention active travel in his 2016 election manifesto (Anderson, 2016) whilst his actions as Mayor, such as the scrapping of Liverpool's bus lane network in 2013 were seen as deeply harmful to cycling provision.
Regardless of the cause, this lack of implementation and, perhaps, the absence of local political leadership over a long-term period was perceived by respondents as a lack of commitment by the local authority, with some saying that, as cyclists, they felt like 'second class citizens' whom the local authority did not care about. Whilst this may be an issue of communicationwith many responses suggesting an unawareness of council policy toward cycling, it also represents dissatisfaction with the perceived piecemeal approach followed thus far.

Analysis
Like others before us (Larsen, 2017b, Spotswood et al., 2015, Watson, 2013 practice theory has proved a useful conceptual framework in understanding the variety of factors that influence a choice to cycle or not. In particular, it allowed us to understand the links and interactions between elements to support or limit recruitment to cycling. This allows us to reach one of our fundamental findings: in contrast to 'cycle-friendly' cities like Copenhagen (Larsen & Funk, 2017), the lack of provision of cycle infrastructure in Liverpool places limits upon the recruitment to the practice. This then shapes the largely negative meanings attached to cycling and places a greater need for certain competencies (e.g. riding skills and alertness) to confidently cycle in the city. Practice theory also highlights how competition for materials results in tensions between competing mobility practices, which often perpetuates the negative meanings attached to cycling.
Whilst previous research suggests understanding the local context is important for a successful approach in supporting mass cycling (Larsen, 2017b, Sheldrick et al., 2017, we instead argue that, given the centrality of materials as an explanation for the stubbornly-low modal share for cycling, in urban areas which currently lack the materials identified as supporting the practice, these conclusions will, in fact, have considerable relevance.
The evidence indicates that supporting sustainable mobility patterns in urban areas, when public transport capacity may be limited, requires a reversal of the marginalization of cycling. To do so the provision of on-road infrastructure is essential to reduce the inherent tensions between mobility practices and to shape cycling to place fewer demands upon competencies of would-be cyclists, thereby broadening the recruitment pool.
Whilst in the wake of COVID-19 many urban areas have implemented temporary cycle lanes, even if these are to be made permanent, this study found this alone is not sufficient to support mass cycling. Instead, there is a need to also consider the provision of other materials such as showering and storage facilities in workplaces, residence and public spaces. In the case of Liverpool, the variability of this material provision clearly placed a further barrier upon the recruitment to cycling. Moreover, we reflect upon the secondary role that materials can play in signaling support for cycling. For example, in workplaces they meet a practical purpose (e.g. showering prior to work), but also provide signal support from the workplace, contributing toward positive meanings of cycling. Greater provision of materials may also play an important role in signaling support from political institutions (and the city as a whole), and in this case the deficiencies in visible materials may instead contribute to negative meanings, expressed in this study as 'feelings' that the city was 'unfriendly to cyclists'.
Going beyond the role of materials, and in support of previous work which highlights the central role supportive politicians play in promoting a cycle-friendly city (Larsen, 2017b), our study found that both non-cyclists and cyclists alike saw the local authority as unenthusiastic at best, or anti-cycling at worse, further contributing to the aforementioned 'feelings' of a city 'unfriendly to cyclists'. Such an attitude from local political leaders was also seen to reduce the legitimacy of cycling in the eyes of local authority officials. In the context of COVID-19, there is an additional opportunity for political leaders to advocate for active travel as an altruistic choice, contributing toward increasing public transport capacity as well as reducing environmental and health risks from transport pollutants.
Practice theory has strengths in understanding how the bicycle is not only a material of mobility, but also a practice which can encompass a series of functions including exercise and relaxation, helping the understanding of the 'corporeal experience' of cycling (Sheller & Urry, 2006). These factors aid in the understanding the process by which positive meaning are associated with cycling, even in sub-optimal conditions observed in this study. This is an illustration of how contradictions between meanings are observed, and it indicates that once an individuals is recruited to cycling positive meanings, which are derived from experiences of cycling, may began to displace, or at least counter the negative meanings that they previously attached to cycling.
Despite the widespread knowledge of such benefits, and modeling tools to account for them, for example, HEAT tool (WHO, 2019), there remains a focus upon the utilitarian aspects of time and efficiency in transport modeling (Simpson, 2017), contributing toward low investment in cycling infrastructure (Mulley et al., 2013). Analysis through practice theory can help combat this bias by illustrating show that when many motilities are competing for practitioner's time and resources , Watson, 2013, cycling can 'beat' other modes of travel, not only in time and money but also in capturing the 'corporeal experience' of cycling.
In times of disruption, it is essential not only to adapt to changing needs, but also to pro-actively shape the outcomes of disruption (Reeves et al., 2020). The insights provided through the lens of practice theory can help shape policy interventions to do so, which Watson (2013) describes as the 'motor of change', and which can add momentum and create new positive feedback loops enhancing recruitment to the practice of cycling. We therefore side with Keller et al. (2016) in that practice theory can provide a useful framework through which policy-makers can be informed. In this case, it highlights the centrality of material provisioning in driving recruitment toward cycling, meaning that if policymakers wish to both manage the challenges of sustainable mobility following COVID-19 and to meet their long-term ambitions for cycling (e.g. Liverpool City Council, 2014), then this must be their focus. Practice theory shows that materials firstly reduce the need for certain competencies, and secondly play a key role in shaping the meanings of cycling, which subsequently have a significant influence upon a decision to cycle.

Limitations of practice theory
Practice theory clearly has significant utility as a model of social life enabling researchers to understand practices such as urban cycling, and in doing so it can inform interventions to alter the nature of a practice, thereby broadening the recruitment pool to achieve policy objectives. However as previously discussed some of the themes identified through our analysis did not easily fit into one of three elements (meanings, materials, competencies). In this section, we set this out in greater detail, and consider how we can resolve this.
This challenge is likely to be present when analyzing many other social phenomena. When doing so we cannot escape that all elements of a practice are ultimately connected to the wider 'socio-technical system (Geels, 2005) that cycling, like all practices, lies with. Here, we reflect upon Schatzki's (2001) suggestion that all reductive framework including practice theory run the risk oversimplifying social life, since the social world is a made up of a series of practices, which are all too a greater or lesser extent connected. Nevertheless, like any analytic lens practice theory must simplify the social world in a bid to understand it, but this creates limitations and inconsistencies.
For example, in this study 'driver behavior' was regularly discussed, but could not readily be categorized, instead we discussed such behavior in the context of tensions between completing practices. In this context, the sharing of materials between auto-mobility and velo-mobility (e.g. road space), which then leads to tensions, offers a logic to this approach. Yet, there is a less-clear rationale for considering the wider, contextual factors in the same way, an important example is the political commitment of the local authority.
We might overcome this issue by excluding the political commitment in our analysis, citing ontological challenges in integrating actor-centered action within practice theory (e.g. the endorsement of cycling by local political leaders). Yet, if we believe that practice theory can inform policy-making (Keller et al., 2016), we must accept that our findings are ultimately enacted by the same institutions we are overlooking in our analysis, thereby leading to a contradictory position.
Instead, Larsen (2017b) illustrated the possibility of reapplying practice theory to certain factors, in their case, the actions political institutions of Copenhagen. However, given many of the challenges surrounding cycling in our study centered upon interactions with auto-mobility, should this also be separately re-examined as a practice? In both cases, following this logic creates a fresh divide between the factors which are re-examined using practice theory and those which are not. Ultimately, we ask: how do we identify the factors which are deemed worthy of re-examination as practices in their own right? Therefore, we find that practice theory does not clearly account for these wider, contextual factor without either the consideration of these elements as practices as in their own right, or by overlooking such factors, both of which we view as an unsatisfactory position. To overcome this, we suggest the introduction of a heuristic device in the application of practice theory, which attempts to 'isolate' such factors. The selection of these factor will require care, and we suggest that they must fail to clear fall within Shove et al. (2012) existing framework of element and must directly interact with the elements of the practice under examination. We tentatively call this addition to the framework 'action of others'.
The behavior of drivers is one such example, this cannot fall within the three elements, and also directly affects the availability of materials of cycling (e.g. road space). A second case is the actions of political institutions, again this cannot be seen as an element, but it directly affects the provision of new materials of cycling (e.g. safe cycle parking). In doing so, we make detailed, consistent analysis of factors influencing the elements of the practice under examination, whilst avoiding the arbitrary selection of certain elements for re-examination. It also circumvent the alternative of an iterative process of reexamination of each factor that influences a practice, which given the complexity of social process risks providing an unnecessarily complicated, and potentially less functional model of the practice under investigation. The application of this suggestion, we hope will add even greater utility to practice theory as a means to consider all kinds of activitiesnot least urban cyclingand we would invite other to utilize this suggestion in future research.

Conclusion
Across much of the UK and in many other developed nations there has been a persistently low modal share for cycling, even despite the wealth of evidence the contribution of a modal shift toward cycling can make toward many urban agendas. As policy-makers turn to cycling as a potential solution to cope with changing mobility patterns post COVID-19, there is a pressing need to understand the factors which encourage, or hinder recruitment toward everyday cycling. This study used Liverpool, UK, as a case study to do so, a city we argue is an 'ordinary city' in cycling terms, standing in contrast to studies which focus upon exemplars in urban cycling (e.g. Copenhagen and Amsterdam), as such we believe this results in findings which have broader relevance to many urban areas. The perspectives of both cyclists and non-cyclists allowed for a rich understanding of the factors which influenced a decision to cycle or not, and to explore the potential impacts of various interventions to support a greater modal shift toward cycling.
Like a number of cycling-related studies, we deployed Shove et al.'s (2012) reading of practice theory as a framework to analyze these factors, separating each into materials, meanings and competencies. This analysis led to the conclusion that the limited and variable provision of materials for cycling is the key explanation for cycling's stubbornly-low modal share within many urban areas. Participants consistently linked negative meanings (e.g. danger and fear) to the lack of appropriate materials for cycling in a car-dominated environment. It was clear that the provision of infrastructure (e.g. segregated cycle lanes) is vital to overcome these negative meanings, not least since participants felt this provision shapes the competencies required for cycling. In the absence of this infrastructure, there is a Catch-22 situation which limits cycling's modal sharewhere do potential recruits build the confidence and competencies which they feel are required to cycle?
More positively, the study found that cyclists cited a wide range of personal, and societal benefits of cycling, as well as simply enjoying the corporeal experience, comparing the mode positively in comparison to other mobility options despite the aforementioned issues. This is an indication of how positive experiences of cycling can create new meanings, creating positive feedback loops, providing an individual can be supported in beginning to cycle.
The application of practice theory was particularly useful in understanding the interaction between each of the three elements, and how this shapes the practice of, and recruitment toward cycling. However, we found some difficulty in clearly defining a number of themes raised by participants within Shove et al. (2012) framework. One approach is to manage this is to reanalyze these factors using practice theory, though there is not clarity in how such factor should be identified to avoid the pitfalls of arbitrary selection of factors for reanalysis. We also considered the exclusion of these factor from our analysis, which were largely actor-driven actions, however as we have argued we believe that practice theory can provide useful recommendations which we hope can be implemented by the same actors which this would exclude. Instead we advanced the suggestion of the introduction of a heuristic device, 'action of others', which 'isolates' actions which interact with the practice under examination. We suggest that further application of this framework in other contexts may test, and provide greater clarity to this suggestion.
Overall, the research has shown that whilst Liverpool is far from a cycling utopia, neither is it an environment in which urban cycling is incompatible. Therefore, if cycling is to not only be part of the shorter term solution to COVID-19 mobility challenge, but also in contributing toward tackling the longstanding challenges of ill-health, congestion, climate change and poor air quality then investment in cycling must be increased. In particular, this will require measures which reduce the tensions between cyclists and other road users (e.g. cycle lanes), but also by through ensuring businesses premises and development include supporting infrastructure (e.g. secure storage and shower facilities). The prevailing view of ambivalence, or even outright hostility, to cycling by the political institutions in the city (and for some the city as a whole) illustrates an important role for local political leaders in signaling support for cycling.
Whilst accepting the caveats of the importance of local context (Larsen, 2017b) and limitations of policy transfer (Sheldrick et al., 2017) we believe these findings can be applied to many other urban areas where cycling is a marginalized mobility choice, both in the UK and beyond.