Smart city technologies and figures of technical mediation

ABSTRACT

With ICTs increasingly influencing the built environment, it becomes ever more important to analyse the impact of smart city technologies on citizens’ behaviour. To explore this impact, the philosophy of technical mediation is used. This philosophy implies that our way of living is always, to some degree, constituted and transformed by technologies. This study’s objective is to assess the applicability of Dorrestijn’s model of human-technology relationships when exploring the impact of smart city technologies on citizen’s behaviour. Based on Dorrestijn’s model of figures of technical mediation, an argumentative literature review is conducted and mediating effects of smart city technologies categorized.

KEYWORDS:

• Smart city
• ICT
• technical mediation
• philosophy of technology

Introduction

Technology is key to becoming a smart city because the use of information and communication technologies (ICTs) transforms life and work within a city in fundamental ways (Nam and Pardo 2011a). Therefore, studies on smart city practices often address issues of technological infrastructure and enabling ICTs (Al-Hader et al. 2009). According to Harrison et al. (2010), being a smart city means that real-time real-world data are sourced from both physical and virtual sensors and interconnected across multiple processes, systems, organizations, industries and value chains. Washburn et al. (2009) refer to ‘a new generation of integrated hardware, software, and network technologies that provide IT systems with real-time awareness of the real world and advanced analytics to help people make more intelligent decisions’ (2).

Smart city technologies will impact on citizens’ behaviour. The impact of the built environment on citizens’ behaviour has been studied extensively (Aschwanden et al. 2011). Analysing this impact becomes even more important with the increasing influence of ICT in the built environment.

Analysing the impact of ICT on human behaviour is also the focus of the philosophy of technical mediation (Ihde 2009; Verbeek 2015). The central idea of this philosophy is that technologies mediate the relationship between humans and the world they experience (Ihde 1990; Verbeek 2006). Goeminne and Paredis (2011) explain this philosophy as follows: ‘Technologies help to shape how reality can be present for human beings, by mediating human perception and interpretation; on the other, technologies help to shape how humans are present in reality, by mediating human action and practices’ (102). In other words, humans’ perceptions and interpretations of reality are transformed when the latter is mediated by technology.

Background of the philosophy of technical mediation is the phenomenology. Phenomenology re-evaluates the modernistic separation of subject and object. The source of knowledge cannot be found in either subjective ideas or objective facts but in the intentional relationship between subject and object: ‘The human subject is always directed at objects: we cannot just “see”, “hear”, or “think”, but we always see, hear, or think something’ (Rosenberger and Verbeek 2015, 11). Postphenomenology adds the fundamental mediated character of this intentional relationship between subject and object to phenomenology: ‘there is no direct relation between subject and object, but only an “indirect” one, and technologies often function as mediators’ (ibid., 12). The human-world relation is a human-technology-world relation. The focus in postphenomenology is on understanding the different roles that technologies play in the relationships between humans and the world (Rosenberger and Verbeek 2015, 31). It is in the context of these relationships that the impact of technologies on human beings is elaborated on.

In this study, this recently developed perspective of technical mediation is used to explore the impact of ICTs in the smart city on citizens. We employ the model established by Dorrestijn (2012a, 2012c) that translates the general idea of technical mediation into more specific and applicable figures of technical mediation. The Product Impact Tool (Dorrestijn 2012a, 2012c) is meant as a practical turn to the approach of mediation. Core of this tool is a model which brings together different types of impact of technologies, categorized according to from which side technologies seize the human being. Dorrestijn distinguishes four types of contact between human beings and technologies, corresponding to four zones around the human body: to-the-hand, before-the-eye, behind-the-back, and above-the-head. The objective of this study is to assess the applicability of Dorrestijn’s model of different types of contact in human-technology relationships when exploring the impact of smart city technologies on citizens’ behaviour. By using this repertoire of figures of technical mediation, the mediating effects of smart city technologies are explored through an argumentative review of relevant social science and engineering literature.

First, we elaborate on the theory of technical mediation and Dorrestijn’s model of the human–technology relationships. Next, the role of ICTs in the smart city concept is discussed. Subsequently, the different roles that ICTs in the smart city concept play in human-technology relations are explored through an argumentative literature review of relevant social science and engineering literature. It is in the context of the different modes of mediation and interaction, as mentioned in Dorrestijn’s model, that the impact of smart city technologies is assessed. Finally, this assessment is the basis for philosophical reflection on the impact of ICTs in the smart city concept on citizens and the applicability of Dorrestijn’s model.

Technical mediation interaction and impact

The philosophy of technical mediation argues that our existence and ways of living are always, in varying but important ways, constituted and transformed by technologies (Ihde 1990; Verbeek 2005). Notably, Verbeek (2016) has made technical mediation the central concept of his research. Ihde’s postphenomenology started from the phenomenological approach of studying the human way of being-in-the-world and included technology that relation. The technological mediation approach directs attention to the concrete relationships between humans and technology, and of humans through technology to the world. Exploring the various and changing forms of human–technology relationships is seen as an important focus (Ihde 2009). Verbeek, schematized the reciprocal relations between humans-technology-world by discerning how, on the one hand, technologies mediate human beings’ perceptions of their world and, on the other hand, technologies mediate human beings’ actions towards the world. In other words, technologies, as ‘mediators’ of the relationship between users and their environment play a role in human practices and experiences (Kudina and Verbeek 2019; Verbeek 2005).

The concept of technical mediation considers human–technology relationships in a philosophically profound but non-essentialist way. It enables one to take a middle position between the extremes of instrumentalism (technologies are neutral instruments malleable in the hands of humans) and of technical determinism and substantivism (technology has a direction or essence of its own and determines society).

The technical mediation approach connects generalizing philosophical views on the meaning of technology for humans and society with how designers and engineers consider human-technology interaction. For this reason, there is convergence with the focus in product design on the user interface of technology and human-technology interaction. In an article addressing designers and design researchers, Verbeek (2016) presents the technical mediation approach as a way to deepen the analysis of human technology relationships beyond the notion of interaction. Interaction can be understood as an encounter between two pre-existing entities. The philosophical view of technical mediation adds that humans and technologies co-shape each other. The aim is to denote a composite whole where the two parts only obtain their own characteristics through their relationship. That is, technologies only acquire their function and meaning in the context of human use practices, while these human practices are shaped by technology.

From this perspective, the very notion of impact may begin to sound suspect. If mediation indicates a more profound understanding of interaction, than the notion of impact has the least profound interpretation of the three concepts since it suggests a one-way influence from product to user, where both technology and humans are conceived as separate entities. The notion of interaction better conceptualizes a two-way effect, but is nevertheless still between separate entities. The technical mediation notion provides the most profound understanding by deepening these views by conceiving an interdependency and mutual shaping of technologies and human users.

Despite this, the notion of impact remains useful. The general idea of technical mediation needs to be elaborated in more specific terms if it is to be useful in analysing the role of technology in human practice, and for critical evaluation. The most direct experience of how technology mediates human existence, at least from a critical perspective, is often in the form of impact or influence. A framework that allows one to think critically in terms of impact, but without lapsing into a pervasive technical determinism, would be most helpful. This is indeed the goal of the Product Impact Tool, which takes into account that ‘impact’ can have multiple forms and various degrees of ambivalent ethical value.

Product impact tool

In structuring this research into the effects of smart city technology, we use the Product Impact Tool (Dorrestijn 2012a, 2017). This framework provides an overview of the types of impact and dimensions of human–technology interaction. The Product Impact Tool is based on the technical mediation approach found in the philosophy of technology, and translates this into an applicable framework with modes of interaction and types of impact. In the online version of this framework, these are illustrated and further explained with examples.

The framework established by Dorrestijn (2012b) translates the general idea of technical mediation into a more specific and applicable repertoire of figures of technical mediation. It aims to bring together the three levels (impact, interaction and mediation) that can be used when approaching the relationship between humans and technology. The model as a whole expresses the idea of the technical mediation of human existence, and from various sides. The quadrants in the diagram (see Figure 1) express different modes of interaction: a division into four ways in which technologies seize humans. Further, the tool contains twelve impact types, allocated across the quadrants.

Figure 1. Product impact tool with modes of interaction and impacts (Dorrestijn, Van Der Voort, and Verbeek 2014; Dorrestijn 2012a, 2012c)

In discussions on the theoretical background of the tool (Dorrestijn 2012a, 2012c), the notion of mediation is used extensively. However, for applications in practice, it would seem preferable to refer to the more everyday notion of types of impact. The types of impact are discussed as ‘figures of technical mediation’ that constitute a range of specific instances of technical mediation against a background of technical mediation as a general idea. The four modes of interaction and the associated twelve types of impact will now be briefly introduced.

Before-the-eye: guidance, persuasion and image

The eye is used to symbolize cognitive interaction, where humans perceive and understand information or technological meanings that can impact their decisions about actions.

The first type of influence in this quadrant is guidance towards intended use. In design, this is addressed by aiming for self-evident forms and colours through product semantics (e.g. Norman (1988); Boess and Kanis (2008)), or by adding arrows and text etc. The second type reflects a more intrusive influence on human action in the form of persuasion through design, a term related to ‘persuasive technology’ (Fogg 2002). In this case, technology not only guides towards proper use but is designed to interfere in people’s behaviour, as in the case of pop-up banners on websites that, for example, persuade people to ‘buy today’ or ‘click here’. In both these types, technology addresses the human decision-making process. The third type of effect is in the expression of people’s image or lifestyle by design. For example, products such as clothing and cars (Miller 2010) allow people to shape and express their identity.

To-the-hand: coercion, embodied technology and subliminal affect

The hand refers to physical interaction, where technology influences the human body and gestures directly, bypassing decision-making processes.

The first type of influence in this quadrant is coercion. This is perhaps the most obvious of all the impacts of technology. Simple examples are a fence to control people’s access and a speed bump to force car drivers to slow down (Latour 1992). Another type, embodied technology (Ihde 1990), concerns abilities such as writing with a pencil, riding a bike or playing a musical instrument. Such activities are unthinkable without the associated artefacts, which typically must be handled with skill. Developing techniques of use (Tenner 2003) involves considerable practice, but once accomplished the discipline of learning is soon forgotten. The objects involved come to be experienced as natural extensions of the body and are smoothly integrated into routine behaviours. The third type of effect, subliminal affect, relates to being attracted or repelled by semiconscious sensations, such as when supermarkets introduce the smell of fresh bread and coffee to enhance people’s sense of hospitality and influence their buying mood.

Behind-the-back: side effects, background conditions and technical determinism

The notion of behind-the-back and the symbol of brackets (context) denote indirect effects of technology, stemming from the wider environment or infrastructure, without direct user–product interaction.

First, it is not uncommon for technologies or designs to have side effects. A product may perform its intended function well, but the advantages with respect to the primary function may be undone by disadvantages on another level. Second, the successful functioning of a product is dependent on background conditions. For example, a product may require an infrastructure for maintenance or provisioning, or its operation may require prescience and skills. The third type, technical determinism, reflects how technical developments, rather than responding to existing needs, may have a dynamic of their own and create or transform human values and needs.

Above-the-head: utopian technology, dystopian technology and ambivalent technology

What is, once all these concrete impacts are combined, the meaning of technology as a whole? Does it liberate or control humanity? Is it desirable, or dangerous, to develop behaviour-influencing products? The cloud placed above the head in the figure refers to generalizing estimates and claims about the effects of technology on human existence. The different conceptualizations refer to philosophical schools and historical periods, but are also reoccurring aspects of people’s attitudes.

The first type, utopian technology, reflects a very optimistic belief in progress through technology (a core view in modernity). Dystopian technology refers to the opposite view: the fear of domination (prevalent in the twentieth century with the nuclear bomb and ecological crisis). The third type, ambivalent technology, is the prevailing view of technology in contemporary philosophy. Although a profound hybridity of humans and technology is acknowledged, this is not evaluated with either euphoria or despair, but always with ambivalence.

Overall, the aim of the Product Impact Tool is to provide a broad overview of concepts across a range of disciplines. This categorization is not the only possible one, but it does aim to be both broad and comprehensive. Using this framework when reviewing the range of impacts of ICTs in the smart city concept is the basis for philosophical reflection on understanding the different roles that ICTs in the smart city concept play in relations between users and the world.

Enabling ICTS as prerequisites for a smart city

ICT is mentioned as the main driver for the growth of smart cities (Mora et al., 2019a). Smart cities focus on applying ICTs through embedding sensors and sensor networks in railways, bridges, tunnels, roads, buildings, water systems, dams, oil and gas pipelines and other objects (Su, Li, and Fu 2011). Through these sensors and sensor networks, data and information are collected and analysed (Lim and Maglio 2018). The smartness of cities is related to the increasing share of big data generated by sensors (Batty 2013). A smart city can therefore be distinguished by ‘the use of information and communication technology to sense, analyse and integrate the key information of core systems in running cities’ (Su, Li, and Fu 2011, 1028).

IBM (Harrison et al. 2010) defined the smart city using the following three ICT dimensions:

• Instrumentation: ‘the capture and integration of live real-world data through the use of sensors, kiosks, meters, personal devices, appliances, cameras, smart phones, implanted medical devices, the web, and other similar data-acquisition systems … ’ (ibid., 2).

• Interconnection: information ‘obtained from instrumentation data is integrated throughout an end-to-end process, system, organization, industry, or value chain. In addition, such data may be interconnected across multiple processes, systems, organizations, industries, or value chains’ (ibid., 2). Combining instrumented and interconnected systems connects the physical world to the virtual world.

• Intelligence: analysis of this interconnected information provide a ‘new insights that drive decisions and actions that improve process outcomes or system, organization, and industry value chains’ (ibid., 2).

This definition is a typical example of a technocratic conception of the smart city (Söderström, Paasche, and Klauser 2014): a technology-led vision on the development of smart cities based on a technology push (Angelidou 2015). Mora, Deakin, and Reid (2019a) relates this vision to the corporate development path: ‘urban areas become smart when they are equipped with a platform of digital solutions provided by ICT consultancies’ (ibid., 64). According Mora, Deakin, and Reid (2019b) ICT companies suggest that in their technology-led vision the introduction of ICT solutions is the only factor to be considered when discussing smart city development.

Baccarne, Mechant, and Schuurman (2014) elaborate on various concepts related to the integration of ICT in city planning and management, such as ‘digital cities’, ‘intelligent cities’ or ‘smart cities’. Each of these concepts try to capture different aspects (De Jong et al. 2015) but have also characteristics in common (Lee, Hancock, and Hu 2014). These city concepts share the focus on the effects of ICT (Kitchin 2014). Baccarne, Mechant, and Schuurman (2014) cite Loukis, Charalabidis, and Scholl (2011) stating that ‘digital cities’ are ‘extensive information systems (including network infrastructures and applications running on them) that collect and organize the digital information of the corresponding “physical cities” and provide a public information space for people living in and visiting them’ (144). Aurigi (2005) states that the vision on the city in general influences the vision on the ‘digital city’ and the way such a city will develop. When the ‘city as machine vision’ is dominant, the digital city will be characterized by ‘expert-only’ systems for decision support and city management. In the vision of ‘the accessible city of the open government’ the digital city is actively encouraging its citizens to use digital urban systems. Another notion is ‘intelligent cities’, which aim at ‘uniting, promoting, acquiring and stimulating diffusion of information’ (Baccarne, Mechant, and Schuurman 2014, 160) and combines a number of intelligent networks (Abir and Khaled 2018). According to Mora, Deakin, and Reid (2019a) ‘intelligent cities and digital cities share the same interest for ICT driven urban development’ (ibid., 63) but differences between the two concepts are also mentioned. In intelligent cities ICT facilitates the creation innovation environments. Intelligent cities offer skills, institutions and virtual spaces of cooperation for creation, dissemination and application of innovations (Komninos 2009). The evolution of urban intelligence of these cities is based on the effective combination of digital communication networks, sensors and tags and software (Mitchell 2007). New information technologies change form and development of cities, and systems of physical mobility (Graham and Marvin 2004). In effect, to be an intelligent city, one has to be a digital city. For Lee, Hancock, and Hu (2014), the ‘digital city’ and the ‘information city’ are characterised by ICT delivering innovative online services to the citizen.

According to Nam and Pardo (2011a), a well-functioning ICT infrastructure is not sufficient to become a smart city. Although a wireless infrastructure is a major technological element, in a smart city it is necessary that public institutions, the private sector and other organizations collaborate. ICT in the smart city also enables collaboration among stakeholders (Mayangsari and Novani 2015) and within the government (Viale Pereira et al. 2017). Different approaches to smart city development and implementation are characterised by technological and institutional elements (Lee, Hancock, and Hu 2014).

Comparing different city concepts, Baccarne, Mechant, and Schuurman (2014) state that the smart city focus is on ‘the involvement of all relevant stakeholders, whereas “digital cities”, “wired cities” or “ubiquitous cities” stress the presence of technological infrastructure’ (160). Mora, Deakin, and Reid (2019b). show that the level of citizens’ involvement in the development of smart city projects and initiatives differs per city. According to Caragliu, Del Bo, and Nijkamp (2011), a smart city is one in which ‘investments in human and social capital and traditional (transport) and modern (ICT) communication infrastructure fuel sustainable economic growth and a high quality of life, with a wise management of natural resources, through participatory governance’ (70). This infrastructure facilitates social, environmental, economic and cultural development (Kitchin 2014). Mora, Deakin, and Reid (2019a) speaks about a holistic development path in which digital technologies support participatory governance and community-led urban development.

The perception of technology in smart city initiatives stresses the integration and combination of systems, infrastructures and services, all mediated through enabling technologies (Nam and Pardo 2011a; Pérez González and Díaz Díaz 2015). From this perspective, smart technologies are ‘just a facilitator for creating a new type of innovative environment, which requires the comprehensive and balanced development of creative skills, innovation-oriented institutions, broadband networks, and virtual collaborative spaces’ (Nam and Pardo 2011a, 288).

The technical mediation approach, however, asserts that technology will never be limited to this facilitating role, but has effects beyond the role of ‘means to ends’ as set by humans. Technologies are not neutral, nor only passive. In line with Latour, they should be considered as somewhat closer to humans, as actors and as mediators that influence human courses of action and ways of living (Latour 1992). Verbeek (2005) therefore called for attention to ‘what things do’.

An argumentative literature review

The purpose of postphenomenological studies is ‘not to develop an accurate description of specific technologies, but to investigate the character of the various dimensions of the relations between humans and these technologies, and their impact on human practices and experiences’ (Rosenberger and Verbeek 2015, 31). Therefore ‘technological artifacts are not neutral intermediaries but actively coshape people’s being in the world: their perceptions and actions, experience, and existence’ (Verbeek 2006, 364). With the emphasis on human practices and experiences, postphenomenology views empirical studies (work of others or self-conducted studies) as the basis for philosophical reflection (Rosenberger and Verbeek 2015, 31).

The objective of this study is to make a theoretical contribution by assessing the applicability of Dorrestijn’s model when exploring impacts of ICT in the smart city on citizens’ behaviour. This exploration is the basis for philosophical reflection on understanding the different roles that ICTs in the smart city concept play in relations between users and the world. For this theoretical contribution, the different types of impact are searched for through an argumentative literature review: literature is selectively examined in order to assess the applicability of the different modes of technical mediation. This approach does not seek generalization or cumulative knowledge from what is reviewed (Davies 2000). Attention is paid to studies to that describe a specific instance of technical mediation. ¹

The search for a selected group of studies that constitute specific instances of technical mediation was based on the core concepts used by Dorrestijn (2012a) that describe the types of impact in the different modes of human–technology interaction. A search for peer-reviewed articles on these concepts in relation to smart city technologies was carried out in the Scopus, Web of Science and Google Scholar electronic databases (English language articles only). It was decided to use Google Scholar as major database because it has broader data coverage than Scopus and Web of Science. Publications were collected from this database till May 2018. In this first stage, several phrases and terms were used to identify a comprehensive selection of articles on different types of impact related to the modes of interaction between humans and smart city technologies. The strings of words that constituted the basis of the search process for the before-the-eye mediation effect was a combination of ‘smart city technologies’ in combination with ‘guidance’, ‘persuasion’, ‘nudging’, ‘steering’, ‘controlling’, ‘monitoring’, ‘image’, or ‘lifestyle’. For the to-the-hand mediation effect ‘smart city technologies’ in combination with ‘coercion’, ‘power’, ‘embodiment’, ‘tangible technology’, ‘subliminal’, or ‘consciousness’ was used. Behind-the-back mediation effects were searched for by combining ‘smart city technologies’ with ‘side effects’, ‘indirect effects’, ‘planning’, ‘participation’, ‘democracy’, ‘technology’, or ‘infrastructure’. For above-the-head mediation effects the strings of words used were a combination of ‘smart city technologies’ and ‘utopia(n)’, ‘dystopia(n)’, ‘privacy’, ‘dictatorial’, ‘surveillance’, ‘totalitarian’, and ‘regime’ (see Table 1).

Table 1. Number of publications collected

	Number of publications for each type of impact*
Category of mediation	>800	>200 and <800	<200
Before-the-eye	controlling monitoring	guidance image lifestyle	persuasion steering nudging
To-the-hand	power	tangible technology consciousness	coercion embodiment subliminal
Behind-the-back	planning participation infrastructure technology	side effects indirect effects democracy
Above-the-head	privacy	utopia(n) surveillance regime	dystopia(n) dictatorial totalitarian

*Based on word string ‘smart city technologies’ and a term or phrase for a particular type of impact in Google Scholar (May 2018).

In the second stage, to decrease the number of publications the strings ‘smart city technologies’ and combinations of phrases or terms indicating a particular type of impact within a category of mediation were used, i.e. ‘smart city technologies’ and ‘persuasion’ and ‘nudging’. This resulted in a group of 30 to 40 publications for each category of mediation. The review of these publications consisted firstly of checking the definition of mediation with the topics in the articles. Through an iterative reading of the publications, focusing on themes that could be interpreted as a particular mediation effect, the studies were related to the different impacts of a particular mediation category. Through bibliographic ‘snowballing’, reference lists in a number of articles were used to find other relevant articles, books, policy documents and reports. Besides peer-reviewed articles, the ‘grey’ literature (conference presentations, on-line papers, research reports) was also reviewed. Publications were classified by the two authors to ensure reliability of the classification. This resulted in a final selection and classification of 50 publications which were then reviewed in-depth: journal articles (24), conference papers (14), book chapters (4), books (4), working papers (3) and one policy report (see Appendix). Six publications (Sadowski and Pasquale 2015; Vanolo 2016; Calzada and Cobo 2015; Hollands 2015; Nam and Pardo 2011a; Cardullo and Kitchin 2017) were classified in more than one mediation (sub-)category.

Literature review results

Before-the-eye: guidance, persuasion, and image

In the literature, a large number of studies relate the mediation effects of smart city technologies to different forms of guidance, persuasion and related ‘smart’ lifestyles: before-the-eye smart city technology ‘makes contact with the mind and influences decisions’ (Dorrestijn 2012c).

Guidance

Various technologies are used in smart city systems and designed to guide citizens and to produce particular outcomes (Cardullo and Kitchin 2017). Meijer and Bolívar (2016) speak about smart administration that focuses on how smart city technologies are used to ‘interconnect and integrate information, processes, institutions, and physical infrastructure to better serve citizens and communities’ (400). Intelligent transportation systems (ITSs) constitute an important form of guidance in smart cities and focus on navigating citizens through a traffic network. Sensor technologies and wireless networks can also support intelligent parking management in cities.

Sherly and Somasundareswari (2015) discuss how, in the near future, sensors in cars will monitor road traffic and send the information wirelessly to a central traffic control system. This system then compiles data to feedback information to the vehicles on the road. When a road becomes busy, the central traffic control system will propose or impose speed limits to be followed by the vehicles in the congested area. Forlano et al. (2011) gives an example of a ‘congestion management’ project in New York City where using sensors, video cameras and readers to measure traffic speeds has resulted in a 10% increase in travel speeds.

Parking guidance is another smart city solution. Rather than driving around looking for space, drivers will be informed about the vacant spaces available nearby. San Francisco Metropolitan Transit Authority established a system using wireless parking sensors and smart meters that uses real-time data about parking availability and sets demand-responsive pricing in order to reduce traffic and improve parking (Forlano et al. 2011). Further, drivers will be guided on the shortest route to their destination in order to minimize emissions (Sherly and Somasundareswari 2015). This system will warn drivers about school zones and suggest alternative routes.

Persuasion

A number of studies discuss the persuasive aspects of smart city technologies. Some of them are critical, suggesting that citizens are steered and freedom is lost. Others stress the positive aspects of persuasion or ‘nudging’, in particular in changing citizens’ environmental and social behaviour.

Cardullo and Kitchin (2017) offer a critical perspective: they discuss how a citizen, by providing data, is steered ‘by way of nudging, that is, gently persuaded of how to conduct a way of life contained within optimal or ideal targets – usually around environmentally friendly use of resources or care of own body’ (8–9). When citizens are ‘nudged and steered towards specific sets of behaviour’ (7) one can speak of ‘non-participation’. These authors state that smart city technologies treat citizens as ‘people to be steered, controlled, and nudged to act in certain ways’ (11). The citizen is reduced to a consumer or even a patient who is to be ‘cured’. Klauser and Albrechtslund (2014) also critically discuss how aspects of the smart city concept, such as self-tracking and self-surveillance, are related to the trend of self-optimization, often framed ‘in terms such as “the good life”, “sustainable lifestyle”, “healthy living”, “good learning”, and “work productivity”’ (282). This ‘individual management agenda’ through the use of smart technologies is a typical example of persuasive technology.

A number of technology-focused studies however stress the positive side of the persuasive effects of smart city technologies. Here it is expected that technologies will positively impact on the behaviour and needs of citizens. Khansari, Mostashari, and Mansouri (2014) discuss how smart city technologies are capable of changing the environmental and social behaviour of citizens resulting in reduced energy consumption and sustainable travelling. Through improved access to information on resource consumption, citizens will make better use of their resources, increasing the city’s sustainability (Khansari, Mostashari, and Mansouri 2014). Energy information on various urban services will affect households’ energy behaviour. Carreras et al. (2012) discuss the European SuperHub concept that focuses on a user-centric, integrated approach to multi-modal smart city mobility systems. This concept uses persuasive technologies to raise awareness of the impact of citizens’ transport and mobility habits, thereby fostering more environmentally friendly mobility behaviour. Yigitcanlar and Lee (2014) elaborate on persuasive technologies in the Korean ‘ubiquitous-eco-city’, a Korean variant of the smart city. In u-eco-cities, citizens use an eco-mileage card that monitors electricity, water and gas consumption, as well as public transportation and bicycle use. This card provides incentives for those who use sustainable public transport systems and have a ‘sustainable lifestyle’.

Self-image or lifestyle

The critical or more pragmatic vision of the persuasive effects of smart technologies can also be found in discussions on the relation between smart cities and particular lifestyles.

Cardullo and Kitchin (2017) speak critically about a ‘smart lifestyle’ where citizens live in a smart building or area ‘served by multiple smart city technologies designed to enhance the lives of residents through improved security, energy and waste services, and transportation and parking options’ (11). A ‘smart citizen’ lives in a gated community and is ‘a high-income consumer seeking an exclusive property investment with the latest technological trimmings’ (11). Cardullo and Kitchin relate nudging to ‘lifestyle consumerism’: individuals that are nudged to act in certain ways are turned from active participants into passive consumers.

Hollands (2008) states that labelling an urban area as a smart, intelligent, wired and digital creative city is often used just for marketing purposes. This makes it difficult to separate a city’s actual smart policies from the image building. Nam and Pardo (2011b) also state that image building and marketing is a major aspect of the transition to a smart city: a ‘city’s smartness is increasingly becoming a major selling point’ (189). Marketing a city as smart attracts new talent, resources and investments.

A number of studies relate smart city technologies to a smart lifestyle for citizens in a more pragmatic way. Smart city technologies support citizens in their ‘mobile lifestyle’ (Nam and Pardo 2011a). Budde (2014) simply states that security, energy and access control systems are part of the smart lifestyle in smart homes.

To-the-hand: coercion, embodied technology and subliminal affect

The mediation effects of smart city technologies can also be discussed in terms of coercion, embodied technology and subliminal affects since the influence of to-the-hand smart city technology ‘operates through contact with the body and directs gestures’ (Dorrestijn 2012c).

Coercion

Coercion focuses on the physical influences of smart technologies. Examples of coercion are given by Shepard (2011) in ‘Too Smart City’ where future city furniture forces citizens to behave in the ‘right’ way: the smart bench discourages vagrancy and loitering by ejecting a sitter who sits too long, and a smart rubbish bin throws back the wrong kind of garbage. Street furniture thus becomes a controlling technology through physical contact.

Sadowski and Pasquale (2015) provide other examples of the physical influences of smart technologies when discussing the ‘hard power of policing technologies’. Alongside water cannons, pepper spray and tear gas in the ‘science of protest management’, there are smart technologies such as ear-paining long-range acoustic devices (LRADs) used to disperse a crowd. Drones with strobe lights, pain-inducing speakers and paintball guns are being developed. In the terms adopted by Latour (1992), police activities are delegated to technological systems.

These technologies are used alongside non-physical and psychological interventions, i.e. tracking technologies (Sadowski and Pasquale 2015). Registration and real-time data are used to prevent crowds coming together – so-called crowd dispersal tactics. Registration by the police of protesters attending demonstrations discourages people from joining. Another example of coercion is when trains are forced by the authorities to not stop at specific stations to keep people from assembling.

Embodied technology

Through increasing embodiment, digital technologies may become more an extension of one’s thoughts rather than an external device. Embodiment takes the form of extended cognition. The online and offline selves becomes blurred. This growing embodiment can also be found in the development of innovative user interfaces in the smart city.

David et al. (2011) discuss several types of user interfaces in a smart city. They state that the ‘traditional’ user interface, as used on the desktop computer, is no longer appropriate for new mobile and connected devices. An ‘environment-dependent’ interface is seen as one that ‘is closely connected with in-environment information and markers (on the wall of the corridor, on the door, on the surface of an appliance or any other surface)’ (ibid., 3). Users interact through RFID tags and QR codes.

Another form of user interface seen in smart cities is the so-called Tangible User Interface (TUI). TUIs are based on ‘embodied interaction, tangible manipulation, physical representation of data, and embeddedness in real space’ (Hornecker and Buur 2006, 437). In a smart city, tangible interfaces enable technology-mediated interactions (Behrens, Valkanova, and Brumby 2014). Users manipulate physical objects, surfaces or spaces.

Skinput technology (David et al. 2012) is another innovative user interface, and enables the skin to be used as an input surface, and provides an always available, natural portable finger-input system. By providing input through the skin, the body turns into a touch screen. In some cases, projectors are used to place visual interfaces on the body. These kinds of interfaces are body-interactive and mean that user interfaces are moving from computer keyboards directly to the human body.

Subliminal affect

The subliminal affect is described by Dorrestijn (2017) as the effect of being attracted or repelled by only half-conscious sensations. Here, the influence of technology on perception takes place in unconscious dimensions. Hayles (2012) and Förster (2016) elaborate on the subconscious influence of smart technologies and how humans, through these technologies, are embedded in smart environments. Hayles speaks of a ‘techno genesis of consciousness’ referring to how digital technologies constitute the way humans think prior to a conscious experience. Humans and technologies such as digital devices are co-evolving. Technology constitutes consciousness and cognition in subliminal ways, the so-called technological unconscious.

An example given by Hayles (2012) is the ability to retrieve and access information on a global scale. Through this ability, the entire lifeworld is perceived as being interconnected, which has a significant impact on how one thinks about one’s place in the world. Another example is the effect of self-tracking technologies. Citizens are encouraged to use digital technologies for monitoring aspects of their lives to produce personal data which are used not only for private purposes but also serve the purposes of others (Lupton 2014). These technologies can help people feel more in control of their lives and potentially compensate for the loss of fixed social structures and social ties and the increased choices in conducting one’s life (Lupton 2014). Another effect that could be interpreted as subliminal is what Gabrys (2014) describes as the changing concept of citizenship: rather than active engagement, adequate responsiveness becomes the central form of citizen’s participation.

Behind-the-back: side effects, background conditions and technical determinism

Smart city technologies can also have mediating effects in terms of side effects, background conditions and technical determinism. Behind-the-back effects refer to ‘the material infrastructure that has an impact on our actions and experiences’ (Verbeek 2015).

Side effects

Smart city technologies can, on the one hand, relate to global and multicultural open information platforms. On the other hand, data collected for profit-making platforms may not be truly open and may become ‘digitally gated communities’. Once an IT company has sufficient data, they could use algorithms to optimize several activities in the city, such as routing traffic, stationing police officers, waste management etc. According to Greenfield (2013), subjecting citizens to the logic of algorithms could result in a loss of freedom. A top-down technology-driven smart city may be a threat to democratic liberties and rights and democracy (Viitanen and Kingston 2014). However, the opposite effect is plausible: that, through the decentralizing and distributing effects of smart technologies, more power would come to reside with the citizens themselves.

Following Mumford’s city-as-megamachine concept, Ahlers et al. (2016) discuss the effects of smart city technologies on urban planning. They conceptualize the integration of smart ICTs into the urban space as the ‘Smart-City-as-Social-Machine’. This smart city also contains the Mumfordian tension between authoritarian and democratic tendencies (Mumford 1970). When discussing urban planning, Ahlers et al. (2016) state that ‘Smart-Cities-as-Social-Machines possess the latent potential to push participatory urban development … toward full citizen control’ (2). Smart city technologies may be able ‘to move away from traditional predict-plan-provide models of urban planning toward more decentralized forms of problem solving and strategic planning that more closely resembles the development of open source software where complex planning issues are solved in a distributed, Web-mediated space’ (ibid., 2).

Background conditions

Studies on the practices of smart cities often address the background conditions of technological infrastructure and enabling ICTs. The focus on infrastructure and technology stresses the accessibility and availability of urban systems (such as energy, water, mobility and the built environment) (Nam and Pardo 2011a). For Washburn et al. (2009), the core of the smart city is smart computing technologies applied to critical infrastructure components and services. They refer to ‘a new generation of integrated hardware, software, and network technologies that provide IT systems with real-time awareness of the real world and advanced analytics to help people make more intelligent decisions’ (2). Recent ICT developments, such as smart personal devices, data-sharing platforms and cloud-based solutions, have made it possible to improve the delivery of urban services (Carvalho 2014).

Al-Hader et al. (2009) specify a smart city pyramid made up of technological components: smart interfaces (i.e. shared platforms, integrated web services), smart control systems (i.e. control networks), and smart database resources (databases, database server). According to Zygiaris (2013), smart cities comprise real-time connections between RFID, wireless sensor networks, smart phones and other mobile devices enabling access and utilization of the real-world data and resources. Hu and Li (2012) state that wireless-sensor networks, when connected to the Internet, make observation data accessible anywhere at any time. This availability of real-time data is a major element of smart cities connecting the physical world with the information world. The next generation of Internet technologies will enable communication with devices attached to virtually all human-made objects.

Technical determinism

Baccarne, Mechant, and Schuurman (2014) observe that many smart city projects have a technological-deterministic nature and build ‘upon the belief that (new) media and ICT solutions can improve life in the city and that technology is the main driver to solve the complex societal challenges we face in contemporary cities’ (175). Wolfram (2012) also notes that the digital city has tended towards technical determinism, assuming that ICT will have a positive impact on social and cultural urban practices. Graham (2002) argues that, through advertising and magazine articles, ICTs suggest some ‘value-free technological panacea offering instant, limitless access to some entirely separate and disembodied on-line world’ (35).

Calzada and Cobo (2015) also criticizes the dominant technocratic mode of urban governance. Similarly, Galdon-Clavell (2013) criticizes the technical determinism of smart cities and the belief in the technological fix ‘despite a wealth of academic evidence that questions the ability of technological solutions to manage urban spaces, administrative processes or knowledge and information’ (721).

Patrício (2017) discusses how smart cities can be interpreted as a digital analogue of the Panopticon: ‘the presence of an unseen watcher keeps order more efficiently than physical violence’ (62). Smart technologies enable the surveillance of citizens such that a central authority can continuously monitor citizens, potentially leading to ‘predictive policing’. The movements of citizens are tracked by cameras, sensors, drones and large multiscreen control panels. Smart phones are one of the key sensors since they can track the activities of every person in the city. Decisions are collected and stored in vast databases. The existence of surveillance technologies affects our emotions and our behaviours (Patrício 2017). Security and surveillance technologies may result in a ‘surveillance society’. In this way, according to Vanolo (2016), the smart city could produce ‘political subjects which are suspect and suspicious, ultimately harming our freedom’ (32).

Above-the-head: utopian technology, dystopian technology and ambivalent technology

In the literature, smart city concepts and their attributes are related to utopian cities, dystopian scenarios and cyborg urbanism.

Utopian technology

In the literature, the smart city concept is – sometimes critically – related to the ideal of the utopian city. More generally, cities are nowadays a core issue in discussions on economic development (Sunley, Martin, and Tyler 2017) .

According to Hollands (2015), the smart city is today seen as the solution to ‘traffic congestion, inefficient services and economic stagnation, promising prosperity and healthy lifestyles for all. In short, the smart city symbolizes a new kind of technology-led urban utopia’ (61). Examples of such promised ‘utopian’ smart cities can be found on various IT corporate websites. Large IT and real estate firms, such as Microsoft, IBM, Cisco and Morgan Stanley, have built, in close collaboration with governments, what Calzada and Cobo (2015) refer to as ‘flagship Smart Cities’ including Songdo (South Korea), Masdar (United Arab Emirates), and PlanIT (Portugal).

According to Vanolo (2014), some smart city experiments are created, developed and managed solely by private firms. An example is the smart city of Kochi in India, sponsored by an investment holding company from Dubai. In such situations, the smart city becomes a privatopia. Vanolo (2014) provides another utopian conceptualization of the smart city: so-called ‘sustainable districts’ in ‘which human communities live in futuristic micro-environments ideally living a “zero impact” life’ (892).

Boulanger (2015) compares, in ‘Smart City: Reality or Utopia’, the utopian city ideal at the end of the Middle Ages with current debates on smart cities. He believes the smart city is not only utopian but genuinely able to foster societal needs, create ‘good’ governance in collaboration with all urban stakeholders, improve the management of urban systems and increase sustainability and energy management. Valdez, Cook, and Potter (2018) compare, in a case study, the ‘idealized smart city narrative’ with the actual processes and outcomes when developing smart aspects in a city. They conclude that, over time, smart city concepts can be beneficial for a city but that smart city projects are not universally beneficial for all stakeholders and that realizing these benefits takes time.

Dystopian technology

Several studies relate the smart city to dystopian concepts and scenarios such as the Panopticon, datacracy, technopoly and privatopia. As Vanolo (2016) puts it ‘many utopian projects are revealed to be sort of “degenerated utopias”’ (29). The dystopian impacts of smart city developments concern issues related to privacy, security and control. Radfahrer and da Cunha Pasqualin (2017) fear a ‘paternalistic, and even dictatorial arrangement’ (8) when, in a ‘smart city’, too much power is concentrated in a single, central operator.

Patrício (2017) compares the smart city with the infamous Panopticon, Jeremy Bentham’s prison design that Foucault used to theorize on surveillance. Living in a smart city that resembles the Panopticon can be interpreted as ‘existing in the state of normalized surveillance, instilling the same nature of a totalitarian society in many ways, everything being known to the central authorities’ (57). Technologies, as element of smart cities, can take on the role of a digital Panopticon (Klauser, Paasche, and Söderström 2014). Vanolo (2016) refers in this respect to the 2013 novel ‘The Circle’ by David Eggers in which everyone is constantly monitored by omnipresent intelligent cameras.

Radfahrer and Da Cunha Pasqualin (2017) relate the smart city to the concept of a datacracy: a data-based political regime that aims ‘to transform cities into machines, searching for an ever-increasing “efficiency”’ (7). A datacracy can develop in two directions: a techno-polis (a technology-enabled community) or a technopoly (a society in which technology is deified). A technopoly can be seen as a good example of the dystopian conception of a smart city.

Several authors discuss the implications of the smart city concept for society. Canton (2011) elaborates on several mega-city scenarios. One of them is the fortress city in which ‘the power elites … have created a safe, secure and sustainable city’ (55). This fortress city can also be interpreted as a smart city that has become a privatopia. Radfahrer and Da Cunha Pasqualin (2017) comment that experts and planners fear that smart cities will be developed and governed by powerful actors that override local laws and turn smart cities into enclaves for the rich.

Cyborg

In the literature, the smart city is also analysed in terms of ‘cyborg urbanization’. Given the omnipotent presence of control technologies, the body–city–technology boundaries become blurred (Sadowski and Pasquale 2015). The city dweller is ‘understood as an urban cyborg: one who doesn’t live in the city, but who lives as part of the city’ (13). The person, the building and the smart technology become ‘entangled assemblages of flesh, concrete, and information’ (13).

Gandy (2005) observed that ‘the emphasis of the cyborg on the material interface between the body and the city is perhaps most strikingly manifested in the physical infrastructure that links the human body to vast technological networks. If we understand the cyborg to be a cybernetic creation, a hybrid of machine and organism, then urban infrastructures can be conceptualized as a series of interconnecting life-support systems’ (28).

As Xu (2018) expressed it: ‘technology users become smart sensors capable of generating and acting upon information. The synergy between people and technology transforms each individual and creates cyborgs. Cities traversed by cyborgs become sensible’ (107).

Discussion and conclusions

The objective of this study is to assess the applicability of Dorrestijn’s model of different types of human-technology relationships by exploring the different roles that ICTs in the smart city concept play in relations between users and the world. It is in the context of these relations that the range of impacts of ICTs in the smart city concept is discussed. Using this framework is the basis for philosophical reflection on understanding the different roles that ICTs in the smart city concept play. In exploring the mediating effects of smart city technologies, four types of human-technology interactions have been analysed. The ‘direct’ effects, or interactions, concern the ways in which individual humans encounter technologies: cognitively (before-the-eye: interpreting technology) and physically (to-the-hand: bodily interaction with technologies). The ‘indirect’ impact concerns the effects of technology in the wider environment (behind-the-back: technology without interaction). Finally, there are ‘abstract’ ideas (above-the-head: generalizing claims that go beyond concrete examples).

Before-the-eye relationships and smart city ICT

In before-the-eye effects humans perceive and understand information or technological meanings that can impact their decisions about actions. When focusing on the before-the-eye effects of smart city technologies, one can observe a fluid transition from guidance to persuasion to lifestyle. A variety of smart city technologies are used to guide citizens and to produce particular outcomes. Intelligent transportation systems (ITSs) guide citizens through a traffic network. The transition from guidance to persuasion in smart city technologies is fluid: technologies try to change the environmental and social behaviour of citizens to reduce energy consumption and boost sustainable travelling. Citizens are ‘nudged’ towards living a smart lifestyle. A number of studies discussing the before-the-eye effects of smart city technologies are critical, arguing that citizens are steered and freedom is lost. Others stress the positive aspects of these effects, in particular in changing environmental and social behaviour. Smart city ICT not only guides towards proper use but is designed to interfere in people’s behaviour.

The before-the-eye interaction is similar to the so-called hermeneutic relationship of Ihde (2009), one of the founders of the theory of technical mediation. In a hermeneutic relationship, digital technology is used in order to read and interpret a specific representation that it provides (Rosenberger and Verbeek 2015). Smart city technologies mediate different views by providing multiple representations of different data in different forms. In terms of Ihde (2009), by means of technical mediation a smart city technology may have the ability to direct our actions or change our behavior (Angelidou 2015) by invitation and inhibition. In case of the smart city, certain technologies provide information and raise awareness of, for example, the impact of citizens’ transport and mobility habits, thereby inviting them to travel more environmentally friendly. Big data in the context of smart cities offers more informed decision-making (Batty 2013). Compared to Ihde, in Dorrestijn’s model more variants are mentioned of how technology directs actions. By applying Dorrestijn’s model of how a smart city technology invites or inhibits actions and behaviors, the impact of such a technology can be evaluated in a more differentiated way.

To-the-hand relationships and smart city ICT

To-the-hand refers to physical interaction, where technology influences the human body and gestures directly, bypassing decision-making processes. The first type of influence in this quadrant is coercion. In this type of relationship, the smart city technologies function as a ‘quasi-other’ to which users relate by obeying (intelligent) directions verbalized by a device (Hogan and Hornecker, 2011).

Within the to-the-hand quadrant, effects of increasing embodiment can be found in the development of innovative user interfaces. These innovative user interfaces, based on touch and gestures, show that physical interaction remains important in the era of smart city technologies (see also (Dorrestijn 2017)). In terms of Ihde (2009) these smart city technologies can also be seen as involving an embodied human-technology relationship. In such a relationship these technologies are experienced as natural extensions of the body and are integrated into routine behaviours. Compared to Ihde, however, Dorrestijn’s model offers more variants of how smart city technologies influence the human body and gestures directly. Embodiment is just one mode of physical interaction.

Some studies that discuss the to-the-hand effects of smart city technologies stress that these technologies become an instrument of control. Hayles (2012) sees humans and smart city technologies as co-evolving and that the latter effect the way humans think prior to their conscious experience in what amounts to a subliminal effect. These intelligent environments ‘actively interact with our behavior, most often without us explicitly noticing it’ (Verbeek 2010, 53).

Behind-the-back relationships and smart city ICT

Behind-the-back effects denote indirect effects of technology, stemming from the wider environment or infrastructure, without direct user–product interaction.

When looking at effects, one can see contradictory side effects of smart city technologies. On the one hand, a top-down technology-driven smart city may be a threat to democratic liberties, rights and democracy, resulting in a loss of freedom. Smart city technologies could centralise power and decision making (Kitchin 2014). An opposite effect could be that, through the decentralizing and distributive effects of smart city technologies, more power would come to reside with the citizens themselves. Mora, Deakin, and Reid (2019b) speaks about the dichotomy between a top-down technology-driven and the bottom-up holistic smart city development. According to Lee, Hancock, and Hu (2014) cities need to balance centralized or top-down and decentralized or bottom-up control and coordination mechanisms. In terms of technical determinism, the existence of surveillance technologies affects our emotions and moral behaviour (Patrício 2017) . The smart city could then produce ‘political subjects which are suspect and suspicious, ultimately harming our freedom’ (Vanolo 2016, 32).

Behind-the-back effects of technology are without direct user–product interaction. This interaction is similar to the background form of relationship of Ihde (2009). Intelligent environments based on smart technologies consist of a ‘continually communicating network of devices that are permanently in contact with the environment and respond to it actively and on their own initiative’ (Verbeek 2009, 233). This Internet of Things (IoT) improves the communication between devices and creates an interconnected world of things and services (Gubbi et al. 2013; Mora, Deakin, and Reid 2019b). Typical for such environments is their ‘total or partial invisibility’ (Verbeek 2009). According to Verbeek (2010) these environments are going even ‘beyond what Ihde calls the “background relation” because they actively interact with our behavior’ (53). Smart city technologies merge with the environment, and interact with their users.

Above-the head relationships and smart city ICT

According to several authors (Baccarne, Mechant, and Schuurman 2014; Wolfram 2012) another consequence of technical determinism is the belief that smart city technologies are the solution to several of a city’s societal challenges. Believing this to be true, one above-the head effect is that the smart city becomes seen as a ‘technology-led urban utopia’ (Hollands 2015) promising prosperity and healthy lifestyles for all. For others (Cardullo and Kitchin 2017; Radfahrer and da Cunha Pasqualin 2017), the ‘smart lifestyle’ is a dystopia and only feasible for the rich living in their own safe, secure and sustainable gated communities or privatopias. Some even fear a dictatorial arrangement or a totalitarian society when, in a smart city, everything is known by a single, central authority with a concentration of power.

The utopian view on smart city technology reflects a very optimistic belief in progress through technology (a core view in modernity). Smart city technologies make a city automatically “efficient, where ‘people’s needs would be satisfied instantly and intuitively’ (Angelidou 2015, 98). The so-called canonical examples of smart cities mentioned in the literature provide idealised visions (Kitchin 2015; Shelton, Zook, and Wiig 2015). According to Söderström, Paasche, and Klauser (2014) utopianism is an important aspect of the ‘corporate storytelling’ on smart cities. They speak about left- and right-wing utopias. Smart city technologies can optimize each system. The dystopian concepts assume that smart city technologies accumulate to create a system that dominates humans based on a ‘culture of control’ (Kitchin 2014). This dystopian vision on technology is in line with the classic philosophy of technology (Dorrestijn 2012a). This critical conceptualization of the relationship between humans and smart city technologies and their users is similar to Heidegger’s concept of enframing and Habermas’ colonization of the lifeworld. That is, technology becomes ‘autonomous’ at the cost of human autonomy (Ellul 1964) and humans are absorbed as parts of a ‘megamachine’ (Mumford 1970). Cowley, Joss, and Dayot (2018) challenge this dystopian perspective by concluding that the smart city being citizen centric is not only a rhetorical claim.

Conclusion

The focus in postphenomenology is on understanding the different roles that technologies play in relations between humans and world. In this study, the objective is to assess the applicability of Dorrestijn’s model of different types of contact in human-technology relationships when exploring the impact of smart city technologies on citizens’ behaviour.

In this study, the Product Impact Tool proved helpful in conducting a literature study on the impacts of technology, in this case of smart city technologies. All the four quadrants of interaction modes and all the impact types could be traced in the literature. When evaluating the various roles smart city technologies play in human-technology relations one can conclude that the direct, before-the-eye, cognitive mediation effect related to the interpretation of information provided by the technology is prominent. Guidance and persuasion are very important mediating effects in a smart city context. Less prominent, but still important, are the to-the-hand mediating effects that focus on bodily interaction with technologies. However, the direction of the behind-the-back and the above-the head effects are less clear.

The outcomes of this study can serve as input for future empirical research focusing on specific technologies and particular cities. One line for further research is assess more in detail the behind-the-back and the above-the head effects of smart city technologies. The success of smart city technologies and the liveability of a smart city could benefit from an assessment that adopts the perspective of balancing the different mediation effects in terms of both design and governance. This review can also be used to frame the debate and facilitate discussion about the social and ethical issues related to smart cities, which again can inform their governance and design.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. An argumentative literature review examines literature selectively in order to support or refute an argument or conceptual framework already established in the literature. Argumentative approaches to analysing literature can be legitimate but can also introduce bias because literature is selectively examined and interpretive elements are necessarily. This bias can be dealt with when several researchers select and examine the literature independently.

Appendix Items selected for the literature review

Table

Mediation categories		Authors	Items
Before-the-eye	Guidance	Cardullo and Kitchin (2017) Kitchin, Coletta, and McArdle (2017) Meijer and Bolívar (2016) Sherly and Somasundareswari (2015) Forlano et al. (2011)	Working paper Working paper Journal article Journal article Policy paper
	Persuasion	Cardullo and Kitchin (2017) Carreras et al. (2012) Khansari, Mostashari, and Mansouri (2014) Khansari et al. (2013) Klauser and Albrechtslund (2014) Yigitcanlar and Lee (2014)	Working paper Journal article Journal article Journal article Conference paper Journal article
	Self-image or lifestyle	Cardullo and Kitchin (2017) Hollands (2008) Nam and Pardo (2011a) Nam and Pardo (2011b) Budde (2014)	Working paper Journal article Conference paper Conference paper Journal article
To-the-hand	Coercion	Latour (1992) Sadowski and Pasquale (2015) Shepard (2011)	Book chapter Journal article Book
	Embodied technology	Behrens, Valkanova, and Brumby (2014) David et al. (2011) David et al. (2012) Hornecker and Buur (2006)	Conference paper Conference paper Conference paper Conference paper
	Subliminal affect	Gabrys () Hayles (2012) Lupton (2014) Förster (2016)	Journal article Book Conference paper Working paper
Behind-the-back	Side effects	Ahlers et al. (2016) Greenfield (2013) Mumford (1970) Viitanen and Kingston (2014)	Conference paper Book Book Journal article
	Background conditions	Al-Hader et al. (2009) Carvalho (2014) Nam and Pardo (2011a) Hu and Li (2012) Washburn et al. (2009) Zygiaris (2013)	Conference paper Journal article Conference paper Conference paper Journal article Journal article
	Technical determinism	Baccarne, Mechant, and Schuurman (2014) Calzada and Cobo (2015) Galdon-Clavell (2013) Graham (2002) Patrício (2017) Vanolo (2016) Wolfram (2012)	Book chapter Journal article Journal article Journal article Book chapter Journal article Conference paper
Above-the-head	Utopian technology	Boulanger (2015) Calzada and Cobo (2015) Hollands (2015) Valdez, Cook, and Potter (2018) Vanolo (2014) Sunley, Martin, and Tyler (2017)	Conference paper Journal article Journal article Journal article Journal article Journal article
	Dystopian technology	Canton (2011) Klauser, Paasche, and Söderström (2014) Patrício (2017) Radfahrer and Da Cunha Pasqualin (2017) Vanolo (2016)	Journal article Journal article Book chapter Conference paper Journal article
	Cyborg	Sadowski and Pasquale (2015) Gandy (2005) Xu (2018)	Journal article Journal article Book chapter