Introduction

Education technology feels stuck in a rut. For at least forty years, scholars have documented how new equipment is rapidly deployed but then quickly abandoned when educational reforms fail to materialize. In 1986, Cuban published Teachers and Machines: The Classroom Use of Technology Since 1920, describing how ‘foundation executives, educational administrators, and wholesalers’ fueled an ‘exhilaration/scientific credibility/disappointment/teacher bashing cycle’ with the rise and fall of film, radio, and television in US classrooms (5–6). A decade later, Mayes pointed to the persistence of this boom and bust pattern with teaching machines and artificial intelligence in UK universities, making allusions to the Hollywood film Groundhog Day, where the main character keeps re-living the events of a single day (1995). In 2004, Warschauer similarly claimed that the problems with deploying the internet and computers for education stem from misguided policies related to ‘digital divides’ and ‘occur again and again in technology projects around the world’ (6). By 2018, Sims introduced the concept of ‘disruptive fixation’ to describe how heterogenous, trans-local networks of public officials, philanthropists, designers, corporate executives, parents, and educational experts cling to the false promise that games, personalized learning, platforms, and other digital innovations will radically transform education.

Given this cycle of policy hype, disenchantment, and attendant scholarly critique, is it feasible or desirable to design any kind of technological intervention for educational futures? What sets us apart from the well-intentioned reformers of the past who believed ‘this time will be different’? What sets us apart from those astute critics who have compellingly argued, ‘this time, it’s the same’? Can education technology for ‘social good’ be designed in a postdigital era, with ‘devices, infrastructure, code and software saturating our everyday lives’ (Knox in Jandrić et al. 2019: 166)? In a recent contribution to this journal, Macgilchrist et al. have gestured at the contradictory entanglements that lead them to ‘simultaneously question a sense of design optimism while also optimistically designing educational interventions and research’ (2023: Para. 2). They urge us to consider what postdigital futures in education could be made if: 1) engineering approaches are decentered; 2) care is the main driver of innovation; and 3) the limits of design are critically examined. Taking the ambivalence of these scholars as ‘an imperative to pause’ (Redfield 2016: 176), this essay will look back on technologies that have been designed for social good, in order to ‘think about how to design futures that matter’ (Macgilchrist et al. 2023: Para. 2).

My inquiry into designing postdigital educational futures is informed by an ‘engaged program’ within Science and Technology Studies (STS) which seeks to ‘contribute both to some version of activist projects and to general theoretical perspectives’ (Sismondo 2008: 21). ‘Engaged STS’ recognizes how academic practices intervene on the world in non-innocent ways (Haraway 1988), and offers spaces for reflexive, theoretically-informed studies about the ethics and politics of university research, writing, and teaching (e.g., Callon and Rabeharisoa 2004; Latour 2005; Law and Singleton 2013; Sørensen and Traweek 2022). I draw especially from feminist literature that examines the role of affect in academic interventions (e.g., Verran 2001;) and that uses the concept of ‘care’ to thicken and situate sociomaterial descriptions of technoscientific practice (e.g. Mol et al. 2010; Puig de la Bellacasa 2011; Murphy 2015; see also Lindén and Lydahl 2021). This body of ‘Feminist STS’ research dates back to the 1970s, and has brought feminist, queer, postcolonial, and indigenous scholarship together to interrogate the co-constitution of science, technology, and society (Subramaniam et al. 2017; Landström 2007).Footnote 1

I will use these methodological sensibilities from Feminist STS to revisit two strands of academic work that I carried out at the intersection of Critical EdTech, STS, and Design research. The first strand relates to an empirical study that I completed as a doctoral student at UCL Institute of Education. Weaving together the insights of Callon (1984), Mol (2002), Barad (2007), and Puig de la Bellacasa (2011), my doctoral research used feminist inflected Actor-Network Theory (ANT) to study how participatory design was enacted in an academic-led global development project. This culminated in a thesis entitled ‘Theorising the design-reality gap in ICTD: Matters of care in mobile learning for Kenyan community health workers’ (2018). The second stream of work that will be revisited grew out of my interactions between 2018 and 2021 as a Visiting Fellow at Goldsmiths, University of London. At the time, Goldsmiths was home to both the Centre for Invention and Social Process and the Interaction Research Studio. Working there with designer researchers and social scientists, I hoped to learn how the ‘care-full’ practices of participatory technology design that I had observed in my doctoral research could have been made to be more ‘speculative’, and how such a ‘speculative commitment’ relates to social change. In what follows, I will recount this academic trajectory from doctoral to postdoctoral researcher, with the aim to think with Macgilchrist et al. (2023) about what futures are possible when engineering approaches are decentered, care is the main driver of innovation, and the limits of design are critically examined.

When Engineering Approaches are Decentered: How Did Digital Technology Projects for Global Development Fail?

It might seem frivolous, if not unethical, to have conducted an ANT study of an international development project for poverty alleviation (see Tronto 2021; Winner 1993). Research on the sociomateriality of technoscience does not typically lead to the roll-out or evaluation of a promising new device, program, or strategy for social reform. Frontline workers do not often find sociomaterial accounts of technoscience particularly helpful for their everyday practices. How can research that is, arguably, so theoretical, address the urgent day-to-day challenges of human rights and global development campaigns? What is the use of studying participatory design as a messy, relational, sociomaterial practice when there are teachers, doctors, nurses, and other ‘essential workers’ on the ground, struggling to make better futures through initiatives in education, primary health, clean drinking water, or food security? Giraud argues that ‘simply acknowledging that human and more-than-human worlds are entangled is not enough in itself to respond to problems born of anthropogenic activity’ and provocatively asks, ‘What comes after entanglement’ (2019: 7)?

I did not begin doctoral studies with the intent to engage with sociomaterial research. The initial impetus for my scholarship was related to a certain commitment to social justice and a rather practical concern with the opportunity costs of so many failed deployments of digital technologies for ‘social good’. My doctoral fieldwork began in 2013 and was of course subject to its own set of ‘disruptive fixations’, when dizzying advances in connectivity and mobile telephony were just beginning to allay an ‘acute anxiety with failure’ in the movement known as Information and Communication Technologies for Development (ICTD) (Avgerou 2008: 137). This movement had emerged from yet another prior bout of techno-enthusiasm that had erupted ten years earlier, when dramatic innovations in telecommunications and computing converged with the promulgation of the United Nations Millennium Development Goals (Heeks 2008). These Millennium Development Goals formed the basis of ‘an unprecedented agreement – among developed countries, developing countries, and international agencies – to work towards a world in which sustaining development and eliminating poverty would have highest priority’ (World Bank 2003: 7)Footnote 2. The convergence of technology and policymaking at the start of a new century had created a promising space for ‘making a better world with ICT’ (Walsham 2012: 89), drawing researchers and practitioners from an array of disciplines, such as Information Systems (IS), Human–Computer Interaction (HCI), Participatory Design (PD), Communication and Media Studies, and Development Studies, as well as Anthropology, Geography, and Community Informatics (Avgerou 2010; Walsham 2017).

By the time I started my doctoral research on mobile learning, the ICTD community had amassed a considerable body of practical experience and scholarship related to failures of scalability and sustainability, particularly in relation to projects involving local telecentres, and the One Laptop Per Child program championed by Nicholas Negroponte (e.g., Best and Kumar 2008; Masiero 2011; Villanueva-Mansilla and Olivera 2012; see also Pérez-Bustos et al. 2012 for a postcolonial feminist analysis). The field was addressing issues that resonate strongly with those discussed by Macgilchrist et al. in their recent commentary on postdigital educational futures (2023). ICTD researchers and practitioners were, for example, keenly aware of the shortcomings of engineering techniques that emphasize single optimal solutions for narrowly-specified problems, and had moved away from ‘trickle-down’ forms of innovation towards more socially-embedded design projects (e.g., Tongia and Subrahmanian 2006; Dearden 2008; Dearden and Rizvi 2009). They were also alert to the intractable ‘paradoxes of participation’ in international development projects (c.f. Cleaver 1999; Cooke and Kothari 2001), having confronted the ethical, political, and practical dilemmas in defining who and what constituted ‘the community’ in their participatory technology projects (e.g., Bailur 2008; Heeks 1999; Haikin and Duncombe 2014). They drew from this knowledge to frame ICTD as an ill-structured ‘wicked problem’ (e.g., Tongia and Subrahmanian 2006; Dodson et al. 2013) that involved thorny questions about 1) who should participate in design; 2) who were the optimal end users; 3) who were targeted beneficiaries, and 4) what benefits should be conferred. With its ‘single-answer’ approaches, classical engineering may have been suitable for the closed domain spaces of ‘tame problems’, but proved inadequate for the wicked domains of ICTD, which were open systems that required iterative solution-making for ever-evolving problem spaces (see Rittel and Webber 1973)Footnote 3. A ‘design challenge’ in ICTD was therefore understood as a complex ecology made up of many problems, solutions, design actors, and success criteria that could all shift and expand anytime during the life cycle of a project.

The ‘theory of design-reality gaps’ introduced by Heeks (2002) was widely used to study this ‘wicked problem’ in ICTD as a complex ecology. Drawing from sociomaterial sensibilities in STS research (e.g. Akrich 1992; Suchman 1987; Orlikowski 1996), the theoretical framework conceptualized success and failure in ICTD projects as parameters in a contingent and adaptive ‘sociotechnical system’ (see Bijker and Law 1992). As seen in Fig. 1, ’failure’ was operationalized as a function of the ‘gap’ that persists between a design – or ‘representation of an intentional future’ – and the reality or ‘local actuality of the users’ (Heeks 2002: 104–105). Design was understood as a projected, sociomaterial configuration of information, technology, processes, objectives and values, staffing and skills, management systems and structures, as well as resources. These envisioned socio-technical configurations implicitly and explicitly inscribed the economic, cultural, political, and technical context of the designer(s), as well as their assumptions about users and local environments. According to this model, projects were more likely to fail when large gaps between design and reality persisted or else widened due to changes in either the designers’ strategy or in the context of use. The chances of project success were better if initial gaps were small, and/or when designers made local improvisations to reduce or eliminate gaps.

Fig. 1
figure 1

The design-reality gap in ICTD. Adapted from Heeks (2002)

Full size image

The aim of my doctoral research was to extend this theory of design-reality gaps in order to study participatory ICTD projects that engaged with the poor as both users and designers of new digital technologies. Heeks’ sociotechnical model of success and failure had been widely applied by IS scholars to analyze failed roll-outs of government information systems and community telecentres in countries across the African, Asian, and Latin American continents (e.g. Heeks 2006; Lungo 2008; Hawari and Heeks 2010; Dasuki et al. 2015; Rugchatjaroen 2015). However there were few, if any, studies of the design-reality gaps in ICTDs designed with or exclusively by the poorFootnote 4. Most participatory ICTD projects involving poor users remained within the domain of HCI and focused on computing, user interfaces, and refining participatory methods in design (Maail 2011). Methods to evaluate participatory design projects in HCI remained constrained to transactions at the user interface, incorporating theories from cognitive ergonomics, psychology, and community engineering (Harrison et al. 2007; Ho et al. 2009; Pitula et al. 2010; Dell and Kumar 2016).

Whereas IS scholars were concerned with the sociotechnical entanglements of existing ICTD artefacts (Avgerou 2010; Walsham 2012), HCI researchers were focused on developing ethical design techniques to produce and evaluate new devices and applications (Dearden 2012; Ho et al. 2009; Toyama 2010). Yet there was a growing sense of frustration among HCI scholars who felt that their disciplinary field was too narrowly focused on quantitative outputs such as ‘task error rates’ and ‘time to task completion’, or short term qualitative assessments about user preferences and experiences (Toyama 2010). Many in HCI recognized the important effects of the wider sociomaterial ecosystem, but these insights remained peripheral to their theoretical considerations about the design and use of ICTD (Maail 2011). Narratives about the wider context of success and failure were instead dealt with anecdotally in discussion sections of scientific papers (e.g. Ho et al. 2009), or as separate reflexive accounts of researchers’ field experiences (e.g., Brewer et al. 2006; Anokwa et al. 2009). The success and failure of participatory ICTD projects involving poor users had not yet been theorized beyond the user-device relationship, in spite of the widespread acknowledgement that in order for initiatives to succeed, ‘technology alone is not enough’ (Toyama 2010: 61).

To extend these theoretical concepts related to success and failure, I set out to study the sociomateriality of a participatory action research project. My doctoral research would adopt the ‘material ecosystemic’ approach proposed by Burrell, a sociologist and computer scientist working in ICTD (2016). Like Heeks (2002), Burrell looked to STS and the methodological sensibilities of ANT to ‘dispense with the mere adoption or uptake of a technology as the primary marker and measure of project success’ (2016: 9). A material ecosystem approach, she claimed, could be used to challenge entrenched role hierarchies, engage with the specific materialities of different ICT artefacts, and ‘problematize the role of the designer in contrast to that of the user’ (Burrell 2016: 2). It could also foreground the linkages between technology and desired social change, and thereby ‘recover the possibility of a form of ethical design practice’ (Burrell 2016: 8). This, in response to prominent development scholars like Ferguson (1990), who had critiqued ‘the development industry’ as an ’anti-politics machine’ where vexing political questions about bureaucratic state power, modernization, and poverty were disguised as neutral technical problems.Footnote 5

Given their shared genealogies in STS scholarship, I proposed to use Burrell’s material ecosystemic approach to update and elaborate Heeks’ earlier model of design-reality gaps, rather than generate a completely new model of success and failure of ICTD. As mentioned earlier, Heeks had problematized the sociotechnical context of technology deployments and project failure, but the practices of designers in relation to the production of those material artefacts remained black-boxed and therefore under-theorized. This is perhaps due to the fact that the design-reality model was formulated when deployments of e-government systems were the prevailing interventions in ICTD, with designers in industrialized countries ’who create the dominant IS design’ and users in less-industrialized settings ‘who populate the local actuality’ (Heeks 2002: 104). Socio-technical contexts of design and use were thus considered ‘distant in physical, cultural, economic, and many other ways’ (Heeks 2002: 106), but such clear-cut distinctions between ‘design’ and ‘use’ were no longer evident in the growing number of socially-embedded, participatory interventions at that time. I suggested that in this new ICTD landscape, what was required was not new theoretical concepts about success and failure, but new methodological approaches to study the sociomateriality of ICTD projects when distinctions between design and use were not so clear (cf. Dourish 2021).

From October 2013 to September 2015, I conducted a multi-sited, non-representational ethnographyFootnote 6 of design practice in an academic research intervention led by my doctoral supervisors at UCL Institute of Education and University of Oxford, in partnership with AMREF Health Africa, a large Kenyan non-governmental organization based in Nairobi (Fig. 2). The project was funded for three years by the UK ESRC-DFID Joint Scheme for Research for Poverty Alleviation with the aim ‘to design, develop, implement, and evaluate a mobile phone based learning intervention to train and supervise community health workers (CHWs) in Kenya’. The proposal stated that the intervention would adopt a ‘Participatory Action Research’ methodology to ‘ensure that all stakeholders have a voice in the decision-making process’ and build on prior educational research in ‘mobile learning’ (e.g., Kukulska-Hulme et al. 2009; Kakihara and Sorensen 2002), ‘collaborative knowledge-building’ (e.g., Scardamalia and Bereiter 1994; Roschelle 1996), as well Laurillard’s ‘Conversational Framework’ (2002). The stakeholders were Kenyan nationals working as community health workers, supervisors, public health authorities, development officers, and members of other Kenyan civil society organizations. My involvement as a doctoral researcher was expected to lead to an independent piece of sociomaterial research about participation, design, success and failure that would fulfill the requirements of a thesis, bring critical reflexivity to the work of the wider participatory learning intervention, and thereby generate the project’s collective response to the Joint-Scheme’s dual appeal for ‘a more conceptual and empirical basis for development’ to ‘enhance the quality and impact of social science research which contributes to the achievement of the Millennium Development Goals (MDGs)’ (ESRC and DFID 2011: 2). Ethical approvals for this project were granted by institutional review boards at the Kenyan non-governmental organization partner organization as well as at the two UK universities. More details about the methodology and findings of this ICTD project are documented in my doctoral thesis and two related publications (Henry 2018, 2021; Henry et al. 2019).

Fig. 2
figure 2

Conducting a non-representational, multi-sited ethnography of design practice in ICTD

Full size image

When Care is the Main Driver of Innovation: Matters of Care and the Multiplicity of the Design-Reality Gap

A key takeaway from this ethnography of design practice was that the academic researchers’ mobile learning intervention was not the only ‘sociotechnical future’ in play. As described in Table 1, I observed that the same set of actors – community health workers, mobile phones, software, the ‘community’, and ‘me’Footnote 7 – were implicated in what STS scholar Mol calls a multiplicity (1999, 2002) of design-reality gaps, each involving different learning theories, different criteria for success and failure, and different understandings of social justice. During my study, the ‘active learners’ enacted by the educational research project were also constituted as ‘bottom of the pyramid’ consumers, human resources in a struggling health system, and community activists working on behalf of neglected disabled children. I argued that the mobile phone acted not just as a learning device, but as a commodity, job aid, and an advocacy tool. The software not only served as the pedagogical scaffolding for e-learning, but as a rather unassuming technological feature amongst a suite of more seductive consumer applications (see Donner 2010), a handy set of clinical guidelines to follow during home visits, and as sensitive outreach material to gain the trust of marginalized families in the community. I described how different learning theories were all in circulation – behaviorist, constructivist, and sociomaterial movement learning (see McGregor 2014 in Henry 2021) – in order to enact different competing designs for different futures. Designing and enacting educational futures did not take place in isolation from designing and enacting economic futures or health futures.

Table 1 The multiplicity of the design-reality gap. Matters of care as four sociotechnical systems
Full size table

If the ‘gap’ between ‘design’ and ‘reality’ in this participatory ICTD intervention was no longer simply a question of geographic, sociocultural, or economic ‘divides’, it could instead be approached as a lively and messy entanglement, constituted by sets of sociomaterial practices that enacted a multiplicity of desired futures. This emergence and messiness was deliberate on the part of the academic researchers, and written into the original ESRC-DFID Joint Scheme proposal. The researchers were inspired by De Laet and Mol’s canonical STS analysis of the Zimbabwe Bush Pump (2000), and had been keen to put ‘the concept of fluidity at the centre of the process and the design solution’ (Winters and Underwood 2007: Para. 4). In the mobile learning intervention, ‘fluid technology design for development’ was to be achieved by practicing established participatory research techniques and undergoing several iterative ‘cycles of design, action, reflection and re-design’. The researchers claimed no patent rights for themselves or their institutions, specifying in the grant proposal that ‘the software, applications and activities will be made freely available as open educational resources’. The researchers would eschew the ‘managerial’ approaches of ‘generals, conquerors and other exemplars of strong and solid authority’ (de Laet and Mol 2000: 251–252). Like the ‘modest inventor’ of the Bush Pump (225), they were ‘willing to serve and observe, able to listen, not seeking control, but rather daring to give themselves over to circumstances’ (253).

What emerged from this design approach was not then simply an empty gap but instead a ‘fluid space’ (Mol and Law 1994) of sociomaterial practices that shifted and expanded as the community health workers implemented the mobile learning intervention in their neighborhoods. Wheelchairs, prosthetics, baby diapers, and handmade mud bricks became implicated in a learning intervention involving mobile phones, educational researchers, solar chargers, referral forms, and a child assessment tool validated by global health experts. Research practices in public health and educational technology were conjoined with the delivery of preventive health care, informal professional learning activities, disability services, and a ‘merry-go-round’ micro-credit scheme for mothers. At first, the community health workers and academic researchers were able to juggle the competing priorities of this fluid space. But given the unmet needs of so many disabled children, it became increasingly difficult for the different sociotechnical systems to co-exist. A project that was initially designed to identify early cases of childhood neurodisabilities ended up detecting over a hundred ‘hidden children’ with severe disabilities. It led to new forums where isolated mothers of disabled children could talk about economic hardship and the grueling logistics of caregiving. They talked about the incessant accusations of witchcraft and moral impropriety. They talked about how vulnerable their children were to acts of violence and depravation. These conversations made the imperatives to ‘abandon control’ as ‘fluid designers’ feel rather irrelevant and callous. But how to intervene? Or, in the words of Mol, ‘[w]hat might it be good to do? What might the good be, here and now, in this case …?’ (2002: 169).

The four sociotechnical systems described in Table 1 no longer operated as a fluid space due to heightened emotions and constrained resources. Who should benefit most from this particular mobile learning intervention – the community health workers, the disabled children, or their mothers? I described earlier how the academic researchers had taken great measures to ensure that ‘Kenyan voices’ where incorporated into the design and deployment of the mobile learning intervention. Indeed the ‘voices’ of community health workers were ‘heard’ in their practices as consumers, public health cadres, active learners, and disability activists. What voices mattered now?Footnote 8 What justice mattered? What learning theory really mattered? Was it possible to empower community health workers as consumers, public health cadres, active learners, and disability activists – all at once? Tensions between the four sociotechnical systems grew as the community health workers and the researchers became more and more concerned about plight of the disabled children and their families. They felt empowered to make a difference. I argued that here, the researchers shifted away from the tenets of ‘fluid technology design’ (see Winters and Underwood 2007) and began to actively intervene the multiplicity of the design-reality gap. Drawing from Mol (2002), I described how the researchers engaged in the work of ‘coordination’ to ensure that the fluid space of the design-reality gap would ‘hang together’ as a durable ICTD project.

According to Mol, coordination is ‘a task’ – ’[i]t is what designing treatment entails’ (2002: 70). Coordination involves practices to assemble sets of enactments into a ‘patchwork singularity’, a ‘composite reality that is a judgment about what to do’ (2002: 71–72). I argued that to tame the tensions in the project, the researchers adopted a mode of coordination whereby one sociotechnical system wins: they established what Mol calls a hierarchy (2002: 63). In this mode of ordering, all four sociotechnical systems articulated in Table 1 could potentially be part of the mobile learning intervention, but in instances of conflict, the mobile phone’s enactment as a disability advocacy tool would prevail over its performances as a health assessment tool, as pedagogical scaffolding for e-learning, as a consumer good, and as a job aid. I described how the researchers were cognizant of the political implications of this mode of coordination. In their words, their interactions with the ’hidden children’ moved them to ‘make a judgement’, and ‘take a position on the conditions of abject poverty’. The researchers were thus deliberate in their work to empower the disabled children and their families, and conscious of the political implications of their research.

This political stance was attached to a moral claim. The researchers’ hierarchal ordering of the design-reality gap was inspired by the late physician and anthropologist Paul Farmer and his sustained engagement with liberation theology (2005). The researchers explicitly tied their ‘inherently political’ design moves to progressive Catholic ethics and the work of prominent scholars and activists in Central and South America (cf. Boff and Boff 1987; Gutiérrez 1988). Drawing on visions of ‘social justice’ and the prioritarian ethical principles associated with a ‘preferential option with the poor’, the researchers argued that ICTD should reach ‘beyond the mere analysis of social inequality towards a critique of the structures that perpetuate and produce social injustices’ (see also Winters et al. 2020a, b). Relinquishing their stance as ‘fluid inventors’, they applied an ethical standard to intervene in the fluid space of the design-reality gap, making new design moves to assert what forms of learning mattered, and for what futures.

I came to the study of ‘matters of care’ during my doctoral research because I disagreed with the researchers’ ordering of the design-reality gap. Like my counterparts on the project team, I too was deeply moved by the condition of disabled children who lived in the struggling communities where we worked. I was similarly outraged that their local needs were not acknowledged or well-funded by the powerful global Millennium Development Goals.Footnote 9 But everything I knew as a consumer, a non-Kenyan development consultant, a doctoral student, and a non-Kenyan parent of three children left me feeling unsettled. I continued to wonder about our original commitment to the community health workers. For me, the mode of coordination which had prioritized the mobile phone’s enactment as a disability advocacy tool raised questions concerning the other sociotechnical systems described in Table 1. Community health workers by definition ‘come from the communities they serve’. So even if they were indeed better off than the disabled children in their neighborhoods, their living conditions were still starkly different from my own. Community health workers may not have been the ‘poorest of the poor’ in our study sites, but could it be really that easy for us to dismiss their joy in ‘going from analog to digital’, their ‘focus in terms of financial status’, their pleasure of feeling like ‘that guy [who] knows what he is doing’ (see Henry 2018)? Would it be fair to pull the community health workers’ practices away from the structure of the formal health system and the professional opportunities it might confer? Was it ethical to hope that a small and committed group of Kenyan community health workers and UK researchers working on a three year UK research project would successfully bypass the agendas of the powerful networks promoting the Millennium Development Goals? I was ambivalent about prioritizing the mobile learning intervention’s enactments as an advocacy movement for disabled children in such an unequivocal way. And I felt angry and frustrated that my ambivalence could be subject to moral indignation in a clarion call for social justice.

Puig de la Bellacasa’s writings on ‘speculative ethics’ in ‘matters of care’ (2011, 2017) helped me to think systematically about how such highly-charged affects shaped, and were shaped by, the design-reality gap in this mobile learning project. Rooted in the Pragmatist philosophies of Isabelle Stengers, Gilles Deleuze, Alfred Whitehead, and John Dewey, speculative ethics engage empirically with how care in technoscience is practiced as an ‘an affective state, a material vital doing, and an ethico-political obligation’ (Puig de la Bellacasa 2011: 90). This feminist form of ethical reasoning builds from the same classic ANT traditions as Heeks’ theory of ICTD failure (2002) and the material ecosystem approach proposed by Burrell (2016), but also incorporates affective dimensions such as anger, worry, unease, trouble, concern and passion into its analyses of sociotechnical systems. Studies of ‘matters of care’ are also different from canonical accounts of ‘matters of fact’ and ‘matters of concern’ (see Latour 2004) because they hone in on sociotechnical gatherings of ‘neglected things’ and on how technoscience makes its ‘cuts’ or ‘exclusions’Footnote 10. With speculative ethics, ‘care’ is thus adopted as a method of ethical reasoning as well as a critical social science analytic, and is not to be conflated with motherly love, positive feelings, sentimental attachments, abstract moral principles, or inherently ‘political goods’ (Martin et al. 2015).

When the Limits of Design are Critically Examined: Participation, Design and the Speculative Ethics of Care

Decoupling care from normativity makes it possible to study ethics and politics as they co-emerge in technoscientific practice, both ‘within and against development projects, public health practices, labor stratigraphies, family planning practices, humanitarian interventions, pedagogy, family formations, and so on’ (Murphy 2015: 722). This critical ethicality is closely linked to the study of empirical ethics, where ethics ‘does not only come from outside care practices, but also from within’ (Pols 2015: 82). To study how ‘matters of care’ emerge in the world is to study how ethics have politics – and how these politics have ethics (Puig de la Bellacasa 2010). It is to also develop respectful critique that turns away from the ‘critical barbarity’ bemoaned by Latour (2004: 240) – so that instead of simply ‘debunking’ the researchers’ approach to social justice, or completely ‘dismantling’ their strategy to serve disabled children, analyzing matters of care could result in a ‘generative critique’ of the mobile learning intervention that opened up ‘possibilities for other ways of relating and living’ (Puig de la Bellacasa 2011: 99), ‘a way that things might be done differently to effect futures different from pasts’ (Verran 2001: 20).

During my doctoral fieldwork, I felt that the researchers’ recourse to prioritarian principles of justice had been too hasty, the design moves too categorical, and that not all of the issues had been adequately considered. Having unpacked the design-reality gap as a multiplicity of sociomaterial practices, I wanted to return to the enactments that had been abandoned in the name of the disabled children – enactments that included actors such as mobile phones, the newly-developed mobile application, national health strategies, and health forms. I wanted to consider how the entanglements of the design-realty gap might be tamed through the other ordering strategies that Mol has described (2002) – through design moves related to addition (68), distribution (87), and inclusion (119). What other forms of learning would have emerged? What role would the mobile phones then play? Would other principles of justice come to matter? Would this ICTD project for mobile learning have succeeded or failed? But perhaps most of all, I wanted to use matters of care to think about the tensions that I still felt in my shifting subjectivities as a consumer, a non-Kenyan development consultant, a doctoral student, and a Non-Kenyan parent of three childrenFootnote 11.

By the time I finished writing about these matters of care for my doctoral thesis, the funding period of the mobile learning intervention had long ended. If my ‘generative’ critique could no longer influence the design moves within this project, I hoped that this ‘staging of things’ (Puig de la Bellacasa 2011: 99) would nevertheless stimulate post-hoc reflections about the potentials and pitfalls of different caring relationalities described in Table 1. To my disappointment, it did not. My depictions of this design-reality gap in my thesis and subsequent academic papers and presentations (Fig. 3) did not inspire people to ask ‘For whom?’ … ‘Who cares?’ ‘What for?’ ‘Why do ‘we’ care?’, and …, ‘How to care?’? (Puig de la Bellacasa 2011: 96). In gentle feedback, one researcher from the mobile learning intervention asked, ‘[w]here is the controversy’? It seemed self-evident to most that the ethical course of action would always be to intervene on behalf of the poorest of the poor (cf. Winters et al. 2020a, b). Even if my analysis of the design-reality gap fulfilled the requirements of a doctoral degree, it did not inspire others to make a further ‘speculative effort to think about how things could be different’ (Puig de la Bellacasa 2011: 100). It did not support an engagement with the ‘speculative ethics’ in these politically-charged matters of care (Puig de la Bellacasa 2017).

Fig. 3
figure 3

’Staging of things’ at an international academic conference

Full size image

Both Latour (2004) and Puig de la Bellacasa (2011) remind us that certain aesthetics and knowledge politics go into the ‘staging’ of a technoscientific controversy for effective public debate and speculative thinking. Designers and artists have intervened in the materiality of these aesthetics and knowledge politics in order to convey matters of concern and matters of care in new and provocative ways (see Latour and Weibel 2005; Dantec and DiSalvo 2013; González and Castillo 2018). Given the lackluster response to my ‘generative critique’, I was keen to learn more about these creative interventions in STS. At Goldsmiths, there was a strong and lively tradition of research at the intersection of Design and ANT (e.g., Ward and Wilkie 2009; Wilkie 2010; Kerridge 2015). There was also a long-standing commitment to developing and testing digital technologies that operated explicitly against the grain of industry-driven solutions and design standards, most notably in places like the Interaction Research Studio (see Sengers and Gaver 2006; Gaver and Martin 2000). An interdisciplinary group of STS scholars, designers, artists, and activists at the Centre for Invention and Social Process had been engaging with ‘invention’ and ‘speculation’ as both an object of research and a methodological approach since 2005 (see Marres 2005; Lury and Wakeford 2012; Wilkie et al. 2017; Marres et al. 2018). Through my interactions as a visiting postdoctoral researcher, I hoped to learn how to craft renderings of the design-reality gap that would shift and reformulate debates about learning technologies and social reform. I imagined being able to creatively reconfigure the materials collected from my doctoral research into an academic workshop, a public engagement event, or a research proposal that would gather issues and stakeholders to participate in the ‘speculative ethics of care’ (Puig de la Bellacasa 2017).

The research community at Goldsmiths helped me to understand that speculative design is not necessarily a magic bullet. These projects can fail to open up possibilities and there are lively debates in this field as well. In her doctoral thesis (2022), Pennington describes how certain speculative design artefacts and practices have been critiqued for privileging the perspectives of professional designers (Prado de O. Martins 2014), oversimplifying complex issues (Malpass 2012: 6–7), constraining effective public engagement and debate (Kerridge 2015; Russell 2017), and can be ill-suited for unstable ’borderlands’ such as refugee camps, where ‘one person’s dystopia is another person’s reality’ (Healy 2020: 108). She argues that limited attention has been given to the ethics of speculative design projects themselves, noting that although the aim of speculative design ’is to open up public debate on the ethical implications of scientific and technological research, the ethical repercussions of the design work that is made, or the scenarios that are imagined, are rarely addressed’ (Pennington 2022: 23). Pennington explores all these critiques in her doctoral thesis, ‘Care-politics in design: Towards an inventive feminist research practice’ (2022), which investigates what a feminist STS sensibility of care might mean for the ethics of ‘design processes that are conceived as issue-gathering, problem-making and speculative’ (13).

Pennington suggests that the current debates in speculative design reflect an over-arching tension between ‘tradition and transcendence’ that has been observed in the Participatory Design community (Ehn 1988: 128–131). There is, for example, speculative work that is deliberately ‘problem-solving’, such as financial forecasting and risk analysis, which works to maintain capitalist futures. To unsettle those dominant narratives, there are also speculative processes that are ‘problem-making’, in the sense that they attempt to generate new kinds of questions and unanticipated logics for living with technology. One of the most familiar approaches to ‘problem-making’ can be found in Speculative and Critical Design (SCD) (Dunne and Raby 2013; Auger 2013), where designers’ renderings of oppositional socio-technical futures are presented in formal exhibits that aim to inspire ‘social dreaming’ ((Dunne and Raby 2013: vi). Problem-making projects associated with the Interaction Research Studio take design artefacts out of formal exhibit spaces and into communities to provoke speculative thinking about everyday technologies (Gaver and Martin 2000). Pennington’s design research focuses on a third, perhaps lesser-known orientation described as ‘inventive problem-making’, which finds speculative potential not only in design fictions and bespoke artefacts, but in what is already happening in the world (Marres et al. 2018; Wilkie et al. 2017).

Drawing from the same Pragmatist philosophies that inform Puig de la Bellacasa’s matters of care, inventive problem-making treats speculative processes as an emergent and dynamic constellation of situated, knowledge-making activities (Debaise 2017). Pennington quotes Whitehead (2022: 107), who likens speculative thinking to ‘the flight of an aeroplane’ that ‘starts from the ground of particular observation; it makes a flight in the thin air of imaginative generalisation; and it again lands for renewed observation rendered acute by rational interpretation’ (1978: 5). Inventive problem-making is thus not only concerned with the movement from the status quo to unexpected alternative futures, but also with the return back to ‘the ground’, and with the capacity to register subsequent shifts in the status quo (Fig. 4). Designing speculative events therefore involves the negotiation of the tension between ‘tradition and transcendance’, ‘steering a course’ back and forth between dominant ‘existing orders’ and ‘ungrounded speculation’ (Pennington 2022: 122). ‘Speculation’, she asserts, ‘is not an activity of the imagination that can think of any possibility; it is tied to the possibility of an order’ (109).

Fig. 4
figure 4

Amelia Earhardt’s Lockheed Model 10 Electra. See https://commons.wikimedia.org/w/index.php?curid=1308910. Accessed 26 September 2023

Full size image

Having worked 15 years as a curator of speculative projects with the Interaction Research Studio, Pennington may be more attuned to how everyday materials other than novel design artefacts – lighting, mounting, walls, tables, chairs, text, or bodies in motion – play into the successful staging of a ‘speculative flight’. As a design educator, she finds speculative potential in the materialities of routine university practices that are all too familiar to academics both within and beyond the Design Academy – practices related to lesson plans, tutorials, assessments, timetables, conferences, and deadlines, as well as departments, faculties, research proposals, projects and papers. These too, she argues, can be studied as ‘designed objects’ with potentially productive constraints that might be disrupted and then reformulated to transform their intended uses. She asserts that an ethics of care in speculative design invites us to keep noticing the materiality of those ‘things that are neglected or unfinished’ (2022: 113) so that we might continue to repair them, reconfigure their relationalities, and attune to the ‘shifts of what ‘we know’ in our practice’ (173).

Although I could not stage the speculative events that I had imagined to explore the ethics and politics of this mobile learning intervention, Pennington (2022) and Macgilchrist et al. (2023) create new openings to revisit and recuperate what has felt neglected and unfinished in my academic work. Turning ‘a caring eye’ on the mundane materialities and limitations of my rather inconsequential early career research (see Coopmans 2020), I can locate new possibilities for the theory of design-reality gaps, an ’academic artefact’ now thickened and situated with empirical data and the concepts of multiplicity (Mol 2002) and matters of care (Puig de la Bellacasa 2011, 2017). Whereas Heeks’ original design-reality framework was applied in low and middle income countries to study and prevent project failure, I propose that an updated model can be used more widely in a postdigital society to problematize the boundaries of ‘design in flux’ (Dourish 2021) and to frame contemporary debates on the Objectives and Values of successful projects. The framework in Fig. 5 illustrates how ‘Objectives’ and ‘Values’ in design can always be analyzed as components of adaptive, affectively-charged sociotechnical systems. I have slightly tinkered with the original Heeks model to emphasize that in an era where boundaries between design and use are blurry, the design-reality gap is now better-understood as a fluid space of multiplicities, and that what is arguably most pressing is to study the differences in competing ‘Objectives & Values’, rather than the disparities in ‘Information’ and ‘Technology’.

Fig. 5
figure 5

A postdigital theory of design-reality gaps (adapted from Heeks 2002 and Burrell 2016)

Full size image

Figure 5 is a reminder that the same set of actors may be enrolled in multiplicity of configurations related to ‘doing justice’ (refer to Table 1). Each sociotechnical configuration of ‘justice’ inscribes different economic, cultural, political, and technical contexts of design, as well as distinct assumptions about designers, users and their local environments. When these enactments of justice conflict, the ethical question of ‘how to intervene’ in design futures is no longer simply a technical matter of decreasing the distance between a single ‘design’ and ’reality’. Rather, the ethical question is whether to reconfigure the multiplicity of ‘justice’ - and if so, by what means of coordination (see Mol 2002). In Fig. 5, I have replaced the straight arrows in Heeks’ original model with a messy, astral formationFootnote 12 to emphasize that the design-reality gap is a sociomaterial ‘cosmos’ made up of multiple ‘contingent sets of orders, ordering processes, disorders and unknowable fluidities’ (Law and Lien 2013: 364), and that the spatiality of this cosmos can be investigated through topological, rather than Euclidean mathematical concepts (Mol and Law 1994; Law and Singleton 2005; Michael 2014). This updated conceptualization of the design-reality gap can then align with Tuck’s longstanding call for educational interventions that turn away from deficit and damage-based theories of change to ones based on desire – approaching difference ‘not [as] an absence – not something that is blocked or missing, so therefore wanting … not a hole, not a gap, not a lacking, but an exponentially growing assemblage’ (2010: 639). Putting Deleuze into conversation with indigenous knowledge systems, Tuck has forcefully argued that ‘[d]esire, multiplicitous, complicated, paradoxical, is a way to begin to explain’.Footnote 13

But dear reader, is this conceptual framework for analyzing and designing postdigital futures too burdened with specific genealogies to fit with your academic practices and intellectual loyalties? Then perhaps Fig. 5 can just serve as a speculative lure during your own moments of doubt and ambivalence – like a small pebble that hits the surface of a pond, creating ‘ripples that release the thought’ so that ‘thought augments and distorts the ripples’ (Whitehead 1968: 36 in Wilkie et al. 2017: 184). Maybe these ripples and thoughts will entice you to ‘stay with the trouble’ a little longer (Haraway 2016), ‘not letting a situation or a position – or even the acute awareness of pervasive dominations – define in advance what is or could be’ (Puig de la Bellacasa 2017: 60). Maybe ambivalence will then become a place for you to imagine multitudes of (im)possible futures, enriching your own ‘hands-on, ongoing process of recreation of ‘as well as possible’ relations’ in more than human worlds (6).

Conclusions – Towards What Justice, Towards What Educational Futures?

In response to the invitation from Macgilchrist et al. to ‘think about how to design futures that matter’ (2023), I have shown how decentering engineering approaches and designing educational technologies with care complicates and enriches our understanding of what justice and ethics entail. I have argued that in a postdigital era when ‘gaps’ between ‘design’ and ‘use’ can no longer be treated as geographical, cultural, socio-economic, or digital ‘divides’, studying design practice as a sociotechnical matter of care can bring new insights on the boundaries of design and the take-up of digital technologies for social change. It can foreground how a multiplicity of sociotechnical systems enacts justice in different and often contradictory ways, raising new kinds of ethical questions about how care is designed and practiced in our academic interventions for social reform. Studying design-realty gaps as affective matters of care rather than an empty void can support more nuanced and complex engagements with marginalized communities, by turning away from deficit and damage focused research towards ones based on desire. As Tuck asserts, ‘even when communities are broken and conquered, they are so much more than that—so much more that this incomplete story is an act of aggression’ (2009: 416). Desire-based frameworks for educational research attune to the ‘complex personhood’ of marginalized communities and acknowledge that ‘all people remember and forget, are beset by contradiction, and recognize and misrecognize themselves and others’ (Gordon 2008: 4 in Tuck 2009: 420). Tuck insists, ‘[w] can desire to be critically conscious and desire the new Jordans, even if those desires are conflicting’ (2009) (emphasis added).

This paper also aims to contribute to Knox’s cogent appeal for ‘EdTech ethics’ that engage more substantively with social justice concerns and that work to promote constructive tensions and debates rather than universal consensus (2023). Departing from the more prevalent deontological approaches to ethical reasoning in the EdTech field, my analysis of a mobile learning project demonstrates how speculative ethics from Pragmatist philosophical traditions can used be as an analytic to trace how moral principles circulate through the sociomateriality of affective, politically-charged sociotechnical systems. In this ‘ethico-political’ approach, ‘what care is and if it is good is not defined beforehand’ and ‘may be contested by comparing it to alternative practices with different notions of good care’ (Pols 2015: 81). In seeking to articulate rather than represent collective concerns around gender, race, and coloniality in technoscience (see M’charek 2010), speculative ethics of care can help move STS studies of educational technology beyond descriptive accounts of ‘entanglement’ towards the ‘ethic of incommensurability’ advocated by decolonial scholars such as Tuck and Yang (2012: 28). In educational technology as in all technoscience, ethics have politics – and these politics have ethics (see Puig de la Bellacasa 2010). By articulating rather than eliding the sociomaterial differences between our decolonializing agendas, civil rights movements, human rights campaigns, and critical research methodologies, the approaches described in this paper can contribute to the making of stronger allyships which give ‘space for a critical conversation of what justice is, or more precisely what justice wants, what it produces, whom it fails, where it operates, when it is in effect, and what it lacks’ (Tuck and Yang 2016: 3).

Finally, this reflexive essay has highlighted what Redfield describes as ‘the complicated allure of engagement in academic work’ (2016: 161), showing how the ‘field of affect’ around technology design for social good ‘stretches into a complex of contradictory reactions, mixing passion with ambivalence, hope with doubt or even despair’ (176). With the boom and bust history of educational technology, it can be tempting to either look back with anger, cynicism, and exasperation with our academic governance regimes or to just keep moving blindly ahead. This paper is inspired by STS scholars who have taken a third path, and attempted ‘care for’ their own past scholarly work while holding on to a sense of ambivalence (e.g., Singleton 1996; Verran 2001; Jerak-Zuiderent 2019;). Resonating with Pennington’s claim that there is speculative potential in mundane scholarly practices, Coopmans writes,’ [t]he speculative question then is what happens if we would, for a moment, treat the fruits of past efforts as ‘others’ to care for and discover something about ‘how to care’ with’ (2020: 146). She suggests that caring for ‘work we feel ambivalent about, is a way to be kind to ourselves and open to learning about what moves us; it also has potential for relating differently individually and collectively – to the job of being an academic’ (2020: 151). In his allusions to the film Groundhog Day, Mayes makes the related argument that to escape cycles of disappointment and hype in educational technology, the research community needs to learn how to comprehend and register these ebbs and flows in novel ways (1995). Holding on to ambivalence can create new and different opportunities for critical and open deliberation, where unforeseen possibilities might emerge out of uncertainty and hesitations (Roberts 2018). As Redfield observed (2016), social critiques of technology may duly inspire indignation and denunciations – but ambivalence that leads to sustained inquiry is a good alternative to despair, if hope still feels elusive.