British Educational Research Journal

Introduction

Innovative learning spaces emerged in response to the influx of educational technologies and new social practices associated with increasingly participatory forms of learning. Using them has resulted in calls for better alignment between the design of educational environments and transformation in school and university curricula (Woolner, 2010; Blackmore et al., 2011; Leiringer & Cardellino, 2011; OECD, 2013; Cardellino et al., 2017; Woolner et al., 2018). However, many educators are struggling to align their pedagogical models with these new spaces for learning (Beetham & Sharpe, 2013; Scott, 2015; Singh & Hassan, 2017). Despite widespread recognition that ‘the “transmission” or lecture model is highly ineffective for teaching twenty‐first century competencies and skills’ (Scott, 2015, p. 1), direct instruction, inauthentic assessment and rote learning are still common across many contexts, although these practices are not specific to particular sectors or subject areas. For example, in universities, lectures are still common in the early stages of undergraduate courses, and in schools some subject areas are more likely than others to adopt project‐based or inquiry‐based approaches. What is more, deterministic accounts of tools and spaces for learning tend to focus on identifying generic and decontextualised properties of tools or spaces, without considering the qualities of the objects themselves, and how these, in turn, may influence people, their values and purposeful action.

In this article, we argue that those involved in educational design (e.g. teachers, space planners, architects, instructional designers) need analytical tools capable of increasing the correspondence between (a) pedagogy, place and people and (b) theory, design and practice. When we speak of correspondence, we draw on the work of Tim Ingold (2013), who contrasts interaction with correspondence, which he illustrates with a simple sketch: two fixed points with an arrow between them—representing interaction— and two lines issuing from each of these points that flex in response to movement in the other—representing correspondence. Understanding how we use this term is important for two reasons. First, our work is deeply rooted in material accounts of situated learning activity, which means we are not satisfied by descriptions of learning in which the role of materials is either absent or overly deterministic. Second, our aim is to understand the learning whole in action, which means we are always concerned with how changes in one aspect of the learning environment are expressed in another.

Materials, once unquestioningly described as unchanging, are now more often described as fluid. But this general state of flux is not new. Things (objects, thoughts, practices) and people have always been subject to cycles of growth and decay (Hodder, 2012, 2014). What is new is the rate at which they change, which has altered our perception of their relative stability (Thomas & Brown, 2011; Sewlyn, 2014). It is therefore not surprising that many of these new things have been termed disruptors, because they have radically reordered how we make our way in the world—whether working, travelling, shopping, banking, reading, writing, resting, staying connected, learning to stay healthy, learning to solve problems or just learning to muddle through. What is more, these alterations in daily life are not neutral. They reveal values and highlight the challenges associated with acting in accordance with these values. This values‐based underpinning of design is often overlooked when it comes to educational design—or design for learning—and this makes bridging the theory–praxis divide difficult (Goodyear et al., 2006). Our aim, when working with educational designers, is to support the articulation of a shared epistemology of learning and develop creative ways of keeping it visible through the design process, in the final design, and in ongoing teaching and learning practice.

Learning has traditionally been described in terms of a change in behaviour or cognitive processes, with a focus on demonstrating a unidirectional transfer of a stable body of knowledge. However, this definition no longer reflects what is known about how people learn, nor does it reveal the complexity of orchestrating or navigating the diverse assemblage of tools, tasks and people required to demonstrate knowledgeable action in the world. In calling for a richer conceptual repertoire, Säljö (2009) highlights the role of time, situatedness and reciprocity between individuals and cultural practices in learning. New theories of learning—such as embodied or extended cognition (Clark, 2011)—build on the situated nature of learning (Lave & Wenger, 1991), emphasising the value of learning through ‘doing’ in the company of others (Wenger, 1999), and acknowledge both the sociocultural (Vygotsky, 1978) and sociomaterial (Sørensen, 2009; Fenwick et al., 2011) nature of learning. Säljö (2009) suggests that the value of learning theory lies in its explanatory power relative to a set of issues, and not in its ability to reflect a particular view of the world. Whilst we heartily agree with the first sentiment, our commitment to situated ways of knowing leads us to challenge the second. Theories shape our thinking—they are tools for thinking—but our thinking is also profoundly shaped by our view of the world or what we value.

In this article, we introduce an analytical approach to framing learning entanglement that combines work from networked learning (Goodyear & Carvalho, 2014; Yeoman, 2015, 2018), architecture (Alexander et al., 1977) and archaeology (Hodder, 2012, 2014). We take an activity‐centred approach that can best be described in ecological terms. That is, our focus is on the learning whole (people, place and pedagogy) and our aim is to support the analysis and design of complex networked learning environments. As such, we conceptualise learning as an emergent phenomenon and it is therefore fitting that we start by defining emergence. Alexander's (2002) analogy of the whirlpool is helpful in this regard. In describing the whirlpool, he notes that it is not an object, but a momentary vortex induced by the passage of water through a particular configuration of riverbed, riverbanks and rocks. That is, the vortex did not exist along with the riverbed, riverbanks and rocks but was induced in the action of the whole. In exploring learning activity, through the lens of emergence, we face two challenges: first, to identify the elements—the equivalent of the riverbed, riverbanks and rocks; second, to understand how their properties and spatial configuration give rise to the emergent phenomenon we wish to stimulate.

This article starts with a brief overview of the learning theories that inform our work, followed by a review of some of the frameworks currently used to model the learning landscape and an introduction to two analytical tools—the activity‐centred analysis and design (ACAD) framework (Goodyear & Carvalho, 2014) and the ACAD wireframe (Yeoman, 2015, 2018). Having highlighted the importance of gaining a deeper appreciation of the qualities and temporalities of material and digital elements of innovative learning spaces, we draw on Hodder's (2012) theory of entanglement to extend the analytical power of the ACAD framework and wireframe, before demonstrating the power of this approach by using it to analyse a vignette from an ethnographic study conducted in an innovative Kindergarten to Year 12 school (Yeoman, 2015). In conclusion, we argue that framing learning entanglement in this way supports the identification of elements open to design and increases correspondence across scale levels and dimensions of design—bringing pedagogy, place and people together—through the enactment of knowledgeable action in the process of designing for learning.

Theoretical lenses: Framing learning activity

Learning theories, such as distributed and embodied cognition (Hutchins, 1995; Clark, 2011), inform our understanding of the learning landscape. In distributed cognition, cognition emerges through interaction with a range of distributed elements in a system (Hutchins, 1995). Minds are understood as extended outwards, not contained within brains or bodies, with artefacts playing a significant role in reasoning processes. From this perspective, minds and the environment are recast as a system that connects bodies, minds and technologies (Clark, 2011; Kirsh, 2013). People's conceptions and beliefs are therefore grounded in their perceptual‐action experiences with artefacts, and so ‘the more we have tool mediated experiences the more our understanding of the world is situated in the way we interact through tools’ (Kirsh, 2013, p. 3:3). Essentially, it is not only cognition that influences our behaviour, but our perceptual system as well. Both cognition and perception work as a system to find alignment between our actions and our predictions about the environment, which continuously co‐evolve (Markauskaite & Goodyear, 2017). This systemic view of cognition has important ramifications for design and learning, since ‘in building our physical and social worlds, we build (or rather, we massively reconfigure) our minds and capacities of thought and reason’ (Clark, 2011, p. xxviii).

We acknowledge learning as epistemically, physically and socially situated and foreground the social nature of knowledge through the notion of ‘networks’. Networked learning stresses the importance of collaboration and participation in fostering co‐creation and supporting people's engagement in knowledge‐building processes. Networked learning is about connections—between people and resources, often promoted via technology (Goodyear et al., 2004). Moreover, learning—understood as sustained or persistent change in behaviour based on experience, as distinct from development or maturation (Illeris, 2009)—is extended to include change that is incidental, unintended and sometimes not directly visible (Damşa et al., 2010). Bringing these learning theories together with the notion of emergence and drawing on the work of Ingold (2011, 2012, 2013), we offer a description of learning that emphasises the role of perception in learning or in developing sensitivity to cues in the environment and a corresponding ability to match one's actions to alterations in the environment without disrupting the flow of one's actions. We argue that this sensitivity is crucial, in conjunction with an appreciation for different forms of knowledge and ways of knowing, in supporting knowledgeable action and actionable knowledge (Markauskaite & Goodyear, 2017) in a world that is perpetually in motion.

Technological innovations alter our social practices (Thomas & Brown, 2011). These alterations in practice have implications for how we perceive, use and equip physical spaces for knowledge‐oriented activity. Acknowledging this is vital, because learners need opportunities to practice and refine their ability to collaborate and maintain connections in networked structures, something many traditional pedagogical approaches are ill‐equipped to support. This is concerning, given the types of challenges they will face upon entering the work force, including energy–water–land issues, income–health–education inequality, globalisation and human migration. Addressing these issues will require both individual insight and aggregated perspectives from a range of disciplines. As such, digital fluency, innovation, creativity and collaboration are more than the touchstones of current thinking, they are necessary for making our way in a world that is perpetually in motion (Gatt & Ingold, 2013).

Human knowledge and interaction cannot be divorced from the world. To do so is to study a disembodied intelligence, one that is artificial, unreal, and uncharacteristic of actual behaviour. What really matters is the situation and the parts people play. One cannot look at just the situation, or just the environment, or just the person: To do so is to destroy the very phenomenon of interest. (Norman, 1993, p. 4)

In this instance, the phenomenon—or spatiotemporal object of our attention—is learning activity. What learners actually do at learntime (when learning activity unfolds) is what matters most, not what educators plan in advance for them to do. Educational design needs to be understood as related to learners’ activity—because their plans may influence but do not entirely determine what learners will do, as learners have agency to reconfigure what is proposed or reshape their learning environs. This is a subtle but important difference, which is based on our understanding of learning activity as an emergent phenomenon. This way of conceptualising learning relies on the ability to distinguish between that which is open to alteration through design, and that which is not. Our focus is on understanding how ‘designable elements’ can be said to support ‘emergent phenomena’ in learning. As the demands and complexity of how and where we learn have increased, there have been attempts to map the learning landscape—that is to say, we are not without cartographers.

Modelling the learning landscape

There are conceptual models that offer stylistic representations of the constituent parts of learning environments, such as the school climate model (Owens & Valesky, 2007). Some of these models begin to describe relational elements, such as the nested nature of designed environments for learning—from the classroom to the city (Nordquist & Laing, 2015). Other models show how new technologies are shaping spaces for learning, such as Radcliffe's (2009) pedagogy–space–technology (PST) framework, and how different modalities of learning map to particular communities of learners and different degrees of formality in the built environment, such as Wilson's (2009) places for learning spectrum. Each of these representations attempts to scaffold the exploration of a specific subset of challenges. The school climate model (Owens & Valesky, 2007) used by Gislason (2009) is derived from organisational theory and offers visual cues to consider the needs of different stakeholders. The networked learning landscape model (Nordquist & Laing, 2015) reminds designers that learning is never contained within a single set of four walls, and the PST framework (Radcliffe, 2009) aids in designing, embedding and extending the built environment for learning through the use of new technologies. Each aims to map the learning whole, but most fail to clearly articulate how their epistemology of learning shapes their representation or model.

Some of the challenges associated with creating representations of thought include how one theorises the elements, the relations between the elements and the degree to which proximity on the page represents proximity or coherence on the ground. A pyramid suggests that what is represented at one level is necessary for the next level up, which inadvertently communicates causal relationships. A Venn diagram asks only that you name the three or four circles and work towards the point of maximal overlap, which obviates the need to justify or theorise the relations between the categories. A matrix offers cross‐mapping and scale levels but cannot account for time, and a flow chart or process map can be subject to the challenges associated with causality, theorisation and time.

Despite these challenges, in our work we chose to use a framework and a derivative grid for representation to guide both the analysis and design of complex learning environments. This choice is a function of our commitment to producing knowledge that is actionable. The etymology of frame is ‘framian’ (to be useful or make ready for use), and a framework is defined as a basic structure underlying a system, concept or text (Oxford English Dictionary online). As such, we present the activity‐centred analysis and design framework (Goodyear & Carvalho, 2014) with the aim of illustrating both its explanatory power and its ability to support design for learning by connecting learning activity to the designable elements of any learning situation.

Framing learning entanglement

The clearly articulated heart of ACAD is emergent learning activity—what people actually do, their thoughts and feelings—which cannot be predicted in advance. Drawing on an extended body of learning theories, including embodied and distributed cognition, ACAD foregrounds how activity is shaped by tools, tasks and social arrangements (Goodyear & Carvalho, 2014), designed in advance but variously enrolled in activity at learntime. In framing this complexity, ACAD acknowledges that learning is socially, physically and epistemically situated. There are four structural dimensions to the ACAD framework (Figure 1); three are open to alteration through design and the fourth is not. This is an important distinction to hold onto, because it is the link through which we can begin the task of connecting emergent learning activity to designable elements of the learning environment.

Building on the analytical concepts of the ACAD framework (Goodyear & Carvalho, 2014), in combination with Goodyear's (1999) earlier notions of pedagogical frameworks and the work on pattern languages by Alexander et al. (1977), we now turn to the ACAD wireframe (Yeoman, 2015, 2018). A wireframe is usually an unadorned outline used by interface designers to map elements of design to functionality before prototyping. The single‐view outline invites designers to consider the designable element to be placed in each blank space, how this selection relates to the whole and how it supports valued learning activity. Wireframes are shareable representations that can be annotated and revised by both designers and users. The act of sharing them with users has been shown to improve the quality of feedback by using cognitive walk‐throughs (Roth et al., 2016), and developing them in partnership with users increases acceptance and use of the final design (Morson, 2014). The ACAD wireframe provides a grid within which one builds a representation, a single view of the designable elements of any learning ecology (Table 1). The first three dimensions of design (as per the ACAD framework) are represented from left to right (set, epistemic and social) and the scale levels across which they operate (macro, meso and micro) from top to bottom. The shorthand used for the scale levels (macro, meso, micro) is useful, but for a deeper understanding of the distinctions between these levels we draw on Alexander et al.'s (1977) notation: region (macro), shape (meso) and detail (micro).

The work of Alexander et al. (1977) originated in the field of architecture where, through the pairing of recurrent architectural problems with potential solutions, they created what they referred to as design patterns. Each pattern contains a written narrative of a single problem‐solution, and a rationale and description of one way to address the issue of concern. Together, a collection of patterns forms a pattern language. In laying out their pattern language, Alexander et al. (1977) started with the global, because these are the patterns that conceptually order scale, shape and connection. These Level I (macro) patterns frame the broad context for the design and are built through the cumulative acts of a community over time. Level II (meso) patterns detail the shape or structure of things, or groups of things, and the spaces between them. Level III (micro) patterns provide the details for Level II patterns and can be built in a single act by an individual.

In Table 1, under the header for each dimension of design is a prompt to articulate a specific high‐level philosophy of learning, with particular reference to that dimension of design. This part of the wireframe enacts ideas from the pedagogical framework proposed by Goodyear (1999), where correspondence between philosophy, high‐level pedagogy, strategy and tactics is discussed. The positioning of the high‐level philosophy in the first cells of the ACAD wireframe is intentional and addresses something that is often overlooked in educational design work: the fact that design is inherently value‐based. Unless this is explicitly addressed, it can be difficult to bridge the theory–praxis divide (Goodyear et al., 2006). This is particularly relevant when working with heterogeneous teams of educational designers, including a range of teachers, working alongside space planners, instructional and architectural designers and others.

Our intention is to support correspondence between dimensions of design (left to right) and scale levels (top to bottom). However, reaching consensus about a shared epistemology of learning in these diverse teams is the crucial first step, and failing to do so often results in dissonance across scale levels and/or dimensions of design. A common example of this type of dissonance arises when, instead of clearly articulating epistemologies of learning upfront, project user groups start with aspirational visions of newness (macro‐epistemic) that drive design briefs for innovative buildings and technology (macro‐set) which, if enacted in the presence of hierarchical organisational forms (macro‐social), can result in aesthetically pleasing environments that speak of newness but fail to give rise to the desired quality of learning activity. That is, when working towards coherence, everything is on the table, and a desire for collaboration and a need for compliance are often at the heart of dissonance. We argue that this challenge can be addressed through an exploration of the material properties of the designed environment and the quality of learning activity they support.

At its simplest, dissonance is easy to identify when imagining a project‐based tutorial held in a raked lecture theatre for 250 people, or a lecture delivered in a flat‐floor computer lab for 120 people. But face‐to‐face learning activity is increasingly less homogeneous. It is not difficult to imagine the need to switch between periods of individual reflection, whole‐group presentation, small‐group bench work, individual/group computer use and face‐to‐face collaboration, without having to allocate chunks of time for students to move between venues. Increasing demand for access to learning environments characterised by diverse space typologies highlights the importance of understanding how lines of sight, acoustics, group orientation and a shared internal or external focus of attention can be said to support valued learning activity.

In summary, the ACAD wireframe (Yeoman, 2015, 2018) is a representational tool for thinking about where and how we learn. It offers a simple outline that calls for a clear articulation of learning theory and directs attention to those elements of any learning situation that are open to alteration through design, across multiple scale levels. On a practical level, it acts as a visual aid to investigate a single element of one dimension, without losing sight of the learning whole. At a theoretical level it acts as a translation device (Bernstein, 2000) that helps designers to operationalise the conceptual underpinnings of the ACAD framework (Goodyear & Carvalho, 2014), the pedagogical framework (Goodyear, 1999) and Alexander et al.'s (1977) pattern language. Having examined the theoretical evolution of this approach and how the properties and spatial configuration of this tool support a certain quality of design activity—be it analysis of what is or design of what is to come—in the next section, we extend the reach of this approach by exploring the role of materials and the entanglement of humans and things in activity, before illustrating the power of our approach in action.

Representing entanglement

People's interactions with tools are not neutral; they are influenced by goals and guided by implicit suggestions embedded in the design of tools that support perception, thought and action. But to stop at affordance—what this tool can do for me—is to perpetuate what Ingold (2011) refers to as object blindness or what Devall and Sessions (1985) refer to as a shallow ecological perspective. Materials are neither mute nor inert; their properties influence the quality of any interaction, and this results in complex webs of (enabling) dependence and (constraining) dependency between humans and things (Hodder, 2012). This is important when it comes to design for learning, because if we subscribe to situative and embodied theories of learning, then we need good ways of theorising how the properties of a particular tool can be said to support a desired quality of learning activity. What is more, it is not only learners who are influenced by the properties and spatial configurations of the tools available to them. Educational designers are subject to similar influences. These ideas are foundational to our research, which involves creating tangible analytical tools to facilitate the work of educational designers.

We draw on perspectives peripheral to education to help us understand what it really means to say that learning is situated, embodied and distributed. After all, archaeologists and anthropologists have well‐established methods for studying the role of tools in individual and collective human activity. Describing the efforts of others to examine the increasing complexity of human life in terms of networks, meshes, mixes, chains and engagements, Hodder (2014) observes a tendency for archaeologists to speak in terms of the enchainment of humans and things, and for sociologists to speak of interpersonal relations. In contrast, those working under the banner of actor–network theory (Latour, 1988; Knorr Cetina, 1999; Law, 2002) reveal how things—such as engines, measuring instruments and laboratory probes—are enrolled in the structuring of social relations. This work has affected a widespread shift towards relationality more generally (Law & Hassard, 1999; Latour, 2007). As a consequence, the dualisms of agency and structure, human and non‐human, knowledge and power, before and after, material and social are no longer taken as given or fixed, but as the effects or outcomes of assemblages. This shift has been so marked that ‘it is now accepted that human existence and social life depend on material things and are entangled with them: humans and things are relationally produced’ (Hodder, 2014, p. 19). However, both Hodder (2012, 2014) and Ingold (2011, 2012) remain critical of purely relational approaches, because such approaches often demonstrate a cultivated disinterest in the very things they study, their relations to other things and the ecologies of things within which they exist and function. That is, many of these relational approaches to materiality work within a shallow or use‐value notion of ecological or systems thinking. Rather than networks or meshworks, Hodder (2012) proposes a dialectical tension between enabling dependence and constraining dependency, resulting in what he calls sticky entrapment—a state in which choices, once made, limit the range of subsequent opportunities for future action. This sticky entrapment is a function of asymmetrical relations between humans and things, and it is this relational dependence that is at the heart of Hodder's (2012, p. 217) theory of entanglement, which can be summarised as follows:

Entanglement + fittingness + conjunctural event → problem → fixing → selection → E′ (total entanglement)

In summary, entanglement starts with the dialectic between (enabling) dependence and (constraining) dependency, between humans and things, where the sum of all dependences between humans (H) and things (T) in their many forms (HT, TT, TH, HH) is described as giving rise to entanglement. Fittingness is described, not only in terms of function or affordance, but also in terms of fit or the coherence of the whole. The centrality of time, and the effects of order and sequence, are acknowledged in human–thing entanglements. Combinations of circumstances give rise to conjunctural events, which create problems that require fixing, and solutions are selected from what is to hand that is contextually appropriate, resulting in an alteration to the entanglement of the whole.

Hodder's (2012) theory has far broader application than in archaeology alone. The ‘foregrounding of material stuff, not just as material meanings and social processes but also as matter that affects us, is a key part of an adequate social theory’ (Hodder, 2012, p. 211). Oliver (2013) makes a case for non‐deterministic theories of technology in educational research. We agree, but argue for a broader conceptualisation of materials in learning, noting that materials do not determine human action either by material necessity or practical convention. But in focusing on flows of matter, energy, information and human–thing dependences, one renders visible entanglements that are heterogeneous and not materially determinative. Moreover, having untangled our small bit of the world, we should always remember the intention is not to pull things apart but to ‘explore entanglement itself, engaging in thick, rich, contextual analysis’ (Hodder, 2012, p. 218).

Networked learning involves multi‐layered assemblages of artefacts, tools, places, ideas and people (Carvalho & Goodyear, 2014) and networks are, by their very definition, relational. In what follows we illustrate how the ACAD framework and wireframe, in combination with Hodder's (2012) theory of entanglement, supports analysis and design for learning in the twenty‐first century. We start with a rich description of a moment of learning activity before analysing it in two moves. In the first, we use the ACAD wireframe to scaffold our description; in the second, we trace learning entanglement using Hodder's (2012) theory of entanglement.

Framing a moment of learning activity: Theory in action

The vignette presented below is taken from an ethnographic study (Yeoman, 2015) conducted in an innovative Kindergarten to Year 12 school in Sydney, Australia. The fieldwork for this study was carried out over the course of a year and involved 549 hours of observation in open Year 5 and Year 6 learning space that was home to 181 students and their team of seven teachers. Whilst this moment is certainly remarkable, it shares many of the distinctive qualities characteristic of learning in this place: a measured pace, acceptance of difference, appropriate levels of support and independence, a willingness to reach or dig deep, and respect for tools that did not limit their appropriation in creative teaching and learning practice.

A lesson in 24‐hour time. It's the end of a two‐week numeracy cycle. This means concepts have been presented, set work has been completed, and most are working independently on extension tasks selected from their online learning environment. To my left, Ms Talbot is working with a small group of eight children still struggling with the concept of 24‐hour time. She gets them up on their feet in pairs and asks each to describe what they would be doing at x am and x pm respectively. The pairs get side tracked. Their linear arrangement leads to confusion about where morning ends and night begins, and they are more interested in talking about variations in bedtime than they are about the distinction between 8 am and 8 pm. She thanks them, sits back down and waits for them to settle. Looking from the students seated in front of her, towards her Caddie (portable storage) in the corner, it appears she wants to fetch something but doesn't. Following her gaze, I see that Ms Collier is talking to a student who is visibly upset and they are standing alongside the Caddie. Turning back, I see Ms Talbot is biting her lip and I feel sorry for her. Not for want of trying, this group has made little progress. Remarkably undaunted, she stands and walks across the carpet to the wall behind me, where she helps herself to the clock. Armed with the clock and three whiteboard markers, she sits down amongst the group and slowly makes her way around the perimeter writing the ‘missing’ 24‐hour equivalent at each hour marker (see Figure 2). Visibly caught up in the moment, the group is sharply focused as they annotate this newly inscribed clock face with actions relating to either the 12 or 24‐hour time. After a time, she cleans the clock and writes four questions across the centre, the students quietly answer them in their workbooks. Once they are finished, she asks different individuals the 24‐hour time at which they do each of these things. This time they are less distracted by discrepancies in individual routines and are clearly engaged with the mechanics of calculating the 24‐hour equivalent of each response. The lesson comes to an end. It is recess. The clock is placed face up on an ottoman and as students from other groups make their way out they pause to have a look, consider the current time in 24‐hour time, and move on. The clock sits there for a number of days before it is cleaned and returned to its home on the wall, without any fuss. (pp. 306–310)

To demonstrate how the ACAD framework and wireframe help to (a) identify elements of design that are open to alteration and (b) explore how their properties and spatial configurations give rise to emergent learning activity, we completed Table 2 based on the moment described in the vignette. Completing the entire grid is a form of discipline. One's analysis may focus on a very specific question at the macro‐social level, but taking the time to add something to each cell honours a commitment to more holistic accounts of learning and lays the groundwork for connecting the properties and spatial configurations of materials to emergent learning activity. One could start with any one of the nine cells in this three‐by‐three grid, but to illustrate our point we will start with the micro‐set. This is the point at which the teacher captures her students’ attention and works towards understanding by annotating the face of the clock with a whiteboard marker. The crux of this moment is when the teacher combines narration with annotation using her voice, her hand, the marker and the analogue clock face (micro‐set). However, the corresponding details of the tight‐knit group formed around her—with bated breath and laser focus—as she ‘defaces’ the clock (micro‐social), and the realisation that these students have come to the end of a two‐week cycle and are still struggling with the concept (micro‐epistemic), complete the picture.

The presence of the analogue clock and whiteboard markers was not remarkable. What was remarkable, was the interplay between the elements as this moment played out. Given the group this teacher was responsible for (micro‐social), she was able to reframe an abstract mathematical task (micro‐epistemic) using a concrete analogue representation (micro‐set). These actions were supported by access to empty space for just‐in‐time workshops (meso‐set) and a suite of competency‐based independent online tasks (meso‐epistemic) that freed teachers to work with groups of students as and when they needed assistance (meso‐social). All of this was supported by correspondence at the macro level in terms of set design (mobile, flexible learning environment), epistemic design (research‐based practice in teams driven by interest) and social design (a distributed leadership style that valued innovation).

This episode illustrates correspondence across dimensions of design and scale levels, but it is not difficult to anticipate where the sticking points might be on another day in this space, or in an altogether different space. Something as simple as not being able to find a working whiteboard marker, or something a little more complicated—such as different social norms governing the use of writable surfaces, could stifle the emergence of this quality of learning activity. But it is not only access and availability, or use value, that is illustrated here, because there is something fundamentally material at play in this example. Had this particular clock face been dark rather than white, digital rather than analogue, or permanently fixed to the wall, this critical moment of learning activity would not have emerged. In exploring this vignette using the ACAD wireframe, we have begun the task of connecting elements of design to emergent learning activity. What the matrix structure is less able to accommodate is the notions of time and the complex skill set required to navigate conjunctural moments in which repair work is necessary to restore or establish a valued quality of learning activity. For this, we will return to Hodder's (2012, p. 217) theory of entanglement:

Entanglement + fittingness + conjunctural event → problem → fixing → selection → E′ (total entanglement)

Table 3 is helpful in stepping us through the sequential emergence of a problem and its resolution based on what is to hand. Finally, the tanglegram (Figure 3) brings it all together by mapping the dynamic flows of humans and things and their contingent dependences, in a manner that demonstrates the value of understanding entanglement when designing for learning.

Designing for learning involves complex multi‐layered considerations about the nature of innovative learning spaces and their relations to emergent learning activity. The ACAD framework, ACAD wireframe and the theory of entanglement offer analytical tools to help educators consider this multi‐layered complexity and how various elements come together when designing for learning. These analytical tools assist educational designers in tracing the interplay between elements across a number of dimensions (set, epistemic, social), as they consider design choices across scale levels (macro, meso, micro), helping them to anticipate problems or explore unintended consequences with respect to the learning whole, illuminating how alterations in the designed environment can be said to shape emergent learning activity.

Conclusion

Fenwick et al. (2011) note that sociomaterial studies of education challenge the centrality of human processes in learning, in favour of the materiality of learning. This claim, they say, does not come at the expense of the personal but seeks to treat the material and the human symmetrically in order to explain how entities, knowledge, other actors and relations of mediation and activity converge in learning. As such, sociomaterial studies of learning explore relations between entities through which activity occurs, rather than the individual entities themselves. In doing so, they trace the ever‐shifting web of interaction that holds these processes together, all the while shaping their properties and interactions without relegating the environment to an inert backdrop to the main act of life or, in this case, education. What becomes clear, through the lens of a sociomaterial approach, is that increased connectivity and participation in networked structures gives rise to an increased dependence, which drives the need to understand the relations between constituent parts. However, in this article we have argued that if all we do is explore the relations between constituent parts, we run the risk of failing to explain how the parts relate to the whole and how these parts can be said to support valued learning activity. What is more, in returning to our depiction of learning as emergent, the challenge lies in understanding how the properties and spatial configuration of materials can be said to give rise to emergent learning activity.

In response, we have presented a set of conceptual and analytical tools capable of supporting those called on to participate in the (re)design of innovative learning environments in ways that reflect shared epistemologies of learning and accommodate increasing levels of complexity and diversity. Key aspects of design for learning include the careful alignment of theory and practice, and correspondence across dimensions of design (set, social and epistemic) and scale levels (macro, meso and micro). We have demonstrated the practical application of our analytical approach and suggest that it is particularly helpful for educational designers working in heterogeneous teams, because it is both deeply theoretical and eminently practical. That is, it offers well‐theorised representations that facilitate discussions about part–whole relationships and connect properties of materials with desired qualities of learning activity. Our more recent research examines the practical application of these ideas through the development and evaluation of additional tools (ACAD cards) to support the work of educational design teams (Carvalho & Yeoman, 2017, 2019, forthcoming). Taken together, the ACAD cards, ACAD wireframe and various visual and descriptive artefacts are being used in workshops in Australia and New Zealand. We are currently collecting evidence and analysing how our ‘Toolkit for Action’ supports the work of educational designers through mediated conversations about what it means to design for learning.

Acknowledgements

Foundational work for this article was carried out by both authors under the guidance of Professor Peter Goodyear and with the generous support of the Australian Research Council (ARC) Laureate Fellowship (FL100100203). In addition, Pippa Yeoman's contributions to writing this article were supported by the ARC Discovery Grant (DP150104163).