The Materiality of Method: The Case of the Mass Observation Archive

Introduction

1.1 The Mass Observation Archive is known to be methodologically problematic to those more accustomed to traditional modes of social research. As many have shown (e.g. Nettleton 2011, 2013; Bloome et al. 1993; Hubble 2006; Sheridan 1993a, 1993b, 1996; Sheridan et al. 2000), the sample is not representative of the UK as a whole; key social indicators for the correspondents, such as location, occupation and age, are often missing, and it is hard to detect when or why; the data are not easily amenable to any analytical software tools; and it can be difficult to conduct any kind of systematic comparison between the cases.

1.2 This article highlights a further characteristic of the archive: its unwieldy materiality. Focusing on the sensory-material experiences of the Archive and its particular types of data, we show how this unwieldy materiality is intrinsic to conducting Mass Observation Archive research, offering both constraints and opportunities. We begin by outlining some of the logistical practicalities involved in using the Archive and how these shape the way that researchers engage with the responses, and then turn to a discussion of what we see as some of the currently unexplored opportunities associated with its dense materiality. Drawing attention to the sensuous cues and hints offered by the Archive materials, we explore different ways of acknowledging and responding to these features, and their implications for the type and status of outputs produced. We conclude with a brief discussion of the implications of these ideas for the likely future digitization of many of the Archive's contents.

1.3 Our aim in discussing the materiality of method is not merely to add to the list of methodological problems already noted above. Rather, we take the Mass Observation Archive to capture the importance of the role of materiality in the research process in general. Indeed, we argue that it is the importance of the materiality of the social, as opposed to the quality of one or other type of data, that needs to be acknowledged. By implication, it is also through paying attention to the materiality of method that we learn about the production of data (and the production of knowledge) and how data literally come to matter in the course of the research process. We use the term 'method' here broadly to refer to techniques and practices throughout the entire research process, and not bounded activities that lie in discrete sections of a linear sequential list of events. As a way of providing concrete examples of some of the more subtle and abstract aspects of our argument, we ground the discussion in our own experiences of using the Archive. Thus, although some of the claims about social research are general in nature, we draw on empirical illustrations from two studies. The first comes from Liz Moor's analysis of responses to the Archive's 1982 Directive on pocket money, whereas the second study led by Emma Uprichard refers to her use of the 1982 Food Winter Food Directive.

1.4 The Archive has its own idiosyncratic kind of materiality, but by exploring both what is specific and what is generic about the way materiality matters, the importance of the materiality of method in general becomes more visible. Paying attention to the materiality of method has significant advantages when it comes to studying social change and continuity, since this is always and necessarily intrinsic to social research, however much or little social research changes because of the data or research techniques. In what follows, then, we summarize our approach to the materiality of method and how we link the material form of the Archive to social research in general. From there, we move to three concrete and common domains of work within the Archive, namely: (1) questions of access and sampling; (2) issues linked to working with paper; and (3) reading handwriting and type. In doing so, we consider the ways in which we might empirically capture how materiality matters in social research, however subtle or nuanced the digitization of the research process becomes.

Thinking through objects

2.1 The relationships between materials and people – between human agents and 'brute matter' – have preoccupied social scientists for some time, but have been explored most exhaustively by archaeologists and anthropologists. While an archaeological focus directs our attention to the ways in which a knowledge of the materials of a given time (the availability of particular materials during wartime, for example, or the persistence in the use of 'outdated' types equipment) can illuminate contexts of writing, and place respondents in social and political as well as geographical and historical contexts, the argument we make here rests more closely on social anthropologists' work on material culture. Specifically, we draw on the ideas of anthropologists in the Maussian tradition who have theorized the identification or overlap between persons and things (see for example Henare et al. 2007: 16-17), the extent to which material things may be said to have – or to partially constitute – forms of 'agency' (Gell 1998), and the centrality of objects to everyday cosmologies and forms of order (Miller 1987). In the context of the current article, we are particularly interested in the ways in which the encounter between material objects and human researchers may contribute to new modes and forms of knowledge creation.

2.2 Central to this encounter between researchers and material artifacts is sensory judgement. Exploring the relationship between sensory history and memory, on the one hand, and broader cultural, political or social shifts, on the other, anthropologist Nadia Seremetakis (1994) observes that it is through the senses that historical change is most clearly experienced at the level of the everyday. In Greece, for example, she notes that the disappearance from shops and markets of many regional foods that followed from European Union (EU) membership was one of the key ways in which such political and historical shifts were registered. Changes in the appearance of common foods, and the disappearance of others (along with their specific sensory experiences and associated memories) were key to how ordinary people 'knew' the EU – a way of 'knowing' historical change that involves both remembering and forgetting (Seremetakis 1994: 3-7). Building on these insights, she proposes the senses as 'meaning-gathering apparatuses' that work 'beyond consciousness and intention', and often against the grain of literal, textual readings, to produce new or alternative ways of knowing objects of study. Thus, where 'literal' readings of objects (perhaps, in this case, formal analyses of changing volumes of goods produced, imported and exported) tend to produce 'limited, functional and repetitious' accounts of their meaning and meaningfulness, sensory approaches have the potential to offer different versions of history, and different ways of knowing an object, place or environment, in which what appears to be the case from one kind of analysis may be relativized, contradicted or confirmed' (1994: 6-10).

2.3 If we apply these insights to the Mass Observation Archive – a site in which the aim is also to understand history as experienced 'from below' – we are presented with an alternative way of thinking about the relationship between the ostensible 'texts' (the words in the responses) and the way that we respond to their material support and organisation – the papers, boxes, handwriting, bindings, and so on. As Seremetakis indicates, the risk of focusing primarily or exclusively on 'manifest' content (on what is said in the content of responses, but not their formal or sensory properties) is that aspects of respondents' social or emotional experience that could be traced in objects more easily than in words get ignored or overlooked – 'left unmarked, unvoiced and unattended to as a banal element of the everyday' (1994: 19).

2.4 Our argument has some resonances with Savage's (2007) point about the importance of exploring content and form together. He used the Mass Observation Archive to examine the changing nature of British social class and argues that it is 'it is the form, rather than the content, of class talk which is important' (para. 6.4). However, whereas Savage mostly uses form to refer to aspects of the data such as length), we extend this to refer to the materiality in all aspects – from the logistics of access, to the paper, to the handwriting. Similarly, our argument has echoes with literature that deals explicitly with the ways that archives shape knowledge and how that knowledge is changing through processes of digitization (e.g. Cook 2013; Ernst 2006; 2013; Featherstone 2006; Manoff 2006, 2010; Taylor 2003). Indeed, much of that literature discusses the ontological status of archives precisely because the form of the archive is not static and keeps changing, as it responds to the transformations of the social world it seeks to archive. Whilst we consider these issues to be relevant, our focus, by contrast, is on the material aspects of using the Archive specifically for social research purposes.

2.5 In many ways, then, we bring different genres of scholarship together – the anthropological approach to material culture, the qualitative approach of studying the form as well as the content of data, and debates about social research in general. In doing so, what is new in what we are proposing is an explicit argument that a close examination of the material form of the data offers an alternative way into interpreting data for social research purposes. More specifically, as will be seen, our argument rests on the importance of more mundane material 'sense-encounters' with empirical data and how these necessarily seep their way into the research process, whatever research we do.

2.6 Paying greater attention to the diverse materiality of (and in) the Mass Observation Archive data is also, we suggest, entirely in line with the anthropological roots of the project, and in particular with its interest in the potential of a 'surrealist ethnography' (Highmore 2002). The paper responses located in the Archive's famous brown boxes juxtapose lives, and offer a way of knowing respondents and their world that is distinct from both digitized and transcribed versions of the same materials – a claim whose consequences we will discuss further below. This link between the specific material form of the archived responses and the kinds of knowledge that can be created with them comes from the capacity of the formal properties of the responses (paper, layout, writing and type) to supplement, modify or even contradict the communication 'inside' the responses. Researchers are well aware of the capacity of archive materials to surprise or move the reader. What is less common is for them to produce work that reflects explicitly on the meanings they have made from these material qualities, and it is this aspect of using the Archive for social research that we wish to highlight.

2.7 Making use of the 'vast social unconscious of sensory-emotive experience' (Seremetakis 1994: ibid.) requires that researchers' own experiences and encounters with documents receive greater attention. Our view is that approaching empirical data through our senses may lead us in interesting directions away from literal, 'rationalized' meanings and instead allows us to explore the data in its material everyday context. Doing this not only has the advantage of explicating some core ideas relating to the importance of matter in method, but it also explores the extent to which the objects of study that matter to social researches may matter precisely because they are intricately intertwined with our senses. Furthermore, it is because time is itself is matter, as Prigogine (1980) argues, that a focus on materiality allows a way of approaching data not only to study what people have said, thought or done at a particular point in history, but also to reflect on the 'dating' and the 'timing' of social action at both individual and collective levels. For example, the fact that letters were written on paper rather than sent in electronically situates the responses in a particular point in time in social history. Likewise, that the data were housed in the library rather than The Keep also places a time-stamp on the Archive itself. The sheer materiality of data and the possibilities available in terms of how we interact with that data analytically (e.g. with software or not) say as much about the time of the research as they do about how the data are produced. By implication, a focus on the materiality of method, we suggest, also has the capacity of preserving the singularity of the individual biography as it is situated within a specific socio-historical context (Mills 1959).

2.8 Much of this will become clearer throughout the article as we move to discuss, in more concrete terms, how grappling with materiality is intrinsic to using the Archive. We focus specifically on: issues of access and sampling; how we 'read' paper; and ways of approaching handwriting and type. The choice of examples is in some ways arbitrary, since we might have chosen other material or physical/sensory aspects of the Archive (e.g. the design of the building, the process of taking notes). There is also important work to be done in attending more carefully to the material objects 'in' the responses (the mundane objects – clothes, cheque books, furniture, and so on – that people include in descriptions of their lives). Each example highlights material and/or sensory aspects of archive research that are manifestly present throughout the research process, but which do not often feature in social research discussions. Indeed, regardless of the research topic or background of the researcher, accessing the archival materials, rummaging through piles of paper and reading handwritten or typed responses are inevitable parts of the research process. Each domain, therefore, serves to reveal the ubiquitous but mundane materiality of the data in the Archive and, we suggest, elsewhere. We start with the problem of access since this is in some ways where the story begins, before moving on to discuss issues in paper, handwriting and type. Finally, we reflect on the extent to which the digitization of the archive may or may not disrupt our claim about the importance of materiality to the research process, and consider how we might capture empirically the ways that the materiality of the research process matters in research.

Access

3.1 We want to begin with the opportunities and constraints involved in simply accessing the Mass Observation Archive materials. The data in their physicality are always situated in time and space. Therefore, the material logistics involved in 'access' necessarily shape the research process. Whether primary or secondary qualitative or quantitative data, real or simulated, this is always the case. Now, one might think that the sample of an archive is sat there in a box waiting to be opened, posing no particular issues. The data certainly are 'there' already, in the same way as they always are in traditional secondary research. However, accessing them is not without any material consequences, and it is this aspect of using the Archive for research that we make explicit here.

3.2 Part of the problem, and excitement, of working with any archive is that researchers often need to dis-place themselves physically to actually read the material. That is, they need to travel to the location where the archival material is housed. In the case of the Mass Observation Archive, the archive is located in Brighton, in the south east of England; at the time of our research, it was located in the University of Sussex library (instead of The Keep). Like clambering up a tree or hiking up a mountain to obtain a better view, the journey to the archive is a necessary part of the preparation involved in exploring the data. At the time the research, the Archive was closed to visitors on Fridays. It is always necessary to book in advance and to indicate which Directives one wants to access. Once there, personal items, such as bags, coats and the like, are put in a small locker. The documents in question are held in boxes, slightly bigger than a box file, which are delivered by the reception archivist on duty. However much material there is in any one directive and no matter how long or short your time availability is, it is only possible to access two boxes at a time. The notion of 'browsing at leisure' is, therefore, immediately curtailed by the basic material logistics relating to where in the world the researcher is travelling from, how long they have to be able to sit with the data, how big their budget is if they need to stay overnight in Brighton, and so on and so forth.

3.3 Once in possession of the relevant boxes, the treasure trove of data can be unlidded and the exploration can begin. Right from the outset, the material logistics ridicule the very notion of sampling according to classic systematic strategies. As with most Directives, the respondents' letters are catalogued according to what and how much is returned, and the letters are organized alphabetically. In the case of the 1982 Food Directive, there was one 'male box' (Box 46) and 'two female boxes' (Boxes 47 and 48). The first research trip took place Wednesday-Thursday, squeezed around other teaching duties. This meant travelling from York to Brighton on Wednesday, leaving just the afternoon to begin the research. Given the time constraints, it was decided to start with the male box with what was left of the Wednesday and then move to focus on the two female boxes all day Thursday, and to do so methodically according to how they were indexed and therefore moving from A-Z across the three boxes.

3.4 For certain research questions, it might have been preferable to proceed differently, i.e. to work from variables of the respondents and then retrieve the targeted documents. However, although the list of respondents is digitally available upon request, it is still difficult to 'search' or 'sort' the data electronically, since the electronic details of the respondents are separate from their non-electronic data responses. Yet, even the non-electronic browsing is shaped by the materiality of the documents themselves. After all, there are many other ways the data might have been organized. The fact that the men's and women's responses were kept physically separate is interesting in itself, and reflects the materiality of social ordering at work. (Indeed, in more recent Directives, men's and women's responses are not kept separately.)

3.5 Another example of how the material ordering of data matters relates to how the data are housed in folders within an A-to-Z sequence within a particular numbered box. The pages for each individual are paper-clipped together. The process of sifting through the layers of paper is itself disrupted by each metal paperclip used to bind the paper of each respondent, with the researcher having to carefully 'un-bind' the case without damaging the paper. Like being offered a key to gently enter these unknown paper worlds, these paperclips for each person quickly become a way of 'greeting' and 'leaving' each volunteer, one by one. Stumbling upon an extra paperclip or discovering that one is missing can be confusing as we grapple to make sense of the dis-located materiality. Occasionally, the 'official' paper clip is replaced by the respondent's own relatively forceful attempt to 'take over' the binding process by using string office tags or even plastic spiral binding. In both cases, the material organization of the data matters to how we then access or sample them.

3.6 In turn, whilst we might approach the documents with particular hopes and ideals about how to explore them, the reality is we are constantly wrestling with the ordered way of 'binding' the data in files, 'caging' the documents in sets of boxes, enveloping and casing the documents in particular ways. As Uprichard 2011: 6) notes, 'the material reality of the cases necessarily and contingently places limits on which cases are selected and how they are selected.' We cannot get around this problem, regardless of what kind of data we are accessing, whatever the research, whichever methods are used, problems of access are intrinsic to empirical social research. What we highlight here is that the material logistics involved in accessing data are not 'outside' of data. Rather, they are ontologically and epistemologically intrinsic to data. Paying attention to the mundane materiality of method of accessing data allows us to have a deeper understanding of that data – its material organisation and constitution. As we shall see, the epistemological advantages associated with paying attention to the materiality of method are echoed throughout each example discussed here.

Paper

4.1 Once access is gained, what becomes immediately evident is that all responses share (at least for now) the quality of being written or typed imprints made on paper, which is our second illustrative example of the materiality of method. Where respondents diverge quite markedly, however, is in the type of paper used: in any box of the Archive materials, it is common to find paper of varying sizes, shapes and colours. Tiny blue sheets of notepaper, old-fashioned foolscap, ultra-thin white typing paper, graph paper, 'notelets' decorated with flowers, paper with holes punched in the sides, mismatched sheets of 'scrap' paper that look as though they might be used to take phone messages, all sit together in a box, testifying to respondents' singularity (Pollen 2013) and inviting us to form images of their place of origin.

4.2 Paper is traditionally understood as a 'space-binding' rather than 'time-binding' medium (Innis 1951; see also McLaughlin 2005), whose defining property is its portability and capacity to travel across space, rather than its durability and capacity to endure over time. Archives work the opposite way; making sense of paper in the archive, therefore, requires that we see it as something more than a medium for communication, and instead as having a communicative potential in its own right (Dworkin 2013). Paper, in other words, is an artifact: the sheer substance of data rather than merely its material support. It is as such – as a historical artifact in its own right, and not only because of the words that it bears – that it lends itself to the more sensory, tactile and reflexive methodology that we outline here.

4.3 Yet, how might we 'read' paper as data? Anthropological approaches emphasize ethnographic immersion as the means by which the role of material culture in organising and expressing everyday cosmologies and forms of order can emerge (Miller 1987). Thus, in an ethnographic study of a London local government office, Pellegram (1998) traces the role and functions of different types of paper expressing the relationships and forms of order underpinning the life of the office. She uncovers rules regarding how paper is used, with different types for internal and external communications, for typewritten versus handwritten messages, and different colours of paper for different tasks. Letterhead stationery is used for official correspondence and is distinguished by the paper's grain, thickness and quality as well as its logo (1998: 106). Paper can also delineate nature and type of work – writing a memo, submitting an invoice – and paper qualities can be taken as messages about the 'regard one person shows for another' (105). Hence, paper plays a crucial role in ordering the symbolic world of the office and its relations with the outside world.

4.4 However, researchers using the Archive do not have access to this kind of ethnographic 'big picture' – to the ways in which a given response fits within the broader material environment of a respondent's home or workplace. This lack of context is perhaps part of the reason why there is reluctance to make explicit inferences based on such material features. Our contention, however, is, first, that these material features are of genuine interest and deserve to be acknowledged as part of the 'worlds' that respondents present to the Archive – that is, for their value as material culture; and second, that we cannot help but absorb, and be influenced by, these materials in our work in the Archive, and that our research would be improved by acknowledging and reflecting upon this fact. The fact that someone writes on American-style yellow A4 paper with a green pen while no one else did tells us something about their world – even if it is hard to discern – and also affects how we respond to them. It tends to make that person stand out and it draws our attention, shaping our affinity or otherwise for the author. Latent messages are real and unavoidable; the question is how, not whether, to acknowledge them.

4.5 There are, we suggest, two ways in which researchers might acknowledge the latent messages of archive materials and their possible impact upon the research process. The first has to do with the practical effort to record these material features systematically, that is, to treat the material properties of documents in their own right rather than simply as 'substrate' (Dworkin 2013); the second has to do with the attempt to represent the materiality of correspondents' worlds more fully. Both require reflexivity on the part of the researcher and both have the potential to enrich our research. Let us begin with the first point. It is relatively straightforward to catalogue the formal properties of archive responses. For those who use coding frames or schemes, it is possible to add another box or node, and to decide always to note whether something was handwritten, typed, word-processed and so on. The Archive itself keeps simple records of the number of responses to each directive that were handwritten versus typed, although there is scope for this kind of cataloguing to be extended. For those who keep a research diary or longer-form notes, it is possible to add more detail: colour and size of paper, observations about handwriting, and so on. Researchers might also attempt to make more explicit the ways in which these material features prompt particular thoughts and responses. We might, for example, notice ourselves making certain inferences about the person's occupation (for example, graph paper may suggest someone who works in mathematics or engineering). We may also make inferences about their home environment or habitus (scrap paper might bring to mind images of clutter or disorder). We might find ourselves making assumptions about class or status, based on the type and quality of paper used, as well as handwriting and spelling (see below). We may also feel that we can discern a particular orientation to the Archive itself: using 'scrap' paper might be taken to indicate a different form of engagement than the use of 'office' paper; the use of flowered notelets might suggest a more 'intimate' or informal relation, and so on. Our point is not that we 'should' (or should not) make such inferences, but rather that we very often do make them, and that our work would benefit from acknowledging this.

4.6 If we wish to deepen the reflexivity of our research process, we might take this documenting process a step further and attempt to record those times when one is particularly moved, surprised or affected by a response. This connects to our second way of acknowledging the latent messages of archive materials, which has to do with how we can represent the worlds of respondents more fully by referring to the materials in and of their responses. What does it mean, for example, if an elderly correspondent who carefully types her very 'proper' and polite responses suddenly adopts a more playful, even girlish, tone when she recalls the hat she bought with her first pay check (as in an example from the pocket money Directive)? Or if she goes to the trouble of drawing a picture of the hat in ink at the side of the page once she has taken the paper out of the typewriter? We are surprised, and moved, because we see a different side to the individual, and because we sense that this particular memory has interrupted her flow. The change of tone, and the introduction of a new material trace (ink on typing paper), syncopates the response, giving it texture and feeling and allowing us access to a previously hidden part of the person.

4.7 What we are hinting at is that these sometimes discordant material traces – the interplay between 'formal', office-style typing paper and hand-drawn ink images – combine with shifts of tone (a more playful written style) to give us unexpected access to feeling and emotion, to points of emphasis and points where people are taken, momentarily, out of themselves. How respondents' use of materials comes to embody and preserve their thoughts, feelings and orientations to the world is something, we suggest, that might be observed and noted, both in our research process and in its final outputs, as a way of doing justice to the singularity of the correspondents and their reflections (Pollen 2013; Highmore 2002).

Handwriting and type

5.1 Our final example of the way in which the materiality of the archive shapes the knowledge we produce relates to the traces respondents leave on the page – that is, to handwriting and type. Much has been said about the genres of writing found in MO responses – about, for example, 'subjective' and 'social reportage' styles of response (Sheridan 1993), or respondents' concerns about spelling and technique (Sheridan et al. 2000). However, there is less discussion of the significance of working with typescript, manuscript (handwriting) or emailed responses. Yet in practice this is one of the most striking aspects of Mass Observation Archive work: responses bear the physical, material imprint of their authors, and of their authors' engagement with a set of media (pens, typewriters etc.).

5.2 Handwriting is especially compelling because of the 'constant identification of handwriting with the self that produces it' (Thornton 1996: xii). There is a romantic sense, bound to modern conceptions of the individual, in which handwriting is seen 'to reveal one's character by proceeding from the unconscious as an automatic gesture' (Grusin 2006: 102). Yet, equally, forms of penmanship appropriate to their gender, occupation and rank have also been deliberately taught (Thornton 1996). Handwriting has been understood as a reflection of physiological uniqueness, and yet bodies were carefully trained to produce a 'muscular' style or 'delicacy' of motion (ibid.). Indeed, what makes handwriting interesting is that it seems to capture both mind and body, and 'straddles the cusp of individuality and conformity' (Thornton 1996: xiv).

5.3 Like other aspects of the Archive, then, handwriting simultaneously reflects both the social and historical located-ness of respondents (handwriting and typing conventions change over time, and also reflect social determinants like gender and education) and their individuality and idiosyncrasy. In fact, the idea that handwriting gives unique access to the person – albeit a necessarily 'enculturated' one – seems to gain strength in an era of electronic writing. It is felt that we 'inevitably lose something relating to the physical dimension of the 'ductus', i.e., the 'movement' of the hand leading a pen' (Neef 2011: 18); letters on a screen cannot reflect the force with which keys were struck, and standardized fonts means that word processed scripts seem to 'lack a body' (ibid.). Interestingly, similar anxieties were felt when the use of print first became widespread. As typeface designs moved away from 'calligraphical, somatic models' chosen for their suitability for certain types of text (religious, legal, fictional, etc.), people became concerned about the 'unsettling blankness' behind 'expressionless' typefaces (Thornton 1996: 31).

5.4 All of this would seem to bode ill for an era in which responses to directives are likely to arrive in more or less standardized email formats. Yet, most of us are familiar with electronic documents containing excessive amounts of bold, capitalized or underlined script, and we perhaps unwittingly make assumptions, or ask questions, about the person that chose them. We have a strong sense of what is and is not appropriate, and come to expect that word-processed documents for a particular purpose will take a particular form. In the future, therefore, it may be all the more important to preserve a sense of respondents' choice and size of typefaces and the use of bold, underline and italic functions as a way of acknowledging the singularity of the person's response both to the task at hand and to the set of standardized commercial word processing systems available to them.

5.5 At the same time as these details deserve to be preserved as part of the material culture of respondents' worlds, handwriting and type share with paper the capacity to promote particular forms of recognition or interest in the researcher. The male correspondent who writes in thick flamboyant script on brightly coloured paper prompts, rightly or wrongly, a sense of intrigue about the observed difference relative to the other batch of handwritten responses. Similarly, one author recalls one of the letters on crisp white A5 paper had a beautiful drawing of a bird on a gate with a house on a hill in the background. Both the drawing and the handwriting (in thin black ink) reminded her of her grandmother. She read it with a particular feeling of proximity to the stranger who had written it, remembering being close to her grandmother, while also making an effort to create distance and to read the response with a more detached 'researcher's cap' on. Such experiences, we suggest – of remembering, of feeling proximity or distance, or simply of being more or less interested – may be common, but tend not to be made explicit in critical discussions about social research.

5.6 Making sense of handwritten or typed responses requires accounting for the affordances of different media and their effects on the writing (and remembering) subject. One area in which this is particularly striking has to do with respondents' errors and self-corrections (which in turn relates processes of remembering, as well as processes of self-presentation). Indeed, one of the ways in which modern word-processing technology differs most notably from earlier forms of handwriting or typed script is in the ease of correcting errors. In handwritten (and some typed) responses, respondents leave traces of their thought processes on the page for the researcher to see – crossed out words, additions in the margins, or Tipp-Exed spots underneath type. Some of the most compelling moments in reading the Archive materials are those where one sees a respondent erasing their own words or adding in new ones as they re-read their response. These moments are arresting precisely because we are brought into greater proximity with the person's process of recollection, reflection and self-editing. We are also reminded once again of the traces of the physicality of composition. All of these have the capacity to make us feel closer to the respondent, and also to infer something about the feelings they might have about a particular topic. As with other material features of the Archive, these physical traces tend to move us as researchers, but that experience does not often find its way into our research outputs – partly because of the difficulty of naming or articulating what are almost pre-conscious aspects of our response to sensory features.

Conclusion

6.1 Throughout this article, we have drawn attention to the materiality of the Mass Observation Archive as both a challenge and an opportunity. We have outlined the ways in which various material features of the Archive present obstacles while also 'pre-formatting' aspects of the research process. We have shown how the material qualities of paper and type/script make the research process a sensory and emotive experience rather than purely cognitive one, and have made a series of suggestions as to why these material features are important and how they can better be acknowledged in research procedures. Our discussion has focused on what we see as vital, but under-explored, material details of the Mass Observation Archive.

6.2 As part of this, we have already hinted at some of the changes already taking place in an era of digitization, including the fact that more documents are being scanned and that respondents increasingly email their responses rather than sending them by post. This raises the questions: as the Archive becomes digitized, and the material logistics of accessing the 'form and content' alter, to what extent might the processes and outcomes of research themselves shift? What happens when 'search' and 'find' facilities become the norm and researching the Archive no longer requires booking ahead, travelling to Brighton, and when documents can be accessed in any order that the 'find' facility allows? How will the Archive's commitment to particularity and specificity be challenged by technical developments that allow respondents to more easily be treated as data points?

6.3 Our answer is to treat the form of responses as seriously as the content. This is partly, as we have indicated, to preserve a sense of the material culture of the respondents (and the social and personal worlds from which it stems). However, it is also, as we have shown, because of the ways in which the formal properties of responses bring additional layers of meaning to the ostensible content of the words. Furthermore, if we accept that (social, economic, political, cultural) change and continuity will manifest themselves in the archival materials, then preserving the specificity of the individual within the responses becomes essential, as an index of the balance between broad social experience and particular individual lives that the Archive celebrates. As we have also noted, there is considerable scope for researchers to think more carefully about the medium as well as message in Mass Observation Archive work, and about how material supports may change the content as well as the form of that message.

6.4 We have suggested some concrete ways in which researchers might reflect on and incorporate the significance of the materiality of respondents' lives in their work at the Archive. The argument does not change because the data are digital (or quantitative or simulated for that matter). We can still examine stylistic differences, choice of fonts, font sizes, formatting arrangements and even colours of ink. In the future, the fact that someone responds in 14 point Comic Sans, or sends their emails in green type, will illuminate something about both the set of available resources for communication at the time and the way that the respondents located themselves among them. This, as we have suggested, is part of a broader commitment to retaining a sense of the specificity and singularity of the respondents – something that is always stressed as important in researchers' responses to the Mass Observation Archive, but which risks getting downplayed or subtly flattened in an era in which forms of writing and typing appear to be returning us to an earlier moment in the history of 'imprint and trace' (Neef 2011), in which forms of writing/typing are becoming standardized again.

6.5 We have seen this happen already with the standardization of digital data more generally and quantitative data specifically, from the demise of the Fortran log sheet, to the shift to the home PC the portable CVS file and the multi-dimensional and 'live' relational databases (see Uprichard et al 2008). Whereas the paperclip may bind the pages of each person, here we see the 'row' standing in for the individual case, binding, splintering and caging the individual into a sequence of movable variable grid of case-based cells. Yet, there too, authors have recently stressed the importance of preserving the qualitative changes to quantitative information (e.g. Callon and Law 2005; Desrosières 1998; Thrift 2004; Uprichard 2011), since it is by studying the im/material aspects of change and continuity that the temporal-spatial micro-macro dynamics can be traced. That is to say, studying how change 'sounds', 'looks', 'smells' as well as what is written and counted over time may provide windows into why the content of the data has (or has not) changed in the way it has.

6.6 A focus on changing material forms allows a way of tracking change and continuity through the Archive in a way that content alone is unlikely to allow. However subtle or nuanced the digitization of data, the materiality of the research process matters both to using the Mass Observation Archive and to research in general. One of the gains from this approach is the ability to notice where respondents show a sense of agency in expressing their individuality - their 'biographical fingerprint', if you like. It will be significant, for example, if no more responses are hand written. Likewise, it will matter if the content is identical across the respondents, just as much as it will if no differences in length or format of responses can be detected. The archival responses represent part of the material culture of respondents' worlds – a negotiation between the modes of communication made available socially, and their own distinctive ways of expressing themselves. The ways in which respondents use these materials offer a series of cues and hints that may allow us to make more subtle or nuanced readings of respondents' feelings and experiences, while also requiring that we reflect more systematically on how such materials and expressions do or do not prompt feelings of identification, surprise, memory, and so on.

6.7 Our concern throughout this article has been to highlight the importance of matter as a way of preserving both the historicity of the documents and the specificity of individual responses. C. Wright Mills (1959) argued for the preservation of both in his biography-society-history triad, arguing that we need to pay attention to the singularity of biography in society in its historical moment, even as the march of history often tries to subsume individual biographies into 'society' and 'history'. Both historicity and specificity, we suggest, lie in the materiality of the data as much as the epistemological facilities available to researchers at any point in time. The 'form' of what people send in and not merely the 'content' is of value. The actual paper, pens, typing equipment that are used, the ways that the data are accessed, read and analyzed become key to understanding the idiosyncratic singularity of the individuals and how this is nested within more generic aggregate social changes as well.

6.8 We recognise that these suggestions present a number of challenges to the researcher: the same materials – a particular type of paper, a smell, a style of handwriting – are likely to evoke different memories and forms of recognition for different people, and may leave others altogether unmoved. Greater sensitivity to the 'latent messages' of materials is likely to be productive of new forms of knowledge, but will compromise methodological reliability. Our response to this would be that while these difficulties in reading materials should be acknowledged, they do not constitute a reason for ignoring them. We tend to take it for granted that different researchers will alight upon different details of the content of the written response – that they will, to some degree, 're-make' the person by seizing upon some themes, phrases or ideas while ignoring others. Our argument is that it is no less reasonable for researchers to do this in relation to the materials too.

---

Acknowledgements

This research was supported by the Economic and Social Research Council [RES-061-25-0307].

---

References