Social Media and the Politics of Small Data: Post Publication Peer Review and Academic Value

Introduction

As Roger Burrows (2012) has cogently argued, the complexity and stresses of what counts as academic value and its various measures and metrics are now a ubiquitous part of life in the neoliberal academy. Where Burrows outlines some of the key measures of value which structure academic life (including the H-index, various research assessment exercises, national student surveys and so forth), one recent addition to this assemblage of data are measures of the success of social media strategies by academics, publishers and editors to increase the impact of published journal articles and books. These strategies are situated within the changing landscape of academic publishing, and the ‘digital disruption of the publishing industry’ (van Mourik Broekman et al., 2014: 36). These disruptions are contributing to a range of anxieties over what, where and how one should publish. These anxieties are heightened for individual academics under pressure to increase the chances of one’s article being read and cited or, conversely, who want their work to be ‘read, appreciated and re-used’ (van Mourik Broekman et al., 2014: 43).

Journal editors and publishers who have attracted high-impact measures, and whose journals are ranked in various hierarchies of value, know that impact measures are influenced by the likelihood that many if not most journal articles are zero-citations; that is, they are rarely if ever cited. These anxieties are distributed across publishers, editors and individual authors, increasingly finding themselves in a publishing arena where it is not enough to secure a publication; one is now encouraged and sometimes incited to increase the visibility of their article through the use of social media. The aim of such strategic activities is to improve the likelihood that the article will be downloaded, read and cited, and there are various measures of effectiveness, such as altmetrics, which can be used to audit such activities. The holy grail of these practices is an increase in visibility and possible impact and citation. Savvy academics are increasingly turning to social media to draw attention to a published article or book, as at the same time they find themselves negotiating potential accusations of self-promotion and self-branding, as well as sometimes (and especially in the case of feminist academics) attracting abuse, harassment, trolling and flame-wars. ¹

This article will explore the challenges of these new publishing environments by turning to science publishing and the difficulties facing scientists who have to tread a path between attracting attention and visibility and negotiating the problems that might arise from post-publication practices, which are sometimes akin to strategic self-commodification, personal branding and practices of micro-celebrity (see Papacharissi, 2012). The anxieties and challenges of such practices will be situated within the emergence of what has come to be known as post-publication peer review. Post-publication peer review (or PPPR) refers to the afterlives of journal articles and other forms of writing as they circulate within and across social media. This movement affords the potential to extend review by allowing post-publication contributions to comments sections of blogs and websites, discussion forums, and increasing circulation and traction of an article or book across Facebook, Google+ posts and Twitter. In some cases this extended and hyperlinked commentary can court the attention of broadcast media and contribute to the public communication of science.

Although it is recognized by scientists that PPPR can extend the article’s afterlife, allowing a bigger readership and more publicity, for many this comes at the expense of the integrity of science. PPPR is an inevitable part of new publishing cultures, but carries the dangers of opening academics and scientific theories up to scrutiny, judgement, reflection and evaluation that exceed those criteria that have been invested in as the cornerstones of scientific truth and objectivity, such as reliability, validity and replicability. There is also the very real danger that it might not hold up legally, exposing the academic to the possibility of legal action and being sued. ² PPPR can also provide a corpus of data that make visible in new ways the citational politics that govern what it might mean to be highly cited and the ambiguities, conflicts and contradictions that participate in the making of various citational metrics.

These issues and their significance for humanities and social science scholars will be explored by following the fate of a highly-cited journal article from 1996, published by the Yale cognitive scientist John Bargh. This article became the subject of a social media controversy in 2012 highlighting some of the issues shaping PPPR, its possibilities and discontents. The article in question is one of the most highly-cited articles in the area of priming (currently cited over 2300 times), which crosses cognitive science and social psychology. It is an area that was popularized by Malcolm Gladwell in his book Blink: The Power of Thinking without Thinking (2005). The article suggests that experimental subjects can be primed to walk more slowly to an elevator following being shown words associated with ageing on a scrambled language task. Priming is an already controversial area of study as it has close historical links with psychic research and is an area that is characterized by puzzles, anomalies, contradictions and what many have described as attention-grabbing studies. The article in question recently gained an after-life as it was revitalized and became the subject of an ongoing controversy, which took shape across various social media platforms and produced a corpus of data available for analysis and inspection (see Blackman, 2015).

What can we learn from the after-lives which this article accrued as it moved on from the original event of its publication and became an attractor for a series of contemporary debates confronting scientists writing within new publishing contexts shaped by digital media? As much as humanities and social science scholars have inherited various measures of value from science, I will argue that we are also being forced into a post-publication environment that scientists have already confronted through the emergence and rise of digital, distributed forms of communication, sometimes, but not only, linked to open-access science journals, and the architecture and topology of such relationships. This means that published articles are often hyper-linked to formal and informal networks of commentary, response, review and judgement, which introduce processuality into the review process, meaning that the review process is always subject to change post-publication. In some cases these distributed forms of PPPR are opening science up to controversies which in the past have been occluded, obscured and contained by positivist forms of science writing. What does it mean to court controversy and attract visibility within this context, and what can we learn about the cultural and social life of data?

The anxieties that this is creating for scientists draw attention to the politics of academic publishing, the politics of citation practices, the social and cultural life of data, and the mystery of what attracts attention and gains a traction and visibility across social media and why. PPPR entangles multiple actors and agents, including social media and open-access platforms, websites, communities of academic scholars and researchers, journal editors, science journalists, media broadcasters and particular interested publics. These entangled relations are simultaneously affective, material, technical, cultural and political, which extend PPPR beyond discussions of what peer review means as it is extended within digital environments. As we will see, if PPPR is approached as a material-semiotic-affective apparatus for producing, shaping and entangling entities, objects and processes, then it provides an interesting repository for analysing the politics of data and its social, cultural and historical life. The analysis that follows will show how academic value and its various measures and metrics co-exist. They are co-produced and entangled with relations of status, prestige, hierarchy and worth. I will argue that these relations silence, omit, cover over and disavow the asymmetries and inequalities that structure academic life and its citational practices.

The Social Life and Cultural Politics of Data

A recent special issue of Theory, Culture & Society on ‘The Social Life of Methods’ (Ruppert et al., 2013) has encouraged humanities and social sciences to confront the politics of data and the challenges and possible resistances to the development of post-positivist methods that move beyond the well-rehearsed qualitative/quantitative divide (see Savage, 2013). These methods will need to be able to confront metrics as well as analysing the ‘social life’ of data that are obscured by numbers and the scale of so-called ‘big data’ (Beer and Burrows, 2013). One focus of such data-analytics is social media and how to determine the reach and traction of what circulates – sometimes referred to as ‘networked virality’ (Sampson, 2012). In this context, following the reach and traction of a highly-cited science journal article across time provides an interesting corpus of data for analysis. The article became hyperlinked to an open-access science article published in 2012, a series of blog posts, and commentary, retort, review and analysis, which spread across a network of academics, commentators, journalists and interested publics. The reach and traction would not meet the criteria of ‘big data’ in terms of scale but it is a good example of the importance of retaining a focus on what boyd and Crawford (2012) term ‘small data’.

As the authors suggest, big data is a poor term (boyd and Crawford, 2012: 663), as it is ‘less about data that is big than it is about a capacity to search, aggregate, and cross-reference large data-sets’. In relation to social media analysis, for example, big data – or the capacity to map, aggregate, condense, measure and represent data – is often linked to the actions of specific social media platforms or APIs (application programming interfaces) – Facebook and Twitter, for example. In this sense the API affords and shapes the possibility of what counts as data, thus bearing the imprint of all those human and non-human agencies which have already entered into the shaping of specific data-sets. As Gitelman and Jackson (2013: 8) argue, these aggregated patterns and their algorithmic supports often obscure ‘ambiguity, conflict and contradiction’, engaging in acts of erasure in order to associate, connect and produce what we might identify as ‘collective phenomena’.

My approach to the politics of small data will be an attempt to develop the principles of hauntological analysis within digital environments. This is captured by the concept of ‘haunted data’ (see Blackman, 2015), which is designed to disrupt the distinction between big and small data and to explore what leaves the frame if we focus solely on metrics, quantification and digital methods based on counting, measuring and aggregating numbers. This is part of a bigger book-length project, Haunted Data: Social Media, Weird Science and Archives of the Future, which is analysing two related social media controversies in the area of science communication. The controversies, which were carried by particular journal articles, all gained a reach and traction across social and broadcast media, including TED lectures, Facebook tags, comments and groups, blogs and micro-blogs such as Twitter, Google+ posts, online open-access science magazines and journals, as well as attracting broadcast media attention, including newspaper periodicals and even American comedy shows such as The Colbert Report.

The controversies blur the distinctions between fact and fiction, between the real and the imaginary, past and present and science and fantasy. They are all controversies which shape, police, animate and disrupt the borders and boundaries between cognitive science and the anomalies, puzzles, paradoxes and phenomena that enact a curious form of kinship between psychic phenomena and brain-based neuroscience. This queer kinship, which links cognitive science to archives populated by psychic machines, humans and animals, and to medial concepts, such as suggestion and telepathy (see Andriopoulos, 2008, 2013; Blackman, 2012; Derrida, 1977; Sconce, 2000), is one that cognitive science and scientists have to continually police.

The controversies are also therefore populated by key protagonists: science journalists keen to re-animate science’s forgotten pasts and cognitive scientists who are fixed in the role of sceptics, repeatedly challenging, discrediting, overturning or lampooning the various claims that are made. The areas linked by these controversies are replete with fantasy, fears, anxieties and cultural fascinations, which operate at the borders or limits of what is considered intelligible. In this article I will focus specifically on what I have termed The ‘John Bargh priming controversy’, which speaks to these concerns and also has particular relevance for cultural theorists interested in affect theories. Priming, the subject of the controversy, demonstrates how people can be made to move by experimental apparatuses which are consolidations, subtractions and intensifications of the supposed everyday ways in which we are open to being affected and affecting others (see Blackman, 2014). Priming techniques are seen to operate within registers below the threshold of conscious attention and awareness and to bring about change in thought, feeling, belief, action and perception, for example.

The ‘John Bargh priming controversy’ has multiple intersecting points of interest for cultural theorists; these include the politics and ethics of open access journal publishing, the integrity of science in the context of increasing PPPR, the question of how we might re-imagine concepts of time, duration and temporality that govern digital archives with their capacity for re-mixing and remediation (see Derrida, 1995), and how to analyse the agency and autonomy of data as it accrues ‘after-lives’ (Fuller, 2009), as data moves on from the particular event that originated it and becomes active.

Social Media and the Politics of Post-Publication Peer Review

Mark Deuze (2012: x) has coined the term ‘media life’ to argue that ‘who you are, what you do, and what all of this means to you does not exist outside of media. Media are to us as water is to fish’. The refrain repeated throughout his book is that media mediate everything and have become intertwined with ‘every single way of being, seeing, moving and acting’ (p. 3). Deuze gives numerous examples of how media are organizing and re-organizing politics, celebrity and science, where people, he suggests, might perform themselves in terms of media – a politician imagining him or herself as a photograph, soundbite or speech, for example. Deuze uses the term ‘media self’ to refer to the possibilities of being captured by media and how this possibility might change behaviour, thought, feeling and action. The selfie is one such example of the intertwining of media with self-performance, which perhaps illustrates this process. There are problems with equating mediation solely to media, particularly if one assumes that mediation is simultaneously historical, social, political, technical, somatic and affective, for example (see Blackman, 2012; Kember and Zylinska, 2012). However, Deuze’s thesis is useful as it does open up the question of how digital and social media might be transforming science publishing, particularly for scientists who are writing with an awareness of PPPR. PPPR creates contradictions and ambivalences which shape the desire of scientists to increasingly engage in the self-conscious staging of mediality; that is, to enhance and even write articles that are more likely to gain a traction across social media, and the contradictions which exceed the frame and control of the individual author as the article potentially accrues after-lives (Fuller, 2009).

There are two aspects to the self-conscious staging of mediality in the context of PPPR that I want to explore in this article: its affective and hauntological dimensions (Derrida, 1995; Gordon, 2008). The affective dimensions of PPPR refer to the kinds of discourse which are shaped within online environments, where communication often becomes uncivil (Ashley et al., 2013), emotional, and in the case of science writing exceeds the norms and conventions of positivist science writing. This is leading to anxieties amongst scientists and online open-access trade journals such as Popsci, which is a self-described 141-year-old science and technology magazine, committed to stimulating intellectual debate. It has argued that comments sections, which it has made available to enhance PPPR on science journal articles, are primarily carriers of emotion and all that should be excluded from norms of rational deliberation. In the editor’s words, they ‘tend to be a grotesque reflection of the media culture surrounding them’. They have consequently closed comment sections to articles to mitigate against this possibility. ³ This is a cause for concern for many scientists. Sanjay Srivastava, in an entry in his blog The Hardest Science on 25 March 2013, suggests: ‘These comments aren’t just like comments on some random website or blog – they become part of the published scientific record, linked from the primary journal article’.

The equating of PPPR to new media cultures and their potential for shaping forms of communication which exceed the norms of positivist science writing are constituted as a negative and potentially damaging aspect of PPPR. Certainly this is one after-life that accrued to John Bargh’s (Bargh et al., 1996) highly-cited article as it was hyperlinked to a recent open-access journal article published in Plos One (Doyen et al., 2012). This article became hyperlinked to a blog post written by a respected science journalist, Ed Yong, who intervened in his blog for Discover Magazine, Not Exactly Rocket Science, with the title ‘Primed by Expectations: Why a Classic Psychology Experiment Isn’t What It Seemed’ (18 January 2012). Ed Yong is considered influential in terms of his ability to court attention due to the large number of his followers on Twitter and his extended Google+ readership. Doyen’s article and Yong’s subsequent blog post motivated John Bargh to respond in his blog for Psychology Today, which had laid dormant for two years prior to this. Bargh wrote two blog posts, ‘Angry Birds’ and ‘Nothing in Their Heads’, which were later taken down. If the reader follows the URLs relating to these posts ⁴ they will be taken to Psychology Today and the response: Page Not Found.

Despite the myth or ‘fiction of the universal database’ (Gitelman and Jackson, 2013), where it is assumed that every action and transaction leave a trace, this dystopian myth of complete dataveillance does not stack up. Data can be redacted, removed, accumulate, leave traces and also disappear. In previous writing I have used digital methods to recover the redacted blog posts and re-stage the haunted data (see Blackman, 2015). The re-animation and removal of the ghost-data not only illustrate the difficulties for scientists and academics of engaging in self-conscious mediality but also make visible the regimes of remembering and forgetting which characterize science as it becomes conjoined, shaped and re-shaped by software platforms, algorithms and programmes (also see Mackenzie, 2013). This includes prominent scientists, such as John Bargh, writing blogs for influential magazines, such as Psychology Today, to promote and brand their own academic theories and standing.

Many prominent academics, for example, have Wikipedia pages that might be considered forms of personal branding. They tell a particular story about the academic’s life and career and establish relations of influence, status, and prestige within Google’s Page Rank algorithm (Rieder, 2012). Using a Wikipedia Edits Scraper and IP Localizer, ⁵ it is possible to see the edited and erased content of John Bargh’s Wikipedia page. This includes attempts by other users to include references to the controversy and what has been erased by the removal of Bargh’s blog posts from Psychology Today. The edited history discloses the dynamism of this controversy and the work that has to be done in order to remove it from particular regimes of visibility. Although the Wikipedia page ‘forgets’ the erased content, the Wiki Scraper is not so forgiving and performs different regimes of remembering and forgetting (also see Blackman, 2015). An example of this is provided in a screenshot seen in Figure 1. OPEN IN VIEWER

Data can disappear, are removed, become submerged or displaced, are lost, overlooked, deemed irrelevant, make accidental connections (rather than aggregated patterns) and can remain alert or lifeless (Gitelman and Jackson, 2013). Despite Bargh’s efforts, which mined his standing, prestige and status as a star-academic, he was unable to completely redact the blog posts, and the capacity of these posts to remain as an absent-presence led to Bargh’s final attempt to close down the controversy through staging a narrative of redemption, where Bargh draws on his status as a Yale scientist in a published article in The Chronicle of Higher Education, entitled ‘The Power of Suggestion’ (13 January 2013). This was published some 12 months after the redacted blog posts and reveals how much scientists are having to engage in new forms of storytelling in order to manage reputation and the unpredictability of PPPR. I will argue that such strategies reveal the extent to which scientists are having to confront and engage in forms of storytelling that might be described as trans-medial. The concept of trans-medial storytelling usually refers to forms of multi-platform storytelling which have emerged and are situated within practices of media convergence and the emergence of networked media (Jenkins, 2006). One might, on this basis, find reference to transmedial narratives and texts (Leavenworth, 2011), to transmedial television (Evans, 2011), to transmedial technologies, performances and even transmedial worlds. My approach to transmedial storytelling is more hauntological, as I will go on to develop in the article.

The importance of new forms of storytelling shaping science and its public communication is also partially recognized by the editor of Popsci who, although closing down comments sections to articles, recognizes that social media has the potential to extend scientific debate and discussion. The editor argues that interested actors and agencies should ‘talk back’ – ‘through Twitter, Facebook, Google+, Pinterest, livechats, email, and more’. But the goal or aim of such communication should be ‘vigorous and intelligent discussion’. The injunction to intelligent discussion in this context individualizes the problem and reduces it to ‘incivility’ and the ethics of using social media, rather than involving wider questions as to how PPPR might help to innovate science and open up to what is often erased, forgotten, actively policed or simply deemed irrelevant or of no consequence in relation to science’s multiple pasts and possible futures. This relates to the more hauntological dimensions of PPPR, as articles move and are re-moved across time.

Hauntologies and Media Time(s)

Hauntology is often associated with Jacques Derrida (1994) and his meditation on the fate of Marxism following the so-called fall of communism at the end of the 1990s across Europe, which was articulated within a ‘discourse of the end’. This was a melancholic moment for many left academics who lamented such a fall and wondered at the costs of this supposed ending to critical (Marxist) thinking. The refrain that is central to Derrida’s reflections is the feeling of time being out of joint – what he terms a ‘disjointed now’ (p. 1). Derrida sets out to explore this ghostly and melancholic feeling and to conjure and summon the spectre of Marxism in the present by engaging with various ghostings – the ghosts within Marx’s writing itself, as well as the way in which fictional writing (Shakespeare’s Hamlet, for example) has staged apparitions and called forth, interpreted and interrogated ghostings.

Derrida’s main site of reflection is the spectral quality of language itself, which he mines to perform the different temporalities that language can evoke. The radical untimeliness of language, Derrida suggests, performs the myriad of ways in which a feeling of radically disjointed time can be conjured. This, he suggests, can create the sense of ‘being-with spectres’ (1994: xvii). This might include time being ‘off its hinges, time is off course, beside itself, disadjusted’ (1994: 20), it ‘does not walk straight, or goes askew’ (1994: 23). As well as ghosts or revenants always returning and coming back, language itself is structured, Derrida suggests, by relationships of condensation, displacement, metonym and metaphor, which perform temporalities ‘beyond the living present’ (1994: xix). Derrida’s writing itself is akin to a magical incantation which calls forth the spectral qualities of language to question what it might mean to live and to learn to live with ghosts. This question is aligned to a ‘politics of memory, of inheritance, and of generations’ (1994: xvii), to those, he goes on to suggest, who are not present, ‘those who are not yet born or who are already dead’ (1994: xvii), or perhaps exist as an absent-presence.

Hauntological approaches extend well beyond Derrida’s immediate concerns and are suggestive of an ontology of media time made possible by social and digital media. One dominant argument within media studies in relation to the ubiquity and pervasiveness of digital media is that time is experienced as ‘living immediacy’ (Deuze, 2012: 6). Deuze cites the work of Manuel Castells and his concept of ‘timeless time’, where instantaneity and what Deuze also calls ‘placeless space’ (2012: 8) are seen to define the new media times of the 21st century. This argument is suggestive of a perpetual present, where users are ‘always on’, and where immediacy becomes the primary modality through which time is lived. There are many versions of this argument, which assumes that such a logic of immediacy erases the forms of mediation which are shaping such experiences (see Bolter and Grusin, 2000, for example). These arguments become more insidious when it is deemed that there is no ‘outside’ to media ‘that would allow for a comparison between media and life’ (Deuze, 2012: 62).

Rather than assume that media time can be captured by immediacy or instantaneity, or is primarily one of duration framed as flow, movement and change (see Kember and Zylinska, 2012), I will explore the different temporalities and media times that are disclosed when a hauntological imaginary is brought into the analysis and mediation of data associated with the fate of John Bargh’s highly-cited journal article. I am calling this an example of transmedial time (Chow, 2012), a time that Chow suggests is performed hauntologically and characterized by ‘non-linear temporalities’ (Carsten, 2007: 2; Cho, 2008; Gordon, 2008). Chow (2012) describes transmediality in the context of digital and social media through the concept of entanglement (influenced by Karen Barad’s writings), which is extended by figurations such as knots, masks, traps, mazes, shadows, etc. She develops performative methods of capturing such ‘topological looping’ (2012: 1), including montage, collage and tableau. In the next section I will develop the concept of transmedial time by engaging with the work of Hans-Jörg Rheinberger (1994, 2010), a scientist and philosopher who has developed a hauntological approach to scientific controversies. Rheinberger is influenced by Derrida and develops an ‘epistemology of time’. His work adds a useful supplement to the metrics of citation by situating science controversies within a longer period of time, and is illustrative for engaging with how what becomes submerged or displaced by classic or highly-cited studies can always return or re-move, to use Rheinberger’s term.

Controversy Analysis and Hans-Jörg Rheinberger

Hans-Jörg Rheinberger is a significant German science studies scholar who, until his retirement, was based at the Max Planck Institute in Berlin, and whose work, like that of many feminist science studies scholars (Haraway, Barad, Franklin, for example), has produced new objects, entities, methods and ways of thinking at the intersection of science and philosophy. His work was very influenced by Derrida, Haraway, Bachelard, Foucault and Canguilhem, among others. He has been described as a leading historian and philosopher of the biological and life sciences (Lenoir, 2010). His philosophy of experimental practice is one that has many shared ontologies with those taken up within anthropology, sociology and literary studies (those which foreground process, enaction and relationality, for example) and is what Lenoir (2010: xii) refers to as an ‘exercise in historical epistemology’. His work and historical method present a critique of scientific positivism and explore the entanglement of science, the technical and cultural in the production of scientific objects and entities, or what we might term, following Karen Barad (2007), phenomena. Phenomena are akin to what Rhineberger terms ‘epistemic things’. Rheinberger’s approach foregrounds recursion or the patterns of repetition and difference which underpin both the invention of new scientific objects and also the epistemological foreclosure of specific materialized interpretations. Scientific objects are always mediated and become agents in ‘the process of making knowledge’ (Lenoir, 2010: xiii). They are part of ‘experimental systems’ or apparatuses which are performative; they invent rather than discover. However, the processes of what becomes stabilized are always haunted for Rheinberger in terms of displaced and suppressed narratives, which threaten to surface and come back; they exist as traces or deferrals in the Derridean sense (see Derrida, 1995).

The concept of ‘epistemic things’ captures the patterns of difference and repetition that are characteristic of scientific objects and entities. Although science controversies might be considered settled at particular times, Rheinberger (1994) shows how they have the tendency to resurface in new ways and forms. This is something he cogently shows when following the controversies surrounding chicken tumour agents within oncology across time. This is what Rheinberger (1994), following Derrida, refers to as the historical movement of a trace (its haunting, perhaps) – the tension between persistence and transformation. He argues this process is not captured by Kuhn’s (1962) more totalizing notion of a paradigm and a paradigm shift to understand change and transformation within science. Experimental systems are haunted by traces of the past, and these traces, those ‘half-private, half-public conjurations’ (Derrida, 1995: 57), also open to what Derrida termed ‘archives of the future’ – those lost futures of science or science-yet-to-come. Where Derrida asks how archival machines change the object(s) of science, ⁶ I argue that new forms of animation and automation in the form of PPPR and its often volatile displays of affect, emotion and feeling (see Ashley et al., 2013) provide a way of responding to Derrida’s prescient question. The time lags, time shifts and multiple media times and temporalities afforded by transmedial time(s) are forms which can orient attention to the uncertainty and indeterminacy that characterize experimental systems. They present opportunities to analyse the hauntological dimensions of social and digital media and science controversies more generally.

Rheinberger’s influence from Derrida is most telling in the neologisms that he constructs as heuristics, which shape his approach to science and scientific forms of experimentation. This includes the concept of historiality, which draws attention to the multiplicity of times that intrude within experimental systems. The concept also draws attention to science as a storytelling machine where, as he argues, ‘an experimental system has more stories to tell than the experimenter at any given moment is trying to tell with it’ (Rheinberger, 1994: 77). He equates this dynamic potential to older narratives that persist in the future, as well as ‘fragments of narratives that have not yet been told’ (1994: 77). As is apparent from this discussion, controversies are scenes of entanglement where the past and future crisscross, intervene, intrude and open up the potential for something new to emerge. Rheinberger also characterizes this potential dynamism as an excess which escapes definition, that has different momentums, and which allows for a potential tinkering. He also characterizes this process as a form of ‘re-moving’ (1994: 78).

The concept of re-moving expresses what I take to be the hauntological vision of this thesis, the re-moving or animation of frozen time or temporalities which haunt what has come to pass (as scientific truth or certainty, for example). Controversies are potential scenes of entanglement which re-move, or have the potential to set traces in motion – traces which perhaps are not visible or knowable in relation to the intelligibility of particular experimental systems. Re-movings have the potential to perform retro-active re-shapings, but importantly such dynamism is not reducible to scientific research and practice itself. Re-movings are not internal to science or to the internal times and practices of formal scientific laboratories. This dynamic process of re-staging multiple temporalities and the gaps, contradictions, events, figures, objects and entities which might set such a process in motion is captured by the concept of haunted data. Haunted data works with what remains foreclosed despite its absent-presence. I will illustrate this now by exploring the narrative archive that has become associated with the ‘John Bargh priming controversy’, and which takes prominence on Google PageRank.

A Narrative Archive

If you’ve not been following the wires, when a research team couldn’t replicate Bargh’s study everything kicked off and handbags were flailed around by Bargh, a Belgian research group, Nature, a Nobel prize winner and the internet.

Bin your copies of Kuhn people, this is how science really works. ⁷

The experiment which failed to replicate Bargh’s classic study was carried out by a young Belgian post-doctoral researcher and his team (Doyen et al., 2012), with some changes made to the original material culture of the experimental apparatus. ¹⁰ Doyen found it impossible to publish the study in any of the established proprietorial journals in the field. ¹¹ After a history of rejection from these journals it was eventually published in PLoS One, an open-access science journal. The article came to the attention of Ed Yong, who posted the blog referred to earlier, which in turn led to John Bargh’s response.

In the redacted blog posts, Bargh took issue with the experimental conditions of the Doyen study, with the business model of PLoS, and with the practices of science journalists such as Yong (see Blackman, 2015). The interested reader can find hyperlinks to Bargh’s post, commentaries on the post and responses to the responses, but the actual posts now remain elusive. ¹² The subsequent erasure of Bargh’s posts has been met with humour, puzzlement and parody, with the erasures substituted by various images mocking Bargh’s decision to attempt to erase and cover his tracks.

However, this is not all there is to say, or that has been said about this controversy. Yong’s intervention also introduced a different temporality to the proceedings, bringing the controversy to a wider attention and setting in motion submerged narratives, displaced actors and some of the multiple pasts and temporalities which haunt priming as a field within cognitive science. These exist as traces, deferrals, absences and ghost-data, contained within a corpus of data associated with Yong and Bargh’s blog posts. The haunted data are consolidated and carried by the image of Clever Hans, the subject of an earlier 20th-century controversy within psychology, which revolved around a horse who appeared to be able to tell the time and solve multiplication puzzles by stamping his hooves. Yong opens his blog entry with Clever Hans and compares John Bargh to William Von Osten, the owner of the horse and unwitting experimental actor. As Yong suggests, the legend of Clever Hans has largely been consigned to history, although it has caught the attention of many contemporary cultural theorists interested in embodiment and affect (see Despret, 2004). Yong ends this evocative or perhaps provocative opening with the words: ‘But history, as we know, has a habit of repeating itself’.

Yong revitalizes this earlier controversy about a horse who could be made to move and how this controversy was settled within experimental psychology as an example of experimenter expectations, explained as unintended minimal communication from experimenter to experimental subject – in this case Von Osten to Hans the horse (Pfungst, 1911). Yong re-draws this historical association to amplify the conclusions that Doyen had made about his ‘failure to replicate’; that is, that his experimental subjects were not being primed by the semantics of words (associated with ageing) but by the experimenter’s expectations. On page 4 of the article Doyen et al. argue: ‘we were indeed able to replicate Bargh et al.’s priming effect, but only when the experimenter expected the primed participant to walk slower after having been primed’. Although this experimental non-event produced a particular narrative outlined above, Yong’s interventions and the comparison of Bargh to Von Osten produced an archive of ghost-data that followed Bargh and his redacted blog posts, which were kept alive by a network of users and publics (see Blackman, 2015).

A Logic of Pre-emption and Anticipation

Daniel Kahneman, a prominent Noble Prize winning cognitive scientist, sent a letter by email on 26 September 2012, at 9.30am, to a range of undisclosed recipients who Kahneman suggested were named by John Bargh as current students of priming. These students were invited to pass it on to anybody else they might think relevant. It has found its way onto the internet and can be easily found as a PDF copy. ¹³ I am interested in how this letter became an object that entered into the narrative framing of the controversy and its after-lives. ¹⁴ The letter urges students to openly acknowledge the controversies surrounding priming (rather than deny or defend them) – particularly the failure to replicate – and suggests that the problem is more severe within the priming field ‘because every priming study involves the invention of a new experimental situation’.

Rather than invention being conceived as an important component of the dynamism of science (see Rhineberger, 2010), it is rather framed as an insight that needs intense scrutiny via the development of new protocols for examining the replicability of studies and to increase the reputation of replicability within and outside the field of priming research. Kahneman develops nine protocols to do this and urges students to adhere to them in order to re-brand their research field, which he positions pre-emptively as a field that is under threat from ‘outsiders’. This threat, he suggests, might result in a vastly decreased job market for young researchers and in a potentially severely eclipsed field ‘yet-to-come’ (Massumi, 2009). This is affectively consolidated in the form of an image of a looming train wreck, which requires urgent action to remedy. In this respect the address to his intended recipients acts as a form of premediation (Grusin, 2010), a logic which perpetuates ‘an almost constant, low level of fear or anxiety’ (2010: 1) in relation to unspecified threats, which generate the requirement and adherence to particular practices of securitization (in this context nine ‘experimental protocols’). Kahneman presents his invitation to premediate ‘a future before it turns into a present’ as one that perhaps acknowledges how the internet and social media proliferate media transactions and hence the potential shaping of controversies, but closes this down by positioning future researchers as mediators – those who can be ‘actively involved in changing whatever they mediate’. In the next section I will contrast this response to PPPR and the John Bargh priming controversy with a more confessional storytelling practice as it was enacted by John Bargh some 12 months after the priming controversy first erupted.

A Redemption Narrative and Confessional Strategies

Google’s PageRank algorithm currently ranks an article based on an interview with John Bargh in the Chronicle of Higher Education as number 2 in relation to the ‘John Bargh priming controversy’. This article was published on 30 January 2013 and is cast as a sober evaluation of events over the last year (see: http://chronicle.com/article/Power-of-Suggestion/136907/). The beginning of the article is framed by a photograph of John Bargh in his office at Yale taken in subdued lighting (adjacent to a floor-lamp) where he gazes at an angle to the camera. He does not meet the camera’s gaze but looks into the distance, creating a feeling of composed reflection. This is a redemption narrative, one where he regrets his actions (taking down the blog-posts) and where the controversy is framed as a personal assault on his academic career. As the article suggests:

I met Bargh on a Thursday morning a couple of weeks before Christmas. He was dressed in cable-knit and worn jeans with hiking boots. At 58 he still has a full head of dark, appropriately mussed-up hair. Bargh was reclining on the previously mentioned moss-green sectional while downing coffee to stay alert as he whittled away at a thick stack of finals papers. He rose to greet me, sat back down, and sighed.

Bargh’s redemption narrative acknowledges that his removed blog posts were a response to being compared to Mr Von Osten by Ed Yong in his blog post of 18 January 2012. It was this comparison which fuelled Bargh’s anger and his subsequent redacted blog posts for Psychology Today; as Tom Bartlett, the writer of the article for The Chronicle of Education, says:

The post, on the blog Not Exactly Rocket Science, compared what happened in the experiment to the notorious case of Clever Hans, the horse that could supposedly count. It was thought that Hans was a whiz with figures, stomping a hoof in response to mathematical queries. In reality, the horse was picking up on body language from its handler. Bargh was the deluded horse handler in this scenario. That didn’t sit well with him.

In the subsequent 12 months following the controversy, and despite redacting the initial blog posts, the ghostly trails of this comparison have extended across the internet and social media. Various committed and interested users have recovered the posts, circulated them via Twitter (available for recovery by the open access Twitter analytic Topsy), commented upon them, replaced the gaps with various animations mocking Bargh’s responses, and attempted to keep this association alive. The data has been subjected to particular practices of remembering and forgetting and exists within specific regimes of visibility made possible by digital software practices, often used against the grain of predictive analytics. This controversy also discloses the multiple temporalities of science, its submerged pasts and lost futures, those which open up the dynamism of science and also what is policed, managed and actively disavowed within particular scientific fields, knowledge practices and paradigms. Despite the complex politics that surround this data, Bargh’s storytelling retains prominence in Google PageRank, and his confessional story remains as one last attempt to ‘set the story straight’.

Conclusion: Altmetrics, Academic Value and Social Media Impact

As we can see from the fate of a particular highly-cited paper and its after-lives, the new publishing cultures generated by PPPR in science are not without contention or controversy. The new media cultures generated by PPPR exceed the usual debate in academic publishing, which is often polarized between proprietary journals and those that are open access and free to use, and often antagonistically pitted against each other. I would argue that this misses the point. Open access does not guarantee attention and visibility and the use of social media to increase impact is bound up with issues that exceed the actions of particular scholars. These are frequently related to already-existing relations of prestige, status, value and worth underpinned by the social networks the academic belongs to and the social capital that this generates. However, there are other issues which shape visibility and the capacity of an article or book to capture attention. In this sense following the fate of articles as they enter and shape controversies is a useful case-study in highlighting some of the issues at stake.

As with the John Bargh priming example, Bargh’s highly-cited article was revitalized by the actions of a particular science journalist who picked up on the failed replication study by Doyen et al. and intervened to shape the controversy in a particular way. This created a corpus of data which became entangled and shaped within specific regimes of remembering and forgetting. Despite the efforts of particular users to keep Bargh’s redacted data alive, Bargh’s status as a prominent Yale scientist maintains his position in Google PageRank, influencing how the controversy is likely to be remembered in the future. This is despite or even in spite of those cultural intermediaries, such as Ed Yong, who attempted to attach different histories of controversy in psychology to Bargh’s article, and to keep these associations alive.

Despite the incivility and tone of John Bargh’s attack on PLOS-One, and on Doyen and the science writer Ed Yong, it seems it is business as usual. Bargh’s narrative of events is established and authorized by the Google algorithms which enact John Bargh’s Wikipedia webpage as the key player and actor shaping the controversy. This is confirmed by his number 1 page-rank status in Google searches of the priming controversy and by his redemption narrative as a close second. The politics of page-ranking in the context of this controversy confirms the importance and conclusions of Rieder’s (2012) genealogical approach to software studies. His focus is on exploring the conditions of possibility for Google’s automated logic of page-ranking to take form, what he calls its ‘conceptual a priori’. Rieder cogently analyses how software programmers have drawn on and materialized a particular ontology of the network, which is derived in part from sociometry, enacted by a specific algorithm known as Lawrence Page’s PageRank. This algorithm, as analysed by Rieder, shows the exchanges between mathematical theory and the social sciences, and particularly social psychology and the theories of Leon Festinger.

Sociometry is a broad area of study which purports to analyse and, importantly, visualize the psychological characteristics of populations, linked to Kurt Lewin’s topographical approach to psychology, for example. As Rieder argues, the concept of the network which was central to these theories staged connections and links based on status, authority and influence. These concepts were put to work in particular ways where calculations of social status and social power were made from sociometric data, rather than other kinds of connections and influences, for example. The conceptions of social influence that this enacted provided the normative model for various metrics of ranking to be performed. These are based primarily on relations of prestige, hierarchy and canonization, rather than popularity per se. The Google PageRank algorithm therefore makes commitments to certain conceptions of the social, and I would add enacts specific forms of capital (social, educational and cultural) that are part and parcel of scientific networks, paradigms and traditions. In this sense, algorithms are part of a new micro-physics of power, which entangle material, psychological, cultural and historical processes as part of their automated logics. These extend beyond and have the potential to govern and shape the actions of individual users.

Where does this leave the humanities scholar negotiating new publishing and media environments and who is under increased pressure to expand the impact of an already-published article or book? As we have seen, PPPR in science publishing entangles multiple actors and agencies, including social media and open-access platforms, websites, communities of academic scholars and researchers, journal editors, science journalists, media broadcasters and particular interested publics. These entangled relations are simultaneously affective, material, technical, cultural and political. These relations ultimately disclose the new forms of power taking shape in new publishing cultures, which are governed by relations of status, prestige, hierarchy and already-existing norms of academic worth and value.

Perhaps at this conjuncture the humanities and social sciences are at a cross-roads where there is the potential to invent and shape new forms of PPPR, which attend to asymmetries of power and their racialized, gendered, classed and sexed operations. There is an opportunity to learn from the sciences and create something new with the possibility of a more humane post-publication environment. In this respect we can learn from our science colleagues who are all too aware of both the possibilities and problems of PPPR and the issues surrounding the politics of citation that fail to go away. These issues coalesce in relation to the increasing need and desire to disseminate our work to wider publics as part of an impact agenda, and to manage the tensions and anxieties associated with what, where and how to publish. This is all the more vexed as publishing landscapes, publishers, higher education funding bodies and associated business models and legal frameworks change in response to the potentially more open, distributed and decentralized networks of communication disrupting higher education and academic publishing (see van Mourik Broekman et al., 2014). ¹⁵ These issues, I fear, will not be resolved by pitting open-access and proprietorial journals against each other.