The fifth workshop of the "How Well Do 'Facts' Travel?" project was held on Thursday 27 and Friday 28 March 2008. The main focus was on the labels, packages and vehicles needed to make small facts (such as data) travel to and be used in new scientific domains.
Despite the crucial importance of circulating small facts across research contexts, the means and consequences of such travel are not well understood. On the one hand, travelling strategies need to be efficient, so as to make results available to any researcher who might be interested in them.
On the other hand, the labels, packages and vehicles used for travel might compromise the content of small facts, insofar they involve the modification and re-interpretation of facts to fit the new research context in which they are received and used. Invited participants discussed and compared various ways of packaging small facts, including classification systems and taxonomies; standardisation procedures and modelling techniques; material artefacts such as digital databases and repositories; and various kinds of agency involved in the construction and maintenance of these packaging tools.
About the Workshop
Scientific research establishes many facts about the world. Some of these are big facts: that is, general claims about phenomena such as 'the earth revolves around the sun' or 'smoking causes lung cancer'. Obtaining big facts is one of the main goals for both science and society at large. These are the types of facts that science is expected to contribute to society: they contain knowledge that helps us to deal with the world. However, they are not the type of facts that scientists usually work upon. In their everyday activities, researchers are mostly concerned with obtaining and interpreting small facts about the world: measurement, data points, observations derived from experiments or fieldwork.
Small facts are always obtained under specific conditions and in a particular epistemic culture. At the same time, they need to be made available to researchers in other contexts, so as to allow for comparisons, inferences and interpretation. Much of scientists' efforts are directed towards finding strategies and tools to disclose their results and share them with their peers, so that they can be used as evidence for big facts. The advancement of science depends on how small facts are made to travel.
This workshop focuses on the labels, packages and vehicles needed to make small facts travel to and be used in new scientific domains. Despite the crucial importance of circulating small facts across research contexts, the means and consequences of such travel are not well understood. On the one hand, travelling strategies need to be efficient, so as to make results available to any researcher who might be interested in them. On the other hand, the labels, packages and vehicles used for travel might compromise the content of small facts, insofar they involve the modification and re-interpretation of facts to fit the new research context in which they are received and used.
(Programme can be found here.)
Abstracts and Speakers
- Thomas Stapleford: Marketing Statistics: Commodifying Information on Family Purchases and Incomes in the United States
Beginning in the late 1930s, US government surveys of family incomes and expenditures became important data sources for market research. Although several contingent factors made the 1930s a ripe period for such a convergence, this talk will explore the structural factors that helped to build and sustain a pipeline of information from individual families to government statisticians to private market research firms. I argue that this story provides an illuminating case study of the production of commodified small facts: bits of information that can be taken, reassembled into new forms, and (crucially) resold by other actors.
- James Griesemer: Tracking and Colligating Work Make Scientific Facts that can Travel
I argue that much of scientific activity can be helpfully interpreted as tracking work: marking and following processes, documenting trails of association with working representations, and producing and propagating facts. As scientists track processes, they create, represent, and propagate the trails that colligate (bundle together) associations, representations and facts. A key consequence of this 'colligation' activity is that scientific facts are made in the same movements through which they are represented; they are made mobile in the same movements through which they are made, to some extent, immutable (Latour 1988). For a fact to travel (e.g. among specialties or disciplines or contexts of application and policy), it must retain some sense of validity and identity immutability across contexts. And for a fact to remain immutable, it must take on those aspects of abstraction and generality that resist change (through tinkering in a local context). I draw upon a series of vignettes from historical and contemporary biology to illustrate tracking activities in which abstractions from representations in local contexts of production render facts that are able to travel and resist change. The aim of the paper is to characterize some ways in which 'big facts,' such as the central dogma of molecular genetics, the causal separation of genotype and phenotype, and the identification of the genetic material with DNA, arise from and then travel as the representatives of 'small facts' created in tracking work. Key examples from the history of biology of tracking work that produced small facts which are now represented by big facts bundled with associations and representations resistant to change include: (1) tracing the point of differentiation of germ cells to the 9th cell division in the worm, Rhabditis Nigrovenosa (Weismann), (2) the inability of natural selection to alter seed weight in a pure-breeding line of the common bean, Phaseolus vulgaris (Johannsen), the same plant to which Mendel found that his 'small facts' about peas could not easily travel. Each of these 'small facts,' discovered at the laboratory bench (or experimental garden), whether produced in a moment of diagram abstraction, experimental control, or mathematical idealization, became 'big facts' that guided vast areas of research until the late 20th century.
- Geoffrey Bowker : The Fog of Data: Memory, the Past and Computers
We hear that memory is cheap nowadays. However, cheap isn't necessarily the issue. We are reconfiguring our social and natural landscapes as we database them. I discuss dimensions of the social, political and ethical values that get written into databases and structure our access to the past of the earth, of species and of society.
- James McAllister : Middle-Sized Facts Travel Best
If 'small facts' are the outcomes of individual observations and measurements which make up empirical data sets and 'big facts' are high-level scientific findings about universal regularities and causal relationships, then it is necessary also to acknowledge a category of 'middle-sized facts': these are patterns that scientists pick out in empirical data sets. The role of patterns in data is crucial in empirical research: patterns have greater stability than largely ephemeral and idiosyncratic data points, and they provide the empirical evidence for high-level scientific findings. These 'middle-sized facts' have a greater facility for travel than either 'small' or 'big facts.' Individual data points are rarely transferred between research approaches or disciplines: they have little evidential significance outside the context in which they have been gathered and without the interpretation placed upon them by researchers in that context. Likewise, high-level scientific findings are often too intrinsically the product of a specific research approach or discipline-for example, too laden with perspective-to be easily transferred. By contrast, patterns identified in data sets are good travellers: they carry enough interpretation to even out the variability of empirical data, but not so much as to predetermine their use or significance at their destination.
- Steffan Müller-Wille: Indexing Nature: Carl Linnaeus and the Origins of Binominal Nomenclature
Early modern naturalists were faced with what has been termed the 'first bio-information crisis.' A key figure in resolving this crisis was the Swedish naturalist Carl Linnaeus (1707-1788), who in 1751 introduced 'trivial names' which provided a universal and stable index of plant and animal species. In hindsight, the virtues of this innovation seem obvious. Trivial, or binominal names, as they were later called, allowed naturalists to refer to animal and plant species unambiguously, and thus to communicate and accumulate incremental pieces of information about these species. These names, that is, worked like universal labels, that were able to connect information often gathered from widely disparate contexts. This paper will take a fresh look at the origins of binominal nomenclature, by focussing on Linnaeus's day-to-day working routines on the basis of manuscript material held at the Linnean Society (London). What this material shows is that Linnaeus had to manage a conflict between the need to bring observations into a fixed order for purposes of retrieval, and the need to integrate new observations into that order. A way out of this dilemma was to keep information on particular subjects on separate sheets, sometimes folded to form small fascicules, which could be complemented and reshuffled. Surprisingly, binominal names played a very minor role in this. Linnaeus rather relied on practices of numbering and bibliographic referencing, which had a much longer historical legacy in natural history. What I thus hope to show in my paper is that one of the main cognitive advantages commonly assigned to writing - the possibility to abstract words from their linguistic context - had to prevail over considerable practical and psychological obstacles. What seems so obvious in hindsight, namely that names can perform the function of labels, had to be learned in a painstaking process of information gathering.
- Filippa Lentzos: Communicating security-sensitive life science research
This presentation will consider how the labels, packages and vehicles used to communicate security-sensitive life science research shape its reception in various contexts. Drawing on two controversial articles - one on engineered mousepox virus published in the Journal of Virology in 2000, the other on polio virus synthesis published in Science in 2002 - it will follow the journey from experiment to publication to reception by various audiences. It is hoped the presentation will provide a helpful starting point for further discussions on how security concerns impact the way in which facts travel.
- Sabina Leonelli: Making Small Facts Travel: Databases in Model Organism Biology
Bioinformatics is devoted to creating tools and strategies to disseminate small facts about model organisms. This is a task complicated by the sheer size and diversity of the data to be circulated, and the fragmentation of biological research into countless specialised communities. Once small facts are produced, whether and to whom they become accessible depends on how they are disclosed and circulated beyond their context of production. This paper examines the labels and vehicles created by bioinformaticians to package facts for travel across research communities. I argue that there is much to be learnt from bioinformatics about what may count as 'good packaging': tools such as bio-ontologies and databases are built to satisfy the diverse competences and needs of their users, as well as to accommodate the variation in the usability of small facts as evidence within different contexts. At the same time, the case of bioinformatics highlights some of the thorniest issues involved in the travelling of facts: the difficulties in checking their reliability; the need for users to trust 'packaging specialists' such as database curators; and the excessive centralisation of packaging expertise, which leaves little scope for dissent over the definition and use of labels and vehicles.
- Joan Fujimura: Population Labels and Software Programs in Studies of Genomics and Diseases
In the practical production of knowledge, some scientists take into account past, present and future contingencies of traveling facts. In collaborative research across disciplines and fields, this accounting oftentimes produces difficulties and debates. This paper addresses future accounting and multiple audiences through a study of a recent genome wide association study. I focus on a disagreement between two collaborators on a large project as they take account of past, present and future audiences in their different disciplines and professional situations. The disagreement focuses on their different naming of outcomes of a software technology devised to neutralize race and race labels in the conduct of disease genomics research that involves the use of human genetic variation and different populations. The software package is a device for both doing "good science" and avoiding concerns and critiques from bioethicists and various publics about the use of race in genomics research. The introduction of different names and interpretations of the software package's results in this case reduces the package's advantages for insuring robust travel through wider worlds and increases its robust travel to a narrower disciplinary audience.
Speakers and Participants
- Jon Adams, LSE
- Marcel Boumans, University of Amsterdam
- Geoff Bowker, Univeristy of California, Santa Clara
- Joe Cain, University College London
- Jane Calvert, University of Edinburgh
- Hasok Chang, UCL
- Albane Forestier, LSE
- Joan Fujimura, University of California, Berkeley
- Miguel Garcia-Sancho, Imperial College London
- James Griesemer, University of California, Davis
- Midori Harris, Cornell University
- Janet Higgins, John Innes Centre
- Peter Howlett, LSE
- Christopher G. Larminie, University of Manchester
- Filippa Lentzos, LSE
- Sabina Leonelli, LSE
- Harro Maas, University of Amsterdam
- Erika Mattila, LSE
- James W. McAllister, University of Leiden
- Julia Mensink, LSE
- Martina Merz, University of Lucerne
- Ashley Millar, LSE
- Mary Morgan, LSE
- Staffan Mueller-Wille, University of Exeter
- Ed Ramsden, LSE
- Max-Stephan Schulze, LSE
- Tom Stapleford, Harvard University
- Chris Taylor, University of Manchester
- Stephan Turner, University of Florida
- Aashish Velkar, LSE
- Patrick Wallis, LSE