Talking infrastructures
An email conversation between Johanna Rivano Eckerdal and Fredrik Åström, Lund University, Sweden.
Introduction
Why are we talking infrastructures when we could be talking big data? One answer would be that the data doesn’t exist in a vacuum — data is collected and made available, there is machinery for storing and using, as well as analysing, data; and there are people and organizations involved in all of these processes. We would like to claim that there is an immediate association between big data and infrastructures; from the big science machines like those at CERN, creating huge datasets, to servers and networks making Google possible.
We’d like to think that the infrastructure perspective we’re discussing has great potential for understanding, maybe not so much the content of big data; but the context of big data, in terms of the environment in which big data is created, managed and analysed, as well as the people and organizations involved in big data. The following is a correspondence between us on the topic of infrastructures and how they can be understood, in part sparked by a seminar paper presented at the Department of Arts and Cultural Sciences at Lund University, in part by a series of workshops called “Big Data - small meaning - global discourses”, organized in collaboration between the Department of Arts and Cultural Sciences at Lund University and the Royal School of Library and Information Science at the University of Copenhagen.
_____________________________________________________________
From: Johanna Rivano Eckerdal [johanna.rivano_eckerdal@kultur.lu.se]
Sent: 18 May 2016 08.47
To: Fredrik Åström [fredrik.astrom@ub.lu.se]
Subject: Infrastructures
Fredrik,
When reading the paper that you wrote together with Björn and Joacim I was intrigued by your use of infrastructure, and I began to think a bit further about infrastructures as a concept. For some time now, there seems to be some kind of hype around infrastructures, and I have some ambivalent feelings about that. Infrastructures have been used as a descriptive concept for a variety of systems for communication and distribution of goods for a long time. From the traditional ones like railroads and electricity to more recent ones like Internet but also systems related to research like the planned European Spallation Source in Lund. They are all labelled infrastructures, and I suppose that it is to render some kind of aura of importance to them.
At the same time, to me and to you too I suppose, as researchers within information studies we are familiar with infrastructures as an analytical concept for understanding systems distributing information of various sorts and their interconnectedness with their surroundings: The work of Susan Leigh Star and Geoffrey Bowker plays a very prominent role for thinking about infrastructures as including not only the systems but also the practices surrounding them and to “become aware of the social and political work that the infrastructure is doing” (Star & Bowker, 2010, p. 242). Getting back to your paper, you adopt infrastructures in this analytical sense there and I really think it is a useful analytical tool.
Infrastructures enable distribution of goods like cargo but also distribution of good things. It is a word that has positive connotations. You cannot say no to investing in infrastructures, as it would be the same thing as saying no to progress and a prosperous future. Malicious infrastructures, have you ever heard of them? Infrastructures are rather smart, facilitating, allowing speed and success. As Sally Wyatt wrote concerning the metaphor of the Internet as a highway on which it is assumed that everyone wants to be driving a fast car, when not driving a car, as well as not using the Internet, might be a conscious, well-informed and preferred choice.
But, how to think about this abundance of labelling almost everything infrastructure around us? Why is that happening? I just wanted to know if it is something that you have noted too and if so, what are your reflections about that?
All the best,
Johanna
_____________________________________________________________
From: Fredrik Åström [fredrik.astrom@ub.lu.se]
Sent: 20 May 2016 15.07
To: Johanna Rivano Eckerdal [johanna.rivano_eckerdal@kultur.lu.se]
Subject: RE: Infrastructures
Hi Johanna!
Yeah, when there is a bit on Comedy Central’s Last Week Tonight with John Oliver on infrastructures (https://youtu.be/Wpzvaqypav8), I think it’s safe to say that it is an issue receiving quite a lot of attention; and yes, it seems like it is a concept that is being used both more, and in more situations than before, perhaps also taking on different meanings depending on who you are listening to. On a general note, it is kind of interesting to note that typically, infrastructures is something that we don’t really pay attention to until they stop working; and the question is if it’s a buzz around the concept or if the increased use is depending on many infrastructures increasingly being in pretty bad shape?
It is kind of interesting to compare how the word is used in different situations. Taking a look at the Swedish government website, the area of interest for the Minister of Infrastructure is “ensuring the provision of economically efficient, sustainable transport services for the general public and businesses throughout the country.” And it looks more or less the same when looking at the agendas of the other political parties as well: there are one or two parties adding ICTs to the agenda when describing their politics in terms of infrastructures. But in all, the focus is on a very narrow definition of infrastructure.
The Swedish Research Council (VR) describes infrastructures for research as “research facilities, facilities for register research, comprehensive data networks and resources for calculation which are used nationally or internationally by several research groups”; and also present a long list of infrastructures with support from VR, where we find large ‘science machines’ like particle accelerators, networks to provide access to data on for instance geodiversity or meteorological information, as well as large scale survey programs. However, within VR there is also talk about infrastructures for research information, including publications, bibliographic databases and so on; albeit separately from other research infrastructures.
If we go to the Oxford English Dictionary (OED), we find the following definition: “A collective term for the subordinate parts of an undertaking; substructure, foundation; spec. the permanent installations forming a basis for military operations, as airfields, naval bases, training establishments, etc.” Then, the OED goes on, listing the use of the word historically, mentioning the typical things related to infrastructures like tunnels and bridges, but also for instance “the infrastructure [as] the regularly produced two- or four-beat meter that characterizes any jazz performance” and “infrastructure of scores of vernacular languages”.
It is kind of interesting to note that the main definition is very specific in terms of both its use and in what setting the term is used, but also how the term has been ‘travelling’ over time.
Since me, Joacim and Björn wrote our paper, I’ve been working some more on using infrastructures as some kind of analytical framework for understanding research evaluation systems and practices. As a starting point, I use Star and Bowker’s ideas on infrastructures and their description of infrastructures, not as specific technical systems, but as relations between technical systems and social aspects, including the people and/or organizations involved.
So, there are widely different ways of thinking about infrastructures; and it seems to be something we hear more about these days. To some degree, it might be something of a buzzword right now, of which I guess I also to some extent take part in. But I also think that there are a lot of infrastructures that are facing challenges: from bridges in the US being on the verge of collapsing, to bibliographical databases increasingly being used for analysing the performance of scholars and scientists, that are not originally being intended for such use, and with a number of problems associated with this practice.
All the best to you too,
Fredrik
_____________________________________________________________
From: Johanna Rivano Eckerdal [johanna.rivano_eckerdal@kultur.lu.se]
Sent: 7 June 2016 08.49
To: Fredrik Åström [fredrik.astrom@ub.lu.se]
Subject: RE: Infrastructures: analysing a quality register
Fredrik,
I think that you have a point in that one reason why we hear so much about infrastructures today is because so many of those that we have been taken for granted as part of our welfare society are cracking. And of course, we too are part of creating the hype around the concept, but I think that we in different ways can offer a critical stance; at least I hope to do so. Recently, Karolina Lindh and I have written an article about national quality registers within the Swedish public health care sector. We find them interesting as research infrastructures, that is, we are interested in pointing out that these systems fruitfully can be construed as infrastructures, in the analytical sense of the concept. Quality registers started out as initiatives from professions within the public health care system as a means to improve the quality in their medical practice. By aggregating data about a particular diagnosis or a treatment it was possible to get a bigger picture and draw conclusions of the quality in a way that was not possible for the practitioners on their own. The number of registers in Sweden is increasing. The Swedish Association of Local Authorities and Regions [Sveriges Kommuner och Landsting] as well as The National Board of Health and Welfare [Socialstyrelsen] realised that the registers could be useful for improving the quality and have made investments in these registers during some five or six years. The initiative has gone from being driven by the professions to being driven by the public health authorities. Besides improving quality another driving force then entered on the arena: using the registers as a basis for saving money.
Karolina and I found another conflict when we analysed a couple of news on Swedish public radio. Data from one register, The Swedish Neonatal Quality Register (SNQ), have been used within a different context: as an argument for a possible change of the limit of permitting late abortions in Sweden. When understanding infrastructures as relational it opens up for discussing the different people and settings that may be involved with the system and that it may hold different roles and meanings in different circumstances. In this way I think that our analysis of quality registers within the Swedish public health care as infrastructures is a case in point of how such a system“ never stands apart from the people who design, maintain and use it” (Star & Bowker, 2010, p. 230).
/Johanna
_____________________________________________________________
From: Fredrik Åström [fredrik.astrom@ub.lu.se]
Sent: 20 June 2016 09.34
To: Johanna Rivano Eckerdal [johanna.rivano_eckerdal@kultur.lu.se]
Subject: RE: Infrastructures: analysing research evaluation systems
Johanna,
You make a really interesting point when talking about how the use of information and infrastructures changes over time, from one intended function to something quite different. In my field, I guess the change from using information in the ISI databases (Science Citation Index and such) for finding scholarly literature to the use of the Web of Science databases for evaluating research - and today the government allocation of funds to the universities in Sweden - would be an obvious example. And I guess the ‘boundary object’ concept proposed by e.g. Star and Bowker that kind of set off our correspondence to begin with, is of some relevance here. And another thing that you touch upon that I think is of great importance when it comes to understanding infrastructures are the people involved.
I mentioned earlier some work I’ve been doing on using infrastructures for understanding research evaluation practices, where I try to discuss them as an analytical framework. More concretely I’ve been trying to operationalize Star & Bowkers notion of infrastructures as “…a number of possible distributions of tasks and properties between hardware, software and people“ (Star & Bowker, 2010, p. 232). I try to do this by suggesting “people” being the stakeholders involved, the “hardware” being the systems and data used in bibliometrics based evaluations, and the “software” being the actual evaluation systems. So, the stakeholders would be: the ones being evaluated -individual scientists and scholars, research groups and universities; the ones evaluating - other individual scholars (for instance assessing grant proposals), university and/or faculty management, and research councils, but also evaluation being done for the aforementioned evaluators by independent consultants that can be either private/commercial entities or independent institutes; and those that I’m calling auxiliary stakeholders, that is, those supplying bibliometric analysts with tools and data (the hardware), such as Elsevier and Thomson Reuters, providing citation data, but also tools for making bibliometric analyses.
When looking at the evaluation landscape and those involved - tying back to your last letter and the issue of the different people or organizations involved - I find it fascinating to see how the different roles between evaluees, evaluators and auxiliary entities are fluctuating. Just to take myself as an example: on one hand I work as a scholar, being evaluated for instance when submitting grant proposals and contributing to the literature from Lund University being analysed in the national Swedish funding allocation; as well as advising people at Lund University how to relate to different bibliometrics based indicators, in my role as an expert in the field. On the other hand, I also perform bibliometric analyses for evaluation purposes, both commissioned by Faculties and central management here at Lund University, and externally as an independent consultant. I guess the only thing missing for me, would be to start developing systems for bibliometric analyses, to also become part of the ‘auxiliaries’, albeit I do work in the group at the University Library managing LUCRIS, and as such, at least by proxy - being part of supplying publication data to for instance SwePub, which of course is being developed to be a better data source for bibliometric analyses (potentially to be used in a new national fund allocation model).
/Fredrik
_____________________________________________________________
From: Johanna Rivano Eckerdal [johanna.rivano_eckerdal@kultur.lu.se]
Sent: 20 June 2016 13.09
To: Fredrik Åström [fredrik.astrom@ub.lu.se]
Subject: RE: Infrastructures: personal experience of interlinked systems
Fredrik,
Your letter really got to me, especially what you write about the fluctuating roles that we have as information studies academics studying the research infrastructures that we also, unavoidably, are involved with; realising that the presentation of the research we are doing in publication systems and research databases is linked to other systems and that the systems in themselves are used as means to evaluate research in problematic ways. Instead of being ways to distribute research the infrastructures are controlled by commercial actors.
When I applied for a research grant in the Swedish application system PRISMA this spring I had to create an ORCID-account. I understand that it is very good to be able to identify individual researchers internationally, to avoid misunderstandings, but I want to know the context that I am connected to. Therefore I checked the ORCID website and I found it to be a non-profit organisation so I agreed to setting up an account. The decision was a bit hasty, but I was ok with it. Then, in the next step, I was required to create a Researcher ID and connect it to my ORCID identity. Researcher ID is a Thomson Reuter product and realising that it is a commercial company that provides the identification and that I just do not have any control over how it will be used made it very hard for me to create that account. But when the choice was between being able to apply for funding or not I did push that button, but it was an effort. I created a path for checking my research, related to a company that do not favour the discipline that I am working in. The feeling of giving assistance to trapping myself is still lingering…
/Johanna
_____________________________________________________________
From: Fredrik Åström [fredrik.astrom@ub.lu.se]
Sent: 18 August 2016 11.03
To: Johanna Rivano Eckerdal [johanna.rivano_eckerdal@kultur.lu.se]
Subject: RE: Infrastructures: evaluation as a neutral practice
Johanna,
I can see why you start feeling a little less than comfortable, feeding into an increasing number of systems that are more or less communicating with each other. Michel Foucault’s ideas on surveillance and governance don’t feel all that far-fetched. They are all made with the motivation of making things easier for us, but at some point, a surveillance aspect of it all becomes more and more pronounced; not the least when combined with evaluation programs and assessment processes such as using Web of Science data (also a product of Thomson Reuters) for resource allocation in research. And it also ties into how infrastructures often are seen as being something neutral, serving all of us without a vested interest, while we actually find a great variety of stakeholders with a great number of things on the agenda. Today, Google might be perceived as being ‘just that place where you search for information’, whereas they actually are a commercial enterprise with specific organizational goals; and in addition to that: algorithms for e.g. personalized search and sponsored links are examples on how the technology being used in itself operates in a way that is far from objective or neutral.
The rhetoric of neutrality is really obvious in the context of research evaluation, where people often talk about evaluation in terms of subjective peer review assessment and objective evaluation using bibliometric indicators. On a general level, this is related to the notion of statistics as something neutral, as opposed to the judgement of an individual person. This can of course be juxtaposed to the Darrell Huff quote: “If you torture the data long enough, it will confess to anything” (Huff, 1954). But there are also more specific aspects. One is the policy decisions behind the use of bibliometric indicators for research evaluation or resource allocation: the choice of data sources and indicators to use has effects on what kind of research and what kinds of publications are being rewarded. For instance, when the Swedish national system for resource allocation was implemented (using Web of Science data for field normalized citation statistics), getting Swedish academics to increasingly publish articles in international scientific journals was an explicit goal.
Another aspect is going back to the issue of the different stakeholders and their various vested interests. Earlier, I mentioned how different roles within the evaluation process can be situated in the same stakeholder, but there is also the issue of how the scholarly communication and evaluation infrastructure to some extent is increasingly being provided by an increasingly smaller number of commercial enterprises. Companies like Thomson Reuters and Elsevier are providing communication channels and databases with publication and citation data; but also tools for bibliometric analyses as well as Current Research Information Systems (CRIS).
Going back to the notion of research evaluation using bibliometrics - and the infrastructures surrounding it - and systems in some way being neutral entities, I think it’s interesting to take into account the ideas on infrastructures presented by Star together with Bowker and Griesemer; and consider how the infrastructures per se relates to those developing, providing and using the infrastructures, as well as those choices being made by the stakeholders.
/Fredrik
_____________________________________________________________
From: Fredrik Åström [fredrik.astrom@ub.lu.se]
Sent: 9 September 2016 13.31
To: Johanna Rivano Eckerdal [johanna.rivano_eckerdal@kultur.lu.se]
Subject: RE: Infrastructures: neutrality in quality registers
Fredrik,
I find the assumed neutrality of infrastructures’ extremely interesting and important too. I told you already about the article that Karolina and I have been writing. A starting point for us when looking into national quality registers within the Swedish public health care was to analyse them as infrastructures. When infrastructure is used as a purely descriptive concept a consequence is that the system seems to become a part of the background in any given situation. It becomes transparent and neutral when research within e g information studies and STS has shown that the systems are both shaped by and play a part in shaping the situations in which they are used or enacted. In the article where they introduced the concept “boundary objects”, Susan Leigh Star and James R Griesemer also defined “infrastructure” where embeddedness and transparency are two significant features of an infrastructure that they put forth. To analyse the variety of objects that are described as infrastructures analytically as infrastructures entails bringing the practices and people involved, engaged and engaging the system, into the analysis. With this broader picture, with a lot more players present, we can show the effects that the interactions with the system have, and that they may be questioned, altered. The systems are not neutral and by offering an analysis that shows the conflicts, matters of power, it is possible to discuss which uses that are the wanted ones. And by whom they are wanted or sought for.
I am fascinated by the rhetoric around infrastructures and it struck me that one reason for widening the scope of what is possible to include as and described as infrastructures are a connection to the earlier, very tangible and concrete uses related to vehicles for transporting people and goods. To build roads, railroads, naval- and airports demands huge economical resources and it is paired with the positive connotations of progress, and prosperity. If something is generally accepted as an infrastructure it has already from the start a connotation of being a huge monetary investment that will be worth the investment in the future.
If infrastructures are the winners what and who is excluded? When research infrastructures are discussed, university and research libraries tend to be left out, when they are as a matter of fact part of the infrastructure for knowledge in society since way back. This is worrying. We should perhaps include the word infrastructure more often when we ourselves talk about and describe libraries. So, getting back to where our discussion started, we should insist on not considering our own adoptions of the concept infrastructure as contributing to the hype but as a pro-active wording, an articulation for change.
/Johanna
_____________________________________________________________
From: Fredrik Åström [fredrik.astrom@ub.lu.se]
Sent: 9 September 2016 15.55
To: Johanna Rivano Eckerdal [johanna.rivano_eckerdal@kultur.lu.se]
Subject: RE: Infrastructures: concluding comments
Hi again, Johanna!
Just a couple of short comments on your last mail. The first one is related to the last paragraph, on the topic on what gets defined as an infrastructure and how it might be seen as a way of legitimizing a certain topic or system. In the Swedish Government Budget Proposal, made public earlier this week, funding for the Royal Library in relation to open access issues, as well as the development of SwePub, was actually brought up under the headline “Infrastructures”, together with large science machines like ESS, large data set management and computational networks.
My second comment relates to the notion of ‘transparency’ - and to some extent maybe also our discussion on neutrality: when investigating various resource allocation systems used at local Swedish universities, we found that there is very little documentation surrounding the use of bibliometric indicators. Although indicators are being used for resource allocation at different levels at most Swedish universities today, it is very hard to find even a formal, documented decision on the use of one indicator or the other, much less descriptions of the systems used, even less so, any attempt at relating the indicators to wider organizational goals; and not to mention, any plans for quality assurance or evaluation of the use of the indicators.
This makes me wonder if the assumed neutrality of the system or infrastructure to some extent makes the notion of transparency seen as, maybe not unnecessary, but implied (given that we are dealing with ‘neutral’ statistics and data). What we find within the evaluation system, or in the infrastructure, is taken for granted to the extent where any accountancy for the system per se is not necessary, which in terms might bring us back to the idea of infrastructures being invisible for as long as they are working. To me, this only make me even more interested in digging into the concept of infrastructures, as well as what we find in terms of stakeholders, the systems per se, and ways they are being used.
/Fredrik
About the authors:
Johanna Rivano Eckerdal (PhD) is a senior lecturer in information studies at the Division of Archival studies, Information studies and Museology at the Department of Arts and Cultural Sciences at Lund University. She belongs to the research group Information practices: Communication, Culture and Society. She works as a teacher and researcher with an interest in information practices, specifically related to sexual and reproductive health, and cultural policy related to public libraries.
Fredrik Åström (PhD) is a reader in information studies and works as a specialist in bibliometrics and research evaluation systems at Lund University Library. He belongs to the research group Information practices: Communication, Culture and Society. He specialises in research on bibliometrics and scholarly communication; recently with a focus on effects of the use of, and policy aspects of, research evaluation systems based on bibliometrics.
References
Åström, F. (2016). Outlining an analytical framework for mapping research evaluation landscapes. Paper presented at the STI/Enid Conference 2016, Valencia, September 14–16, 2016.
Åström, F., Hammarfelt, B. & Hansson, J. (2016). Scientific publications as boundary objects: theorizing the intersection of classification and research evaluation. Paper presented the 9th Conference in the Conceptions of Library and Information Science (CoLIS 9), Uppsala, Sweden, June 27–29, 2016.
Bowker, G.F. & Star, S.L. (1999). Sorting things out: classification and its consequences. Cambridge, Mass: MIT Press.
Government Offices of Sweden (2016). Central government budget.
http://www.government.se/government-of-sweden/ministry-of-finance/central-government-budget/
Government Offices of Sweden (2016). Transport and Infrastructure. http://www.government.se/government-policy/transport-and-infrastructure/
Hammarfelt, B., Nelhans, G., Eklund, P. & Åström, F: (2016). The heterogeneous landscape of bibliometric indicators: Evaluating models for allocating resources at Swedish universities. Research Evaluation, 25(3), 292–305. dx.doi.org/10.1093/reseval/rvv040
Huff, D. (1954). How to lie with statistics. New York: Norton.
Lindh, K. & Rivano Eckerdal, J. (2016). Livsdugliga data i gränslandet mellan vård, styrdokument och teknik, Socialmedicinsk tidskrift 93(3). pp.297–305. Available: http://lup.lub.lu.se/record/83a1ce70-8346-4158-ae6f-8564b41c0225
Oxford English Dictionary (2016). Infrastructure. http://www.oed.com/
Star, S.L. & Bowker, G.C. (2010). How to infrastructure. In: L.A. Lievrouw & S. Livingstone (Eds), Handbook of New Media: Social Shaping and Social Consequences of ICTs (pp. 230–245. London: Sage.
Star, S. L. & Griesemer, J. R. (1989). Institutional Ecology, ‘Translations’ and Boundary Objects: Amateurs and Professionals in Berkeley’s Museum of Vertebrate Zoology, 1907–39. Social Studies of Science, vol. 19(3), 387–420.
Swedish Research Council (2016). Research Infrastructure. http://www.vr.se/inenglish/researchinfrastructure
Wyatt, S. (2003) Non-users also matter: The construction of users and non-users of the Internet. In N. Oudshoorn & T. Pinch (eds) How Users Matter: The Co-construction of Users and Technology, Cambridge, MA: MIT Press (pp.67–79).
