An Introduction to the social situation of SRM research
What the SRM knowledge-network looks like, how we can map this out, and my attempt to utilise some of the STS concepts I spent the summer learning
The purpose of this substack is primarily to explore the downstream impacts of SRM research and other upstream actions. Despite this, the first set of posts will primarily describe the social situation of research, because this social network fundamentally defined the substrate where these relevant interactions occur. The relevance of social structures in the organisation, imagination and perceived success of a scientific endevour has a long history, ranging from early works including those of Fleck’s thought collectives and Kuhn’s paradigms, through actor-network theory , epistemic communities and knowledge networks. Suffice to say, there is a lot.
Much focus of work, particularly in Science and Technology Studies, has been about demystifying the privileged epistemic status of science1 by opening up the black-box of social interactions2 that construct science and scientific fact. By highlighting the normative and discursive underpinnings of the scientific enterprise, a variety of normative aims are commonly pursued, most commonly a call for democratisation and a criticism of the power dynamics going on.
Whilst this is laudable, it is not the project I am interested in. Rather, the aim here is to understand the workings of this black-box in order to both understand how actions we take now within this structure might effect downstream actions, or to understand how we ought to remake these structures to achieve our relevant normative ends. Those normative ends, as discussed in the previous post, are chiefly about reducing existential and global catastrophic risk (not that it is my sole normative end, but the most relevant for this project). As the chief focus is on the actions that can be taken, the description of the social world is not intended as critical, and in that sense flirts with positivism, but is intended to inform action to pursue specific normative ends, which may inform reconstruction of this social world.
Knowledge Networks
Moller 20233 has done a far more professional job than I will at exploring the knowledge network of SRM; however, the utility of a substack like this is the ability to examine some of those features of the social world that don’t fit the strict methodology of studies like that.
Knowledge networks, in the way that Moller uses them refers to a combination of the concept of epistemic communities with the concept of transnational advocacy networks. Epistemic communities are, according to Cohendet "groups of knowledge-driven agents linked together by a common goal, a common cognitive framework and a shared understanding of their work4", thus sharing methodological, epistemic and commonly normative values. However, this concept fails to appreciate the role of advocacy and other non-knowledge-driven agents in the network formed around and in opposition to SRM research and deployment. Transnational advocacy networks are networks of activists, with shared values or norms, shared discourses, and with dense flow of information. The concept here of a knowledge network combines the knowledge driven epistemic communities with the normatively driven transnational knowledge networks.
I will differ slightly from Moller’s definition, and from the canonical definition of knowledge networks. For Moller, knowledge networks have to be engaged with the shared normative values, shared policy enterprise and a shared understanding of how the world works. I think for the SRM knowledge network universally shares none of this between every member. However, every member shares normative values, policy enterprises and worldviews with enough people in the network that in turn share this with others in the network that it is relevant to conceptualise this as a single interconnected network. Moreover, all in this network are orientated enough towards a shared governance object, that of SRM, such that their worldviews and interests are sufficiently similar that they must interact with other members of the knowledge network. It may be argued that it would be better for me to refer to the SRM knowledge networks with an S rather than knowledge network, however, the level of interconnection between these networks is enough that I am happy with the singular.
Formal Institutions
The knowledge network of SRM is fundamentally an informal arrangement, however, it contains both formal institutions and informal communities. The formal institutions form a more traceable and understandable part of the knowledge network of SRM. These are institutions or projects with a well defined spatio-temporal existence, with often legal existences, including research insitutions,5 conferences6 (which indeed Moller uses the attendees of to map out the knowledge network), civil society groups78 and learned societies9.
These formal institutions are clearly act as important sites for the organisation of the knowledge network, as well as a site where discourses can form and strengthen. Oomen 202110 explores the different imaginings present in both the Harvard Solar Geoengineering Program and the German SPP group, highlighting how the discourses present in formal institutions differ, as well as can be vital in forming the discourses of the SRM knowledge network as a whole. Similarly, Moller 202311 uses the attendees of conferences to map the knowledge network as a whole, again identifying these institutions as key sites of the formation of the knowledge network. Important informal connections happen at formal institutions, which may remain important linkages in informal networks after the actors are no longer at the same site. Prestige can be conferred by formal institutions onto individual actors, and criticism can be directed at formal institutions as a proxy for the broader knowledge network.
However, as well as viewing these institutions as sites where actors communicate and co-create science and knowledge, they can also be viewed as actors in themselves, contingent but non-reducible to the sum of their parts. Institutional actors can be seen to have identities connected to SRM to very different degrees, and this is important for positioning them in the knowledge network. Some institutions are explicitly set up in relation to SRM, and can be seen as having their core identities in relation to the knowledge network, whilst other institutions have primary identities elsewhere, but nonetheless apply themselves to the governance object of SRM and so become a part of the relevant knowledge network. Meanwhile, the identities of actors within these institutions may be constituted differently still, but this is obscured enough that the institutional identity often has the stronger discursive effect.
Institutions act to flatten the variation of their members, and present a quasi-unified face through the appointment of representatives, whom whilst having their own interests (and clearly pursuing them), do not speak on behalf of them. Thus, when Janos Pasztor speaks in a capacity as the head of C2G, he doesn’t really speak as solely Janos Pasztor, but rather as a spokesperson for a diversity of people within C2G, and other stakeholders that the institution can reasonably be seen as representing. In this way, he becomes both a powerful and essential intermediary, and his own views get obscured and flattened. Similarly, examining the C2G website, there are a variety of guest blogs and videos of people representing themselves, as individual actors, but the organisation of these into one place, translated through the website, has a distinct discursive effect independent of the sum of its parts, in this case legitimising discussion into SRM.
The NAS report into SRM can be seen as a good example of this. The NAS has an institutional identity outside the SRM knowledge network, however contains a variety of actors with differing relations and ideas on SRM. The report itself set itself up as a unified representation of the NAS12, thus obscuring the different beliefs of the scientists that make it up13, instead representing them as a spokesmen. In turn, as an institution, the NAS can be seen to represent the much large stakeholder of ‘scientists’14. Thus, a report represents an institution which represents its members which represents scientists at large. In this way, I hope it is illustrated how formal institutions as actors can be conceived of as non-reducible to the people that make it up15.
Informal Communities
Informal communities are pretty diverse, and can exist at a variety of different scales. These refer to essentially tighter knit informal associations with some sense of shared identity that distinguishes them from the broader knowledge network. This shared identity can be due to a shared ‘thought style’, or due to demographic or social similarities. These informal communities are much less well defined, much harder to trace, and whilst some approaches, such as citation studies, network analysis to twitter communities or interviews could function, it seems unclear this would capture the full texture of the informal communities that exist in SRM.
Some of these informal communities are sub-units of formal institutions, whilst others are super-units, encompassing individual actors and formal communities within a single informal community. In many ways, these are just finer grained versions of the overall knowledge network, which can be seen as the largest ‘informal community’ that it is useful to define as ‘SRM related'.
One obvious way these informal communities organise themselves is via the concept of invisible colleges16. This is groups of scientists publishing and communicating with one another, often collaborating, on a subject of shared interest despite being geographically disperate and a lack of formal affiliation. Whilst this could be used to describe the whole knowledge network, the concept lacks some of the relevant conceptual richness. Rather, it perhaps better describes certain aspects of the knowledge network, such as the academics who regularly communicate on the Geoengineering Google Group, as incorporating an invisible college. Ideas get generated, discussed and debated, as if in a seminar room, research and social connections are made, and ideas gain currency through mediation by the group. Invisible colleges can also be seen as forming around institutions, as people join, leave and enter into the orbit of these institutions, regularly publish together and maintain social ties. The Brown Climate Social Science Network17 might also be considered an example of one of these invisible colleges gaining a sense of quasi-solidity.
Another example of an informal community might be the SRM twitter communities. There are a number of accounts on SRM twitter (eg Geoengineering1, PeteIrvine, DanVisioni, OliverMorton etc) whom semi-regularly post and reply to one another, predominantly on SRM but sometimes on issues that are more diverse. I have personally benefited from this informal twitter network, with my initial contact with Aaron Cooper, Olaf Corry and Kevin Surprise all being mediated through this SRM twitter community/ies. Whom one is in the same community as might differ for different people, and these networks are constantly in flux.
Others can form for short periods of time before essentially disbanding. The early career researcher journal club was essentially one of these, forming and running for a few months with changing people in it, before essentially stopping functioning. Nonetheless, this acted as another communicative tool, helping construct new relations within the knowledge network. Similarly time limited research projects, such as the LEISE report I was part of, can also highlight some of the formation and then disbanding of specific communal structures.
Some of these informal communities can also have passive members, including chiefly passive membership. The community that listens to the Reviewer 2 Does Geoengineering podcast18 is unstable, unknown and do not know each other. Nonetheless, this may be an important way people make connections (eg through hosting an episode of this19 I first got in contact with Daniele Visioni) and to get their name known as well.
Some of these informal connections can also have a clear ideological bent. There is, to some extent, an unspoken divide between those who broadly support SRM research and are at least open to deployment to some extent, and those highly critical of SRM research and deployment. It may be said that there is an emerging faction of people pushing for immediate or near-term SRM deployment even without research and governance, although this faction is at this stage small, unorganised (and perhaps unconnected to one-another) and somewhat seperate from the knowledge network, although with the action of Make Sunsets, this may become considerably more relevant. This can be seen as mapping onto a variety of the imaginaries that Oomen 2021 explores. He explores this primarily with reference to a clear formal institutional situation, but their also exists more diffuse social networks centred around them as well. Like sides on the political spectrum, it makes sense to speak about those who are ‘pro-research’ and ‘anti-research’ not just as in-depth imaginaries as Oomen does, but social camps as well. These are generated from social connections made by working together20, speaking the same language, and having broadly similar worldviews, with the social networks both producing and being produced by the worldviews associated with them. In some ways, these politco-social coalitions may be soft versions of Fleck’s thought collectives21, communities of thinkers mutually exchanging ideas with shared thought styles that allow for the establishment and construction of facts about reality.
Sometimes, these politico-social coalitions do have shared imaginaries as complex and rich as the ways of seeing that Oomen elaborates on. On the other hand, sometimes the coalitions are orientated towards shared goals irrespective of ways of seeing, or between people with different ways of seeing and normative aims but coming to similar conclusions about near term goals, such as the desirability of specific small scale research projects. Moreover, sometimes these coalitions can be at least partially due to historical and personal linkages; for example, despite my relatively large differences in ways of seeing with Andrew Lockley, we nonetheless have strong social linkages, chiefly for historical reasons, allowing the building of mutual trust and respect.
Sometimes, as expected, the diversity of motivations for linking these broad politico-social coalition can lead to conflict or division over the goals of shared projects. Such can be seen in the Non-Use agreement, where there were seemingly divergence between the views of different signatories as to whether a ban on public funding of research is desirable22, whether SRM deployment is ever desirable and more. Similarly, the same project can be interpreted by different people, all broadly supportive, as having different intentions or effects. I know people who broadly support SRM research but nonetheless see deployment as dangerous and undesirable and see research as uncovering that fact, whilst others see research as allowing for safe deployment to happen, and others still seeing SRM research as purely there to legitimise SRM deployment. Some look at experiments like SCoPEx as useful due to the data they would gather being essential to rational, model-based decision making around how and whether to deploy, whilst others seem them as useful political tools for legitimising SRM research. Others opposed the process, either because they see it as part of a slippery slope to legitimisation, or as overly focused on public consent for a basic science project in a way that was counterproductive to its aims. Some see encouraging broad discussion of SRM as an important part of a democratisation process that is a fundamental goal in its own right, others see it as a means to an end of promoting SRM, and others see it as a means to an end of opposing SRM
This means that these coalitions ought to be constantly in flux, as the actors that make them up have goals (based on both normative ends and ways of seeing) that are divergent from others in their coalition. Moreover, they are likely to have different futures they perform23 despite having shared near term goals; similarly, people who oppose one another may have similar futures performed differently. If we hold the social world has to be continually performed to remain even somewhat fixed24 , and divergent goals would lead to difference performances by different actors, and so not remain fixed. Tribalism (often, knowingly or unknowingly, enforced due to power imbalances), social ties, the lack of resolution of important discussions that might lead to rearrangement, and the existence of these informal coalitions in the orbit (and often shadow) of formal institutions has meant that the same coalitions continue to be performed, and are likely to remain as such for the foreseeable future. This makes attempting to understand these social networks as they are, treating them as a quasi-fixed object of the social world, fruitful. They seem in some sense ‘sticky’, and thus understanding the way in which this landscape responds to any action taken may be very important to achieving my normative ends.
An aside- Remaking the social world around SRM
Nonetheless, I am also aware that another approach might be fruitful, that of attempting to remake the social world around SRM by creating (or describing) new and fruitful coalitions. It is indeed arguable this is what I am doing through the RESILIENCER Project. I am trying to work alongside people from all sorts of different ‘camps’, and inviting them to collaborate together. My Utrecht workshop25 had a diverse attendees list, from people who are sceptical of SRM and its research as currently constructed (eg Jeroen Oomen), those broadly favourable to SRM generally (eg Andrew Lockley), those favourable to SRM research but not deployment, those sceptically favourable to certain types of SRM research, those outside but adjacent to the SRM knowledge networks (eg Detlef Van Vuren). Discussions did involve those discussions which are associated with performing existing coalitions (such as the "what could convince me that I’m wrong" session), but others involved discussions that are inherently cross-cutting.
This may reflect some of the key features of the existential risk question. Some of these are reflecting of the core questions that different coalitions are concerned with: the ethics of research26, the linkage of research to deployment27, the ethics of deployment, desirability of different governance regimes28 etc. But other questions that are perhaps key are somewhat different to these; questions around SRM and its relation to other catastrophic events, questions around how to act given the unpredictability and unquantifiability of SRM’s relation with XRisk specifically, how to minimise SRM’s contribution to conflict. Of course, these discussions interact with coalition-orientated questions, but the uniqueness of the XRisk case may allows for some rearrangement to form, as it could be that what is useful for reducing XRisk is very different from what is good for other reasons. The world is exceptionally complex, and thus we shouldn’t necessarily assume what we hold as likely for the median case applies at the heavy-tail. Thus, the shared beliefs and normative aims that have helped form, and continue to perform, the different coalitions in the general SRM discussion needn’t be an insurmountable obstacle for forming new coalitions around SRM + Climate Change’s interaction with GCR and XRisk.
Nonetheless, there are features of analysing the relevance of SRM research with XRisk which interact with these questions that in their answering help perform the current social world. Moreover, their are other features which require some degree of prediction of how that social world will react to research. The extent to which one has power to reshape the social world perhaps informs which mode of analysis, taking the social world as more plastic or more rigid, one takes. As an individual actor, I have the potential to reshape the social world where their aren’t ‘sticky’ coalitions already, where there aren’t ingrained practices being continually re-performed to remake the structures we see again and again. However, in general, be it looking at the SRM knowledge-network or how it relates to the wider world, I’m not in a position of so much power, so treating these features as quasi-fixed, and using that to inform what action I should take to achieve my normative goals, is more fruitful.
Limitations
The SRM knowledge network does not exist in isolation, and other ‘fixed’ facts of the social world, including structural issues such as inequalities between the Global North and Global South, and imaginaries such as securitization of climate politics, map onto the SRM community. Both formally and informally, the SRM knowledge network is very Global North centric, and this seems to be true irrespective of ideological bent; for example, whilst those interested in the carrying out of SRM research have been criticised for being Northern-centric29 (although the Degrees Initiative30 is certainly doing large amounts of very admirable work to try to address this), less than 20% of the signatories of the Non-Use Agreement31 were based in the Global South, despite prominently claiming to have Global South representation32. This is clearly a feature of the SRM Knowledge Network, although not solely a feature of this, as this Global North bias exists in much of the world, including in climate science and ethics specifically, and international negotiations and academia at large.
Moreover, the knowledge network interfaces with the rest of the social world in vital ways. The “Make Sunsets” fiasco has shown this most clearly. It is probably false to group Luke Iseman and Andrew Song as part of the the SRM knowledge network (certainly prior to the public announcement of the company), and yet nonetheless they have been profoundly influenced by research from within the knowledge network33, as well as the political impacts of their actions leading to a potential rearrangement of the fortunes of actors within the knowledge network34. So the purely 'internal' view of how the knowledge-network works is clearly limited.
Thus, this modelling of the knowledge network in quasi-isolation, as containing quasi-fixed structures and imaginaries is only valid under a relatively restricted set of conditions. This is:
When the knowledge network is suitably rigid in relation to the relevant actor (me) such that whole-scale reshaping of the social world around SRM seems infeasible. Put another way, where patterns of performance become fixed enough that the most effective way to have change is through engagement with those patterns of performance rather than the creation of whole scale new patterns of performance independent to those
When the knowledge network is suitably influential with regards to a particular issue that the social world outside the knowledge network primarily interfaces with this issue through mediation by the knowledge network. This means that understanding how the knowledge network interfaces with an issue becomes a suitable proxy for how the social world as a whole interfaces with this proxy.
Both of these conditions are very defeasible, and as SRM becomes more salient as a governance object35, the degree to which these (in particular point 2) are valid will change. Similarly, as I get more influential (assuming I do), the degree to which point 1 is valid will also change. Nonetheless, these are the assumptions as to what the rest of this blog series will focus on.
Bloor, D. (1984) ‘The Strengths of the Strong Programme’, Scientific Rationality: The Sociological Turn, pp. 75–94. Available at: https://doi.org/10.1007/978-94-015-7688-8_3.
Pinch, T.J. and Bijker, W.E. (1984) ‘The Social Construction of Facts and Artefacts: or How the Sociology of Science and the Sociology of Technology might Benefit Each Other’, Social studies of science, 14(3), pp. 399–441.
Möller, I. (2023) ‘The Emergence of Geoengineering: How Knowledge Networks Form Governance Objects’, in Elements in Earth System Governance. Cambridge University Press.
Cohendet, P. et al. (2014) ‘Epistemic communities, localization and the dynamics of knowledge creation’, Journal of Economic Geography, 14(5), pp. 929–954.
Harvard’s solar geoengineering research program (no date). Available at: https://geoengineering.environment.harvard.edu/ (Accessed: 11 August 2022).
CEC (no date) Climate Engineering in Context 2021. Available at: https://www.ce-conference.org/ (Accessed: 18 July 2022).
ETC Group (no date) Climate and Geoengineering, ETC Group. Available at: https://www.etcgroup.org/issues/climate-geoengineering (Accessed: 18 July 2022).
C2G (no date) Carnegie Climate Governance Initiative, Carnegie Climate Governance Initiative. Available at: https://www.c2g2.net/ (Accessed: 18 July 2022).
National Academies of Sciences and Medicine (2021) Reflecting Sunlight: Recommendations for Solar Geoengineering Research and Research Governance. Washington, DC: The National Academies Press.
Oomen, J. (2021) ‘Imagining Climate Engineering’. Available at: https://doi.org/10.4324/9781003043553.
Möller, I. (2023) ‘The Emergence of Geoengineering: How Knowledge Networks Form Governance Objects’, in Elements in Earth System Governance. Cambridge University Press.
National Academies of Sciences and Medicine (2021) Reflecting Sunlight: Recommendations for Solar Geoengineering Research and Research Governance. Washington, DC: The National Academies Press.
Seen most obviously in the public criticism Michael Mann gave of the report, despite being a member of the NAS: https://michaelmann.net/content/my-comments-new-national-academy-report-geoengineering
Harvey, C. and E&E News (By Chelsea Harvey, E&E News on March 26 2021) ‘Solar Geoengineering Should be Investigated, Scientists Say’, Scientific American. Available at: https://www.scientificamerican.com/article/solar-geoengineering-should-be-investigated-scientists-say/ (Accessed: 6 February 2023).
Much of the previous paragraphs have drawn on Actor-Network Theory, although a fuller ANT description of events that occur within the SRM knowledge network is for another time
Zuccala, A. (2006) ‘Modeling the invisible college’, Journal of the American Society for Information Science and Technology , 57(2), pp. 152–168.
https://cssn.org/climate-social-science-network-launched-with-hub-at-brown-university/
Lockley, A. (no date) Reviewer 2 Does Geoengineering, Apple Podcasts. Available at: https://podcasts.apple.com/us/podcast/reviewer-2-does-geoengineering (Accessed: 18 July 2022).
Daniele Visioni and Gideon Futerman discuss the Event Solar Geoengineering: Warnings from Scientists, Indigenous Peoples, Youth and Climate Activists (no date) Reviewer 2 Does Geoengineering. Available at:
(Accessed: 2 August 2023).
eg, I only met Iris de Vries and Martin Jannsens whilst working on: Futerman, G. et al. (2022) ‘Governance and science implications of low environmental impact outdoors solar radiation management experiments’, in, pp. EGU22–7797, and both attended the RESILIENCER workshop: Futerman, G. (2023) Workshop Report: Ramifications of Experimentation into SRM In Light of its Impacts on Existential, Negative-state and Civilisational Endangering Risk.
Fleck, L. (2012) Genesis and Development of a Scientific Fact. University of Chicago Press.
eg
Oomen, J., Hoffman, J. and Hajer, M.A. (2022) ‘Techniques of futuring: On how imagined futures become socially performative’, European Journal of Social Theory, 25(2), pp. 252–270.
Latour, B. (2007) Reassembling the Social: An Introduction to Actor-Network-Theory. OUP Oxford.
Futerman, G. (2023) Workshop Report: Ramifications of Experimentation into SRM In Light of its Impacts on Existential, Negative-state and Civilisational Endangering Risk.
Biermann, F. et al. (2022) ‘Solar geoengineering: The case for an international non‐use agreement’, Wiley interdisciplinary reviews. Climate change, 13(3). Available at: https://doi.org/10.1002/wcc.754. and Bodansky, D. and Parker, A. (2021) ‘Research on Solar Climate Intervention Is the Best Defense Against Moral Hazard’, Issues in science and technology, 37(4), p. 19–.
Reynolds, J. (2015) ‘A critical examination of the climate engineering moral hazard and risk compensation concern’, The Anthropocene Review, 2(2), pp. 174–191. and McLaren, D. (2016) ‘Mitigation deterrence and the “moral hazard” of solar radiation management’, Earth’s future, 4(12), pp. 596–602.
Reynolds, J.L. (2019) The Governance of Solar Geoengineering: Managing Climate Change in the Anthropocene. Cambridge University Press. and Corry, O. (2017) ‘The international politics of geoengineering: The feasibility of Plan B for tackling climate change’, Security dialogue, 48(4), pp. 297–315.
Stephens, J.C. and Surprise, K. (2020) ‘The hidden injustices of advancing solar geoengineering research’, Global Sustainability, 3, p. Global sustainability. Volume 3 (2020).
Our work (2018) The DEGREES Initiative. Available at: https://www.degrees.ngo/about/our-work/ (Accessed: 14 August 2022).
Signatories (2022) Solar Geoengineering Non-Use Agreement. Available at: https://www.solargeoeng.org/non-use-agreement/signatories/ (Accessed: 12 August 2022).
https://makesunsets.com/blogs/news/calculating-cooling
https://www.technologyreview.com/2023/01/20/1067146/what-mexicos-planned-geoengineering-restrictions-mean-for-the-future-of-the-field/
Corry, O. (2013) Constructing a Global Polity: Theory, Discourse and Governance. Palgrave Macmillan UK.